550/9-73-008

     PROCEEDINGS of the
INTERNATIONAL CONGRESS
         on NOISE as a
  PUBLIC HEALTH PROBLEM
       DUBROVNIK, YUGOSLAVIA
           May 13-18, 1973
 U.S. ENVIRONMENTAL PROTECTION AGENCY
         Washington, D.C. 20460

-------
                 PROCEEDINGS of the
            INTERNATIONAL CONGRESS
                     on NOISE as a
             PUBLIC HEALTH PROBLEM

                DUBROVNIK, YUGOSLAVIA
                     May 13-18, 1973
               SPONSORS of the CONGRESS
          Union of Medical Societies of Yugoslavia
        Environmental Protection Agency, U.S. Gov't
          American Speech and Hearing Association
                 World Health Organization
                       Prepared by
THE U.S. ENVIRONMENTAL PROTECTION AGENCY
        Office of Noise Abatement and Control
             For sale by the Superintendent o! Documents, U.S. Government PrintinK Office
                      Washington, IXC. 20402 - Price $U.«

-------
                                     Foreword

     In  1968, a Conference on Noise as a Public  Health Hazard was organized by the
American  Speech and Hearing Association. At this  conference, an attempt was made to
bring together a group of speakers who could present summaries of the current state of
knowledge on all aspects of the "noise problem", ranging all the way from fairly technical
treatises to completely non-technical statements of personal opinion.  Such a wide-ranging
representation was judged to be necessary for the purpose of that conference, which was to
present a broad overview of what "noise pollution" was all about, to government personnel
and other intelligent laymen who saw that it was probably going to become a hot issue, and
give at least a few examples of the scientific evidence underlying arguments about just what
effects noise does have.
     At that time  it was  realized that  as  the  environmentalist movement gathered
momentum, a rapid development of public concern could be expected, and so a permanent
Committee of ASHA was established, one of whose charges was to plan another conference
when it was judged appropriate.
     The  burgeoning of interest in noise in the intervening 5 years has clearly met, if not
surpassed, our expectations  at that time. In the developed areas of the world, millions of
dollars or their equivalent are being spent  on surveys of noise levels and  exposures, and
increasingly stringent noise regulations are being imposed by all levels of government. And,
although  the measurement  of  the effects  of noise  is nowhere near as simple  as the
measurement of the noises themselves, many laboratories, mostly with federal support, are
engaged in full-time research on the hearing losses, sleep disturbance, speech interference,
alteration of physiological state, and annoyance caused by noise.
     Accordingly, in 1971 we began looking for a sponsor for a second conference—one
who would  agree, we hoped, to fund attendance  by a substantial number of researchers
from abroad, so that certain areas of knowledge less intensively studied in the USA could be
included in the subject matter. Fortunately, the head of the newly-created Office of Noise
Abatement  and Control (ONAC) of the Environmental Protection Agency, Dr. Alvin F.
Meyer, had need of just such a conference, as a source material for a document summarizing
all known criteria that might be used to establish national standards for noise control—that
is, provided that the Congress passed the bill, then being duly debated and amended,  that
would make such a document necessary. Furthermore, certain PL 480 funds (money  that
must be  spent in  other  countries)  were  available, which  meant  that  the  degree of
participation by foreign scientists might  be even greater than we had hoped. Not only that,
but the particular PL 480 funds in this case were in Jugoslavia, the country that includes
one of the garden spots of the world, Dubrovnik.
     On the assumption that our Congress would pass some  form of the bill in question
(which it did on October 27,  1972), we forged ahead with plans for  our meeting, now
upgraded  to an International Congress. With the help of Dr. Grujica iarkovic, the energetic
President  of the Jugoslavian Medical Association, and Dr. Mario Levi of the University of
Sarajevo, a planning meeting was held to which we invited a representative from most of the
countries  in which noise research was being done (I say "most" because we could not quite
afford to  pay for attendees from Japan, Australia, and South Africa because of the distance
involved, even though considerable research is being done there). At this meeting the formal
agenda was decided  on, and the list of invited participants prepared. It was agreed that we
would try to limit the Congress content strictly to the effects of noise on health, thereby

-------
excluding discussions of engineering aspects of noise reduction and control, descriptions of
methods for legal control, and presentation of viewpoints of special-interest groups. There
was some debate about how much time  to allot to public opinion surveys of annoyance,
some of us contending that annoyance, as measured in that manner, is not a health hazard at
all in the ordinary sense of the term.  However, proponents  of the WHO definition of
"health", in which any deviation from "optimum well-being"  is regarded  as undesirable,
carried  the  field,  and the  final day of the Congress was therefore given  over to the
sociologists.
    Despite a series of crises precipitated by governmental red  tape originating both in
Washington and Belgrade, the Congress was held on May 13-18, 1973 at the Libertas Hotel
in Dubrovnik. We had two major disappointments; one was  the failure of our Russian
invitees to appear due to  the fact that  our official invitations  had not been sent early
enough. The other was that the Xerox machine at  the Libertas  was out of  commission.
However, the general success of the Congress can be gauged by the fact that the audience
was as large on the final afternoon as at any other time.
    A side benefit of the Congress (or so we hope)  was the formation of an international
organization consisting of 5 "teams" who will try to accumulate and coordinate knowledge
about the effects of noise on (1) temporary and permanent hearing loss; (2) extra-auditory
function; (3) speech; (4) sleep; and (5) community reaction. The  parent group, or "basic"
team, will attempt to consolidate this knowledge for use by governmental  agencies, and will
make plans for the next Congress. Although the organization is now alive,  its name is still in
question. At the moment  it is  still the "International Scientific Noise  Teams", but the
resulting acronym has a negative connotation that pleases  few of us. Other names are being
considered.
    I regret that the length of the invited papers made it impracticable to publish at this
time any of the short contributed papers that were presented  at the Congress, many of
which were excellent, or the often-lively discussions that followed each session. It is hoped
that these can be included if another printing of the Proceedings is to be made.
    An enterprise of this scope cannot be a success without hard work on the part of many
people. Without doubt the most effort of all was  put forth by Dr. Levi, who managed all the
mechanical  details  of the  Congress, with the help  of his  and Dr.  3Larkovic's staff,
particularly, Felih Vesna.
    Official thanks are extended to our sponsoring organizations: The Jugoslavian Medical
Association, The American Speech and Hearing Association, the World Health Organization,
and of course most of all the Office of Noise Abatement and Control.
                                         u

-------
                         SPONSORS OF THE CONGRESS

    Union of Medical Societies of Yugoslavia
    World Health Organization
    Environmental Protection Agency, U.S. Government
    American Speech and Hearing Association

                      CO-PRESIDENTS OF THE CONGRESS

Grujica Zarkovic, M.D.      President  of the  Union of Medical Societies of Yugoslavia

Alvin F. Meyer, Jr.          Deputy Assistant Administrator of the U.S. Environmental
                              Protection Agency

Robert Goldstein, Ph.D.      President  of the American Speech and Hearing Association

                            PROGRAM COMMITTEE

    Angelova, M., Bulgaria                Kylin, B., Sweden
    Borsky, P. USA                      Levi, M., Yugoslavia—Standing Secretary
    Curlee, R., USA                      Nixon, C., USA
    Cerkez, F., Yugoslavia                Pearsons, K., USA
    Dieroff, H., DDR                    Pinter, L, Hungary
    Dzumhur, M., Yugoslavia              Raber, A., Austria
    Gulian, E., Rumania                  Spoor, A., Netherlands
    van Hattum, R., USA                 Sulkowski, W., Poland
    Hinchcliffe, R., UK                   Thiessen, J.G., Canada
    Jansen, G., Germany                  Tobias, J., USA
    Jokic, J., Yugoslavia                  Ward, W. D., USA-President
    Konig, E., Switzerland                Webster, J., USA
    Kryter, K., USA                      Whitcomb, M. USA


             SESSION ARRANGER AND PROCEEDINGS EDITOR

                              W. Dixon Ward

-------
                             TABLE OF CONTENTS

                                    Session 1

Introduction and Masking Effects of Noise
Chairman: Henning von Gierke (USA)                                           Page

     A Preview of the Congress Content, W. Dixon Ward (USA)  	    3
     Systems of Noise Measurement, Karl S. Pearsons (USA)	    7
     The Effects of Noise on the Hearing of Speech, John C. Webster (USA)	   25
     Reception of Distorted Speech, Jerry V. Tobias, F. Michael Irons (USA)  	   43
     Hearing Loss and Speech Intelligibility in Noise, Jerzy J. Kuzniarz (Poland)  ...   57
     The  Long-Term  Planning of a  Noise Control Program,  Michael  J.  Suess
         (Denmark)  	   73

                                    Session 2

Noise-Induced Hearing Loss (NIHL)—Empirical Data
Chairman: D. Robinson (UK)

     Basis for Percent Risk Table, Aram Glorig, William L. Baughn (USA)  	   79
     A Critique of Some Procedures for Evaluating Damage Risk from Exposure to
         Noise, Karl D. Kryter (USA)	103
     The Incidence  of Impaired  Hearing in Relation to Years of Exposure and
         Continuous Sound Level, (Preliminary  Analysis of 26,179 Cases), A. Raber
         (Austria)  	115
     Some Epidemiological  Data  on  Noise-Induced  Hearing Loss in  Poland,  Its
         Prophylaxis and Diagnosis, Wieslaw Sulkowski (Poland)	139
     On  the Problem  of Industrial  Noise and Some Hearing Losses in Certain
         Professional Groups Exposed to Noise, J. Moskov (Bulgaria)	  157
     Noise-Induced Hearing  Loss from Exposure to Intermitant  and Varying Noise,
         W. Passchier-Vermeer (Netherlands) 	169
     Evaluation of the Hearing Damage Risk from Intermittent  Noise According to
         the ISO Recommendations, B. Johansson, B. Kylin, S. Reopstorff (Sweden)  201

     Noise-Induced Hearing  Loss from Impulse Noise: Present Status, R.R.A. Coles,
         CG.  Rice, A.M. Martin (UK)   	211
     Hearing Loss Due to Impulse Noise.  A Field Study, Tadeusz Ceypek, Jerzy J.
         Kuzniarz, Adam Lipowczan (Poland)  	219
     Hearing Damage  Caused by  Very Short,  High-Intensity Impulse  Noise,  H.G.
         Dieroff (DDR) 	229

-------
                                    Session 3

Noise-Induced Hearing Loss—Mechanism
Chairmen: H.G. Dieroff (DDR), R. Hinchcliffe (UK)                             page

     Behavioral, Physiological  and  Anatomical  Studies of  Threshold  Shifts  in
         Animals, Donald H. Eldredge, James D. Miller, John H. Mills, Barbara A.
         Bohne (USA)	237
     Presbyacusis in Relation to Noise-Induced Hearing Loss, A. Spoor (Netherlands)  257
     Noise Exposure, Atherosclerosis and Accelerated Presbyacusis, Z. Bochenek, W.
         Bochenek (Poland) 	267
     High-Frequency Hearing and Noise Exposure, John L. Fletcher, (USA)  	271
     Susceptibility to TTS and PTS, W. Dixon Ward (USA)  	281
     Growth  of TTS and Course of Recovery for Different Noises; Implications for
         Growth of PTS, Wolfgang Kraak (DDR)	293
     Experiments  on  Animals Subject to Acute Acoustic Trauma, Wiktor Jankowski
         (Poland) 	301

                                   Session 4 A

Interaction of Noise with Other Noxious Agents in Production of Hearing Loss
Chairman: E. Lehnhardt (BRD)

     Influences of Chemical Agents on Hearing Loss, M. Haider (Austria)  	307
     Hearing Loss of Forest Workers and of Tractor Operators, (Interaction of Noise
         with Vibration), Istvan Pinter (Hungary)	315
     Infrasound and Hearing, Charles W. Nixon, Daniel L. Johnson (USA)   	329
     The Effects of Airborne Ultrasound and Near Ultrasound, W.I. Acton (UK)  . . . 349

                                   Session 4 B

Performance and Behavior
Chairman: D. E. Broadbent (UK)

     Psychological Consequences of Exposure  to  Noise, Facts, and Explanations,
         Edith Gulian (Romania)	363
     Similar and Opposing Effects of Noise on Performance, L. Hartley (UK)  	379
     The Effects of Different Types of  Acoustic Stimulation  on Performance, C.
         Stanley Harris (USA)	389
     Behavioral  Effects and Aftereffects of Noise,  David C. Glass, Jerome E.  Singer
         (USA)	409
     Effects  of Noise on a Serial  Short-Term Memory  Process, G. Wittersheim, P.
         Salame (France)	417
     The Effect of Annoying Noise on Some Psychological Functions During Work,
         Irena Franszczuk (Poland)  	425

                                       vi

-------
                                    Session 5

Non-Auditory Physiological and Pathological Reactions
Chairmen: E. Grandjean (Switzerland), S. Kubik (Czechoslovakia)                   Page

     Non-Auditory Effects of Noise, Physiological and Psychological Reactions in
         Man, Gerd Jansen (Germany)  	431
     Industrial Noise and Medical, Absence, and Accident Record Data on Exposed
         Workers, Alexander Cohen (USA)  	441
     Factors Increasing and Decreasing the Effects of Noise, D.E. Broadbent (UK)  . .  455
     Examples of Noise-Induced Reactions of Autonomic Nervous System during
         Normal Ovarian Cycle, Barbara Griefahn (DDR)	.^	459
     The Influence of Noise on Auditory Evoked Potentials, J. Gruberova, S. Kubik,
         J. Zal5ik (Czechoslovakia)  	469
     Some Data  on the Influence of Noise on Neurohumoral Substances in Tissues
         and Body Fluids, Lech Markiewicz (Poland)   	473
     Stress  and  Disease  in  Response  to Exposure  to  Noise-A Review, Gosta
         Carlestam, Claes-Goran Karlsson, LennarJ Levi (Sweden)   	479
     Some  Laboratory  Tests of Heart Rate  and Blood Volume  in Noise, Karl  D.
         Kryter (USA)	487
                                    Session 6

Sleep and Its Disturbance by Noise
Chairmen: B. Metz (France), M. Levi (Yugoslavia)

     Effects of Noise on Sleep-A Review, Harold Williams (USA)	501
     Predicting the Response to Noise During Sleep, Jerome S. Lukas (USA)	513
     The Effects of Noise-Disturbed Sleep on Subsequent Performance, M. Herbert,
         R.T. Wilkinson (UK)  	527
     Effects on Sleep of Hourly Presentations of Simulated Sonic Booms (50 N/m2),
         William E. Collins, P.P. lampietro (USA)	541
     Prolonged Exposure to Noise As a Sleep  Problem, Laverne C. Johnson, Richard
         E. Townsend, Paul Naitoh, (USA), Alain G. Muzet (France)  	559
     Relationship   between  Subjective   and   Physiological   Assessments  of
         Noise-Disturbed  Sleep,  A.  Muzet,  J.P.  Schieber, N.  Olivier-Martin,  J.
         Ehrhart, B. Metz (France)   	575
     The Effects of Aircraft Noise on  Sleep  Electrophysiology as Recorded in the
         Home,  Gordon Globus, Joyce Friedmann, Harry Cohen, Karl S. Pearsons,
         Sanford Fidell (USA)  	587
     Noise and Mental Health—An Overview, W. Hausman (USA)  	593
     Observations of the Effects of Aircraft Noise Near Heathrow Airport on Mental
         Health, C.F. Herridge, L. Low-Beer (UK)   	599
                                       Vll

-------
                                    Session 7

Community Response I
Chairmen: G. Thiessen (Canada), P.N. Borsky (USA)                              Page

     Methodological Aspects of Studies of Community Response to Noise,  EHand
         Jonsson, Ola Arvidsson, Kenneth Berglund, Anders Kajland (Sweden)  ...  611
     Decision Criteria Based on Spatio-Temporal Comparisons of Surveys on Aircraft
         Noise, Ariel Alexandre (OECD)  	619
     Psycho-Social Factors in Aircraft Noise Annoyance, Aubrey McKennell (UK)  . .  627
     A Survey on  Aircraft Noise in Switzerland,  Etienne Grandjean, Peter Graf,
         Anselm Lauber, Hans Peter Meier, Richard Muller, (Switzerland)  	645
     Aircraft  Noise  Determinants for the Extent of Annoyance Reactions, Ragnar
         Rylander,  Stefan Sorensen (Sweden)  	661
     Reaction Patterns  in Annoyance Response to Aircraft Noise, Stefan Sorensen,
         Kenneth Berglund, Ragnar Rylander (Sweden)	669
     The Reduction of Aircraft Noise  Impact  Through  a  Dynamic Preferencial
         Runway System, Martin Gach (USA) 	679
     A  Causal  Model  for  Relating Noise Exposure, Psychosocial  Variables and
         Aircraft Noise Annoyance,  Skipton Leonard, Paul N. Borsky (USA)	691
     Community Responses to Aircraft Noise in Large and Small Cities in  the USA,
         Harrold P. Patterson, William K. Connor (USA)   	707
                                     Session 8

Community Response II
Chairman:  R. Rylander (Sweden)

     Measurements of Street Noise in Warsaw and Evaluation of Its Effect on the
         Acoustic Climate of  Dwellings,  Schools, Offices, Hospitals,  Hotels and
         Parks;  the Degree  of Offensiveness to Inhabitants in the  Light of  a
         Questionnaire, Aleksander Brodniewicz (Poland)  	721
     A New Field Survey-Laboratory Methodology for Studying Human Response to
         Noise, Paul N. Borsky, H. Skipton Leonard (USA)   	743
     An  Interdisciplinary  Study  on the  Effects of Aircraft Noise on Man,  B.
         Rohrmann,  R.  Schiimer, A. Schiimer-Kohrs,  R. Guski, H.-O.  Finke
         (Germany)  	765
     Rating the  Total Noise Environment. Ideal or Pragmatic Approach? D. W.
         Robinson (UK)  	777
     Motor Vehicle Noise: Identification and Analysis of Situations Contributing to
         Annoyance, William J. Galloway, Glenn Jones (USA)  	785
                                       vui

-------
                                   Session 9

Summary and Integration
Chairmen: G. Zarkovic (Yugoslavia), W.D. Ward (USA)                            Page

    Summary, I.J. Hiish (USA) 	807
                                       IX

-------
            SESSION 1





INTRODUCTION AND MASKING EFFECTS




     Chairman: H.E. von Gierke, USA

-------
                     A PREVIEW OF THE CONGRESS CONTENT

                                   W. Dixon Ward

                             Hearing Research Laboratory
                               University of Minnesota
                            Minneapolis, Minnesota 55455
     Although the story is too long to recount in detail, I think that most of my fellow
participants here know that we owe our presence in this historic  and delightful city of
Dubrovnik not only to the hard work of Dr. Zarkovic and Dr. Levi, but also to a series of
lucky coincidences, culminating in the Noise Control Act of 1972, enacted by the Congress
of the USA on October 27, 1972. Two  provisions of this Act are as  follows: (1) The
Environmental Protection Agency shall "within nine months of the  date of the enactment
of this Act, develop and publish criteria with respect to noise. Such  criteria shall reflect the
scientific  knowledge most useful in indicating the kind and extent of all effects on the
public health  or welfare which may be expected from differing quantities and qualities of
noise." (2) It shall also, 12 months after enactment, "publish information on the levels of
environmental noise the attainment and maintenance of which in defined areas under var-
ious conditions  are  requisite to protect the public health and welfare with an adequate
margin of safety."
     Because of the obvious urgency of the charges (and in spite of their vagueness), EPA's
Office of Noise Abatement and Control was willing to subsidize our efforts to get together a
truly international meeting devoted exclusively to the effects of noise on human health and
welfare. It seems only fair, therefore, to look a little more closely at the task they have been
assigned.
     Now the term criteria, as used by Congress in the first provision above, consists of a
specified effect  or set  of effects that are set up as some sort of target—generally, a set of
conditions not to be exceeded. These criteria, in  general,  can be of two different types,
depending on whether they reflect concomitant  effects or after-effects of noise. If the
former, they may properly  be termed "noise criteria"; however,  the latter are more accu-
rately called "noise exposure criteria" because after-effects depend  not only on the charac-
teristics of the noise but also on the duration of exposure of a person to it. I believe that
this distinction is of paramount importance, though legislators do  not always understand it.
     For example, noise criteria could be (a) a certain degree of masking of ordinary speech,
or of radio or television perception; (b) a specified degree of vasoconstriction; (c) a definite
degree of probability of shifting the sleep stage  from a deeper to a lighter level; (d) an
average "comfortable loudness" as judged by some fraction of the population; or, conceiv-
ably, (e) the point at which aural pain is felt.
     On  the other hand,  noise exposure criteria could be  based  on a specified degree of
temporary or permanent threshold shift, or on a certain amount of hair-cell damage, or on a
change in circulatory problems in a specified fraction of the population, or a similar definite
change in any aspect of health.

-------
     There are, of course, certain effects that may be  both concomitant and residual. A
decrement in task performance—that most elusive of the many effects of noise that seem so
obviously  real but which generally vanish into thin air when one  looks for them in the
laboratory—might serve as a criterion for either noise or noise exposure. The same holds for
annoyance; we could generate criteria based on the arousal of the feeling of annoyance in a
specified fraction of the individuals concerned by  a specific noise, or we could use an
expression of integrated annoyance, as displayed by complaints and legal action.
     Even in the latter cases, however, it remains important to keep the two types of criteria
separate, and to try to educate the public—and particularly lawmakers—to the distinction.
The  latter, naturally, want to regulate noises, because they are relatively easy to measure.
However, the most important effects on human health come not from noise but from noise
exposure.  Noises per se are not hazardous, sleep-disturbing, or annoying-only noise expo-
sures can be. Therefore in this symposium our attention will be focussed primarily on noise
exposures: a duration of exposure must usually be involved, in addition to the intensity and
spectrum of the noise, when one calculates a noise "dose".
     Ideally,  in order to specify the relation between noises or noise exposures  and their
effects,  what is needed is a scale for measuring noise that  would result in each noise being
assigned a specific number whose magnitude  would reflect the relative noxiousness of that
noise. Let us assume for the  moment that such a scale could be found; since in science the
units of scales are  often named for famous men in the  field  concerned (for  example,
newtons of force, watts of power, amperes of current),  the unit of  this scale might well be
the peyser, in honor of one of the early pioneers in the noise field, Alfred Peyser. A noise
whose rating was 50 peysers would be twice as noxious in all respects  as one rated at 25
peysers, five times as noxious as a 10-peyser noise, half as noxious as a 100-peyser noise, etc.
     Furthermore, if noise exposure consisted of the instantaneous  value of the noise inte-
grated over time, noise exposure could then be expressed in peyser-hours. For example, a
man who worked 8 hours in a noise of 10 peysers, went home and  cut his lawn for half an
hour in  a 40-peyser noise, and then listened to his son's music group practicing at 50 peysers
for 2 hours before retiring for the night would have had a total noise exposure that day of
8x  10 + ^x40 + 2x50 = 200 peyser-hours. His effective exposure on that day would be
the same as that of another person whose noise exposure at work consisted of 10 hours at 20
peysers  and negligible the rest of the day, or that of a third man whose work environment
was  quiet, but who spent an hour at a rifle range in a 200-peyser noise without wearing any
ear protection. If the  size of  the "peyser"  had been defined  in such a  way  that  100
peyser-hours  were the maximum tolerable daily noise-exposure "dose",  then each of these
three individuals would have experienced  twice as much noise as he  should have, and if this
were continued day after day for many years, then he would be expected to show twice as
much hearing loss as the person exposed to only 100 peyser-hours each day.
     Unfortunately, nature has not been so obliging as to furnish us with such a scale, nor
indeed has she provided the uniformity of degree of effect that would make such a scale
even possible. That is, a noise mat is twice as likely as another to cause a person to awaken
is not twice  as annoying nor twice as  hazardous to hearing, nor does it produce twice as
much of a change in the circulatory system nor interfere with twice as much speech. In fact,
many noises that  are highly irritating and  hence should have a high  peyser index may

-------
produce no effect on hearing whatsoever, and so should be rated on that basis as being at
near zero peysers.
     Furthermore, it is known that a noise exposure of say 1 hour at 10 peysers will have
quite a different effect on temporary threshold shift measured immediately afterward than a
cumulative one-hour exposure again at 10 peysers, but in which 5-min periods of noise are
separated  by say  30 min  of quiet.  Therefore  if TTS  were  our criterion, the index  of
noxiousness would have to include some factor  that takes such intermittency into account.
     I fear, therefore, that the search for a single index of noise exposure as an indicator
that a given criterion effect has been  reached is foredoomed to failure. Perhaps I am
unnecessarily pessimistic—or perhaps some first-order approximation such as the concept of
"equal  A-weighted energy"  with appropriate  correction  factors will prove  to be close
enough to reality to justify its use in the absence of a true unifying principle. 1 know that
we will hear something more of this concept during the rest of our symposium.
     Most of us,  however,  find  it difficult enough  to cope  with the complex relations
between noise or noise exposure and our own favorite effect—in my case, for example, TTS.
Thus a host of specific questions will come under scrutiny in the some 90 papers that follow
this one,  questions whose answers are mostly still debated rather hotly. The following are
some that can be expected to appear:
     (  1)  Is hearing above 3000  Hz important to the  perception of speech? If so, under
what conditions?
     ( 2)  What frequency  weighting scheme, such as A-weighting and D-weighting, gives
the closest prediction of the speech-masking ability of a noise?
     ( 3)  Can there be damage to hearing without a change in sensitivity?
     ( 4)  What single exposure (8 hr or less) will just  produce a "significant" permanent
threshold shift (PTS)?
     ( 5)  What relatively steady-state  exposure, 8 hr/day, for many years, will just produce
PTS that exceeds that ascribable to presbyacusis plus sociacusis?
     ( 6)  Is there any  way  to correct audiometric  data for  presbyacusis-plus-sociacusis
other than simple (and probably incorrect) subtraction?
     ( 7)  Under what conditions does the  equal-energy hypothesis hold for steady expo-
sures?
     ( 8)  Can individual differences in susceptibility to PTS be predicted?
     ( 9)  Can this susceptibility be changed by drugs or diet?
     (10)  What is  the evidence for and against  the microtrauma theory as opposed to the
critical-incident hypothesis in the production of PTS?
     (11)  To what extent does it make any sense to speak of a "critical intensity" or even a
"critical exposure" for a given ear?
     (12)  Is a damaged ear more susceptible to further damage than a nondamaged one?
     (13)  In such case, what is "equal further damage" in the first place?
     (14)  Is some aspect of the TTS produced in a group of listeners a valid index  of
average expected PTS after years of exposure to that noise?
     (15)  If so, which parameter—initial TTS, recovery time, or  what?
     (16)  To what extent is the  auditory hazard from  noise enhanced by other noxious
influences such as vibration, fumes, exertion?
                                         5

-------
     (17)  To what extent does intermittency reduce the hazard from a given (cumulative)
noise exposure?
     (18)  In recovering from ITS, what noise level constitutes "effective quiet"?
     (19)  Is 4000 Hz  the place to first look for auditory  damage, or are the very-high
frequencies more susceptible?
     (20)  Does infrasonic noise or ultrasound at commonly-found intensities pose a hazard
to health?
     (21)  What are the effects of repeated awakenings or forced changes in depth of sleep
every night (by noise or by any other agent)?
     (22)  How much does simple reaction time to visual stimuli change in noise?
     (23)  Does such a  change persist after exposure (or, under what conditions does it
persist)?
     (24)  If so, is there any evidence that a permanent change might ensue?
     (25)  Does chronic arousal of the vegetative system lead to circulatory problems?
     (26)  Does chronic noise exposure increase mental problems?
     (27)  Is there any evidence in humans for changes in the adrenals due to noise expo-
sure, as commonly found in rats? Or in any of the other stress-reaction indicators?
     (28)  Is there any way to measure the "fatigue" that many workers complain of after
noise exposure?
     (29)  Does chronic  noise exposure increase the incidence of gastro-intestinal problems?
     (30)  What task performance is adversely affected by noise?
     (31)  Do  workers in high levels of noise show a significantly higher absentee and illness
record?
     (32)  To  what extent do individuals "adapt" to noise that does not pose a hazard to
hearing?
     (33)  To  what extent  can  such  "adaptation" be   manipulated by  propaganda
techniques?
     (34)  What constitutes a significant increase in complaints about neighborhood noise,
Le., how  much greater than the baseline of chronic  complainers must a  complaint  level
attain before a practical problem exists?
     (35)  Is utter silence seriously advocated at being the "best" acoustic environment?
     I  hope that  by our careful consideration of the evidence, answers to at least some of
these questions can be reached and accepted  by the majority of us here. I also hope that we,
at least, can keep from confusing noise and noise exposure.

-------
                        SYSTEMS OF NOISE MEASUREMENT

                                   Karl S. Pearsons
                           Bolt, Beranek and Newman, Inc.
                                Los Angeles, California
     Probably the most universally used system of noise measurement is something we all
carry with us all the time, our ears. It's an extremely versatile device that normally measures
sounds over a range of 120 dB and a frequency range from about 15 Hz to 20,000 Hz.
However, my talk today  is not  about our ears, nor is it a history of noise ratings,  but rather
a summary of the noise ratings which are currently in use today. Much of the material which
I will speak on is contained in a Handbook of Noise Ratings. This Handbook was prepared
for  the  National  Aeronautics  and  Space  Administration, Langley  Research  Center  in
Hampton, Virginia. I had hoped to have  copies of this available at  this time; however,
preparation delays and printing delays probably will prevent the Handbook from  being
available for 6 to 8 months.
     Hopefully my talk will prepare you for the kinds of noise ratings and measurements
which will be discussed in relating certain effects of noise to some noise measures.  Details of
the calculation procedures for  determining the various noise ratings will not be  presented
here since there is not enough time. Rather  the approach is to summarize the various classes
of noise ratings and provide some indication of the type of jobs that rating is supposed to
do. Some comparisons will also be made among the noise ratings but remember that each
noise rating is individualistic and cannot be translated directly to another noise rating except
for perhaps a particular sound which is being measured. To facilitate the discussion of the
noise measures  let us consider them  in five groups:  1) direct measures, 2) calculated
measures, 3) calculated measures for long term exposure (community response measures), 4)
graphical measures, and 5) measures specifically related to hearing level.
     The last category is not contained in  the Handbook of Noise Ratings which I men-
tioned earlier. However, because of the nature of this meeting it seemed important to briefly
touch on the nomenclature to facilitate the understanding of those sessions concerned with
hearing damage.
     Before we discuss the various measures, let me first mention some of the terms which
will  be employed in describing the various measures even  though most of you are familiar
with them. Although there are special units for measuring certain aspects of noise,  in general
noise is measured in decibels. This is a logarithmic quantity chosen because of the  very large
range of sounds  which people perceive. The decibel, usually abbreviated (dB), is  a measure
of a magnitude of a particular quantity such as sound pressure, sound power intensity with
respect to a standard  reference value. This standard reference value is usually 20 micro-
newtons per square meter. This  is about the  threshold of hearing for young ears at 1000 Hz.
     The other major aspect of  sound is its  frequency content. This is measured in terms of
hertz (Hz), formerly called cycles per second.  As I mentioned earlier, the range of hearing is
about 15 Hz to 20,000 Hz. This is really the number of times which something oscillates or
vibrates per second. For an example, the musical pitch "A" is an oscillation of 440 times per
second. A truck passing  by may have energy in the vicinity of 200 Hz. The high pitched

-------
whine of a jet engine would be about 3000 Hz. To further describe a sound in terms of both
its level and  frequency content the latter is sometimes divided into various bands. Such
bands as octave bands are sometimes employed. An octave band is a frequency band whose
upper and lower  cutoff frequencies have a ratio of 2. It is characterized by its upper and
lower frequency bounds or its center frequency which is the geometric mean of the upper
and lower bounds. Noise or sounds may be measured in terms of octave band sound pressure
level as shown  in Figure 1.  This is the  sound pressure level which is contained within an
octave band.  Finer resolutions may be made by employing third-octave bands or one-tenth-
octave bands  or even narrower bands.

DIRECT MEASURES

     The first measure is the overall sound pressure  level or sometimes simply the sound
pressure level with  various abbreviations such  as OASPL or SPL  or L or Lp. The overall
measure, which is approximated by the C-weighted network of a sound-level meter, provides
equal weight  for  all frequency components of the  noise. It is primarily used by engineers
who need a  measure which  indicates the total noise energy. Weighted sound levels are
measured on  sound-level meters in  terms of fast or  slow response.  These terms refer to the
speed with which an indicating meter follows the fluctuating sound. The approximate time
constants of this sampling procedure are about 1/10 of a second and 1 second respectively.
     The most  common and  widely used  sound measure in the world  is the A-weighted
sound pressure level or more simply the  A-level. This measure is also quantified in units of
dB although a shorthand technique has been employed to eleminate the necessity for saying
A-level each time a  measurement is quoted. This shorthand is to consider the unit a dB but
with an (A) following the dB and is usually read dB (A). It should be emphasized that dB (A)
is not an  actual  unit but rather a shorthand method to tell the  reader which weighting
network was employed to make the measurements. Figure 2 shows a diagram of A weighting
along with other weightings which we will  discuss shortly. Notice that the low frequencies
are attenuated.  The reason for this is to  more closely approximate the way people perceive
sounds. Originally it was designed for sounds of less than 55 dB in level; however, currently
it is used for all level sounds.
     The B-level sound does not discriminate  as much  against  the low frequencies. It is
shown  also on this figure, but currently is not widely used. The C-level as mentioned  earlier
provides an indication of the  flat response. Its frequency range was somewhat dictated by
the  frequency range of the ear but originally was influenced by available instrumentation.
Essentially the C-weighting limits the high and low  frequency response, but in spite of this
limitation  it  still  provides a reasonable  measure  of overall sound  pressure level for most
common noises.
     A relatively new addition to the weighting levels is the D-level. The weighting network
shown  in Figure 2 used for this measure is more complicated than the earlier ones and tries
to incorporate more accurately the frequency response of the ear. Actually it was originally
developed  as  a simple approximation of  perceived noise level PNL which I will discuss later
under calculated measures. Originally the  D-level was described as N-level with the  differ-
ence between the D and the N being 7 dB. In other words if 7 dB is added to the D-level,
one should obtain approximately the perceived noise  level of a given sound. Modifications

                                         8

-------
    no
    100

-------
 to D-level have been suggested by Kryter to account for new data. Mainly they affect the
 low frequency portion of the weighting.
     E-level has also been suggested and is fairly similar to the D-level. It was suggested by
 Stevens as an approximation to what he termed perceived level. Differences between the
 E-weighting and D-weighting are relatively small. At this time, the E-level is not standardized
 nor is it available on any sound level meter.
      -o
      c

      •o
      c
      o
      to
     JJ
      o
 20

 10


  0


-10


-20


-30


-40


-50
          20     50   100  200    500  1000
                            Frequency  In  Hz
                                                               5000     20,000
                  Figure 2. Frequency response of various weighting functions.
GRAPHICAL MEASURES

     The two graphical measures which will be described today are the noise criterion curves
and the preferred noise criterion  curves. Other measures such as Zwickjer's calculation of
loudness and some of the community response measures also employ graphical techniques,
however they will be discussed later under the calculated measures. The noise criteria curves
shown in Figure 3 were developed to provide a single number rating for octave band spectra.
They are, mainly employed by architects and engineers to specify the maximum noise levels
                                         10

-------
permitted in each octave band. In using them, the octave-band spectrum is plotted and an NC
value assigned to the noise, a value that corresponds to the highest NC curve to which the
spectrum is anywhere tangent. Thus the NC rating is almost always determined by the sound
pressure level at a single octave band frequency. For example, in Figure 3, an NC rating of
59 characterizes  the noise spectrum.  These NC curves were originally developed for office
spaces; however, they have been used in other environments such as auditoriums, sound
studios, restaurants, etc.
    Recently the  noise  criteria curves have been modified both to accommodate more
precisely the new octave band center frequencies  and also to answer the many objections
made  about the adequacy of the NC curves. Mainly changes are made in the higher and
lower frequencies since spaces designed in accordance with the previous NC  curves were in
some  cases  too  hissy or too rumbly. The new curves are shown in Figure 4. Here the
previously-mentioned noise spectrum would  have  a value of PNC 61  or 2 higher than the
rating given by the NC curves.

CALCULATED MEASURES FOR INDIVIDUAL EVENTS

     Table I provides a list of various calculated measures. For the most part, these measures
utilize octave- and  third-octave-band levels  of  noise which  are  employed  in  various
calculation schemes to come up with a single number rating of that noise. LLS stands for
loudness level, "s" refers to its originator, S. S. Stevens from the United States, who devoted
a good deal of his life refining the techniques for predicting the loudness level of sounds.
The scheme is intended to provide a level of the sound which is numerically equal in level to
that of a 1000-Hz tone which is judged  equal  in loudness to the sound being rated.  The
technique now is a calculation procedure which essentially transforms octave band levels to
a loudness quantity called sones that are added up  in a particular way and transformed back
to a decibel-like quantity known as phons.
    Another  scheme for calculating loudness  level  identified  as LLZ, "z" for Zwicker,
employs graphs and also allows for the upward  spread of masking, (the masking of higher
frequencies by low frequencies). This technique uses  one-third-octave band data and the
result is  intended to represent the level of a one-third-octave band centered at 1000 Hz
judged equally loud to the sound being rated.
    PNL or perceived noise level is similar  to  the loudness level by Stevens except  that
noisiness is employed instead of loudness. The units for this measure are PNdB or perceived
noise decibels. The numerical value was intended to represent the sound pressure level of an
octave band of noise at  1000 Hz which would be  judged equally noisy to the sound to be
rated. Equally noisy means that in a comparison of sounds one would just as soon have one
noise as the other at his home during the day or night.
    Stevens continued improving on his loudness  level calculation and came up with a new
rating technique  called perceived level which was similar in concept to the loudness level but
utilized more  information and included  noisiness  as  well as loudness judgment tests.  The
main difference between perceived level and the loudness level and perceived noise level is in
the numerical value. This time the levels were lower by approximately 8 dB than the earlier
loudness level calculation scheme. Also the  units are PLdB for'perceived level decibels

                                        11

-------
                                            1200
                                            2400
2400
4800
4800
9600
                                              NOISE CRITERIA

                                               NC  CURVES
                                                       IMC-40
                                                       NC-30
   APPROXIMATE
   THRESHOLD OF HEARING
  UFOR CONTINUOUS NOISE
   REF: ACUSTICA 14 (1964)
   PAGE 33, FIG. 14
10
 31.5     63     125      250      500     1000     2000    4000
         OCTAVE BAND CENTER  FREQUENCIES IN Hz (Cps)
            Figure 3. Noise criteria curves with noise spectrum example.
          8000
                              12

-------
                     1971 PREFERRED NOISE CRITERIA
                               PNC-CURVES
 Approximate threshold of
 hearing for continuous noise
 Ret:  Acustica 14 (1964)
      Page 33, Fig.14
H.5
63     125    250    500    1000   2000   4000
OCTAVE-BAND  CENTER FREQUENCIES IN  Hz

     Figure 4. PNC curves with noise spectrum example.

                   13
8000

-------
                                                 TABLE I
                    SUMMARY OF CALCULATED MEASURES FOR INDIVIDUAL EVENTS
      MEASURE
ABBREVIATION
UNITS
                   PROCEDURE
Loudness Level - Stevens
Loudness Level - Zwickei
Perceived Noise Level
Perceived Level
Tone Corrected Perceived
Noise Level

Effective Perceived Noise
Level
Single Noise Exposure
Level
Articulation Index
Speech Interference Level
    LL«
    LL,
    PNL
    PL
    PNLT
    EPNL
    SENEL
    Al
    SIL
Phons
Phons
PNdB
PLdB
PNdB
EPNdB
dB
(none)
(none)
Using 1/3 octave* band SPL's, loudness index (S) values
are determined from tables**. Total loudness S, is then
determined by S, = Smax + F (£S - Smax), F = .15

then LLS = 40 + 10 Iog2 S,

Graphical procedure using 1/3 octave band SPL's; includes
upward spread of masking effects

Using 1/3 octave* band SPL's, noy values (n) are deter-
mined from tables**. The total noy value (Ntot) is then
determined by Ntot = nmax + F (En - nmax), F = .15

then PNL = 40 +33.22 log ,0{N,ot)

Using 1/3 octave or octave band SPL's, sone values (S)
are determined from tables**. The total sone values (St)
is then determined by S, = Smax + F (ES - Smax) where
F is a function of level and bandwidth

then PL = 32 + 9 Iog2 St
PNLT = PNL + Tone correction
                        PNLT
                                          2d
EPNL = 10 log £   antilog (
             r=0         10
                             ) - 1 3
where PNLTj is value of ith .5 second sample
                n        AL
SENEL = 10 log (E  antilog - ) fit
                i=l        10

where AL is level of ith 1 second sample
      At is time interval between samples in seconds
      n  is the number of events

AI can be calculated from one-third octave band or octave
band differences in speech and background noise levels
where a weighting correction is applied to each band to
account for the relative contribution of each band to
speech intelligibility

SIL is the arithmetic average of the sound pressure levels
of the noise in the four octave bands with center frequencies
lying between 5 00 and 4000 Hz
                                                        *Octave band data may be used by employing !•' = .3
                                                       **Tablesare available in "Handbook of Noise Ratings".
                                                         Sec acknowledgement.
                                                     14

-------
instead of phons. The summation procedure for combining the octave or third-octave contri-
butions is more sophisticated and accounts for masking as a function of level.
     With the advent of discrete frequency components or tones in aircraft flyover noise it
seemed advisable to add a correction for the presence of these tones. Perceived noise level
does not  attempt to include these and therefore additional  calculations were employed to
account  for  the increased noisiness caused by  these tones.  The tone-corrected perceived
noise level, PNLT, is the current method for applying the corrections.  This method essen-
tially adds a correction to  the final value of the perceived noise level depending on the
amount that  the third  octave band containing the tone exceeds its adjacent bands. A fairly
complicated  procedure is actually employed utilizing computer techniques to determine
whether  or not a tone exists in a spectrum. This technique  essentially determines what the
noise floor is by reiterative  averaging process. The technique can be accomplished by hand
but is fairly cumbersome and a reasonable approximation can be made by averaging the two
adjacent bands and  subtracting this average from the band  level containing the tone. This
difference then is divided by three for frequencies between 500 and 5000 or divided by six
for other frequencies  and the result added to  the numerical value of the perceived noise
level. The limit of the tone correction is 6.7 for third octave bands between 500 and 5000
and 3.3 for all other third octave bands in the frequency range of 100 Hz to 10000 Hz.
     Since long-duration flyovers appear to be more annoying or noisier than short-duration
flyovers, a duration  correction was applied to the tone-corrected perceived noise level and a
new quantity called effective perceived noise level (EPNL) came into being. The method for
applying the duration  correction is essentially one which integrates or sums the PNLT levels
in half-second periods. This is equivalent to adding 3 dB for every doubling of duration of
the sounds. Currently  this measure is employed  in the aircraft noise certification procedures
in the United States. The units for the measure are EPNdB.
      SENEL  or single  event noise equivalent level is another measure of single events-in
particular, individual aircraft flyovers.  In this sense it is similar to effective perceived noise
level, but has no tone correction and employs the A-level weighting instead of perceived
noise level.  Presently its main use is in conjunction  with  the determination of community
noise equivalent level which will be described later.
     Moving  into the area of speech related noise measurements, we find two main calcu-
lated measures. One which will be mentioned in the following paper is articulation  index,
AI. Essentially it is a measure between 0 and 1 which purports to indicate speech intelligibil-
ity. It is based on the proportion of the normal speech signal that is available to the listener.
An articulation index  of  .6 or greater indicates reasonably  good intelligibility while levels
less  than .2  indicate  poor intelligibility. The technique  used in  determining articulation
index is  to divide the speech and sound into 20 bands from 200 Hz to 6100 Hz. The bands
are specially  selected such that for speech signals each contributes equally to intelligibility.
The value of articulation index is then  determined by the  sum in dB of the  differences
between the  peak speech levels and the noise spectra in each of the 20 corresponding bands
relative to an ideal speech to noise ratio of 30 in each band. Approximations are available
for determining articulation index from third-octave and octave band data using appropriate
weighting factors to account for the relative contribution of each band to speech intelligibil-
ity.  Many tests have been conducted using steady state noise to determine the percent of

                                          15

-------
various  types of speech material correctly understood for various levels of articulation
index.
     A simplified method, useful for engineers to determine approximate effects of noise on
speech,  is the speech interference level (SIL). Currently this is an arithmetic average of 4
bands centered at 500, 1000, 2000 and 4000 Hz, although other bands have been suggested.
No measures are made of the speech but in utilizing the speech interference level various
levels of voice are  assummed such as "normal", "raised", etc. Graphs can then be made
indicating the  distances over which speech is reasonably understood. Articulation  index
values for these various speech levels characterized by normal voice, raised voice, etc. have
been about 0.5. More about this matter will be described in the paper by John Webster.

CALCULATED MEASURES FOR MULTIPLE EVENTS
(COMMUNITY RESPONSE MEASURES)

     Probably the  biggest proliferation  of noise measures exists in the various schemes to
rate community noise. Several of these measures are in use in the United States and other
measures are employed in various countries around the world. The measures discussed here
do not include all those in use today but rather  should provide an indication of the various
ways in which community noise is assessed. All of the measures attempt to relate in some
fashion  to the noise impact on the community. As such, they eventually come to some type
of descriptive meaning for the various levels. Table II provides a summary of these measures.
In most cases the units for these  measures or ratings are in "dB-like" units. This means that
they are not actually in terms of decibels in the  normal sense of the word but they are in
logarithmic quantities which relate somewhat to decibels. Thus for example the measures
would increase by 10 units if the level of the contributing signals went up by 10 dB.
     CNR stands for Composite  Noise Rating or sometimes Community Noise Rating. It
was one of the  earlier attempts to evaluate community  reaction to noise in 1952. The
technique assumes that sounds were measured  in octave bands and that the values are
obtained by averaging over a reasonable time interval for critical locations in the commun-
ity. It utilizes a  family of curves that ranks the noise level on  a scale from A through M,
Thus a noise level rank is determined by plotting the octave band spectra on a set of level
rank curves in a manner similar to that used in determining NC levels described earlier. The
level rank thus determined is corrected for:
     1)  Discrete frequency components
     2)  Impulsive nature of the sound
     3)  Repetitiveness of the sound
     4)  Background noise level in the community
     5)  Effect of the time of day
     6)  Previous community exposure to the noise
     After all, corrections were applied, the final CNR, is determined as a new letter. The
letters are then  converted to  various  community  reaction  such as no annoyance,  mild
annoyance, mild complaints, strong complaints,  threats of legal action, and vigorous legal
action.  These categories have  been changed slightly throughout the years but remained
essentially the same. Later the CNR was employed for rating aircraft noise and at this time

                                        16

-------
the letters were replaced by dB-Iike numbers. The numbers were still derived from the same
level rank. The scheme also tended to shift from one which evaluates community noise to
one which predicted it on the basis of aircraft flyovers. Special contours were developed for
landings, takeoffs, and ground run-ups of various  types  of aircraft.  The end result still
remained the same which was to categorize the CNR into various community responses.
     As methods for rating aircraft noise improved, it was decided to create a new rating for
impact of aircraft noise on communities rather than continually change or update the CNR
rating method.  Thus the  noise exposure forecast NEF was born, which employed E, the
effective perceived noise level,  as its basis rather than the perceived noise level as mentioned
earlier. This measure accounted for both the additional noisiness  of discrete frequencies or
tones and the effect of duration.  A correction for nighttime operations of 10 dB was still
employed. In other words, for the same average number of aircraft operations per hour, the
NEF value for the nighttime operations would be 10 units higher than for daytime opera-
tions. The NEFs around  an  airport were lowered  in absolute value  by  subtraction  of  a
constant of 88  to avoid confusion with the previously developed CNR. An example of NEF
contours are shown in Figure 5. The final NEF values were converted to 3 levels of com-
munity response as follows:

     NEF          Description of Community Response

     Less          Essentially no complaints would be expected. The noise may, however,
     than 30       interfere occasionally with certain activities of the residents.

     30 - 40        Residents  in the  community may complain, perhaps vigorously. Con-
                   certed group action is possible.

     Greater       Individual  reactions would likely include repeated, vigorous complaints
     than          and recourse  to  legal action. Concerted  group action would  be ex-
     40            pected.
                                                             NEF 25,
                            Figure 5.  Example of NEF contours
                                         37

-------
     Before going on to the next type of community noise rating let us discuss briefly a very
general measure. This is called the equivalent sound level, or Lgq. Very simply, this is the
energy average of the noise level (usually in A-level) for some specified amount of time. This
measure is used to allow quantification of fluctuating sounds over a long period of time. In
essence it is that numerical value of the fluctuating sound which is equivalent in level to a
steady state sound with the same amount of total energy. If, for example, one had a sound
which was 80 dB and it was on for half an hour and then went down to a level of 40 dB for
half an hour, then  the Lgq would be the energy average, or 3 dB less than the maximum
value.  Thus the energy Lgq for this sound  sample would be 77 dB and not 60 dB as we
might  expect if we just averaged the dB levels. Leq is usually approximated by taking several
samples in time of A-level, then averaging the samples by first dividing by 10 and taking the
antilog and actually averaging those quantities, then reconverting back to a dB level by
talcing the log and  multiplying by 10. Since this measure is employed in several measures
which  will follow, I  felt it was important to briefly discuss it at this time.
     One measure which does indeed use this Leq is CNEL or Community Noise Equivalent
Level. This rating represents the average noise level on an  energy basis determined for a
24-hour period with different weighting factors for noise levels occuring  during the day,
evening, and nighttime periods. Essentially, then, this is an L eq for the day for the 24 hour
period  but with  special weightings of 5 dB  and  10 dB respectively to account for the
increased disturbance caused by noise events during the evening (1900-2200) and nighttime
(2200-0700) hours. To facilitate these calculations, an hourly noise level or Leq for an hour
is employed and weighting factors are applied directly to  this measure.
     The next group of measures, which includes  the Isopsophic index designated as (N)
which  is used in France, the mean annoyance level (Q) which is used in Germany, the noise
and number index (NNI) used in England, the noisiness index (NI) used in South Africa, the
total noise load (B) which is used in the Netherlands, and the weighted equivalent contin-
uous perceived noise level,  WECPNL which was suggested  by the International Civil Aviation
Organization are  all somewhat similar to the  measures of the CNR or CNEL measures
already discussed. They do differ in some detail.
     First the isopsophic index: This is a noise rating which  takes into account the energy
average maximum perceived noise level of aircraft noise and the number of events. Another
French measure, the classification index (R), is identical in all aspects to the isopsophic
index.  The big difference  between this and other measures is in the  handling of nighttime
events. In the first place, the nighttime events are broken into early night (2200-0200) and
late  night (0200-0600) time periods.  The early night period is viewed as three times  as
significant as the second or late night time period. Also, the effect of doubling the number of
operations at night  is not as great as during the day, since doubling the number of opera-
tions increases the nighttime portion of this measure by less than 2  units as opposed to 3
units for the normal daytime operations. The measure is used to determine zones for various
types of buildings. The zones include areas where all buildings are prohibited down to levels
for which no building restrictions apply.
     The mean annoyance level Q is another noise rating for aircraft noise impact on a
community. It uses sampled A-level to  provide an average noise level  for a specified tune
period-for example, day,  night  or 24 hours. Again, it is similar to CNEL except there is no
nighttime  weighting and a doubling of the number of events increases the measure by 4

                                         18

-------
units as opposed to 3 units for CNEL. The Q is also employed to designate 4 zones as did
the previously discussed isopsophic index.
    NNI uses the average perceived noise level (averaged on an energy basis) in combination
with the number of aircraft heard within a specific period. Unlike the previously discussed
measures,  this rating increases by 4Vz units for a doubling  of the number of events. No
distinction is made between daytime and nighttime in the calculation procedure; however,
different levels of NNI are employed in determining reasonable levels for daytime or night-
time operations. Thus a level of 50 to 60 NNI is assumed  to be unreasonable during the
daytime whereas an NNI of 30 to 45 is intolerable during the nighttime.
    The total noise load, B, employs maximum A-level  and  number of aircraft with ap-
propriate  weightings for time of day. B was developed to be  numerically equal to the
percentage of mean relative nuisance. The Dutch authorities have chosen a B rating of 45
which  is  equivalent to  45%  mean relative nuisance score. The main  difference  in  this
measure from the previous ones is the fact that a doubling of the number of events increases
the rating by 6 units rather than 3 for normal energy summations. The number of time
periods for which weightings are given is increased to include  10 different periods during the
24-hour day, with nighttime hours weighted  much more heavily  than other measures. The
noisiness index (NI) used in South Africa is the energy average noise level  based  on a tone
corrected  A-level for a 24-hour period. Appropriate weightings are applied for time of day
and also season of the year.  The tone corrections for A-level are determined from third-
octave-band levels before summing to obtain  a corrected A-level. The actual tone-correction
procedure is taken from the techniques employed for EPNL or PNLT tone corrections. Two
sets of weightings for day and nighttime activities are provided for two different  groups of
periods; for example, if the day is divided into two periods, there is  a 10-db weighting for
nighttime  events occuring during the hours of 2200 and 0700, while if the day is divided
into three periods, then  a weighting of 5 dB for evening hours of  1900 to 2200 is employed
and a 10 dB weighting for nighttime events occuring between the hours of 2200  and 700.
Seasonal corrections are based on the number of hours in a  month which the temperature
falls in the range of 20 degrees centigrade to 25 degrees centigrade. This is done to provide
more weighting for the situations when the windows are open  in the summer months.
    The weighted equivalent continuous perceived noise level, WECPNL is an attempt to
provide a standardized measure for the impact of aircraft noise on the community. This is
quite similar to the CNEL described earlier but uses tone corrected perceived noise level as
its base for energy averaging rather than A-level. Also, weightings are included for  season of
the year, and time-of-day corrections  for only two periods rather than three three periods
used in the  CNEL calculation procedure are  employed. To provide an indication  of the
approximate values of the various measures we have discussed in terms of number of aircraft
operations per day, see Figure 6, which shows the levels for the various rating techniques.
The figure assumes a flyover noise of 110 PNdB with an effective duration of 10 seconds.
Approximations had to be made in certain cases in order to make this comparison possible.
For example  in the WECPNL and the NEF a flyover of 110 EPNdB  instead of 110 PNdB
was employed. Notice that the lines on the graph are not all  parallel to one another. This is
because the number of operations per day is not always summed in an energy fashion for all
of the  measures as discussed earlier. If we are to pick an average number for CNR  of 110

                                        19

-------
I J
c
      !      o    <3
      
-------
this would provide an NEF of approximately 35 and a NNI of 50, an isopsophic index of
92, a Q of 75, total noise load B of 50, a noisiness index of 70, and a weighted equivalent
continuous perceived noise level of 83. It has been assumed  that the number of aircraft
operations per day fall in the range of about 10 to 20.
    The rating sound level or Lr  is similar in  many  respects to the original CNR rating
method, except that L equivalent or A-level is used instead of level rank curves. Corrections
are provided for such things as the impulsive nature of the sound, the duration of the sound,
and whether  or not a whine or tone is present. In addition, although corrections are not
applied directly to the Lj., other corrections  are employed to the basic criterion associated
with Lj. to include such effects such as the time of day, the background noise and previous
exposure to the background noise.  If octave-band levels are employed in this procedure, a
set of noise rating (NR) curves are available which are somewhat similar to the original level
rank curves. This measure is included in the ISO recommendation on noise assessment with
respect to community response.
     Because of the feeling that the variation of noise levels was not adequately accounted
for in  the measures described above, in particular for the normal variations observed in
traffic  noise, another measure was  developed called the traffic noise index, or TNI, which
uses LIQ  and Lgg. These indices represent the levels which on the average were exceeded
10% and 90% of the time. The difference between the values provides an indication of the
variability  of the  sound. Actually,  the TNI  as shown in  Table II is  then equal to 4 times
this difference plus the background noise level  which is represented by Lp0. In using this
measure as a limit the problem exists that for sounds with very small variation a fairly high
permitted background level  would result.
     An improvement over the TNI developed by Robinson of England is the noise pollu-
tion level,  or NPL. The noise pollution level is a little more sophisticated than the traffic
noise index but tries  to accomplish the same sort of thing. In this case it uses the energy
mean of Leq of the sound and to  this is added the standard deviation of the noise (noise
level, not  the noise energy) multiplied by some constant. Typically the formula is as shown
in Table II. An approximation of this is provided by formulas utilizing LJQ and Lgg; thus,
the noise  pollution level is equal to  Leq plus the  quantity  L\Q  - L^Q.  Still  another
approximation is shown in  Table II.  The latter approximations  are  only  valid if the
distribution of noise levels is reasonably normal.

MEASURES RELATED TO HEARING LEVEL

     Table III shows some abbreviations that are employed in  research concerning hearing.
The first is hearing level (HL) in dB, sometimes referred to as "hearing loss". Essentially it is
the level of an individual's hearing relative to a standardized hearing level determined for
young  adult ears. It is the measure of hearing threshold. Thus a hearing level of 40 dB would
mean that the person's sensitivity to sound is 40 dB less than the standard or average level.
A hearing level of - 5 dB would mean that the person had hearing of 5  dB better than the
average young adult ear. Hearing levels are established at various frequencies usually,  starting
at 125  Hz or 250 Hz and proceeding in octaves and half-octaves up to 8000 Hz. Hearing level
sometimes  refers to the  average  of the levels at various frequencies. For example,  hearing

                                         21

-------
                                             TABLE II
                               SUMMARY OF COMMUNITY MEASURES
         MEASURE
ABBREVIATION
                   PROCEDURE
Community Noise Rating
    CNR
Noise Exposure Forecast
    NEF
Community Noise Exposure Level
    CNEL
Equivalent Sound Level



Isopsophic Index
    N
Mean Annoyance Level
Originally determined from level rank curves plus corrections.

Presently,

CNR = PNLmax + C

where:   PNLmax is maximum perceived noise level
        C       is sum of corrections for time of day,
                frequency of flights, and season of
                the year

                     EPNL
                         n
NEF = 10 log [£ antilog (	) + 16.67 £ antilog
                       10
                     Daytime
                     Events


                                    EPNL
                                          n
                                    (	)]-88
                                       10
                                   Nighttime
                                    Events

            where:  n is the event number
                          HNL

              £ w • antilog (-TT—)
CNEL =10 log {      24           ]

                         SENEL
                                n
              £ w • antilog (—rr—)
CNEL =10 log [   864000      U    ]

where:


w is the tone of day weighting factor (1,3,10)
h  is the number of hours (0-2 3)
n  is the number of events
                         10 log   Q/  antilog AL(t)dt
                   "night =
                                                                  6 Iog10 (3ni + n2 - 1) - 30
                                                   24 hours - 10 1Q8 t«tiog ef) + antilog
                   Q> 13.3 log [
                                                             n       AL_
                                                             £ antilog 13.3

                                                             M
                                                                   T
                                                 22

-------
                                            TABLE II
                          SUMMARY OF COMMUNITY MEASURES (Continued)
         MEASURE
ABBREVIATION
                   PROCEDURE
Noise and Number Index
Noisiness Index
Total Noise Load
Weighted Equivalent
Continuous Perceived
Noise Level
 Rating Sound Level
 Traffic Noise Index
 Noise Pollution Level
    NNI
    NI
                                 WECPNL
    TNI
                                  NPL
NNI = PNLmax+ 15 log N - 80
_
NI = 10 log 2 antilog
                                                                        101og(t/t0)-f-C
                                                                           10
                                        AL
                   B = 20 log [S w • antilog ^-] - C
                                                                    EPNLi  .„.  To
                   TNEL = 10 log £ antilog ——+ 10 log -—
                                          10        °
                                                 ECPNL = TNEL-10 log f-
                                                                5       ECPNL (D)2
                                                 WECPNL = 10 log [5 antilog
                                                                O
                                              10
                                                 •,       ECPNL (N), + 10
                                                 j- antilog	ft	]  + S
 Lj = L A* Impulse Noise Correction

 L,. = Leq + Fluctuating Noise Conection


 TNI = 4(L10-L90) + L90-30


 NPL = Leq + 2.56 a
                                                                         (L  0-L90)2  Assuming Normal
                                                     • L50 + 
-------
levels associated with speech are sometimes the average of hearing levels at 500,  1000 and
2000 Hz.
     After exposure to some levels of noise, hearing levels change. This change is referred to
as threshold shift. It is the difference in the hearing level before and after exposure to some
stimulus. The abbreviation TTS or temporary threshold  shift is the amount of shift that
occurs at a given time after exposure. For example, typically TTS2 refers  to the threshold
shift occuring 2  minutes after cessation of the noise. If no recovery exists after  exposure,
the  threshold shift is considered permanent and is dubbed permanent threshold shift PTS.
Sometimes this threshold shift is further described as NIPTS or noise induced permanent
threshold shift.
                                      Table III

              SUMMARY OF MEASURES RELATED TO HEARING EVALUATION

 Hearing Level                                                               HL

 Hearing Threshold Level                                                      HTL

 Temporary Threshold Shift                                                   ITSf*

 Permanent Threshold Shift                                                    PTS

 Noise Induced Permanent Threshold Shift                                      NIPTS


 *Subscript "t" refers to time after exposure threshold was determined.


                              ACKNOWLEDGEMENT

    The material contained in this paper is primarily taken from the "Handbook of Noise
Ratings" by  K.S. Pearsons and R. L. Bennett. This Handbook was prepared under contract
for the National Aeronautics and Space Administration, Langley Research Center.
                                        24

-------
               THE EFFECTS OF NOISE ON THE HEARING OF SPEECH

                                   John C. Webster
                          Naval Electronics Laboratory Center
                             San Diego, California  92152


     To cover this subject matter, I will talk about  the Articulation Index or AI, and the
Speech Interference Level, or SIL, and in particular the relation between them. I will show
that  the best octaves to choose in calculating the SIL depend on what Articulation Index
(AI) you want  to work at or design for. And to make this meaningful, I will have to show
you  what scores you can expect to get on syllable, word, or sentence tests at various
Articulation Indices.  Beyond this, I will discuss what sort of  tests  can be  used to  test
systems or listeners  operating at high-AIs, that is, in relatively quiet environments. If this
seems off the subject, I will relate these types of tests to methods of evaluating hearing aids
and/or different kinds of hearing losses including noise-induced hearing losses.
     To talk of these things intelligently, I will have to spend a little bit of time discussing
the pros and cons of efficient intelligibility tests.  At the first of these conferences (Webster,
1969) I traced the early history of intelligibility testing. I will not repeat it here, but I would
like  to stress a  single distinction made by the early Bell Telephone Laboratory investigators,
namely, articulation  testing as opposed to intelligibility testing. Articulation testing involves
the use of nonsense syllables to determine what single speech sounds, phonemes, distinctive
features, or consonants are misheard.  Once any aspect of redundancy or language enters
the  testing it is no longer articulation  but  intelligibility  that is being tested. Articulation
testing centers  on speech sounds per se. Intelligibility testing involves  both the ear and the
brain or involves both speech sounds and language.
     To summarize very briefly the problems associated with speech testing, I must mention
that the construction of speech intelligibility tests varies along two dimensions—the redun-
dancy and/or vocabulary size of the input stimulus (language) and the constraints or number
of possible  choices in  the output or response.  Within vocabularies of the same size the
relative familiarity  of the word and the number of syllables in  a  word and the context
within which it is imbedded influence  its intelligibility. The constraints on the response,
open vs. closed sets,  also affect intelligibility scores. Closed set  or Modified Rhyme Tests
(MRT) (House  et al., 1965; Clarke, 1965; Kreul et al.,  1968) and pseudo-closed set rhyme
word tests (Fairbanks, 1958) are largely replacing the open-set Phonetically Balanced (PB)
(Egan  1948) and Spondee tests and other  multiple choice tests  at the present time.  The
major reason is the time and effort required to train both talkers and  particularly listeners in
the open-set PB-type word test.
     This  is not the document to trace out in any more detail the history, the rationale, the
strengths  and weaknesses, nor  the actual listings of syllables, words, phrases, or sentences
used this  century to evaluate the effects of noise on  speakers, listeners, communication
components and systems, etc. Recently, however, Webster (1972) has  compiled 24 lists of
word,  phrase,  and  sentence  tests in English. A  very  good reference  for more details on
intelligibility tests is Clarke, Nixon, and Stuntz (1965)  because it has abstracts of over 160
earlier references.

                                          25

-------
     Table 1  shows the relationship between intelligibility scores and Articulation Index (a
special form of speech-to-noise ratio) of many standard speech tests. The generalizations to
be made from Table 1 are that the smaller the stimulus vocabulary and/or the size of the
response set,  or the more redundant in terms of context, the higher the score for a given AI
or speech-to-noise ratio.


                                       Table 1

 Expected Word  or Sentence  Scores  for  Various  Articulation Indices (AI)
A!
0.2
0.3
0.35
0.40
0.50
0.60
0.80
PB*
22
41
50
62
77
85
92
MRT**
54
72
78
86
91
94
98
SENT*
77
92
95
96
98
98
99
 *From Kryter and Whitman (1963)

 **From Webster  and Allen (1972)
     So far I have mentioned only the printed stimulus and response variables that affect
intelligibility testing. The talkers, listeners, and the noise environment around them  have
very large effects on test validity and reliability. For example, Dreher and O'Neill (1957)
had  15 naive speakers read in 5 different noise levels. When the words and sentences  were
played to listeners at a constant speech-to-noise differential the speech originally recorded in
noise was the more intelligible. Pickett (1956) shows, however, that if vocal effort measured
one meter in front of the lips exceeds 78 dB, intelligibility drops.
     It should  be apparent by now that intelligibility  test results require some interpre-
tation. It is neither  simple nor straightforward to assess the affects of noise on speech using
word testing methods. It  would be advantageous to specify the effects of noise on speech in
terms of the spectra and level of the noise and of the speech. Two  such physical schemes

                                        26

-------
exist-the Articulation Index, AI, and the Speech Interference Level, SIL. The Articulation
Index or AI assumes that there are 20 bands in the speech spectra between 200 and 6100 Hz
that differ in bandwidth such that each band contributes 1/20 of the total articulation. Each
band  contributes linearly  to the extent that the speech peak level exceeds the  RMS noise
level by from 0 to 30 dB. The AI is a specialized method of specifying the speech-to-noise
ratio.  It is a non-dimensional numeric that varies from zero to one, but it can be considered
to be a decibel scale ranging from zero to 30 such that for example an AI of 0.5 corresponds
to a complex signal-to-noise ratio of 15 dB, 0.8 to 24 dB, etc.
     The AI was introduced by French  and Steinberg (1947), generalized and simplified by
Beranek (1947a), and refined and validated by Kryter (1962a, 1962b). The AI was discussed
at the first Congress of Noise as a Public Health Hazard by Webster (1969) and by Flanagan
and Levitt (1969) in sufficient detail that it will not be belabored further here.
     Almost simultaneously with the introduction of the AI, Beranek (1947b)  proposed a
simplified substitute for it, the Speech Interference Level (SIL) of noise. The definition of
SIL is the arithmetic  average of the decibel levels in  three or four selected octaves. The
choice of octaves will be discussed later. The SIL is only a measure of noise, and to interpret
it in terms of permissible distances between talker(s) and listener(s), reference must be made
to a  table (Beranek (1947b)) or a graph (Botsford (1969), Webster  (1969)).  An updating,
Mark II, of the Webster (1969) graph is shown as Figure 1, It differs from Mark I unveiled at
the first of these  conferences by (1) adding  two new physical measures,  the four octave
PSIL (.5/1/2/4) and  the proposed  SI-60  weighting which will be discussed in more detail
later; (2) appending an AI scale to help orient people in the real meaning of the figure; and
(3) a droopoff in the communicating voice level curve to reflect the fact that at voice levels
above 78 dB intelligibility does  not increase as fast with vocal effort as at lesser levels. The
gist of the figure  is that  for an AI of 0.5 using "normal" vocal effort (65 dB  at 1 meter)
conversation at  16 feet or 5 meters can take place in noises as high as 50 dB as measured on
the A-weighting network of a sound level meter.
     The one aspect of AI that has been alluded to by many (see Webster (1965)) but not
fully  appreciated is that as  the AI and its correlate, word intelligibility, increase, the most
important speech  frequencies  and/or the  frequency range of noise  that masks the speech
most  effectively increases from between 800 and  1000 Hz to between 1700 and 1900 Hz.
This  of course  should be reflected in  the octaves chosen to calculate the SIL, and this
relationship will  be developed in  the next four figures.
     Figure 2  shows a method  of calculating the AI by  counting the proportion of dots
between the noise spectra and the upper limit of the conversational level speech spectrum.
The example shows how it can  be used to specify the AI for a -6 dB per octave (-3 when
measured in octaves) noise.
     This figure was developed from the Cavanaugh et al., (1962) procedure of deriving AI's
from dot patterns spaced in a 30-dB range in the shape of the normal male speech spectrum.
The concentration of dots reflects the relative importance of different frequency bands to
the intelligibility of speech heard in noise. Figures 3, 4, and 5 show the results of calculating
AI's at 0.2, 0.5, and 0.8 for 5 theoretical noises, and show why and how the octaves chosen
for SILs should vary accordingly.
     Note from  Figure 3 that the spectra lines cross each other (with about a 2 dB spread)
at  1000 Hz. Since these  are all well-behaved, theoretical noises with constant slopes, the

                                         27

-------
        SI-60
     SIL(LLB)
    32


    16
LU
u_
5   8
LU
:/
       CO
            2

            1
                  35
                  30
                       45
F       X   X'X
      APPROXAI FOR NORMAL VOICE
       \      \  X \
 0.1    0.8     0.6  X 0,4
     A  LEVEL 30
     SIL  (4) 21
        PSIL 23
                                                                                    NOISE-DISTANCE AREA
                                                                                       WHERE UNAIDED
                                                                                        FACE-TO-FACE
                                                                                      COMMUNICATIONS
                                                                                      AREA IMPOSSIBLE
                NOISE-DISTANCE AREA
                WHERE FACE-TO-FACE
                 COMMUNICATIONS IN
                  NORMAL VOICE IS
                   POSSIBLE AI >0.5
                   31
                   33
                        50
                        41
                        43
60
51
53
70
61
63
80
71
73
90
81
83
100
 91
 93
110
101
103
             Figure 1. Necessary voice levels as limited by ambient noise for selected distances between talker and
             listener for satisfactory face-to-face communication. Along the abscissa are various measures of noise, along
             the ordinate distance, and the parameters are voice level. At levels above 50 dB(A) people raise
             level as shown by the "expected" line if communications are not vital or by the "communicating"
             communications are vital. Below and to the left of the "normal" voice  line communications are at  an Al
             level of 0.5. 98% sentence intelligibility. At a shout, communications are possible except above and to the
             right of the "impossible" area line.

-------
                          OCTAVE  PASS BANDS  IN  CYCLES PER SECOND
                    90  —   ISO  —   55S  —   TIC  —  1400  —  2tOO  —  S600  —  11200
    20
                     100                         1000
                       FREQUENCY  IN CYCLES  PER  SECOND
10000
Figure 2. Al speech region for "conversational level" speech. The number of dots in each band signifies the
relative contribution of speech in that band to the Al. A series of idealized thermal noises with -6 dB/oct
spectra are drawn in 5 dB steps. The number of dots above each noise contour is proportional to the Al of
conversational level speech in that level of noise. (After Cavanaugh  et al., 1962.
                                           29

-------
     II 0
                          OCTAVE PASS BANDS IN CYCLES PER SECOND
                     90  —   IBO  —   355  —  TlO   —  HOO —  Z«00 —  S«00 —  llfcOO
                                 250      500      1000     2000     4000
     40
                      100                        1000
                        FREQUENCY IN  CYCLES  PER SECOND
10000
Figure 3. Allowable octave band sound pressure levels of steady state noises with spectrum slopes of -12, -9,
-6, flat, and +6 dB per octave for an Al of 0.2 and conversational level speech. The superimposed St-70
contour is a proposed frequency weighting network for evaluating the speech interfering aspects of noise at
Al = 0.2.
                                         30

-------
crossing point at 1000 Hz is also the S1L for the octaves centered at 500, 1000, and 2000
Hz (.5/1/2 SIL). Note also that the spectra cross, a hypothetical line at 1414 Hz (.5/1/2/4
SIL) with a spread of about 9 dB and that the spread at 2000 Hz (1/2/4 SIL) is about 18
dB. Obviously, the  .5/1/2 SIL is the measure with the least variability for specifying the
level of diverse-spectrum noises at an AI  of 0.2,  which corresponds roughly to Fairbanks
(1958) Rhyme Test (FRT) and Modified Rhyme Test (MRT) score of just over 50%, a 1000
word phonetically Balanced  (PB) word score of just under 25%, and a sentence score of just
over 75%.
     Interpreting Figure 4 in the same way, it is evident that (1) at 1000 Hz (.5/1/2 SIL) the
spread  is about 10 dB; (2) at 1414 Hz (.5/1/2/4 SIL) the spread is minimal, about 2 dB; and
(3) at 2000 Hz (1/2/4 SIL)  the spread is about 8  dB. It is equally apparent therefore that
the .5/1/2/4  SIL shows the least variability in specifying an AI of 0.5 which corresponds to
a PB score just over 75%,  an MRT (and FRT) score of about 90%, and  a  near perfect
sentence score.
     Figure 5 shows the  1/2/4 SIL to be the  least variable in specifying an AI of 0.8 which
results in near-perfect socres  on all word and sentence testing materials.
     It should now be apparent that the  choice of octaves in calculating SIL is directly
related to the intelligibility required of the system to be evaluated or to the AI expected of
the system. But just to summarize it once more let us look at Figure 6.
     Note for example that  the slope of AI versus SIL decreases with decreasing SIL levels
as the spectral slope changes from -12 to -6, to 0, to +6 dB per octave. It therefore follows
and it is evident from Figure 6 that when  these 4 theoretical noises are equated in level to
give an approximately equal AI of 0.2, as measured by the .5/1/2 SIL, they are not equal at
AIs of 0.5 or 0.8. When equated by the .5/1/2/4 SIL, the noises are generally equivalent in
level at an AI of 0.5 but  not at 0.2 nor 0.8. Finally, if equated by the 1/2/4 SIL they are
generally equivalent at an AI of 0.8 and not at 0.5 nor 0.2.
     Interpolation  shows that a 50% PB  score (AI = 0.35)  could  be about  equally well
specified, over a large diverse  sample of noises by  an .5/1/2 or a .5/1/2/4 SIL. An AI of
0.35, MRT (FRT) score  of  80 and sentence score  of 95% has been recommended as the
minimum acceptable specification  for  certain military  communication equipments (see
Webster and  Allen, 1972) operating in highly adverse environments. Even lower levels for
acceptance have been suggested for use  in  the past (see Webster, 1965) and  thus lend
credence to using the .5/1/2 SIL for measuring the effects of Navy noises. Architects and
others  working in  quieter  environments  and requiring higher levels of communication
efficiency naturally  prefer AIs of 0.5 for which the .5/1/2/4  SIL  is the least variable
measure. Only the perfectionist would need to design or operate at  AI levels of 0.8 and so
there is probably no serious reason for considering the 1/2/4 SIL for practical engineers.
     Probably the best validating data concerning  the  change of SIL frequency with AI are
those of Cluff (1969). Guff equated the spectra and levels of 112 industrial noises to give
one-third octave AIs of 0.1, 0.2—0.9, and then  determined the bandwidth that gave the
best prediction (least standard diviatibn) over all noises for (1) an average level in one third
octave  bands—similar to an SIL—(2) an overall or band level—similar to a Gweighted
(but band-limited) sound level meter reading—as well as broadband measures of (3) the
A-weighting,  and (4) the  proposed SI-70 weighting. He found as the AI increased from 0.1
to 0.9  the center frequency  of the optimum bandwidths  increased  from 848 to 2264

                                         31

-------
                        OCTAVE  PASS BANDS IN CYCLES PER SECOND
                   *0  —   ISO  —  355  —   710  —  1400  —   ?§00 —  5»00 —   11200
   40
                    100                        1000
                      FREQUENCY IN CYCLES PER  SECOND
10000
Figure 4. Allowable octave band sound pressure levels of steady state noises with spectrum slopes of -12, -9,
-6, flat, and +6 dB per octave for an Al of 0.5 and conversational level speech. The superimposed SI-60
contour is a proposed frequency weighting network for evaluating the speech interfering apsects of noise at
Al = 0.5.
                                        32

-------
CD
O
cr
o
CM
O
o
o
UJ
or

CD
T3
UJ
>
UJ
DO

UJ

5
r-
O
O
60
 50
40
       30
       20
 10
                            OCTAVE PASS  BANDS  IN CYCLES PER SECOND

                  —   90   —   ISO  —  555  —   TlO  —   1400 —  2800 — 5600 —  IIJOO
                                                                   \
                                                                       \
                                                                \
                                                                   \
                                                                             \
-9


-12
                   -.'
                           l?5      250
                                           500
                                                    1000     2000     4000    8000
                1

                5
                        IOO                        IOOO

                          FREQUENCY  IN CYCLES PER SECOND
                                                                        IOOOO
 Figure 5. Allowable octave band sound pressure levels of steady state noises with spectrum slopes of -12, -9,
 -6, flat,  and +6 dB per octave for an Al of 0.8 and conversational level speech. The superimposed SI-50
 contour is a proposed frequency weighting network for evaluating the speech interfering aspects of noise at
 Al = 0.8.
                                          33

-------
5
3
        70
                   RELATION BETWEEN  ARTICULATION  INDEX  (Al) AND
                            SPEECH  INTERFERENCE LEVEL (SID
               60         50         40          50         40         50
                            40
1,0

0,8

0,6

0.4

0,2

0.0
       I
re Al ALA
CAVANAUGH ETAL
T
 -12
                                             -!*•
                                           (PSIL)
4 BAMHSIL)
 ,5/1/2/4
                  T
        70
               60         70         60          70         60         50
                PREFERRED FREQUENCY SPEECH INTERFERENCE LEVEL IN oB
Figure 6. Relation between Articulation Index (Al) and Speech Interference Level (SIL) for 4 noises with
spectrum level slopes of -12, -6, flat, and +6 dB/octave. Three different sets of octaves are shown for
calculating  SIL; from  left to right: 500, 1000, and 2000 Hz (.5/1/2); 500,  1000, 2000, and 4000 Hz
(.5/1/2/4);  and 1000, 2000, and 4000 Hz (1/2/4). The overall level of each noise is adjusted to obtain Al
levels of 0.2, 0.5, and 0.8, and then the SIL is calculated for each of 3 sets of octaves. The data points at an
Al of 0.18 are actual experimental points (50% Fairbanks Rhyme  Scores) from Klumpp and Webster
(1963), i.e., these are the SILs for noises No. 1, 4,10, and 15 in their study.
Hz—average—or 709 to 2530—overall. For the average measure (SIL-type) the center
frequencies and bandwidths were 1135 Hz (3.33  octaves) for an Al of 0.2,  1421 Hz (433
octaves) at 0.5 Al, and  1797 Hz (3.33 octaves)  for  0.8. These values compare very well
indeed to those proposed in this paper of 1000, averaged over the three octaves 500, 1000,
and 2000  Hz; 1428, averaged over the four octaves 500,  1000, 2000, and 40000 Hz; and
2000, averaged over the three  octaves  1000,  2000, and 4000 Hz. Cluff also found  the
SIL-type measure gave standard diviations varying from 0.3  to 0,9 (ave.-0.54)  while  the
standard deviation of the A-weighted levels varied from 1.6 to 3.8 with an average of 2.20.*
     *CIuff,  G. L. (1969), "A  comparison of selected methods of determining speech interference
calculated by the Articulation Index," J. Auditory Res. 9,81-88.
                                           34

-------
     The previous analysis has shown that the octaves chosen to calculate an S1L vary
according to what AI the SIL is trying to estimate. Webster (1964a,  1964b) constructed a
set of contours (see Figure  7)  for predicting AIs or SILs that also showed the increasing
importance of the high speed frequencies for increasing levels of intelligibility (and AI). It is
suggested that weighting networks for sound level meters could be built to predict AI levels
of 0.2 (SI = 70 dB); 0.5 (SI = 60 dB); and 0.8 (SI = 50 dB). A good set of noises on which to
test these hypotheses are the 16 noises of Klumpp and Webster (1963).
     Calculations made on Klumpp  and Webster's 16 noises equated  in level at AIs of 0.2,
0.5, and 0.8, comparing 4 sound level weighting networks A, SI-70, Sl-60, and SI-50, and 3
ways of calculating SIL, namely using the 3 octaves 500/1000/2000, the 4 octaves from 500
to 4000 and the 3 octaves from 1000 to 4000 are shown in Table 2. The results generally
confirm everything  that has just been stated, namely, that at a level of intelligibility corre-
sponding to (1) 0.2, the Sl-70 and the 500 to 2000 SIL are the best (lowest a and R) (2) 0.5
and 0.8, the SI-60 and the 500 to 4000 SIL are the best, and (3) 0.8, the SI-50 and the 1000
to 4000  SIL are good. A-weighting  appears slightly inferior to the  proposed SI-60 and any
SIL that included 500 Hz.
     If  the  manufacturers  of  sound level meters are  seriously considering weighting net-
works other than A, B, and C, an SI-60 should be considered. It is appreciably better than A
for predicting speech intelligibility at all AI levels.
     I have shown how the choice of frequencies for SILs or weighting networks is depend-
ent on the level of intelligibility to  be specified. Now we get back to intelligibility testing.
What tests should be used for various levels of AI?
     Efficiency factors in test  design dictate that the functional relationship between the
dependent and independent variable should be steep and linear in the critical testing region.
Therefore,  consideration should  be given  to  using  different  language tests for different
communication effectiveness areas.  For example, for marginal  conditions, AI = 0.2, closed
set rhyme words  (Fairbanks, 1958;  House et al., 1965; Kreul et al.,  1968; Griffiths, 1967;
Clarke,  1965), which  yield scores  of about 50%, would  make very  efficient tests. If a
listening situation—room or communications equipment—required adequate  intelligibility,
i.e., an AI of 0.35,  then open-set, 1000-word PB tests would yield scores close to 50% and
therefore be efficient  in test  design, although inefficient  in  terms  of crew  training, test
scoring, etc. The use of closed response-set rhyme words would be on the border line of
acceptability since the expected scores would be around 75%.
     At AI levels around 0.8, no intelligibility  test is inherently difficult enough to be an
efficient  test. Even  1,000 nonsense syllables have an intelligibility of greater than 90% at AI
levels of 0.8. To discriminate between listening conditions—communication systems, com-
ponents, etc.—at AI levels of 0.8 requires something more than a simple  intelligibility test.
Reaction times, quality judgments, scores on secondary  tests, or interference tasks, such as
competing messages, have been used or suggested. We will have time to discuss only one of
these promising approaches, namely the competing message paradigm. Tillman, Carhart, and
Olsen (1970) show  the decrement in performance on a competing message task due merely
to adding the equivalent of a hearing aid between the sound field and the listener's ears. The
listener's task was to recognize  in turn one  of 50 phonetically  balanced (PB) words from a
loudspeaker in one corner of a  room while  competing sentences at levels 6  or 18 dB down
were coming from  a loudspeaker in the  other corner  ahead of  the  listener,  i.e., the 2

                                         35

-------
     no
                           OCTAVE PASS  BANDS (N CYCLES PER SECOND
             4ft   —   tO  —   ICO  —   996  —  710   —  1400 —  MOO — MOO  —  lltOO
     40
                      100                        1000
                        FREQUENCY IN CYCLES PER SECOND
10000
Figure 7. Noise rating contours for estimating SILs based on different averaging octaves. Use SI-70 for
estimating the .5/1/2 SIL. = Al of 0.2; the SI-60 for the .5/1/2/4 SIL, = Al of 0.5; and the SI-50 for the
1/2/4 SIL, = Al of 0.8. The inverse of these contours could be used as frequency weighting networks in
sound  level meters to measure the speech interfering aspects of noises. The SI-60 is the best compromise
contour.
                                         36

-------
                                                                   Table 2
        Cells  on the diagonal show  the  Mean (M), Standaid Deviation (a), and Range (R) for  the designated paiameteis for AI levels of
        (from left to right) 0.2, 0.5, and 0.8. Cells above the diagonal show the physical statistics for the  16 Klumpp and Webster (1963) noises
        (AI is not a factor). The numbers in the cells (above the diagonal) represent the M, a, and R of the difference distribution between the
        parameters listed at the left and across the top.
                                     Weighting Networks
                A-Wtng.
SI-70
SI-60
SI-50
   3L
.5/1/2
  SILs

    4
.5/1/2/4
A-Wtng
SI-70
SI-60
Sl-50

3L
.5/1/2
.5/1/2/4
3H
1/2/4
 3H
1/2/4
83.1 72.3 62.6
3.7 2.8 3.2
11.3 8.3 9.7

5.6
2.6
10.0
77.4 66.6 56.9
2.4 3.1 4.0
9.0 9.0 11.5

6.6
2.1
10.0
1.0
2.8
9.5
76.5 65.7 55.8
3.0 1.6 2.3
11.0 7.6 10.6

3.2
4.4
14.7
2.4
5.8
18.9
3.3
3.2
9.7
79.7 68.9 59.2
5.6 3.1 2.6 x
16.5 9.8 7.8

9.8
3.3
9.3
4.2
1.9
8.2
3.2
1.9
6.8
6.6
4.6
14.6
73.0 62.3 52.5
1.5 1.6 2.7
6.3 6.4 10.4

10-6
3.2
11.5
5.0
3.5
12.7
4.1
1.5
6.2
7.4
2.7
8.2
0.8
1.7
6.9
72.2 61.9 51.6
3.0 0.7 1.4
9.4 4.5 6.6

11.6
4.8
16.4
6.0
5.8
19.7
5.0
3.2
11.5
8.3
0.9
5.2
1.8
4.2
13.0
1.0
2.2
8.1
71.2 60.4 50.6
5.4 3.2 2.7
17.4 10.9 9.4
                                                                                                           M
                                                                                                           a
                                                                                                           R
                                                                                                           M
                                                                                                           tr
                                                                                                           R
                                                                                                           M
                                                                                                           a
                                                                                                           R
                                                                                                           M
                                                                                                           a
                                                                                                           R
                                                                                                           M
                                                                                                           £7
                                                                                                           R
                                                                                                           M
                                                                                                           a
                                                                                                           R
                                                                                                           M
                                                                                                           a
                                                                                                           R

-------
loudspeakers were 45° to the left and to the right of the listener's nose. Unaided listening
was (1) binaural; (2) monaural direct, in which the speech was on the side of the listening
ear, and the other ear was occluded with a muff; and monaural indirect, in which the speech
was on the side of the occluded ear. Aided listening used an artificial head in the sound field
with two  hearing aids and connections via amplifiers and calibrated attenuators to insert
earphones in the ears of the remote listener. Again three conditions were tested—binaural,
monaural direct, and monaural indirect.
     Four groups  of 12 subjects each were tested  including (1) those classified audio-
logically as normal (average age 22); (2) with moderate hearing losses diagnosed as conduc-
tive (average age 42); (3) sensorineural (average age  51); and (4) presbyacusic (average age
70). The groups will be indicated by N, C, S and P, respectively.
     All listening  was at  a  level 30  dB above  the  threshold for spondee words, 30 dB
Sensation  Level (SL), under each of the 6 conditions. Figure 8 shows the results, which can
be summarized as follows: Compared  to an earlier reference group of 20 normal hearing
subjects, on  the  PB word/sentence competition task (Northwestern University Auditory
Test  2, Carhart et al., 1963) the N and C groups sitting in the sound field (unaided) heard
essentially at reference level; the  S and  P groups required, on average, a  14 dB better
word-to-sentence differential than the N and C groups in the sound  field; the N group
required about the same increase in word-to-sentence  differential when a hearing aid was
interposed between them and the sound field; the C group required an even greater increase
for the aided conditions, about 18 dB more; and the S and P groups, who required a 14 dB
word-to-sentence (W/S) improvement  in the unaided case, required further improvements
which increased as the basic word-to-sentence (W/S) differential increased. Restated, the S
and P groups are worse off than the N and C groups in listening to competing speech signals
30 dB above their speech threshold, whether listening with or without hearing aids.
     These results show both a hearing deficiency penalty and an equipment-imposed
penalty  when listeners are placed in competing message listening conditions. This is bad
news for people  incurring noise-induced hearing losses which are generally sensorineural in
nature. No only do  they have more difficulty than their normal-hearing or conductively-
deafened friends in cocktail  party environments, but they cannot look forward to a hearing
aid to help equalize their relative disadvantage.
     The last point  I want  to make  concerns listening to  speech in noise while wearing
earplugs or muffs. It  has long been established that in noise levels greater than 90 dB, speech
is heard better when wearing hearing protection. This early work of Kryter (1946) was for
young normal hearing subjects. However, there  is at least one study by Frohlich (1970)
which shows that unlike young normal hearing males,  senior aviators with high-frequency
sensorineural losses do not  discriminate digits better in noise levels above 100 dB when
wearing good noise-attenuating ear muffs. He shows that this could be expected by plotting
hearing-level  and hearing-level-under-muff for senior aviators on the speech area and noise
masking area. This procedure shows that the muff cuts out a region of speech  frequencies
where the speech is well above the masking noise. It seems safe to say that acoustic-trauma
listeners have more difficulty than  normals in discriminating speech in quiet, in noise, and
particularly  in competing  message  situations. They do  not get the full benefit  enjoyed by
normal listeners of increased intelligibility in high noises by wearing hearing protectors, and
they cannot expect a hearing aid to help them untangle competing messages.

                                         38

-------
          -12   -60     6     12    18    24
QUIET
      EFFECTIVE  WORD-TO-SENTENCE  RATIO  IN  dB
Figure 8. Percentage of 50PB words correct in the presence of competing sentences at various word-to-
sentence differentials. The parameters are: REF-20 reference normal hearing listeners; IM/C = normal or
conductive pathology experimental listeners; P/S = presbyacusic or sensorineural pathology experimental
listeners. Aided refers to listening via hearing-aid circuitry. Unaided refers to listening normally in a sound
treated room. From Tillman. Carhart, and Olsen (1970).
    In summary, I have tried to tell you in this presentation that the octaves chosen to
calculate the SIL and/or the weighting networks that could be built into a sound level meter
to measure the interference of noise with speech vary as a function of what level of speech
communication you desire to design for. Correspondingly, the tests you use to evaluate a
listener or a system vary in the same manner, sentence intelligibility tests being best for a

                                       39

-------
basically  bad system, word or nonsense syllable tests for a good system, and competing
message tests or judgment tests for an excellent system. Persons with noise-induced hearing
loss cannot hear as well as normals when wearing plugs or muffs in moderate to high levels
of noise nor can they by wearing a hearing aid unscramble competing messages ^at a cocktail
party) as well as normals.

                                    References

 1. House,  A.S., C.E. Williams, M.H.L. Hecker, and K.D. Kryter (1965),  "Articulation
    testing methods:  Consonantal differentiation with  a closed-response set,"  J. Acous.
    Soc. Amer. 37, 158-166, also USAF EST-TDR 63^03.

 2.  Clarke, F.R. (1965), "Technique for evaluation of speech systems," Final Report of
    Stanford Research Institute Project 5090 on U. S. Army Electronics Laboratory Con-
    tract DA 28-043 AMC-00227(E), August 1965.

 3.  Kreul, E.J. Nixon, J.C., Kryter, K.D. Bell, D.W. Lang, J.S. and Shubert, E.D. (1968),
    "A Proposed clinical Test of Speech Discrimination," J. Speech and Hearing Res. 11,
    536-552.

 4.  Fairbanks, G. (1958), "Test of phonemic differentiation: The rhyme test," J. Acoust.
    Soc. Amer. 30, 596-600.

 5.  Egan, J.P. (1948), "Articulation testing methods," Laryngoscope 58, 995-991.

 6.  Webster, J.C.  (1972), "Compendium  of speech testing  material and  typical noise
    spectra to be used in evaluating communication equipment," NELC Tech Doc. 191 of
    13 Sept.

 7.  Clarke, F.R., J.C. Nixon,  and S.E. Stuntz, (1965), "Technique for evaluation of speech
    systems," Stanford  Research Institute  Semi-annual Report of SRI Project 5090 for
    U. S.  Army Electronics Laboratory, Contract DA 28-043 AMC-00227(E) AD 462836.

 8.  Dreher, J.J. and J.J. O'Neill (1957), "Effects of ambient noise on speaker intelligibility
    for words and phrases," J. Acoust. Soc. Amer., 29, 1320-1327.

 9.  Pickett, J.M. (1956), "Effects of vocal force on the intelligibility of speech sounds," J.
    Acoust. Soc. Amer., 28, 902-905.

10.  French,  N.E.,  and  J.C.  Steinberg (1947), "Factors governing  the  intelligibility of
    speech sounds," J. Acoust. Soc. Amer.  19, 90-119.

11.  Beranek, L.L.  (1947a), "The design of speech communication systems." Proc. Inst.
    Radio Engrs., 35, 880-890.

                                        40

-------
12.  Kryter, K.D., Methods for the calculation and use of the articulation index. J. Acoust.
    Soc. Amer., 34, 1689-1697 (1962a).

13.  Kryter,  K.D.,  Validation of  the articulation  index.  J. Acoust. Soc. Amer., 34
    1698-1702 (1962b).

14.  Webster, J.C. (1969), "Effects of noise on speech intelligibility," P. 49-73 in National
    Conference on  Noise as  a Public Hazard. Proceedings, 13-14 June 1968, The American
    Speech and Hearing Association (ASHA) Reports 4.

15.  Flanagan, J., and H. Levitt  (1969), "Speech  interference from community noise", p
    167-174 in  National Conference  on  Noise as a Public Health Hazard, Proceedings,
    13-14 June 1968, The American Speech and Hearing Association, ASHA reports 4.

16.  Beranek, L.L. (1947b),  "Airplane quieting II. Specification of acceptable noise levels."
    Trans Amer. Soc. Mech.  Engrs., 69, 97-100.

17.  Botsford, J.H,  (1969), "Using sound  levels to gauge human response to noise," Snd.
    and Vib. 3(10) 16-28.

18.  Webster,  J.C.  (1965),  "Speech communications as limited  by ambient  noise," J.
    Acoust. Soc. Amer., 37, 692-699.

19.  Cavanaugh, W.J., W.R. Farrell, P.W. Hirtle, and  B.C. Walters (1962), "Speech Privacy
    in Buildings", J. Acoust. Soc. Amer. 34,475-492.

20.  Webster,  J.C.  (1946a), "Generalizes  speech  interference contours," J.  Speech and
    Hearing Research, 7, 133-140.

21.  Webster, J.C. (1964b). "Relations between speech-interference contours and idealized
     articulation index contours," J. Acoust. Soc. Amer., 36, 1662-1669.

22.  Klumpp, R.G.  and J.C. Webster (1963), "Physical Measurements of equally speech
     interfering Navy noise," J. Acous. Soc. Amer. 35, 1328-1338.

23.  Griffiths, J.D.  (1967),  "Rhyming minimal contrasts: A  simplified diagnostic articu-
     lation test", J. Acoust. Soc Amer. 42,  236-241.

24.  Tillman, T.W.,  R. Carhart and W.O.  Olson (1970), "Hearing  aid efficiency in a com-
     peting speech situation", J. Speech and Hearing Research 13, 789-811.

25.  Carhart,  R., T.W. Tillman  and L.  Wilber (1963), "A test  for speech discrimination
     composed of CNC monosyllabic words," Perceptual and Motor Skills 16:680.
                                        41

-------
26.  Kryter, K.D. (1946), "Effects of ear protective devices on the intelligibility of speech
    in noise," J. Acoust. Soc. Amer. 18, 413-417.

27.  Frohlich, B. (1970), "The effects of ear defenders on speech perception in military
    transport aircraft," North Atlantic Treaty Organization (NATO) Advisory Group for
    Aerospace Research and Development (AGARD) Advisory Report 19.

28.  Webster, J.C. and C.R. Allen (1972), "Speech intelligibility in Naval Aircraft radios,"
    NELCTR 1830 of 2 Aug.
                                         42

-------
                        RECEPTION OF DISTORTED SPEECH

                         Jerry V. Tobias and F. Michael Irons
                           Aviation Psychology Laboratory
                           FAA Civil Aeromedical Institute
                           Oklahoma City, Oklahoma 73125
     Noise has direct physiological, psychological, and social consequences. It also has indirect
consequences that are associated with public health and that certainly are not limited to
damage to the auditory physiology, to the psyche, or to the community's acceptance of loud
sounds. Consider the effect  of a bit too much noise on an airline pilot's reception  of an
air-traffic-control message: the physical  well-being of hundreds of passengers and of un-
known numbers of people on the ground  can be changed by the inaccurate understanding of
an instruction. A missed warning in a  steel mill can produce frightening-even  deadly-
effects on personnel; the physiological results are not confined to the temporal bone.
     We  know ways to  measure speech  interference; and we know something about the
acoustic  factors that determine how well a  listener will  be able to understand masked
speech. Dr. Webster covered  those things in detail in the previous paper. However, there is
another kind of influence on speech intelligibility that all of us have had experience  with,
but that no one has measured before: people can leam  to manipulate signal-to-noise ratios
mentally as well as acoustically. The procedure involves no poltergeists, no telekinesis, no
meditative  or metaphysical  manipulations. It only requires that the  listener's brain be
adequately  exposed to  the masked signal. This  exposure allows the signal-selection
mechanisms to search out the best methods for processing the speech-plus-noise, and, after a
time, produces  greatly  improved intelligibility.  One of the questions  that has not been
answered before is how much time it takes to learn that new analyzing process. Now there
are experiments that suggest that it takes  less time than you might have believed.
     Here is a practical illustration of what  the phenomenon is. People often whistle or hum
or sing while they work. Many nod their heads or tap their fingers to keep time with their
music. In offices, you can sometimes see three or four people, each tapping out a different
rhythm, oblivious of the tempo being strummed on the next desk. Sometimes, though, in a
noisy work environment, a bizarre variation of this behavior appears: a group of employees
who could  not possibly hear each other's  humming because of nearby loud machinery all
move in time to the same invisible drummer. The first time we saw such a thing, we asked
one of the  workers what he  was waving  at. He said, "It's the music," and we pretended to
understand. Of course, when  we went into a storeroom a little distance from the machinery,
there was music. Whether for morale or for entertainment or for setting a working pace, the
company pumped recorded  music into  the factory. The workers heard it even though a
visitor could not make it out  above the din of the equipment.
     Similar stories can  be picked up from anyone who measures noise. All the anecdotes
lead  to the same conclusion:  the ability to  hear masked signals that are inaudible or unintel-
ligible to the untrained or inexperienced observer can be improved by listening practice. The

                                        43

-------
anecdotal evidence has been overwhelming. The laboratory evidence has been nonexistent.
Now there are  a few experiments that were designed to quantify this  process that we all
know to exist.

Signals.

     Masking and distortion are similar kinds of operations; each  covers up part of the
otherwise-available information with extraneous matter to make the signal "noisy." They
both decrease the intelligibility of speech signals. The effect of an intelligibility-decreasing
distortion can be nearly indistinguishable from that of masking.  For example, in a masked-
speech experiment, Kryter (1946) showed that the measured intelligibility of highly reverb-
erant speech  that'is masked  by enough noise to raise it to a level 60 dB above threshold is
comparable to the intelligibility of non-reverberant speech raised 80 dB. In that study, the
reverberation had a masking effect similar to the effect of an extra 20 dB of noise.
     A learning process permits man to overcome the change that the  noisiness produces.
Practicing listening to the speech without the noise (or without the distortion), however,
seems not  to help intelligibility much. Exposure to the noise alone or to the distortion of
nonsemantic  signals seems not  to help. But practice  listening to the combination of speech
and its intelligibility-destroying noise leads to rapid improvement.  The available data cover
studies of both masked and distorted speech; the results from experiments with one kind of
signal are similar to the results from experiments with the other..
     Subjects were all taught to shadow (Cherry, 1953) while listening  to recorded speech.
In shadowing, the listener immediately repeats every word he hears, even as he is hearing
new  material. Although the  idea may sound difficult, subjects are quite  adept at learning it,
and  intelligibility-test scores measured by  shadowing are similar to scores earned in other
kinds of tests. Indeed, if anything,  shadowing is a particularly  sensitive measuring  tool
(Pierce and Silbiger,  1972). Most subjects reach 95-100% intelligibility  scores on clear,
continuous speech within a  few minutes. Our subjects were trained in shadowing until they
had scored higher than 95% in five successive one-minute intervals.
     The speech used for the speech-learning experiments was not the same as that used to
teach shadowing. The experimental speech is a series of easy-to-understand  120-word pas-
sages, read by a male talker who monitored himself during the recording session in order to
insure a constant speaking level. Later, slight variations in level were made from passage to
passage in order to  insure that  all would be equally  intelligible in  a  simple masking experi-
ment.  Each  passage  is  approximately 50 seconds long, with a 10-second pause  between
passages. A total of  54 such one-minute segments  was available, and  the segments were
spliced together in many randomized orders.
     For some subjects, the passages  were masked  with a wide-band Gaussian noise; for
others, the speech was infinitely peak clipped; for still others, the signal became a pulse train
whose  spacing was determined  by the  line-crossings of the speech  wave; and finally, in one
series of tests,  the speech became a  carrier  that was  amplitude-modulated by a band of
noise.  Subjects selected their own signal levels; for a 1000-Hz  tone adjusted to the same
peak level as the speech, the  sound-pressure  level  was 75  ± 4 dB, which was near the
optimum  choice  according to  preliminary tests of the relation between level and intelligi-
bility.  Figure 1  illustrates the kinds  of distorted signals that were used.  In the  masked

                                         44

-------
                                     SPEECH SIGNALS
                           LJ
                           Q
                           CL
                           2
                                      Clear-Undistorted
                           UJ
                           Q
                           ID
                                    Infinite peak-clipping
                          UJ
                          Q
                          D
                           Q.
                           S
                           <
                           i
                          T
                          u
                          Q
                                       Pulse modulated
                                       Noise modulated
                                         TIME
                           Figure 1.  Waveforms of test signals used.
condition,  the speech wave is  simply added to the  noise. In the modulated condition,
though, a multiplication transform is used, with the effect that each partial in the original
instantaneous speech spectrum is replaced by a steep-skirted band of noise, 1200 Hz wide,
centered on the partial. In the pulse-modulated procedure, all that is retained of the original
waveform is the time and polarity of axis crossings; infinitely peak-clipped signals look to
have only that same information (Licklider and Pollack, 1948), but they are generally much
more intelligible (Ainsworth.  1967), even when experienced listeners adjust  the levels of
                                          45

-------
 both types for maximum intelligibility. Clipped speech sounds harsh; pulsed speech sounds
 harsher.  Noise-modulated  speech sounds very noisy, but is generally reported to be much
 clearer than one would expect with "that much noise" present.
     The masking level and the modulator bandwidth were  selected  to produce approxi-
 mately .the same maxjmum intelligibility score (80% correct)  for highly trained listeners as
 unmasked pulsed speech does. Clipped speech is a bit easier to understand, and maxima near
 90% are common.

Subjects.

     Each segment of these studies used six university students as listeners. All subjects had
normal hearing, and none had any previous experience  with this kind of task. A total of 13
series of experiments used 78 subjects. Everyone was  trained in shadowing before being
exposed  to  the distorted or masked signals. Most subjects were then  simply instructed to
shadow whatever they could hear. Several groups, though, received special treatment; some
shadowed for a total of only eight minutes in a 54-passage session; another few shadowed
everything, but were informed that they'would be given  a monetary incentive to do well.

Speech Learning.

     The basic outcome of all these experiments is perfectly predictable: intelligibility starts
at a low level and improves with listening practice up to a plateau value. Figure 2 shows a
learning curve for each of the four kinds of signal. The rates of change are fairly similar from
one condition to another, although ;the  plateau values vary  somewhat. The  immediately
apparent pojnt to note about all of these data is that learning seems to be complete within
 15 or 20 minutes. The auditory  system  makes its  analysis of the signal-plus-noise, deter-
mines how to extract  the maximum information,  and makes whatever modifications are
necessary in order to perform the extraction—and it does all that  in less than half an hour.
The listeners are  probably not especially conscious  of what they are doing in order to get
this analytical processing under way; most of them report no special effort to get better, and
generally they have little recollection of how well they performed.
     Although the curves are  similar in shape, it is inappropriate to try to draw inferences
and  conclusions  about the speech-learning mechanism from that fact. Learning curves
simply look  alike. That does not necessarily demonstrate anything about similarities or
differences in the  analysis of modulated and pulsed speech signals.
     The curves  that represent  what happens in  this learning mechanism do have one
particularly  fascinating .aspect, though (Figure 3). They show that subjects returning after
one or two weeks away from the task start the first couple of passages with scores slightly
lower than  their  previous maxima, but  then, almost  immediately, they rise  to a higher
plateau than the one they attained during their first test session. The change from the first
to the second plateau is statistically significant at better than the  .05 level; final scores on
session two are  12 to  15 percentage  units  above those  on session  one, making a  total
improvement that is equivalent to about an 8-dB shift in signal-to-noise ratio. During their
time away  from the laboratory, the subjects had no opportunity  to listen to  the kinds of

                                         46

-------
 LJ
 
-------
oo    iz
IUU
90
80
70
*~* 60
o^ 50
*~* 40
S 30
or 20
R 10
o
CO
1 st Session
-
MASKED
^^
- f
I
2nd Session





1 1

-
i i I I I


i i

i \j\j
90
80
70
60
50
40
30
20
10

1st Session
-
PULSED
2nd Session
^*
f
- x^ If
/
J
'-
-
_
1 1 1 1 1





I i

0 10 20 30 40 50 0 10 20 30 0 10 20 30 40 50 0 10 20 3
t
-J 100
QQ 90
O 8O
^•^ w \J
-J 70
UJ 60
H 50
— 40
30
20
10



1 st Session
— M/*\f\iiiiv"^rf\
MODULATED
Y
i
-
-
-
i i i i i

2nd Session
^^^•••••••••••v
~
/"





i i
100
90
no
ou
70
60
50
40
30
20
10



1st Session
CLIPPED
Y
-
-
-
-
i i i i i

2nd Session
X^





i i
            0    10  20  30  40  50   0    10   20   30        0   10  20  30   40  50   0   10  20  30
                                               TIME  IN   MINUTES
        Figure 3. Mean learning curves for second experimental session. Note the improvement following a week or more without practice.

-------
     Whatever the solution to that mystery, though, one thing is clear. The brain, once it
has organized itself to determine the transformations  necessary  for  analyzing difficult
speech messages, continues to refine the analyzing process. These changes and refinements
are the kinds that allow the listener to generalize or transfer the techniques of decoding one
sort of distortion to the interpretation of other sorts.

Transfer.

     In tests of the transfer of speech learning, subjects were trained during the entire first
session with material that had been treated with one variety of distortion. For the first half
hour of the second session, two weeks later, they continued with that same distortion; by
the end of the 30 minutes, it was certain that they had reached their new higher asymptotic
intelligibility score. Then,  for the last half hour of that session, they were given material that
had been subjected to a  different distortion.  For example, subjects who spent all of their
hour and a half of practice time on modulated speech were tested on pulsed speech (Figure
4). Within three  to five minutes,  they reached plateaus that were at least as high as those
reached  by similar subjects  who  had  listened to nothing but pulsed speech during both
sessions. Transfer is equally good in the opposite direction.
     Although pulsed  speech lacks most of the spectral information of the original wave-
form, and modulated speech lacks most of the temporal information, the transfer of ability
from one to the other seems complete. The suggestion is strong that human observers, once
they have learned how to listen to difficult speech, can successfully understand almost any
form of it (for another example, see Beadle,  1970). Perhaps that idea helps to account for
the fact that  some people are able to  understand  English spoken with many kinds of
dialects,  but that others  cannot get much intelligibility out  of what is said to them by
talkers with just moderately variant speech.
     A practical result of  this finding is that a person probably does not need to be trained
to listen to distorted or masked signals that are  of the precise form that will occur in his
work.  Once he has mastered some types of speech learning, he will be able to assimilate
others rapidly. Should we decide  to make a set of recordings to train aviators to listen to
radio transmissions, those records need not  contain precisely the same  kinds of signal
degeneration that actually arise in the  cockpit; anything similar—or maybe even dissimilar-
ought to do as well.

Passive Listening.

     Must such training  involve  continuous  speech-related activity on the part  of the
listener?  To find out, we ran an experiment  on  the learning that accrues to subjects who
hear the  distorted signals,  but who do not have to shadow them (Figure 5). If the shadowing
activity is contributing to the  learning, then levels of performance ought to be higher for
those who are thus employed during their listening. Twelve subjects were used to test this
idea. Before their first exposure to the  distorted speech, they were trained in shadowing
clear speech, just as all the'subjects had been. Then they were asked to shadow the first pah",
the last pair, and two equally spaced intermediate pairs of distorted passages. For the rest of
the time—46 passages—they were  silent. When these passive listeners' responses are  com-

                                         49

-------
    lOOr-
LU
    20
     I 0
                      I st  Session
                       MODULATED
                                    2nd  Session
                           MODULATED
                                                                    PULSED
                                          r
               i
i
_L
I
I
I
I
I
I
        0    10   20   30   40   50  0     10   20   30   40   50

                              TIME    IN    MINUTES

 Figure 4.  Mean teaming curves for subjects who received all of their listening experience on modulated
 speech. When tested on pulsed speech, their scores were almost immediately at the maximum expected for
 subjects experienced with pulsed-speech listening.
pared with the responses of active, continuously shadowing subjects, two differences are
apparent  from the  data and from the  subjects' reports.  First, the passive group's final
first-session scores are consistently higher than the active group's (second-session curves are
alike). Second, active shadowers have almost no retention of the material that they listen to,
but passive listeners remember  a great deal of what they hear. One interpretation of this
finding is that our active listeners are giving us scores that do not represent improvements in
                                      50

-------
   100

o^  90

to  so
LJ
g  70
u
CO  60
    50

    40

    30

    20

    10
OQ
_J
LJ
f-
2
                              1st Session
                        PASSIVE  PARTICIPATION
                        GROUP A AND  B
        2nd  Session
— — GROUP A  ACTIVE

— GROUP B  PASSIVE
                         20       30      40       50        0
                                         TIME  IN   MINUTES
            20
                                                                                        30
40
            Figure 5.  Mean learning curves for passively listening subjects. In the second session, half were asked to
            shadow.

-------
shadowing technique, but rather real improvements in the ability to understand difficult
speech  signals. The higher  passive scores might be interpreted to mean  that shadowing
somehow interferes with speech learning, and we cannot refute that possibility. However, it
is possible too that the  listener who can understand and retain what he listens to is better
motivated to learn. That possibility can be tested.
Motivation.

     Two groups of subjects were tested for motivation effects-one group on masked and
one  on pulsed speech—but, unlike previous listeners, these were given monetary incentives
to do well. After a subject had worked through the first three passages, the three scores were
averaged, and he was told that for each 1% by which the 54th passage was better than that
beginning average, he would be paid a bonus of $.05. Also, in order to keep him working at
a high level during the entire test  session, he  got an  additional $.10 for each passage on
which the score was above 90%.
     The results (Figure 6) are similar to those for the passive listeners: curves continue to
rise for a longer period of time during the first session, and, within the first hour, they reach
values that are comparable to second-session plateaus. This relation between passive-listening
results and motivated4istening results certainly suggests that the passive subject continues to
improve because he is more interested in the  task than the active subject is. He is able to
relax a bit, and he can actually attend to what the talker is saying (remember that his
retention is better than the active subject's).
     Second-session scores for these subjects are indistinguishable from those of any other
subjects. Changes in ultimate peak scores, if they occur as a result of monetary reward, are
not large enough for us to measure with these techniques.


Masked Speech.

     Experiments with masked speech at a -3  dB signal-to-noise ratio show one kind of
quantitative difference from the other experiments: first-session subjects reach  two quite
different kinds of asymptotes, apparently  as a function  of their  earliest scores. Listeners
who do well in the first few minutes are like most listeners; they improve rapidly to plateaus
of 80% or so. But those  who  start with intelligibility scores  of approximately 10% reach
peaks in the neighborhood of only 50 or 55% (Figure 7). The intention  had been to set a
signal-to^ioise ratio that would give a first-minute intelligibility score of 20 or 30%, but the
selection was not right for these subjects; their initial scores actually ranged from 5 to 29%.
The  low-plateau subjects show greater variability  than might be  accounted for by  the rela-
tively unrestricted range in which  they were working. Their learning curves rise  compara-
tively slowly, sometimes taking 35 or 40 minutes to get to the asymptotic value. The curves
are unlike any others we have seen.
     An explanation may lie in an  evaluation  of the learning experience that each group
receives: the people who start off well get exposed to  large numbers of correctly heard
words;  those  who start poorly  receive relatively little  information that helps  them in

                                         52

-------
   100 r
     90  -

-------
     100

o^    90

cn    80
LU
ct:    70
o
   o
   en

   H
   .j
   DO
   C)
   _J
   _J
   LU
      60

      50

      40

      30

      20

      10
                                       I st Session
                       10         20        30         40
                            TIME    IN   MINUTES
                                                                 50
             Figure 7.  Individual teaming curves for subjects who listened to masked speech.
magnify that success into higher scores. The listener who starts low may spend most of his
training time struggling to hear anything at all. His decoder never gets enough samples of the
difficult sounds to permit the formation of useful hypotheses about how to listen.

Retention.

    Most subjects were retested in a third session one month following the second (six
weeks after the first). Third-session curves  and scores are similar to second-session curves
and scores. One group was retested after a year had passed with no known intervening
practice. Their latest performances are similar to their second sessions, too.
                                        54

-------
Non-Normal Hearing.

     Normal-hearing subjects learn to understand badly mangled speech after a short period
of practice. There is no evidence that people with pathological hearing do as well, but it is
certainly possible—even reasonable-to conclude that they do.
     The plateaus of subjects  in their second sessions, no matter what the conditions of
their  training  (except  those who are  able to receive only small  amounts of information
during the first session), and no matter what kinds of signals they were trained with, are all
improved, and all look similar. That fact may  partially explain why hearing aid users are
reported to do much better at speech discrimination after they have used an aid for a week
than they did  when the instrument was first tried. The early listening presents them with a
kind of sound.that is somehow different to them; they have to learn about it before they
can get maximum sense from it. If this learning process works with some kinds of pathologi-
cal ears as well as it does with normals, we might also expect to find that, for these ears,
experience with any hearing aid will transfer to any other.
     The hard-of-hearing person may have a greater problem learning to understand dis-
torted speech  than  the normal-hearing listener, though, for the very reason that he cannot
hear enough of the signal to work out an appropriate analysis strategy. He could be like the
low-plateau  subjects in the masked-speech experiments. However, usually, the overall sound-
pressure level of his  work environment will be high enough to overcome much of the
problem caused by an elevated threshold, so he can learn as well as  his colleagues.  This
likelihood leads to  the interesting possibility  that  the results of some audiometric tests of
the ability to understand  speech that  is immersed in noise may be more a function of
learning than of hearing.


Training.

     Even two minutes of listening can improve the ability to understand a talker (Peters,
1955). Six to  eight  hours may be needed to teach people to understand speech sounds  that
are transformed by a spectral  inversion (Beadle,  1970), and even then, it takes longer to
learn from an  unfamiliar talker. But for optimum training for the reception of non-inverted
speech, about an hour is needed.
     How should the  time  be  spent?  If you want to improve your reception of distorted
speech, it is not enough to  listen to the right  kind of interfering noise.  It probably will not
help to be exposed  to nonspeech sounds that are subjected to the same sorts of distortions
that affect the speech; the analyzing activities of the brain are quite different for speech and
for non-speech signals (Stevens and House, 1972). Student pilots take far longer than half an
hour  of flight time  to learn to  understand air-traffic-control communications; factory
workers do not begin to understand what is said to them in noise until days of listening have
passed, not minutes. Both groups commonly hear speech-plus-noise for only short moments
at a time, and then return  to listening to noise alone. The requirement for rapid learning,
though, is that the listener be able to hear the combination of signal and noise at a signal-to-
noise ratio that is high enough to permit him some success in interpreting the messages  that
are being transmitted.  If his motivation  to learn is high (or heightened), he can reach his

                                         55

-------
maximum  capacity in an hour. Hell probably learn still faster if his training is done by a
talker whose voice is familiar to him.* And when he has found the knack of how to listen,
he will keep it for a long, long time.
     Once, after a static-filled, phase-distorted, narrow-band,  whistling, short-wave broad-
cast of a concert, Sibelius is reputed to have pointed at the radio receiver and said, "I can't
understand how anyone but a musician could enjoy listening to that thing." In the  same
way, one who is not  trained to listen to speech will not enjoy it. Indeed, in many circum-
stances, he may  not even to able to hear it.

*Schubert and Parker (1955) reported in a paper on a speech intelligibility study, "A puzzling phenomenon
occurs with three of the subjects, who were wives of the talker. In each  case the wife exhibits only  a very
slight dip in intelligibility, if any,	when her husband is the talker, but shows about the average dip for
either of the other two speakers. This obviously falls beyond what has previously been considered the
boundary of auditory theory and the authors, who were two of the talkers, are relieved of the risk of discussing
it further." So are Tobias and Irons.
                                      References

Ainsworth,  W.  A.,  Relative intelligibility of different transforms of clipped  speech. /.
     acoust.  Soc.  Amer.,  41 1272-76 (1967).
Beadle,  K.  R., The  effects  of  spectral  inversion  on  the  perception  of  place of
     articulation.  Doctoral dissertation,  Stanford University (1970).
Buxton, C.E.,  The  status  of research  in  reminiscence. Psychol.  Bull,  40, 313-350
     (1943).
Cherry,  E.  C.,  Some  experiments  on the  recognition  of  speech,  with one and  with
     two ears. J.  acoust.  Soc.  Amer.,  25, 975-979 (1953).
Kryter,  K.   D.,  Effects  of  ear  protective  devices on the intelligibility of  speech in
     noise. J. acoust. Soc. Amer.,  18, 413-417 (1946).
Licklider, J. C. R.,  and Pollack,  I., Effects of differentiation,  integration,  and infinite
     peak  clipping   upon  the  intelligibility  of  speech.  /.  acoust.  Soc.  Amer.,  20,
     42-51  (1948).
Peters,  R. W., The effect of length of exposure to  speaker's voice upon listener reception.
     Joint Project Report No. 44, U.S. Naval School of Aviation Medicine (1955).
Pierce,  L.,   and Silbiger, H.  R.,  Use  of  shadowing in  speech  quality  evaluation. J.
     acoust.  Soc.  Amer.,  51, 121  (1972).
Schubert,  E.  D., and  Parker,  C.  D.,  Addition  to  Cherry's findings on  switching
     speech  between the  two  ears. /.  acoust.  Soc.  Amer.,  27,  792-794 (1955).
Stevens,  K.  N.,  and  House, A.  S.,   Speech  perception.  In  Tobias,  J.  V.  (Ed.),
     Foundations of Modem  Auditory  Theory.   Volume II.  New York:  Academic
     Press  (1972).

                                          56

-------
                           HEARING LOSS AND SPEECH
                            INTELLIGIBILITY IN NOISE

                                  Jerzy J. Kuzniarz
                             Oto laryngology Department
                              Silesian Medical Academy
                                   Francuska 20
                              40-027 Katowice, Poland

     While testing patients with sensorineural hearing loss, I have frequently noticed their
complaints about difficulties with understanding speech in the presence of some background
noise. Similar observations were reported  by many authors during the last 10 years (Harris,
1960, 1963; Kryter, 1963; Watson, 1965; Robin, 1967; Tonkin,  1967; Niemeyer, 1967;
Groen, 1969), recently by Carhart et al. (1970), Tillman et al. (1970) and Lipscomb (1972),
so it has become a fairly well established  clinical fact. For simplicity, that symptom will be
called in the following paragraphs a "noise  distractability (ND) phenomenon" in sensori-
neural deafness. The purpose of my presentation is to  show some results of my studies on
that phenomenon in noise-induced hearing loss (NIHL), especially concerning the frequency
area which influences speech intelligibility in noise.
Tests procedure and results.

Experiment I.

     An  NIHL was approximated by  low-pass filtering  (Fig. 1) of speech tests:  mono-
syllables, PB words and sentences (each filtered test recorded separately).
     The intelligibility of each filtered test was examined on 30 normal listeners. The tests
were presented binaurally at an intensity of 65 dB SPL via Pedersen earphones linked with
"Y"-type connection  (no stereophonic effect) with a tape recorder and an audiometer
(Peters SPD 2). The intelligibility was tested both in quiet and in presence of two kinds of
noise: white noise (Fig. 2) and a low-frequency one (Fig. 3).
     The noises were generated by the audiometer and mixed with the speech materials at
different S/N ratios.

Results.

     In quiet, sentences were fully understood when the speech materials consisted of fre-
quencies up to 1000 Hz; monosyllables were understood at a 90% level when the upper
cutoff frequency was 2000 Hz. It is concluded that frequencies up to 2000 Hz are entirely
sufficient for understanding Polish in quiet.
     In background noise, however, the results were quite different. The intelligibility of
filtered speech was markedly reduced in the presence of noise, even at S/N ratios which did
not impair the intelligibility of non-filtered speech (Fig. 4-7). That effect was particularly
evident in the low-frequency noise (Fig. 6 and 7).

                                        57

-------
oo
              0    W5       OJ50       0,50  0.75  W   15   2    3     4     6    6    ^ KHl
           10
           30
          50
                                   Figure 1. Characteristics of low-pass filters used in this study.

-------
MS
                                                                                                 I—1—I	1—
                  0.1
5  6    9   fO  KHZ
                                    Figure 2. Spectrum of white noise, as measured in Pedersen earphones.

-------
Ch
O
                  0.15    a*
0/0
                                    Figure 3. Spectrum of low-frequency noise, measured in Pedersen earphone.

-------
     The difference between the intelligibility of non-filtered speech and that filtered at
3000 Hz was statistically significant for monosyllables in white noise and for both mono-
syllables and sentences in low-frequency noise.
     It can be concluded that frequency bands lying above  2000 Hz and 3000 Hz carry
information which becomes important for speech understanding when lower speech fre-
quencies become masked by low-frequency noise.

Experiment II.

     Thirty selected subjects of the ages 25 - 40 years with NIHL (mean audiograms are
shown  in Fig. 8,10 persons in each group) were tested binaurally with speech audiometry in
a semireverberant room (reverberation time 0.2  sec) in  quiet and in presence of low-fre-
quency noise (Fig. 9).
     Another 30 people, 20-30 years, with normal hearing, were used as a control.
     Arrangement of the apparatus is shown in Fig. 10. Note that, as in Exp. 1, there is no
stereophonic effect. The speech intensity was held constant at  70 dB, monitored with a
VU-meter and controlled with a Bruel and Kjaer Impulse  Sound Level Meter  (mean peak
level).  The noise level was changed in  5-dB steps to obtain S/N ratios from + 15 dB to -15
dB.
Results.

     All persons with NIHL were evidently handicapped even in presence of a mild intensity
of noise (S/N = +10 and +5 dB) that did not disturb normal listeners (Fig. 11). This effect
was seen even in group III, whose members have normal hearing thresholds at 2000 Hz and
good speech intelligibility in quiet. The curves closely resemble those obtained with filtered
speech.
     The results of both experiments may be summarized in three points:
     1.  Persons with NIHL have difficulties in understanding speech in presence of back-
         ground noise (ND phenomenon) even at S/N ratios close to those of everyday
         conditions (S/N = +10 or +5 dB), which do not disturb normal listeners.
     2.  Similar results may be obtained on normal listeners when tested with low-pass
         filtered speech in presence of background noise ("everyday noise").
     3.  The loss of speech intelligibility was  seen  in both  situations even when speech
         frequencies up to 2000 Hz were perceived.
Comment and conclusions.

     The nature of ND phenomenon is not clear. It seems that three factors may play a role
here:
1. Masking  effect: low-frequency  speech spectrum components,  on which patients with
NIHL  mainly depend, are effectively masked  by background noise, especially the  most
common low-frequency one (Niemeyer, 1967).

                                        61

-------
                                           Mynosulob/es PB
               Quiet     +20     +10     +5      0
                                      S/N   (IB
-5     -10
Figure 4. Intelligibility of phonetically-balanced monosyllables, non-filtered and low-pass-filtered (cut-off
frequency is the parameter) in white noise.

                                     62

-------
duiet
                                             Sentences
                                             Unfilled
                                              3000
                                       0     -5     -10
                                5/N     ctB
Figure 5. Intelligibility of sentences, non-filtered and row-pass-filtered (cut-off frequency is the parameter)
in white noise.
                                   63

-------
                                          Monosylables  PB
                                                                -20     -?5
                                    S/V    dB
Figure 6. Intelligibility of monosyllables, non-filtered and  low-pass-filtered (cut-off frequency as the
parameter) in low-frequency noise.
                                      64

-------
                  Sentences
                    Unfiltered
         S/H   dB
Figure 7. As figure 6 for sentences.
          65

-------
•-
              BOA
                                 250
500
1000
2000
4000
6000   Hi
                   Figure 8. Idealized average audiograms of subjects with noise-induced hearing loss, used  in Exp. 2 (10
                   persons in each group).

-------
figure 9. Spectrum of low-frequency industrial noise used in Exp. 2.

-------
0\
ex
              Y/////////. //,
                 Figure  10. Scheme of apparatus:  1-tape recorder with monosyllable test, 2-audiometer Kamplex DA2,

                 3—examiner, 4—control phones, 5—tape recorder with recorded noise, 6—amplifier, 7—loudspeaker unit,

                 8—subject, 9—control microphone, 10—microphone amplifier.

-------
                                         Monosulab/es  PB
                                                     Normal hearing
                                                 Hearing loss:
                                                     > WOO Hz (ff)
                                                     > 500  Hz (I
  §    50.
             Quiet
                                   S/N   dB
Figure 11. Intelligibility of PB monosyllables, tested in quiet and in noise, on subjects with noise-induced
hearing loss and on a group of normal listeners (control).

                                    69

-------
2. Cumulative effect of speech distortion:  reported by Harris (1960): a single distortion,
which by itself does not produce any adverse effect on the intelligibility of speech, may
induce strong degradation of speech when combined with another similar deformation.
3. Inefficiency of organ ofCorti: due to damage induced by noise. It may not be evidenced
by pure tone audiometry, but becomes evident while testing with a complex stimulus such
as speech test, or better, speech-in-noise test (Lipscomb,  1970).
     The most important clinical implication of these findings is that speech tests performed
in quiet do not provide information on ability to understand speech in everyday conditions
as far as patients with sensorineural hearing loss are concerned. This applies also to NIHL, of
course.
     In consequence, the present concept of so-called "most important speech frequencies":
500 - 2000 Hz, based on speech tests performed in quiet  (Report, 1959), should be changed,
as it does not apply to everyday conditions. That has also serious influence on compensation
for NIHL. As we all know very well, under "everyday" conditions, especially at work, there
is nowadays almost no quiet at all. That is why our Congress takes place.
     According to present rules, supported  by the  latest ISO recommendation, (1971) we
are expected to believe that a subject having total hearing loss at all the frequencies over
2000 Hz will have normal ability to understand speech  at work. Is this true? Certainly not.
     I would therefore  propose  the inclusion of the frequencies 3000 Hz and 4000 Hz into
the list of those that are most important for understanding speech in everyday conditions.
There is a lot  of evidence that  these and higher frequencies participate in carrying speech
information (Mullins and Bangs, 1957; Kryter,  1962, 1963; Harris, 1965; Huizing,  1963;
Palva, 1965; Ceypek and Kuzniarz, 1970). Moreover, frequencies  up to 4000 Hz have been
already used in the AMA method for computing hearing loss for speech from the pure-tone
audiogram (after Harris, 1956),  and  frequencies up to 6000 Hz are still used for computing
the Articulation Index (Kryter, 1962).
     That point  of view was accepted by the Ministry of Health in Poland, and so since
1968 the extent of NIHL has been tentatively estimated  in my country on the basis of mean
hearing loss at frequencies  of 1000, 2000 and 4000 Hz, as being the most important for
speech intelligibility in everyday conditions.
     It seems also that for reliable  estimation of a sensorineural listener's performance in
everyday conditions, speech tests in a noise background, similar to those proposed by Kreul
et al. (1968), Groen (1969) or Carhart and Tillman (1979), should  be applied.
Final conclusions.

1. Persons with high-frequency sensorineural hearing loss suffer from disability to under-
stand speech  in  everyday noise, although the noise may not  be  disturbing to normal
listeners.
2. This disability appears even when the hearing threshold up to 2000 Hz is not changed, so
a new  concept of basic speech frequencies for everyday conditions should be developed,
with special attention to the frequency area of up to 4000 Hz.

                                        70

-------
                                    References

Carhart, R., Tillman, T.W.,  Interaction  of competing speech signals with hearing losses.
     Arch. Otolaryngol., 91, 273-289 (1970).
Ceypek, T., Kuz"niarz, J., Znaczenie ograniczonych  pasm  czestotliwosci dla rozumienia
     mowy polskiej. Otolar. Pol., 24,429-433 (1970).
Huizing H.C., Kruisinga, R.J., Taselaar, M., Triplet audiometry: an analysis of band discrimi-
     nation in speech reception. Acta Otolar. (Stockh.), 51, 256-259, § (1960).
Groen, J.J. Social hearing handicap:  its  measurement by speech audiometry in noise. Int.
     Audio., 8,  182, (1969).
Harris, J.D., Combinations of distortions  in speech. Arch. Otolaryngol., 72, 227-232 (I960).
Harris, J.D., Haines,  H.L., Myers, C.K., A new formula for using the audiogram to predict
     speech hearing loss. Arch.  Otolaryngol, 63, 158-176(1956).
ISO - Recommendation R  1999:  Assessment of occupational noise exposure for hearing
     conservation purpose, May, 1971.
Kreul, E.J., Nixon, J.C., Kryter, K.D., Bell, D.W., Lang, J.S., Schubert, E.D., A proposed
     clinical test of speech discrimination. /. Speech Hear. Res., 11, 536-552 (1968).
Kryter, K.D., Methods for the calculations and use of the Articulation  Index. /. Acoust.
     Sac. Am.,  34, 1689-1697  (1962).
Kryter, K.D., Williams, C. Green, D.M. - Auditory acuity and the  perception  of speech. J.
     Acoust. Soc. Amer., 34, 1217-1223  (1962).
Kryter K.D., Hearing impairment for speech. Evaluation  from pure tone audiometry. Arch.
     Otolaryngol., 77, 598-602 (1963).
Lipscomb, D.M., Noise exposure and its effects. Oticongress 2, 1972.
Mullins, C.J.,  Bangs, J.L., Relationships between speech discrimination and other audio-
     metric data. Acta Otolaryngologica (Stockh.), 47, 149-157, (1957).
Niemeyer,  W., Speech discrimination in  noise-induced deafness. International Audiology, 6,
     42^7 (1967).
Palva, A., Filtered speech audiometry. Acta Otolaryngologica (Stockh.), Suppl. 210, (1965).
Report of the  Committee on Conserv. of Hearing: Guide  for the evaluation of hearing
     impairment. Trans. Amer. Acad.  Ophthalm. Otolar., 63, 236-238 (1959).
Robin, J.G., The handicap of deafness./  Laryngol. Otol., 81, 1239-1252, (1967).
Tillman, T.W.,  Carhart,  R., Olsen, W.O., Hearing aid efficiency in a competing speech
     situation. /. Speech Hear.  Res., 13, 789-811, (1970).
Tonkin, J., The diagnosis and  assessment of peripheral sensori-neural deafness. /. Laryngol.
     Otol.,&\,  1187-1239, (1967).
Watson, T.J.,  Speech audiometry  in varied  acoustic conditions. Int. Audiol., 4, 102-104
     (1965).
                                         71

-------
           THE LONG-TERM PLANNING OF A NOISE CONTROL PROGRAM

                                   Michael J. Suess
                               World Health Organization
                                 Copenhagen, Denmark

     Noise, that is to say, an annoying and unwanted sound, has been recognized as a public
health hazard, and endangers both the mental and  physical state of man. Consequently,
major activities have  already been undertaken by  WHO in the mid-sixties to study the
implications of Noise on human health.
     As part of its concern for the improvement of the human environment, the WHO
Regional Office for Europe published a two-year  study on The Environmental  Health
Aspects of Noise  Research and Noise Control.** However, it was the favorable acceptance
of the Office's over-all Long-term Program on Environmental Pollution Control and its
approval by the Regional Committee of the European Region at its 19th session in Budapest
in 1969  that  led to  the detailed planning  and implementation  of a program for Noise
Control (see figure  1). The first  activity within this program was the convention of a
Working Group in The Hague in October 1971.*** The members of that Working Group
reviewed and assessed the noise situation prevailing in Europe and its control, studied future
trends  and  developments, discussed  needed activities of special  importance, and recom-
mended actions and projects to be undertaken. The Working Group stated that

     "Noise must be recognized as a major threat to human well-being."

     "Available knowledge on  the effects of noise and on methods of noise control is not
     being adequately utilized."

     "Progress in noise reduction can be made by setting specific noise limits. While such
     limits  must  necessarily take  technical  and  financial  constraints into account, most
     existing limits cannot be considered as reflecting the prerequisites for well-being, which
     must be the ultimate goal."

     The various  activities shown in Figure  1  have been developed to  support two major
objectives:

(a)  the implementation of investigations for the study of health effects  from noise in order
     to complement existing data and fill research gaps.

(b)  the preparation of a Manual on Noise Control in order to provide the decision maker in
     national and local government with the necessary information for the development of a
     local noise control program.

 **Lang, J. & Jansen, G. (1970), Copenhagen, (WHO document EURO 2631)
***Development of the Noise Control Programme" report on a Working Group, WHO, Copenhagen (document EURO
   3901)

                                         73

-------
    1211
                            1973
                                        1974
                                                    1975
1977
                                                                                        197B
                        1979
Development
of
programme
we-








Survey and
directory of
Institution* •
CS


Study of
nolle limit!
and reiearch ^
priorltlei
CS


Survey of
legislation
*
CS
(continue)
4
CS
Prtpar
of train
manual
CS
1
Preparation
of *tudy
protocol
^
CS
it ion
Ing





T raining
cour*«
(under
EURO 3101)

Field invettlgatlon* of health effect* from
nolle CS
'
: :.
Ditcuiiion
of *tudy
• protocol
WG
T
(continue)
CS
Co-
ordination
of
activities
C

Development
of criteria
and guide*
• r 1
lor nos*e
control
CS

Preparation
of model
chapter on
noi*e control •
(or building
code*
CS
Di*cu*ilon
of model
chapter
r
WG

Ma.nua
T
(continue)
CS


Di*cu*iion
of chapter
*
WG
j 	 '
Study of
enforcement
of nol*e
control
CS

Co-
ordination
of
activities
C


Evaluation
~* of itudie*
WG

4
Disc in lion
of chapter
WG
Di«cua«ion
of nolle
control in
SY
I 	 1
on Noi»« Control
T
Preparation
of chapter*
on Regional
planning, sur-
veillance of
nolle lour-
cei, econo-
mic aipecti.
etc.
CS
Dl*cu**ion
of
*• chapter*
WG


•*• Preparation
of manual for
CS
Type of




Evaluation
of long-
term prog
ramme
WG
i

activities;
C a Consultant
CS = Contractual Service
SY = Symposium
WG =s Working Group
This programme is periodically  reviewed and as such subjected to changes  as appropriate
                Figure 1. The long-term programme on noise control of the WHO Regional Office for Europe.

-------
     The various activities include a survey and the preparation thereafter of a Directory of
European institutions active in the study of health effects from noise and noise control. This
activity will lead to the  identification of appropriate collaborating institutions for the
investigations mentioned above. A study of noise limits and research priorities has just been
completed in draft and will serve both for the preparation of the mentioned investigations as
well as for the preparation of criteria and  guides. The criteria and guides will, in turn, serve
as a chapter for the Manual. Another chapter will serve as a "Model Chapter on Noise
Control for Building Codes", and additional ones will review and discuss various subjects
related to noise control such as regional planning, surveillance of noise sources, economic
aspects, the setting up of standards, manpower needs, etc. A survey of existing legislation
and administrative regulations related to  noise control is under way and will assist in the
identification of supplementary ones needed. A study  on law enforcement will follow and
eventually serve also as a chapter of the  Manual. The urgent subject of noise control in
Europe will be brought to the attention of the general public and the responsible European
authorities through  the convention of a European symposium, probably in  1978.  The
Office's long-term programme will then be evaluated and re-examined for activities needed
in the future.
                                          75

-------
                    SESSION 2




NOISE-INDUCED HEARING LOSS (NIHL)-EMPIRICAL DATA



            Chairman: D.W. Robinson, UK
                       77

-------
                        BASIS FOR PERCENT RISK TABLE

                                  Aram Glorig
                         Callier Hearing and Speech Center
                                  Dallas, Texas

                                William L. Baughn
                                Anderson, Indiana
    The fact that hearing loss produces an impairment is indisputable. The question is how
to evaluate the degree of impairment with presently available measurement techniques.
    The highly complex sound world we live in forces us to make value judgments on what
elements of our sound world are more important than others. The developmental history of
man has passed through stages which have emphasized  different aspects of the auditory
system.  The  development of speech and language as a means of communication has shifted
the emphasis from a simple  warning system for protection from danger to a highly complex
and unique system for storing and dispensing information. Our present  civilization has all
but eliminated the need for the auditory warning system of early man and the importance
of the auditory function now rests mainly with language acquisition and speech. Obviously
hearing  is involved in many  other listening experiences, such as music, etc.,  but I believe I
can say  without fear of contradiction that none are as important as communication  through
speech.
     If  this  concept is acceptable,  it  would  seem reasonable to  evolve  a method of
determining hearing impairment which correlates with "hearing everyday speech". "Hearing"
is used in its broad sense which includes an appropriate and correct response indicating that
what was said can be repeated by the listener.
    On  the assumption  that hearing speech is a common denominator for determining
"impairment"  or "handicap" due  to hearing loss, let us review the history of the present
American Medical Association-American Academy  of Ophthalmology and Otolaryngology
method  for evaluating hearing impairment from pure tone hearing levels.
     Because of the state of confusion that existed prior to the published recommendation
the Subcommittee on Noise of the Committee on Conservation of Hearing of the American
Academy of Ophthalmology and Otolaryngology arranged a conference on Determination
of Handicaps Resulting from Hearing Loss. The Conference was jointly sponsored by the
National Advisory Neurological Diseases  and  Blindness Council (USPHS) and the  Sub-
committee on  Noise. This conference was held on February 12-14, 1958. The following is a
summary of the proceedings of that conference:
     Although much is known about the measurement of hearing with various stimuli, the
use of these measurements to determine  the amount of handicap produced by hearing loss is
in a state of confusion. The conference afforded the first real  opportunity for individuals
representing various disciplines related to hearing to come together to discuss this important

                                       79

-------
problem.  It was the purpose of the conference to pool information and opinions upon
which policy  makers of the American Academy of Ophthalmology and Otolaryngology
might base recommendations for calculating handicap resulting from hearing loss.
     The  group included men  and women  who  represent the fields of Acoustics, Bio-
acoustics, Bio-communications,  Linguistics, Otology, Physics, Psycho-acoustics, Psychology
and Speech (including Speech Analysis  and Speech Synthesis). The following organizations
were  represented: The American Academy of Ophthalmology and Otolaryngology, Ameri-
can Medical Association, Bell Laboratories, Central Institute for the Deaf, Haskins Labora-
tories, Massachusetts Institute of Technology, United States Naval Research Laboratory,
Northwestern  University, University of Illinois, Purdue University, American Speech  and
Hearing Association and  the United States Naval Electronics Laboratory.

                        THEORETICAL CONSIDERATIONS

     Following presentation of and some  brief discussion of the background material, the
conference was thrown open for free  discussion which  centered mainly around (1) the
problems involved in the transfer of information from person to person through the medium
.of  speech and (2) the kind of investigation necessary to allow us to evaluate handicap for
communication from auditory measures.
     The discussion was based on three questions:
      1.   Which  kind of auditory communication efficiency should  be used to estimate
          handicap?
     2.   Which existing auditory  test is the best predictor of mis efficiency?
     3.   What test or type of test would you like to see used to estimate handicap?
Discussion of the first question included the following comments: (a) The normal speaker of
English has to perform certain tasks. If he is consistently  unable to do this he certainly has
"trouble", but  it is  very difficult  to  estimate the handicap  caused by this  "trouble".
Language  impairment, which is relatively easily measured, it is not necessarily the same as
handicap, (b) The question might be rephrased to say first that there is a "normal" listener
and then  to ask  "by what degree does the subject fall  short of meeting minimum normal
standards  for listening?" There are two dimensions of adequacy to be considered here: (1)
the signal level required  to be heard, and (2) discrimination of fine elements of speech, (c)
An approach similar to  that  used by industrial psychologists in areas other than audition
might be adopted here, namely, to specify that there are two ways of defining the criterion
for handicap: (1) as a job sample which would require a full replica of speech conditions and
(2) as an item analysis which would assess critical features of speech communication.
     Discussed as critical  auditory  efficiencies that might be used to estimate auditory
handicap were phonemic differentiation, auditory communication, performance of auditory
communication  tasks, and deficiency in the  reception of speech signals.  It was also noted
that speech and hearing are not necessarily the only factors responsible for a breakdown in
communication, which, after all, is the ultimate measure of handicap due to hearing loss.
     After some discussion, the conference advanced to a consideration of questions 2 and 3
about test materials. A  test of phonemic differentiation  (the Rhyme Test) was discussed.
The test measures word recognition, but confines the basis therefor to the initial consonant

                                        80

-------
and consonant  vowel transition  and yields a score that is heavily weighted with auditory-
phonemic factors, non-auditory factors being attenuated.
     It was suggested that different auditory abilities are  used in different ways depending
on the amount  of handicap and  that ultimately an intelligibility  test might be used to make
grosd distinctions and then batteries of tests applied within the discrete steps of the gross
scale.
     Also discussed were such questions as: Can everyday speech  be represented by specially
constructed  sentences or will carefully  selected words,  phonemes, or nonsense syllables be
more  useful  for determining the change in  information received  from speech when some
part  of the  auditory system is malfunctioning? What consideration should be given to the
effect of education, intelligence and language background of the subjects under test? Should
hearing be measured at levels above threshold?  What  part is played by  the environment
surrounding  both the speaker and the listener? What, for example, are the effects of noise
level  and noise spectrum? The  consensus was that no single functional  test could  apply
under all conditions. Throughout  the discussion, it was  evident that much more research
must  be done before a completely satisfactory method of determining handicap could be
formulated.

                          PRACTICAL CONSIDERATIONS

     Having  accepted the necessity  of further research, the group turned  to the immediate
practical problem of assessing handicap from results  of existing auditory tests and to a
discussion of the possibility of agreeing on an interim  method of handicap determination.
The  need for an interim method for determining handicap resulting from hearing loss was
great. This need was attested to  in part by the confusion that existed  in  the various states
where compensation for hearing loss is provided; at that  time, no two states used the same
method  of  rating disability due to hearing loss.  Further, there was no agreement on a
method for  rating improvement following surgical procedures used to correct conductive
hearing losses. There was a practical need to provide surgeons, legislators and others with an
interim method, even though that  method might not be completely satisfactory and would
have to be changed several years later when more  information became available.
     The conference members agreed  (1) that  they could not,  as scientists,  designate a
completely satisfactory  method  at that time; (2) that the need for a method to evaluate
handicap was urgent and (3) that it would be better for an authoritative group  to recom-
mend the use of an interim method rather than to condone by default  the continued use of
methods formulated, in  some  instances, by groups with little or no  knowledge of the
subject, (4)  that it was reasonable to propose a method  which would be useful, provided the
limitations  of  the  method were understood. It was  eventually agreed that  if sentence
intelligibility is accepted  as a representative measure  of everyday speech,  there was enough
information  to recommend an interim method  of determining handicap. An objection to
using sentence intelligibility as a measure was that currently available sentence tests do not
take into account the effects of background noises, talker identification, localization, etc.
     It  was  agreed  that sentence intelligibility  depends more  on  hearing level than  on
discrimination;  Acknowledging the lack  of  sufficient quantitative information about tests

                                         81

-------
for discrimination, the group agreed that it would be impossible at that time to recommend
a simple method  that includes an evaluation of the changes in discrimination that accom-
pany hearing loss.

        CONCLUSIONS AND RECOMMENDATIONS OF THE CONFERENCE

I.  Although there was no general agreement about how to define or assess handicap due to
hearing loss, the overwhelming majority of the conferees agreed that from a practical point
of view an adequate assessment of handicap could be made from pure tone measurements. It
was agreed by majority vote of 10 to 6 that (a) if the ability to hear  and repeat sentences
correctly in a quiet environment is accepted as  currently the best representation of hearing
for everyday speech and (b) if the measures used  to calculate hearing loss for this everyday
speech are to  be  weighted equally, then: an average of the hearing  level in decibels at the
three frequencies  500, 1000 and 2000 Hz is an acceptable interim method for determining
hearing  loss for  everyday  speech as an  estimate  of handicap. The minority voted to
recommend the use of the average of hearing levels at 500, 1000, 2000 and 3000 Hz to
determine hearing loss for everyday speech. The  chairman and three conferees abstained
from voting.

II. Not  being completely satisfied with the foregoing methods,  the  conference members
unanimously recommended that research,  such as investigation of the effects of different
talkers, of various types of hearing loss, of variations in environmental noises, of various
types of stimuli  both  at threshold and at suprathreshold levels, etc., be undertaken to
discover the factors which should be employed  in a more valid predictive test of the ability
to receive everyday speech.

III. It was unanimously accepted that the principle of the  high and low fence be incor-
porated in the design of any method for determining handicap due to hearing loss.
     The principle of the high and low fence is essentially the principle that the range of
handicap is smaller than the range of auditory sensitivity measured  by pure tones. The low
fence is the point on the audiometric hearing level  scale at which significant handicap begins
and the high fence is the point on the audiometric hearing level scale at which the handicap
is total. Audiometric zero, which is presumably  the average normal hearing threshold, is not
an acceptable low fence because it is not the point at which significant  handicap begins. The
low fence is definitely higher than audiometric zero. Handicap is total at hearing levels lower
than the maximum  output of the standard audiometer; therefore, the high fence is neces-
sarily  lower than the maximum hearing level. It is  evident that  the American Medical
Association-American Academy of Ophthalmology and Otolaryngology method was evolved
with much consideration and a large amount of background information.
    The following is a direct quotation from the American Medical Association Guide to
the Evaluation of Permanent Impairment: (1)
    "Ideally, hearing impairment  should be evaluated in terms of ability to hear everyday
speech under  everyday conditions. The ability  to hear sentences  and to  repeat  them
correctly in a quiet environment  is taken as satisfactory  evidence for correct  hearing of

                                        82

-------
everyday speech. Because of present limitations of speech audiometry, the hearing loss for
speech is  estimated  from  measurements made with  a pure tone audiometer. For this
estimate, the simple average of the hearing levels at the three frequencies, 500, 1000 and
2000 Hz is recommended.
    "In order to evaluate the hearing impairment it must be recognized that the range of
impairment is not nearly as wide as the audiometric  range of human hearing. Audiometric
zero, which  is presumably the average  normal threshold level, is not the point at which
impairment begins. If the average hearing level at 500,  1000 and 2000 Hz is 25 dB (15 dB
ASA - 1951) or less, usually no impairment exists in  the ability to hear everyday speech
under everyday conditions. At the other extremes, however, if the average hearing level at
500, 1000 and  2000 Hz is over 92 dB (82 dB, ASA-1951) the impairment  for hearing
everyday speech should be considered  total. For every decibel that the estimated hearing
level for speech exceeds 25 dB (15 dB, ASA-1951) 1.5% of monaural impairment is allowed
up to a maximum of 100%. This maximum is reached at 92 dB."
    Impairment  in each ear is determined and binaural impairment  is calculated on a 5
(better ear) to one (poorer ear) basis.
    There are several points in the above statement that need discussing.
     1.   "Hear everyday speech"
         Hearing in this context is used in the broad sense indicated by correct repetition
of what was said. Criticisms  of the formula have included statements that to hear is not
necessarily to understand. Understanding  may be used (1)  to indicate discrimination of
word parts or (2) to indicate correct interpretation of the message content. One may hear all
the component parts of a foreign language and not understand the message.
     2.   "Everyday speech under everyday conditions"
         It would  be difficult to define everyday  speech to everyone's satisfaction, par-
ticularly to  define what is included under  "everyday conditions". We have chosen to con-
sider  speech in sentence form to  be  "everyday speech" and correct repetition of the
sentences in quiet as satisfactory evidence of having heard the speech. After consideration of
the varied ambient  conditions that  prevail from day to day during periods of communi-
cation "quiet" was considered to be the condition which prevails the majority of the time,
and therefore best represents prevailing everyday conditions.
     Attempts were made to estimate an average of the noise levels and noise spectra which
accompany speech communication on infinitely varied occasions. Since we could not solve
the question of a representative ambient noise "everyday conditions" are considered to be
listening in quiet
     Obviously it can be  argued that "quiet" is not representative  of everyday conditions
but in our judgment it was the only condition that could be repeated infinitely without
significant variation and therefore subject to standardization.
     3.    "Low fence" and "high fence"
          Twenty-five dB average hearing level (AHL)  was chosen as the level above which
impairment begins. This level was chosen for several reasons among which are:

          1.   The  output  level of "everyday speech"  is usually  between  60-70  dB.
              Listening at threshold is  generally not the case.

                                        83

-------
          2.   Audiometric zero is the central statistic of a range between plus or minus 15
              dB implying that many of the subjects included as normal "hearers" were 15
              dB worse than zero hearing level.
     The "low fence" principle  is supported by at least two studies, one reported by Glorig
et al. (2) and another by Baughn (3). Both studies included pure-tone audiograms and a
"self evaluation of hearing" based on a response to the question. "Is your hearing good, fair
or poor?" in the Glorig study and "How well do you think you hear?" in the Baughn study.
Responses were restricted to the words "good", "fair" or "poor". Both studies indicate a
gradual change in the  direction of higher AHL (average of 500, 1000 and 2000 Hz) with
increase in age in each of the categories, "good", "fair" or "poor". The  twenty-year-olds
stated  their hearing was "good" when the average hearing level was 10 dB. (Figure 1 from
Baughn).  Figure 2 shows a comparison of self-evaluation and speech reception threshold.
Notice the 10-dB change  from 10 dB (ASA) to 20 dB (ASA) for "good" as age increases to
70. The clinical  assumption that noticeable impairment does  not begin until the AHL
exceeds 15 dB ASA or 25 dB  ANSI is  well supported by these data.  In fact, both "low
fence" and "high fence" concepts are seen to be reasonable. Figure 1, from Baughn.
     A review of the literature indicates that the use of 500, 1000, and 2000 Hz does indeed
prove to be quite adequate to represent "hearing of everyday speech" even in noisy situa-
tions.
     Ward et al. (4), in a study related to characteristics of hearing loss found in  subjects
exposed  to  gunfire and  steady noise, determined differences in discrimination for PB's,
paired  consonant  weighted words and paired phonemes between these two groups.  Subjects
with severe  losses above  2000  Hz showed only small differences in discrimination from
those with little or no loss above 2000 Hz. These tests were done in quiet but it is reason-
able to assume that the additional information provided by speech in sentence form would
easily compensate for this small difference when "listening" in noisy environments.
     Ward says, "These results  constitute additional justification for the practice of con-
sidering hearing losses  only at 2000  Hz and below in estimating "handicap" in everyday
speech perception." (4)
     Myers and Angermeier (5)  and Murry and Lacroix (6) tested subjects  with losses above
2000 Hz in noise  and found 5 to 10% difference between normal subjects and subjects with
losses above 2000 Hz. Thus, at least from these data, it appears that the  difference between
subjects with losses  above 2000  Hz and normal subjects is small and it could be expected to
be compensated for by the redundancy of information in speech in sentence form. Further-
more, the 5-10% lower discrimination scores in the hearing loss group appear to be the same
in "quiet" as in "noise". In my opinion, a recommendation to change a well-established and
much-used method of calculating impairment due to hearing loss should be  based on more
significant advantages than those shown in the present data before a change is made in the
method of calculating impairment due to hearing loss.
     Because the  concept of the "percent risk"  table is based on the determination of
impairment  due  to hearing loss I have included  this discussion of the presently recom-
mended evaluation formula.
                                        84

-------
     70i
     60
     5O
     4O
X
• J,
3O
   LLJ

   o
     20
      10
                                 FAIR
                                   FP
                               J.
                                            POOR
                                                           VP
DEAF
               10     20     30      40      50      6O     7O     8O


                      AUDIOMETRIC HEARING  LEVEL  IN DECIBELS


                       Figure 1.  A graphic representation of the results of a self evaluation study.
                                                                          90
             1OO

-------
•/
       -5

   G  10
   w
   Q

       20
       30
       40
    50
CO
CO


3

o
                                          Good
                                             Fair
                                             Poor
       10-19  20-29   30-39  40-49   50-59  60-69  70-79


                              AGE
                 Figure 2. Self evaluation compared with speech reception threshold scores.

-------
                                "PERCENT RISK"

    The concept of "percent risk" was first suggested by Glorig in 1962 as a possible means
of separating the effects of noise on an exposed population from the non-noise effects on a
non-exposed population as a function of equivalent sound level (7). In order to implement
this suggestion, Baughn organized a program to gather the essential data.
    More than 6000 audiograms  and matching noise-exposure data were  subjected to
regression analysis based on three parameters as follows:
     1.   Hearing level  should be defined by the simple arithmetic average of audiometric
         hearing levels  at 500, 1000 and 2000 Hz.
    2.   Exposure intensity should  be defined  by the  reading displayed by a standard
         sound level meter  set at "A" frequency weighting and "slow" inertial dynamics
         with the microphone in a position to intercept the characteristic sound prevailing
         at the subject's ear in his work environment.
    3.   Exposure duration should  be defined  as  years on  the job, approximated for
         groups by subtracting 18 years  from the center year of the  chronological age
         group.
    The noise studies were taken  from over 15,000 detailed sound analyses of work loca-
tions, covering a period of 14 years. Studies of work records, interviews with employees,
and comparative testing of older and more  recent equipment and processes made it possible
to extend exposures back for nearly 40 years with reasonable confidence. Actually the noise
analyses included octave bands, A, B, and C weightings, SIL's and other computed indices in
both slow and faster meter dynamics and in some  areas  repetitions with a General Radio
impact meter. Only the "A" weightings and slow meter readings were used for this study.
All measurements were made  and/or checked by  competent engineers and  checked for
consistency. Three specific levels were used: 78, 86  and 92 dB(A), because the majority of
the sample  clustered around these  points. The group assigned 78 dB(A) spent 90% of their
time in no more than 81  and no less  than 66 dB(A). None exceeded  82  dB(A). The group
assigned 86 dB(A) spent 80% of their time between 86 ± 4  dB(A). The group assigned 92
dB(A) spent 87%  of their time in 92 ± 5 dB (A). (Detailed distributions of time vs. level are
shown in Table 1.)
     All audiograms were made  by two individuals who had done more than 25,000 audio-
grams over  a period of  eight years, all of which had been  submitted to scrutiny by the staff
of the Subcommittee on Noise Research Center. Samples were subjected to consistency tests
and mathematical analysis by a senior research member of this staff. Similar tests applied to
the data used in this study  confirmed its  self consistency.  Approximately 20,000 audio-
grams were culled for consistency of exposure history leaving 6,835 audiograms.
     The audiometric test  environment conformed fully with  the  specifications of the
American  Standards  Association.  The audiometers were Maico  H-l  models and  were
checked against normal experienced ears before each day's use, and were calibrated in the
laboratory  of the Maico Company periodically. They were never found to be out of the
acceptable calibration range.
     The population under study was a very  stable  one since the work force came from a
relatively small community. The age ranged from  18 to 68 years and many were in the same
job for more than 40 years. Age was used  as  a uniform measure  of exposure in years since
                                        87

-------
TABLE 1. MEAN PERCENT TIME SUBJECTS SPENT IN EACH ASSIGNED NOISE LEVEL.


    dBA               78            86            92
65 - 66
66 - 67
67 - 68
68 - 69
69 - 70
70 - 71
71 - 72
72 - 73
73 - 74
74 - 75
75 - 76
76 - 77
77 - 78
78 - 79
79 - 80
80 - 81
81 - 82
82 - 83
83 - 84
84 - 85
85 - 86
86 - 87
87 - 88
88 - 89
89 - 90
-90 - 91
91 - 92
92 - 93
93 - 94
94 - 95
95-96
96 - 97
97 - 98
98 - 99
99 - 100
100 - 101
.75
1.
1.
1.
1.
1.
2.
2.
3.
3.
4.
10.
12.
16.
20.
12.
5.
2.
1.
.5
















                                    .5
                                   2.
                                   2.
                                   3.
                                   4.
                                   6.
                                   8.              .5
                                  10.            2.
                                  13.            2.
                                  14.            2.5
                                  14.            3.5
                                  10.            4.5
                                   8.            6.0
                                   5.           10.0
                                   3.           12.0
                                                14.0
                                                12.0
                                                10.0
                                                 6.0
                                                 5.0
                                                 3.5
                                                 2.0
                                                 1.0
                              88

-------
attempts to  allow for various interruptions in work life provided no advantages. Subjects
with seriously mixed exposures,  or unknown exposures, were excluded in an attempt to
keep the data as nearly ideal as  possible.  Selections for history of otological disease, etc,
were not made. Studies  of large populations  have shown that exclusions for non-noise
effects (exclusive of age) show only very small changes in statistical distributions. (7)
     Following the exclusions noted  above we were left with 6,835 audiograms matched
with exposure history in terms of three exposure groups identified as 78 dB(A), 86 dB(A)
and 92 dB(A).
     The audiograms were divided into two groups; those with AHL's at 15 dB or less and
those with AHL's above 15 dB (15 dB ASA-1951) (25 dB ISO-1964).
     All audiograms were  done prior  to the end of 1965  and all  were done to American
Standards Association 1959 standard  audiometric zero calibrations. (All data were  entered
on punched cards and all sorting and calculations were done by  electronic data processing
equipment).  There were 852 subjects in the group assigned 78 dB(A), 5,150 subjects in the
group assigned 86 dB(A)  and 833 subjects in the group assigned 92 dB(A). Table 2 shows
the distribution of the sample as a function of five year age groupings and exposure assign-
ments.
     These data were used to construct Table 3 and Table  4. It is  obvious that when these
tables are compared there are discrepancies between them. Careful examination will  reveal a
difference related to beginning age; 18 in Table  3 and 20 in Table 4. When the numbers are
picked from  a curve constructed from extrapolations of the three sound levels the 2 year
difference in  starting age accounts for the  differences seen  in the various published Tables.
There is not time to include a complete discussion of the methods used to  extrapolate the
numbers given in Tables 3 and 4.  Figure 3 shows the extrapolated functions by merely using
the three noise-exposure assignments. Figure 4 shows the slight upward curves produced by
extrapolation and curve fitting procedures. A thorough discussion of  the extrapolation is
given in an Aerospace Medical Research Laboratory Technical Report. (8)
     Since  "percent  risk"  is based  on the percentage difference between  a non-noise-
exposed group and a noise-exposed group, the non-noise-exposed group is a critical factor.
The non-noise-exposed data used in these tables were derived from a study by Glorig et al.
(7) A comparison of these data with a U.S. Public Health Service study  is shown in Table 5.
There are differences between these  studies which are probably sample differences since
both studies were designed and supervised by Glorig.
     Table  5  shows  differences  between  two samples whose derivations  are completely
different. The  Glorig sample consisted of professional men  whose history included  no
occupational noise-exposure and very little non-occupational  noise-exposure. The USPHS
sample was  picked to represent the general population of the United States. Theoretically
the general  population should have  more occupational noise-exposure influence than a
population chosen to exclude this. The logic of this difference is supported by the fact that
the Glorig  total sample  shows a lower percentage of "better ears"  that exceed 15 dB
(ASA-1951) AHL (average of hearing levels at 500, 1000 and 2000 HZ)  than is foundin the
USPHS sample. The age differences noted are quite consistent until the Glorig sub-group N's
become small. Further evidence  of consistency is provided by a comparison of non-noise
exposed samples and the 78 dB(A) group studied by Baughn. (Figure 5).

                                         89

-------
          TABLE 2. NUMBER OF SUBJECTS AS A FUNCTION OF AGE GROUPS AND SOUND LEVEL A.
Age Group                 Exposure I   Exposure II    Exposure  III
 Number      Age Span       78 dBA       86  dBA         92 dBA	Total
1
2
3
4
5
6
7
8
13 -
24 -
30 -
36 -
42 -
48 -
54 -
60 -
23
29
35
41
47
53
59
65
N = 10
66
144
148
183
159
95
45
N - 107
476
544
860
1041
1070
723
329
N « 4
39
76
124
189
197
127
77
121
583
764
1132
1413
1426
145
451
                             852          5150             833           6835

-------
TABLE 3. PERCENT RISK TABLE USING AGE 18 AS YEAR EXPOSURE STARTED.
AGE
EXP. YEARS
EXP. LEVEL
80 dBA
EXP. LEVEL
85 dBA
EXP. LEVEL
90 dBA
EXP. LEVEL
95 dBA
EXP. LEVEL
100 dBA
EXP. LEVEL
105 dBA
EXP. LEVEL
110 dBA
EXP. LEVEL
115 dBA
(AGE - 18)
TOTAL '/. EXPECTED
7. DUE TO NOISE
7. DUE" TO OTHER
TOTAL '/,
7, NOISE
7. OTHER
TOTAL 7.
7. NOISE
7. OTHER
TOTAL 7.
7, 'NOISE
7. OTHER
TOTAL %
7o NOISE
7» OTHER
TOTAL %
7. NOISE
7o OTHER
TOTAL 7.,
7. NOISE
7. OTHER
TOTAL 7.
7» NOISE
7» OTHER
18
0
.5
0
.5
.5
0
.5
.5
0
.5
.5
0
.5
.5
0
.5
.5
0
.5
.5
0
.5
.5
0
.5
23
5
1.7
0
1.7
2.5
.8
1.7
6
4.3
1.7
9.0
7.3
1.7
14
12.3
1.7
20
18.3
1.7
28
26.3
1.7
38
36.3
1.7
28
10
3
0
3
6
3
3
13
10
3
20
17
3
32
29
3
45
42
3
58
55
3-
74
71
3
33
15
4.5
0
4.5
9
4.5
4.5
18
13.5
4.5
28
23.5
4,5
42
36.5
4.5
57
52,5
4.5
75
70.5
4.5
87
83.5
4.5
38
20
6.5
0
6.5
12.5
6
6.5
22
15.5
6.5
34
27.5
6.5
48
41.5
6.5
64
57.5
6.5
84
77.5
6.5
93
86.5
6.5
43
25
9.7
0
9.7
16.5
6.8
9.7
26
16.3
9.7
39
29.3
9.7
53
43.3
9.7
70
60.3
9.7
88
78.3
9.7
94
84.3
9.7
48
30
14
0
14
22
8
14
32
18
14
45
31
14
58
44
14
76
62
14
91
77
14
95
81
14
53
35
21
0
21
30
9
21
41
20
21
53
32
21
65
44
21
82
61
21
93
72
21
96
75
21
58
40
33
0
33
43
10
33
,54
21
33
62
29
33
74
41
23
87
54
33
95
62
33
97
64
33
63
45
50
0
50
57
7
50
65
15
50
73
23
50
83
33
50
91
41
50
95
45
50
97
47
50

-------
            TABLE 4. PERCENT RISK TABLE USING AGE 20 AS YEAR EXPOSURE STARTED.
Age
Exposure
(age = 20)
Total
80 Due to
Noise
Total
85 Due to
Noise
Total
, 90 Due lo
PQ Noise
20
0
0.7
0.7
0.0
0.7
0.0
26
5
1.0
No
2.0
1.0
4.0
3.0
30
10
35
15
1.3 2.0
increase in
3.9
2.6
7.9
6.6
6.0
4.0
12.0
10.0
40
20
46
25
3.1 4.9
risk at this
8.1
5.0
15.0
11.9
11.0
6.1
18.3
13.4
60
30
7.7
level
14.2
6.5
23.3
15.6
55
35
60
40
13.5 24.0
of exposure
21.5
8.0
31.0
17.5
32.0
8.0
42.0
18.0
65
Years
46
40.0
46.5
6.5
54.5
14.5

B
— — « .
 g     Total              0.7    6.7  13.6  20.2  24.5  29.0  34.4   41.8   52.0  64.0
•3  95 1)^  jo
H     Noise   	0.0    5.7  12.3  18.2  21.4  24.1  26.7   28.3   28.0  24.0
       Total              0.7   10.0  22.0  32.0  39.0  43.0  46.5   55.0   64.0  75.0
   100 Due  to                                                                              •$
       Noise	0.0    9.0  20.7  30.0  35.9  38.1  40.8   41.5   40.0  35.0        «
       Total              0.7   14.2  33.0  46.0  53.0  59.0  65.5   71.0   78.0  84.5        f
   105 Due  to                                                                               |
       Noise	0.0   13.2  31.7  44.0  49.9  54.1  57.8   57.5   54.0  44.5        fc>
       Total              0.7   20.0  47.5  63.0  71.5  78.0  81.5   85.0   88,0  91.5       £
   110 Due  to
       Noise	0.0   19.0  46.2  61.0  68.4  73.1  73.8   71.5   64.0  51.5
       Total              0.7   27.0  62.5  81.0  87.0  91.0  92.0   93.0   94.0  95.0
   115 Due  to
       Noise              0.0   26.0  61.2  79.0  83.9  86.1  84.3   79.5   70.0  55.0

-------
c
Ul
I

• I


CD
o
                                 BO
                                                                                    1 15
                                                                                            120
                                                                                                    1 25
                                                                                                            130
                                     90     95      1OO    1O5     110



                                        EXPOSURE   dB(A)


          Figure 3. Extrapolated curves representing percent risk due to noise-exposure using the dB(A) sound level

          assignments 78, 86, and 92. The parameter is age.
                                                                                                                   135
                                                                                                                          140

-------

DO
a
        TO
                     80
                            85
90     95     100    1O5     10


    EXPOSURE     dB(A)
                                                                       1 1 5
                                                                               20
                                                                                     1 25
                                                                                            1 3O
       Figure 4. Same data as shown in figure 3 except for use of further extrapolation procedures. The parameter

       is age.

-------
TABLE 5. COMPARISON BETWEEN GLORIG NON-NOISE SAMPLE AND THE U.S.P.H.S. GENERAL
POPULATION SAMPLE. BOTH STUDIES WERE BASED ON AMERICAN STANDARD ASSOCIATION
{ASA} 1951 AUDIOMETER CALIBRATION LEVELS.
         PERCENTAGE OF MEN  OVER  15 dB-AHL.   BETTER EAR.
All
Years
N =
Glorig
N *
US PHS
1323
2.9
52744
7.6
18-24
Years
120
0.8
7139
1.2
25-34
Years
693
1.7
10281
1.4
35-44
Years
438
3.
11373
3.7
45-54
Years
111
7.2
10034
4.1
55-64
Years
30
23-3
7517
10.6
65-74
Years
1
0
4972
30.5

-------
             40
C
3
         X   30
         CD
         T3
         m
A   20
IT
O
         •I
         _J
         r^
         O
         a
10
                                                                              NON-NOISE
                                                                                         /
                                                                     \
                                                                              80 dB (A)
                              10
                               20
                                                    30             40
                                                         AGE
50
60
70
                     Figure 5. Comparison of percent risk curves from Glorig's non-noise data and Baughn's 78 dB(A) data.

-------
    Table 6 shows a comparison of six studies of median hearing levels as a function of age.
All studies were corrected to zero hearing level at age 20 to rule out sample differences on
the  assumption that  the  majority of  20-year-olds have normal hearing. The Table clearly
shows differences in  the  numbers which must be  related to the population samples. In
general there is reasonably good agreement, at least when medians are compared. Distribu-
tions were not available.
    These comparisons confirm the validity of the Glorig sample for use as a non-noise-
exposed group to determine percent risk. Obviously one should expect differences in studies
of different groups. Perhaps the most important comparison in this  discussion relates to
Table 5. It is difficult to disregard so  large a sample but it  is also difficult to classify these
data as non-noise-exposed or disease-free since no attempt was made to exclude either.
    A careful examination  of Table  6 shows that the median hearing levels of sample B
(Glorig) and sample  F (USPHS) at 1000 and 2000 HZ are very close together. When the
hearing levels at 3000 and 4000 HZ are examined the USPHS data show significantly higher
levels  from age 50 up. The most likely explanation of these differences is the effect of noise
exposure in  the  general population sample. What  to  use  as a non-noise-exposed  baseline
remains a dilemma. It is this baseline that determines the percent risk as a function of noise
exposure.
    An examination of Tables 3 and 4 reveals a reversal in direction of the percentages as
exposure years increase. This reversal  at first glance appears to be paradoxical since longer
exposure duration should increase hearing loss. The explanation lies in the fact that noise-
induced hearing loss  assumes  a decelerating function after 15-20 years exposure while
hearing loss due to age assumes an accelerating function.  Figure 6 is a diagramatic repre-
sentation of the change in percentages of population with AHL's above 25 dB due to age,
(OG)  noise-exposure  plus age (OP) and noise-exposure minus age (OK). It is obvious that as
the effects of age on the baseline (non-noise-exposed) population accelerate with increase in
age and the effects of noise exposure on the subject population decelerate the difference
between these two functions  decreases. This decrease is reflected in  Tables 3 and  4 by a
decrease in "percent risk". Figure 7  shows an idealized set of curves taken from Table 3
showing the effects of noise-exposure versus presbyacusis.

                               GENERAL REMARKS

     The previous discussion  presents the background and data-gathering history  of the
numbers used  to establish the percent risk  due to noise as a function of increasing noise
level.  The data used  were carefully  obtained from a representative industrial  population
under conditions which prevail in above-average industrial hearing conservation programs.
Even  if it were practical to obtain this amount of data under more stringent controls it is
doubtful if the  stringently  controlled data  would provide as useful information as the
present data. Hearing conservation programs and compensation are related to data obtained
under field conditions. Highly refined data may provide false impressions.
     We must agree  that further study should  be attempted to confirm or disaffirm the
non-noise  baseline but repeating such a study is not a simple matter. Finding larger N's at
the higher ages is rather difficult.

                                         97

-------
TABLE 6. COMPARES MEDIAN HEARING LEVELS AT 1000, 2000, 3000, 4000 AND 6000 Hz AS A
FUNCTION OF AGE. DATA BASED ON ASA-1951 AUDIOMETER CALIBRATION LEVELS WERE
TAKEN  FROM 6 STUDIES DONE IN VARIOUS PARTS OF THE WORLD. ALL DATA ARE COR-
RECTED TO ZERO AT AGE 20.

AGE
A
B
C
D
E
f

AGE
A
B
C
D
E
f
AGE
A
B
C
D
E
F
AGE
A
B
r
V.
D
E
F

AGE
A
B
C
D
E
f
A- CORSO
B- GLORIG

20
0
0
0
0
0
0

20
0
0
0
0
0
0
20
0
0
0

0
0
20
0
0
0
0
0
0

20
0
0
0

0
0



30
1
0
2
5
0
0

30
0
1
1
5
0
1
30
0
3
3

4
3
30
8
6
4
14
4
5

30
7
7
3

2
5


1000
40
2
0
2
8
0
1
2000
~4U~
0
4
3
7
1
4
3000
1
6
6
NO DATA
5
8
4000
40
8
10
9
18
5
14
6000
40
5
12
9
NO DATA
7
12
C- HINCHCLIFFE
D- JOHANSEN

50
5
1
4
13
0
5

50
7
7
7
14
5
8
50
11
12
16

9
15
50
20
17
18
27
10
23

50
19
20
21

7
19



60

4
5
19
1
6

60

14
12
24
5
12
60

21
29

10
29
60

27
34
38
14
36

60

33
46

8
35
E - ROSEN
F- USPHS

70

14
12
29
4
13

70

27
25
34
5
24
70

34
37

10
44
70

37
42
46
10
48

70

46
47

13
51


                                     98

-------
t
a

E
o


CO
 a
 o
a.
 0)
 o
 k.
 a>
o.
              A                B                     C

              Exposure  Duration   (Age-18)
      Figure 6. Diagramatic representation of accumulating effects of noise exposure and age.
                                  99

-------
                                                           FIGURE 7

   100
 I
••

 •
 •
 :
 •
 •
Q
'
D
                                LIMIT   f ^,

                                FIGURES     4
           Figure  7. Idealized curves showing accumulated effects of noise exposure and age as a function of sound
           level A.

-------
    We are convinced that attempts to improve on these data will prove difficult and are
some years away. Furthermore, new data will probably not change the interval increases in
"percent risk" significantly. It must be remembered that the  risk numbers can never be
applied  to individuals, only to populations of adequate size and  that they never refer to
"amount" of hearing impairment, only to the  percentage of a population having trans-
gressed a certain criterion.

                                    SUMMARY

1.  Impairment due to hearing loss must be based on hearing and understanding everyday
    speech.
2.  The use of the three-frequency average at 500, 1000 and  2000 Hz has proved reason-
    ably accurate to determine impairment.
3.  The low fence of 25 dB (ISO-1964) or 15 dB (ASA-1951) and a high fence of 92 dB
    (ISO-1964) or 82 dB (ASA-1951) are realistic.
4.  The "percent" risk tables are based on carefully done studies. Any differences among
    tables are due to manipulating such factors as age, allowance for TTS or attempts to
    avoid effects of noise-exposure and/or disease.
5.  Attempts to refine the data should be encouraged but no changes should be made until
    further specific studies unquestionably show valid reasons for change.
                                    References

 1.   Guides to  the Evaluation of Permanent Impairment; Committee on Rating of Mental
     and Physical Impairment, Amer. Med. Assoc., 1971. Page 103.
 2.   Glorig, A., Wheeler, D., Quiggle, R., Grings, W., Summerfield, A.; 1954 Wisconsin State
     Fair Hearing Survey, Published by the American Academy of Ophthalmology and
     Otolaryngology, 1957.
 3.   Baughn, William  L.; How Well Do  You Think You Hear? Proceedings of 13th Inter-
     national Congress on Occupational Health, Book Craftsmen Assoc., N.Y. 1960.
 4.   Ward, W.D., Fleer, Robert E., and  Glorig,  A.; Characteristics of Hearing Losses Pro-
     duced by Gunfire and by Steady Noise. J. Audit. Res., Vol. 21, 1962. Pp. 325-356.
 5.   Myers, C.K., and  Angermeier, C.; Hearing Loss at 3 Kilohertz and the Chaba "Proposed
     Clinical Test of Speech Discrimination in Noise." Naval Submarine Medical  Research
     Lab., Report No. 720.
 6.   Murry, T., and Lacroix, Paul G.;  Speech Discrimination in Noise and Hearing Loss at
     3000 Hertz. Naval Submarine Med. Research Lab. Report No. 719.
 7.   Glorig, A., Summerfield, A.,  and Nixon, J.;  Distribution of Hearing Levels in Non-
     Noise-Exposed Population. Proceedings 3rd  International Congress  on Acoustics.
     Elsevier Publishing Co., Amsterdam.
 8.   Relation Between Daily Noise-Exposure and Hearing  Loss as Based on the Evaluation
     of 6835 Industrial Noise-Exposed Cases. Aerospace Medical Research Lab. - Technical
     Report-73-53-1973.

                                        101

-------
References for source of data used in Table 6:

Corso, John K: Age and Sex Difference in Pure Tone Thresholds. J. Acous. Soc.
Amer. 31: 498-507, Apr. 1959.

Glorig, A., Nixon,  J.: Hearing  Loss as  a Function  of Age. Larngoscope, Vol.
LXXII No. 11, pp 1596-1610, Nov. 1962.

Hinchcliffe, R.: The Threshold  of Hearing as a Function of Age.  Acoustica 9:
303-308, 1959.
Johansen, H.: Loss of Hearing Due to Age, Munksgaard, Copenhagen, pp 165.


Rosen, S., Bergman, M., Plester, D., El-Mofty, A., Satti, M.H., Presbycusis Study
of a Relatively Noise-Free Population in the  Sudan, Ann. Otol. Rhin. & Larng.,
Vol. 71, pp 727-743, 1962.

National  Health Survey. United  States - 1960-1962. Hearing Levels  of Adults by
Sex and Age. Series 11, No. 11.
                               102

-------
      A CRITIQUE OF SOME PROCEDURES FOR EVALUATING DAMAGE RISK
                            FROM EXPOSURE TO NOISE

                                   Karl D. Kryter
                             Stanford Research Institute
                             Menlo Park, California 94025

                                   Introduction

    There  are fundamental  definitions and  measurements that must be made before a
proper evaluation of the damaging effects of noise on the inner ear can be performed and
before valid standards  relevant to  the  protection of man's health in this regard can be
promulgated. First, one must define a criterion of hearing, along with a quantitative means
of specifying degrees of impairment to that hearing. Second, the hearing ability, as defined,
of the population of persons who  have not  suffered noise-induced hearing loss must be
known. Thirdly, of course, hearing-ability data of populations  of persons exposed to noise
of various intensities, durations, spectra and years of exposure are  required.

                          Hearing and Hearing Impairment

    It should be obvious that the nature of so-called damage risk tables (i.e. ISO 1999) are
very much  dependent  upon the specification of what constitutes hearing and the start of
hearing impairment, and a critique of present day procedures  for evaluating noise-induced
damage to hearing must begin with that question. For purposes  of hearing conservation with
respect to noise-induced hearing loss, "hearing" is to be known, according to the rather
long-standing recommendations of the American Academy of Opthalmology and Otolaryn-
gology (AAOO)2, as the  average  of pure-tone hearing levels measured at 500, 1,000 and
2,000 HZ,  this average being taken  as an index or indicator  of an ability of a person to
understand  speech.  Speech is defined by  the AAOO as "everyday" speech in the  quiet heard
with the distance between talker and listener presumably approximately one meter or so.
The AAOO Committee further proposed that:  (1) the threshold of impairment (or lower
fence  as it was called) to the understanding or perception of everyday speech, be taken as an
average HL at 500 1,000  and 2,000 Hz of 25 dB, and (2) that there is an increase of 1-1/2%
in hearing impairment with each 1 dB increase above 25 dB until  an average HL  of 92 dB at
the three test frequencies is reached, at which point impairment for hearing everyday speech
reaches  100%. The AAOO  also has specified ranges  of hearing impairment that  are
supposedly  useful in relating the impairment to a description of "handicap" suffered by the
person with that  impairment.3 Whether the handicap is for general social communication or
in one's occupation is not made clear.

                        Hearing Level and Speech Impairment

    The general relations between hearing levels averaged for 500,  1,000 and 2,000 Hz,
percent impairment and percent of speech material correctly perceived, are shown in Fig. 1.
Also drawn on Fig. 1 are the  relations  between hearing level, average for 500, 1,000 and

                                       103

-------
  100
   60
40
   20
                               I    I
 I    I   I    I    I   I
1000 PB WORDS AT WEAK-
CONVERSATIONAL LEVEL
(50 dB) IN QUIET—Al

SENTENCES AT NORMAL
CONVERSATIONAL LEVEL
(55 dB)  IN QUIET—Al
        SENTENCES AT
        "EVERYDAY"
        LEVEL (65dB)—Al
                                                                 I    I   I
SLIGHT
X^X
^<
ZMILD

1 1
MARKED
DEGREES OF
AAOO
1 1
SEVERE EX
: HANDICAP
GUIDE
1 1 1
TREI
—

                                                                                60
                                                                                80
                                                                                    o
                                                                                    o
                                                                                    c
                                                                                    8
                                                                                    I
                                                                                    IU
                                                                            111
                                                                            OT
                                                                            DC
                                                                            o
                                                                            CO
                                                                            o
                                                                            oc
                                                                            O

                                                                            CQ
                                                                            0.
                                                                                100
    -5   0   5  10  15  20  25  30  35  40  45  50   55  60  65  70  75  80  85  90  95
          AVERAGE OF HEARING LEVELS AT 500,  1000. AND 2000 Hz (Earphones re ISO)

     I   I   I   I    I   I   I   I   I    I    I   I   I    I   I   I   I    I   I   I    I
     5  10   15  20  25  30  35  40  45  50  55  60   65  70  75  80  85  90  95  100  105

         AVERAGE  OF HEARING LEVELS AT  1000,  2000, AND 3000  Hz (Earphones  re ISO)

Figure 1.  Relations between HL (average at 500, 1,000, and 2,000 Hz and 1,000, 2,000 and 3,000 Hz) and
measures of speech understanding as calculated according to the Articulation Index (Al), impairment, and
handicap. Speech level measured in field 1 meter from talker. (After Kryter4**)
2,000 Hz, and percent hearing impairment and hearing handicap proposed by AAOO. It has
been shown by  most laboratory and clinical tests that the average of the hearing levels at
1,000, 2,000 and 3,000 Hz provides a better index to hearing impairment for speech than
does the average of 500 1,000 and 2,000 Hz. Fortunately, at least for statistical purposes, an
average difference of about 10 dB can be assumed between the hearing levels averaged for
500, 1,000 and  2,000  Hz and the average for 1,000 2,000 and 3,000 Hz; this is shown by
the lower abscissa in  Fig.  1.
     Attention is invited  to the fact that at  the lower fence  or threshold  of impairment
proposed by AAOO a person would be unable to correctly perceive individual speech sounds
and even some  sentences or  words whenever the speech level  was less than that normally
present (i.e. if the distance between the talker and listener was greater than about one meter
or if the source  of the speech was of a lower intensity than normal).  In addition, it is seen
that the person whose hearing level averaged 65 dB or more would be unable to perceive any
sentences of everyday speech, that specified by AAOO as the  speech signal for evaluating
impairment to hearing. According to AAOO, however, 100% impairment, the "upper fence"
of AAOO, does not occur until about 92 dB average hearing level.
    A particularly relevant study was reported by  Kell, Pearson, Acton and Taylor6 who
found that some 96  subjects, female weavers, whose HLs average 39 dB at 500, 1,000 and
                                         104

-------
2.000 Hz experienced some difficulty in speech communications under  many  conditions
(approximately  809f  of the weavers  reporting difficulty with  communications whereas
controls  persons of the same age but with normal hearing  for  that age about only  109?
reported  difficulty).  However,  according to  AAOO descriptions and  procedures,  these
persons would  have an impairment  of around 20''.  a  "slight handicap", and should have
"difficulty only with  faint speech". Figure 2 illustrates some relations between hearing level
and percent of sentences and other test materials heard as a function of the hearing level of
the subjects; it  is seen in Fig. 2 that when  the hearing level starts to exceed about 0 dB there
is some degradation in the ability of the subjects to perceive the speech material.7
   C
   u
   f-
   t
   -
   I
100
00
BO
70
50
X
30
10
0
-1
*o-»


o*.




o—
- X— —

S

C?

*



"""-x-.
*- --^,


HEE3



— o
	 X



\



^^V



>Y>
\>™—

AVERAGE HL AT
BOO. 1000. AND 2000 H
AVERAGE HL AT
1000, 20OO. AND 3000
I
ri
3=
~-u,


^
	 ,

I
•HI
^ SEN!
NO C
~ » 5 c
" — — x.
' 	 ^.

tf





ENCES,
JISTOR1
B S/N
***-^
riON

^-0
ALL TESTS
IPB Wordl §ne
AT DIFFEREf
AND WITH S(

— —




_*"7^~-1






SPEECH LEVEL
AT 99 dB
*^

S*nttnc«l)
JT S/N RATIOS
3ME FILTERIN*

1 	 >
















0-5 0 5 10 15 20 25 30 35 40 45 S
HEARING LEVEL — dB TA_, „,
    Figure 2. Intelligibility test scores as a function of the average hearing level, re ISO. at 500, 1,000,
    and 2,000 Hz and at 1,000, 2,000, and 3,000 Hz. Speech level at 95 dB. From Kryter7
      All in all, it would appear that the definitions of hearing and the upper and lower fence
 impairment to the hearing given by the AAOO are not consistent with present-day knowl-
 edge of the relation between puretone hearing levels and the ability to understand speech.
 particularly everyday speech.

                                          105

-------
     It must be emphasized that taking the ability to perceive everyday speech as the proper
criterion of what constitutes the sense of hearing is itself debatable, particularly when the
index to that ability is based on but several audiometric test frequencies. For example, the
ability to perceive sounds above 3,000 Hz, let alone 2,000 Hz, is a capacity that is most
helpful  in  perception of many meaningful sounds besides that of speech. It is also an
important frequency region for the maintenance of abilities to localize objects or sound
sources in space and as a safety guide to persons during locomotion. Perhaps most important
is that it is arbitrary to say that the person who loses sensitivity to weak sounds (that is,
losses up to the threshold of 25  dB of the "lower fence"), has suffered no impairment or
handicap inasmuch as in real life speech can  often be very faint or very weak due  to the
distance  between source and listener or because of weak signals coming from the source, or
due  to certain acoustic conditions. Therefore, the  person with normal hearing ability can
enjoy the perception of many weak sounds,  including speech, that are lost to the person
who falls below the "fence" that has been proposed for defining impairment to hearing.

                       Proposed New Upper and Lower Fences

     In spite of all the above factors it is perhaps of practical importance to use the ability
to perceive everyday speech as a criterion  for evaluating hearing impairment due to exposure
to noise. In any event, for present purposes, and keeping the above aforementioned reserva-
tions and conditions in mind, it is proposed that: (1) noise-induced hearing loss be evaluated
with respect to the criterion of impairment to  hearing for speech sentences heard one meter
from talker using normal conversational effort in the quiet, and as predicted by pure-tone
hearing levels averaged at 500,  1,000 and 2,000 Hz, or, preferably,  1,000 2,000 and 3,000
Hz; (2) the threshold of impairment be  taken as 15 dB for the average at 500, 1,000 and
2,000  Hz, or  25  dB at  1,000, 2,000 and 3,000 Hz; and (3) the  upper fence, or  100%
impairment, be taken as being reached at average HLs of 65 dB at 500, 1,000 and 2,000 Hz,
or 75 dB at 1,000, 2,000, and 3,000 Hz. We have taken the liberty of indicating on Fig. 1 a
linear relation between percent hearing  impairment for speech and the average of hearing
levels for three frequencies.

                          Hearing Levels as a Function of Age

     Before plotting data  that relates  the hearing level  of men who have been exposed to
various amounts of noise, and have therefore  suffered hearing impairment that exceeds the
thresholds shown in  Fig.  1, it is appropriate to determine what the population hearing-level
statistics are for people who have not  been exposed to appreciable amount of noise during
their careers. A logical and useful way  to express the effects of  noise on hearing is to
compare: (1) the percentage of the population exposed to noise whose hearing exceed the
lower fence, or start of hearing impairment as defined (or of some other hearing level value
that one may wish to choose) with (2) the percentage of people having hearing impairments
exceeding the specified hearing levels in the non-exposed population.
     Figure 3  depicts the results' of some studies and estimates of the prevalence  in the
population of HLs exceeding 25  dB, averaged at 500, 1,000 and 2,000 Hz, as a function of

                                         106

-------
            UJ
            tr
            "0 o
            So
                 90
                 80
                 70
                 60
   ~   50
i ;=
z 8
< i-
i <
3 S
CL C?
o <
in tr
a. nj
u. >
O <
            (J
            oc
                 40
                  30
20
                  10
         "NON-NOISE EXPOSED" ISO
                              R1999

         "NON-NOISE  EXPOSED" BAUGHN.
                              AFTER GLORIG
                              AND NIXON     /
          ISO  DRAFT  RECOMMENDATION
          TC431/SC-57E. OTOLOGICALLY
          DISEASED AND UNDUE-NOISE
          EXPOSED EARS EXCLUDED.

                       USPHS '60--62
                       BETTER EAR
                        "NON-NOISE
                        EXPOSED"
                                                                    /  „
                                               NON-NOISE
                                              EXPOSED" —\
                     NIOSH '72
                     "NON-NOISE"
                     EXPOSED
                                          AGE — years
                                                                  SA-5825-9R
             Figure 3. Percent of population with HLs of 26 dB or greater (aver, at 500,
             1,000 and 2,000 Hz) as a function of age and approximate years of
             exposure. Dashed  lines  are extrapolations.  Based on  Refs. Glorig and
             Nixon8; Corso9; Glorig and Roberts1 °; NIOSH'l; ISO/TC12.
age.  The 500,  1,000  and 2,000 Hz average is here used because it is the only measure
presented in some of the studies reported. It is seen in Fig. 3 that there appear to be two
sets of curves,  the upper .set being that of ISO R1999 and the non-noise-exposed data of
Glorig and Nixon as reported by Baughn.  Although the basis of the curve presented is not
given, it  would appear that  ISO  1999 was largely drawn from the Glorig and Ni^on  data
base.
                                          107

-------
     The data plotted by Baughn and labeled "non-noise" exposed men came from data
collected by Glorig and Nixon in various studies of hearing of industrial workers. Some
2,518 industrial workers, including  office workers, were asked questions regarding their
previous exposure to noise. Three hundred and twenty nine of the 2,518 men fell into what
Glorig and Nixon labeled the "non-noise"-exposed category. It is not clear from the pub-
lished literature how Baughn derived the  "non-noise"-exposed statistical data, nor whether
they represent all 2,518 men or the screened population of 329 men. Glorig and Nixon
comment that the two populations differed only at the extreme of the distributions. In any
event, Glorig and  Nixon's data must  be interpreted as  being  non-representative  of the
hearing levels of the general population not exposed to noise in the light of the other more
general survey data shown on Fig. 3, in my opinion.
     Some reservations, however, can also be expressed about each of the other functions
shown on Fig. 3: Corso's study of the hearing of a random screened  sample of men in a
small non-industrial town involves but 237 men; the draft ISO document  does not give
statistical data beyond the 25th percentile and median; the NIOSH study involved only 380
men and, further, considered men  working in less than 80-dBA noise to be in the "quiet";
the U.S. Public Health Survey, while  meeting the requirements of a large, randomly selected
population, did not screen the data for men who had been exposed to intense noise in their
work or who had suffered some otological disease.  In an attempt to at least partially remove
from the USPHS data such diseased ears were suggest that the data for the "better  ear" of
the men be used to represent "non-noise" exposed to hearing for a large population. We will
show later the data, which is not greatly  different, for the average of both ears in the U.S.
Public Health Survey.

                        Hearing Levels for Noise-exposed Men

     Baughn studied the hearing of  some 6,835 men who had been exposed for  various
numbers of years to various levels of continuous 8-hour-per-day industrial noise. Some of his
results are shown in Figs. 4  and 5, along with the data from the U.S. Public  Health Survey
for the "better-ear" and average-of-both-ears, as well as some findings from a  recent NIOSH
study.
     Several comments are in order. First, it is clear that the HLs of Baughn's men appear to
shown significant adverse effects as the result from working in noise at levels as low as 78
dBA after a number of years of exposure, with  the increase in incidence in men having
worse hearing dramatically greater for the 15 dB than for the 25 dB fence. Secondly, while
the data shown for the  NIOSH study are somewhat different from the Baughn data for the
lower noise levels—compare  80 dBA  of NIOSH with 78 dBA fqr Baughn-the data  for the
92-95 dBA noise conditions are quite comparable.  NIOSH comes to the conclusion, because
of their data for the 80 dBA noise as  compared with the hearing of "non-noise" exposed
men (defined by them as people working in less than 80 dBA noise) that 80 dBA noise, 8
hours per day will  not increase the incidence of people with HLs greater than either the 15
or 25 dB fence at 500, 1,000 and 2,000 Hz.
     This is so contrary to Baughn's findings, and to other NIOSH data, (Cohen,1 s) as shown
in Fig. 6 that, I think, the conclusions of NIOSH for the lower levels  of noise must be
seriously questioned. In that regard,  it  is interesting to note that the hearing  of the NIOSH

                                        108

-------
       DC
       HI

       <
       LU
       
-------
                          90
                         60
                       §

                      II"
0-50
                      if 0
                      o <
                      ui K
                      * 5
                         40
                      DC
                      s?
                         20
                             HL'i 16 dB
                             OR GREATER
                             (Average at 500
                            . 1000 and 200O H
                                         I
                                        STEADY
                                        NOISE
                                        8 HR/OAY
                                        (RigM Earl
                                        BAUGHN
USPHS
1960-62
BETTER EAR
(Includes
Nonr Ex-
posed and
DtKated
Ears)
                                             I
                         ALSO.
                         PROPOSED AS
                         GENERAL POPULATION
                         (Non-NotM Exposed)

                         	I	I
                           20
                                30
                                      40


                                       I
                       SO     60
                       AGE — v«en
                       I	I
    80

    I
                                                                    Ml
                                       20     30     40    SO
                                      APPROXIMATE YEARS OF EXPOSURE
                      Figure 5. Percent with HLs of 16 dB or greater (aver, at
                      500, 1000, and 2000 Hz) as a function of age and years
                      of exposure. Dashed lines are extrapolations. Based on
                      Baughn'sdata14.

data, were more substantial,-there being 314 men in the 90-94 dBA noise condition, for
example.
     Figure 7  shows  Baughn's and some NIOSH  data, plotted against noise level, in dBA,
with age or years of exposure as the parameter. Also, we have taken the liberty of extrapo-
lating the  data to lower noise levels, using as the non-noise exposed general population the
U.S. Public Health  Survey, better-ear data. Fig. 8 shows a quite  different picture of essen-
tially the same noise-exposed  data versus dBA noise exposure, the source of the difference
between Figs.  7  and 8 being due to the choice of the  "non-noise exposed" population data
chosen. Botsford1* uses  the  data of Glorig and  Nixon, as discussed above  in relation to
Baughn's  use of the same data, and data from a public health survey made by Glorig in a
highly industrialized  community  with  volunteer  subjects—survey  results which  Glorig
described  as being  significantly biased towards the inclusion of people with noise-induced
deafness.
                                            110

-------

0
V)
2
CD
^
z
UJ
W
_J
O
•I
Ul
T
-.
LJ
2

10
20

30

40

SO

60


70

PO


T)

.

-

.

.
O ttFfiPT mm ir* ^'* 4BA
w wvribt wnOUP
( M«]ion Eiperlcoct— 3>ri )
• - NON-EXPCSEO
i I i 1 1 1
     500  1000 2000 SOOO 4000 6000
                                 40

                                 50

                                 60

                                 7O

                                 80
                    O- m-PLANT GROUP
                                    85 DBA
                      ( Median Eiperienc* — 13 yn.)
                       N = 54
                    • - NON-EXPOSED
                    J	I	i	I	I	I
                                    500  lOOO 2000 3000 40OO SOOO
                                      TEST  FREOUEMCY  (Hz)
                     0

                    10

                    20

                    30


                    40

                    SO

                    60

                    70

                    60
                                                          O- IN-PLANT GROUP  66J "aA
                                                             ( Metfion E»p*fii">ee — '2/rt.l

                                                          • —NOM-EXPOSED
                                                           I	I	i    I	i	i
                                                   500  IOOO 2000  3000 4000 600O
          O
          ul
          S  20
             30
             40
O

E
LJ
i

<
Ul
             50

             60

             70

             80
O-IN-PLANTGHOUP   a86dBA
  I Mra.an E«j>ernnc« — I7jr». )
   N« 6
• -NON-EXPOSED
1	J	I  ..  I	1	
                500 IOOO iOOO 3000 4000 60M

                                      TEST
    10

    20

    30

    40

    50

    60

    70

    80
O-IN-PLANT GROUP  _!-%JJ!iS
  ( Median Eipentnc* —  3in.)
   Nr 20
• -NON-EXPOSED
           l',00

FREQUENCY (Hi)
                                                3000 4COO  «iOOO
                                                                                      TA-M2S-13
Figure 6. Mean hearing levels (re ISO) for paper bag workers in different job locations compared with
non-noise-exposed groups equated in  number, age, and sex composition. From Cohen, Anticaglia and
Jones.15


                                         Conclusion

     Standards, guide lines, etc. are, of course, no better than the validity of the data and
definitions  on  which they are based. It would  appear  that the standards  and guidelines
proposed on these  matters by AAOO,  ISO  and  NIOSH are not predicated on  the more
significant and relevant  sources  of presently available data, but rather have consistently
followed the results of rather unreliable, small population studies and rather arbitrary and
unrealistic definitions of hearing that have all  tended to lead to a significant underestimation
of the damaging effects of noise on hearing.
                                             Ill

-------
                CC
                HI
                     80
                UI
                cc
                  •= 70
                OC
                0
^8


^5
  o
  UJ
  o
                LU
                     60
                     50
                     40
                     30
                     20
                  ui
                  >
                I- -  10
                01
                U
                
-------
                   70
                   60
               O  50
               U
               g
                h  30
                Z
                u
                U
                Q<  20
                U
                CL

                    10












^^^^^>





A AGE 50- 59
A AGE 40-49
0 AGE 2.0-39
• AGE 20-29




i
'

i
A






4.
\

I
•




1

1







^

^^



/
x

X7
>-
, ^t_-«
	 ;
r






.
/
^

A y






V
/

V
^
A
, ,f
*0


•


* V
.
r
A


/
^
o/
Xo

^»


/
^


/


/

s


A  A  A
    *   " z
     w    5
                                           65
                               105
                                             SOUND  LEVEL
                                             AT WORK, DBA
                           i
                          z
                          0
                          Z
I    UJ

O   UJ
Z   O
                Figure 8. Percent of population with HLs of 26 dB or greater
                (aver, at 500. 1.000 and 2,000 Hz) of workers in steady industrial
                noise, and of  "non-noise and general  population groups" (from
                Botsford1").


5.  Kryter. K. D.. "Impairment to Hearing from I-.xposure to Noise," J. Acoust. Soc. Am.
   53. May 1973.
6.  Kell.  R. L. J. ('. (i. Pearson. \V. I. Acton, and \V. I;i\ lor in Occupational I/carinx /Lo55
   D. W. Robinson  (1 d. I. (Academic Press. London. 1971) pp. 179-191.
7.  Kryter. K. D. ( 1963). "Hearing Impairment tor Speech." Arch. Otolaryn. 77. 598-602.
8.  Glorig.  A., and  J. Nixon ( |9(iQ).  "Distribution of  Hearing Loss in Various Popula
   tions."  Anals of Otology and Laryngology d9. 2. 497-516.
9.  Corso. J. F. (1963). "Age and Sex Differences in Pure-Tone Thresholds." Arch. Otolar
   yngol. 77. 385-405.
                                          13

-------
                                          Table 1.


Table 1. Results of questionnaire survey of 96 female weavers with average HL's at 500,1000, and 2000Hz of 39dB and a
control group of same age with average HL's of 16 dB. From Kell, Pearson, Acton and Taylor.*

'The social consequences of this impaired hearing ability were:
     (a)   difficulty at public meetings (weavers 72%, controls 5%)
     (b)   difficulty talking with strangers (weavers 80%, controls 16%)
     (c)   difficulty talking with friends (weavers 80%, controls 15%)
     (d)   difficulty understanding telephone conversation (weavers 64%, controls 5%)
     (e)   81% of all weavers considered that their hearing was impaired (5% controls)
     (0   9% of weavers and no controls owned hearing aids
     (g)   53% of weavers and no controls used a form of lip-reading."
 10.  Glorig, A., and J. Roberts (1965). "Hearing Levels of Adults by Age and Sex, United
     States 1960-1962," National Center for Health Statistics,  Series 11, Number 11. U.S.
     Department of Health, Education and Welfare, Public Health Service, Washington, D.C.
 11.  "Occupational  Exposure  to Noise," National Institute for Occupational  Safety and
     Health (NIOSH), U.S.  Dept. of Health, Education and  Welfare,  Washington, D.C.,
     1972.
 12.  Draft Proposal for Hearing Levels of Non-noise Exposed People  at  Various Ages,
     ISO/TC 43/SC-l (Netherlands) 57E.
 13.  Baughn,  W. L. (1966). "Noise Control - Percent of Population Protected," Inti. Audio.
     5,331-338.
 14.  Baughn,  W. L. (unpublished). "Percent of Population Having HLs Greater than 16 dB
     and 26 dB as Function of Years of Exposure in Steady Noise."
 15.  Cohen, A., Anticaglia, J. R. and  Jones, H. H., "Noise Induced Hearing Loss-Exposures
     to Steady-State Noise." Presented at the  American Medical Association Sixth Congress
     on Environmental Health,  Chicago, Illinois, 1969.
                                           114

-------
           THE INCIDENCE OF IMPAIRED HEARING IN RELATION TO
            YEARS OF EXPOSURE AND CONTINUOUS SOUND LEVEL
                  (PRELIMINARY ANALYSIS OF 26,179 CASES)

                                     A. Raber
                 Allgemeine Unfallversicherungsanstalt, Wien, Austria

    Impaired hearing caused by occupational noise is in  Austria a compensable disease.
Hearing conservation consequently belongs to the legal obligations of the Accident Branch
of the  Social  Security Board. Since 1962, our institution, the Allgemeine Unfallversicher-
ungsanstalt, has been performing mass-audiometric investigations in noisy factories.
    For  this  purpose, we use specially-constructed busses with audiometric booths. Our
three audiometric teams have now taken about  165,000 pure tone audiograms.
    The  main objectives of our investigations are:
    1.   To single out  persons whose noise-induced hearing loss is significantly  above
         average, considering the duration and intensity of their particular noise immission;
    2.   To identify  persons suffering from  ear conditions which make  them unfit for
         further work in a noisy environment;
    3.   To check the audiograms in order to see if they  indicate  such a degree of NIHL
         that the investigated persons might be entitled to compensation by law.
    Audiometric investigations are provided for all people who are exposed at their work-
ing places to a sound level of 80 dB(A) or higher. All apprentices are tested irrespective of
their noise-exposure situation at the moment
    The  workers come directly from their working places to our busses. The interval
between the end of noise exposure and the audiometric test is 20 minutes  on the average.
The tests are  performed by  trained and  experienced  Industrial Audiometrists. Otological
investigations  are not carried out  on this occasion. The response rate is 80 to 85%. Non-
participation is generally due to  absence from work because of vacation or sickness. Persons
refuse the test only very rarely.
    The  quality of the audiograms is checked continuously.
    We have collaborated for  many years with some otologists  to whom we send the
following groups of people:
    1.   Persons with remarkable rapidly  developing NIHL;
    2.   Persons whose  social hearing might already be damaged by NIHL;
    3.   Persons in whom the form of the audiogram calls for further otological investiga-
         tion.
    The  otologist repeats the audiometric test and  performs the necessary clinical investi-
gations. On an  average we find a good conformity between our field work and the audio-
grams taken by the otologists.
    Periodic investigations in the factories are repeated every 3-5 years. The serial audio-
grams are performed without knowledge of former results.
    One member of our team is an otologist. He classifies all audiograms- after a preselec-
tion by the computer— and decides which worker has to be sent for a clinical checkup. On
this occasion he examines also the serial audiograms for significant differences which could
have been caused for instance by TTS or imperfect  cooperation of the investigated person.

                                        115

-------
     More details about the organization of audiometric investigations and determination of
noise levels in Austrian industry has been published by Surbock (1).
     In the following report, the data of our routine mass-audiometry are used to relate the
incidence of impaired hearing to the noise immission.
     From all of our audiometric data we extracted those audiograms that met the follow-
ing criteria:
     1.   The person works (nearly) the whole shift at one place;
     2.   At his working place, a continuous noise was measured;
     3,   The person got his  noise immission mainly in the industry in which he was
         working at the time of the investigation; and
     4.   In  me audiogram, the difference  between air and bone conduction is at no fre-
         quency 15 dB or more.
     Using these criteria we got a sample consisting of the data of 18,059 men and  8,120
women.
     Each data record includes the following information:
      1.  Sound level (10 classes:  less or equal 80 dBA, 85 dBA, 90 dBA, 92.5 dBA, 97
         dBA,  100 dBA,  103.5 dBA, 107 dBA, 110 dBA, 115 dBA).
      2.  Years  of exposure (according to the person's report).
      3.  Sex
      4.  Year of investigation
      5.  Age
      6.  Audiogram of the right ear (10 frequencies)
      7.  Industry in which the person was mainly exposed
      8.  Years of exposure according to ISO-Recommendation R 1999 (Years of exposure
         = Age-18)
      9.  Ear diseases yes/no
     10.  Ear trauma yes/no
     The data were subdivided in classes and subclasses according to this hierarchy:
      1.  Sex (male  or female)
      2.  Method  of determining years of exposure (according to ISO or  according  to the
         report of the person)
      3.  Years  of exposure (10 classes)
      4.  Sound level (10 classes)
     For each of these (2x2x lOx 10=) 400 "cells" of hearing losses we computed:
      1.  Number and  percentage  of right ears whose mean Hearing Level at 0.5, 1.0 and
         2.0 kHz is 25  dB or greater.
      2.  Number and  percentage  of right ears whose mean Hearing Level at 0.5, 1.0 and
         2.0 kHz is 15  dB or greater.
      3.  The 5, 10, 25, 50, 75, 90, 95 centiles of the hearing losses at 0.125, 0.25, 0.5, 1.0,
         1.5, 2.0, 3.0,4.0, 6.0, 8.0 kHz.
      4.  Mean, standard deviation, and mean square deviation of the hearing losses.
     In Table 3 of ISO Recommendation 1999 (Assessment of Occupational Noise  Expo-
sure for  Hearing Conservation Purposes,  1st Edition May 1971), the total percentages of
people with impairment  of hearing  for conversational  speech in relation to continuous
sound level and  years of exposure are to be found (2).

                                       116

-------
    In a note to that Table 3 it is mentioned that the values of the table are based on the
limited data available at  that time and that  they are subject  to revision when results of
further research become available.
    In the above-mentioned ISO Recommendation, the definition of impairment of hearing
for conversational speech is in agreement with the AAOO regulation (3). That is, the hearing
of a subject is considered to be impaired if the arithmetic average of the permanent thresh-
old hearing levels of the subject at 500, 1,000 and 2,000 Hz is 25 dB or more compared
with the corresponding average given in ISO-Recommendation R 389, Standard reference
zero for the calibration of pure-tone audiometers.
    As the first step of the evaluation of our data, we compared the percentages of subjects
with impaired hearing in our material with those predicted by the ISO Recommendation.
    The main results are summarized in Table 1 and Table 2.
    Out of the 18,059  men and 8,120 women of our sample, 16,726 males and 7,529
females were working under noise-immission conditions which are registered in Table 3 of
ISO Recommendation  1999. Applying  this table to our data,  it would be expected that
there would be 3,244 males and 1,452 females with impaired hearing in our sample. However,
if one takes the exposure years according to the persons report instead of computing them
with the formula "Years of Exposure = Age - 18", the forecast would be 1,883 males and
743 females with impaired hearing in our sample.
    We actually observed only 851  males and 142 females with impaired hearing. In other
words, relative  to the  number of expected cases, we found in the male population only

                                     Table 1.
                                                  male
                           female
     Sample
              16.726
             7525
     Cases with impaired hearing
     predicted

     (IS01999)
Age-18
Anamn.
3.244
1883
U52
 743
           observed (AUVA)
                 851
              142
                                    117

-------
                                       Table 2

t
b

-------
    100
 I 80
  o
  Q>
     60
  o


  t 40
     20
Years of

Exposure
     ISO Rec.1999

     AUVA
                        \   1
                        35    45




725.665.761.698.695.724.644.391.468.92


           Figure 1
                     119

-------
    100




 •g 80


 I

 "g 60
 • ^^
 o
 a

 •§ 40
    20
85CIBA

ISORec.1999 85dBA

ISORec.1999 ^80dBA

AUVA 85dBA

AUVA
            r	r	r	\	\	\	\	\	r
Years of

Exposure

Coses of
 AUVA *08J555.654.640.m57640&224225.27
                 Ftyun 2
                  120

-------
             90dBA
Years of
Exposure
15    25    35    45
         165.352.462.423.465386290181.17714.
                   Figure3
                    121

-------
100
 60
 20
         954BA
              15    25    35    45
Years of

Exposure

Coses of
 AUVA  121.248.30&279.264.2I6.175M 104. 13.
               Figure 4
                122

-------
  700
Q>
   60
Q.
.i40
*20
           974BA
Years of
Exposure
                 75    25     35    45
       32. 131.194. 200.193.128.106. 62. 58. 8.
                 Figure 5
                   123

-------
   100
•f 80
 CD
   60
NO 20
«N
Years of

Exposure


Cases of

 AUVA
              100 dBA
                  15    25    35    45
           K5.147138.128.119. 82.37. 48.  7.
                  Figure 6
                   124

-------
  700
2>
•| 80
o
0*
   60
.1.
o
9.

•i 40
I 20
Years of

Exposure

Cases of

 AUVA
            105dBA
                 15    25    35    45
          - 75 104-97 9ft ^ 7a
                  Figure 7
                   125

-------
700
         WdBA
Sao-
fe
"S 60
*6
.1 40
^
JS
*»•
/ears of
Exposure
'•^^ » •«M«l^rV •
I ,
1
!
! j
'
[ , '
T Tyf— ^"^ A
- 	 d i_ i •
f- f r [ 1 1 1 1 1 i
5 15 25 35 45,
Cases of
AUVA ft 75- 36 52 ^ 55- 2ft 72 72 a
              FigureB
               126

-------
    100
 •  80
  6
  Q>

 "g 60
    20
           WdBA
Years of
Exposure
Cases of
 AUVA
      15    25    35    45
17 29. 36. 36. 36. 15. 16. 16.  2.
                    Figure 9
                     127

-------
    100
S
fe
Q>
     80
   50
^^
o
&

S 40
   20
         115CIBA
/ears of
Exposure

Cases of
 AUVA
                 15    25    35    45
                  I  4.  2.  3.  2. 2  7.


                  Figure 10
                   128

-------
Table 3
25 dB-criterion years
of exposure according
to ISO; males.

<80dB
85 dB
90 dB
(AUVA 92)
95 dB
97 dB
lOOdB
105 dB
(AUVA 103)
107 dB
110 dB
115dB
ISO%
AUVA%
Nt.
ISO%
AUVA%
Nr.
ISO%
AUVA%
Nr.
ISO%
AUVA%
Nr.
ISO%
AUVA %
Nr.
ISO%
AUVA%
Nr.
ISO%
AUVA%
Nr.
ISO%
AUVA%
Nr.
ISO%
AUVA%
Nr.
ISO%
AUVA%
Nr.
PERCENTAGES
YEARS OF EXPOSURE
0
1
1
725
1
0
408
1
1
165
1
0
121

0
32
1
0
39
1
0
49

0
9
1
0
6
1
0
1
5
2
1
665
3
1
555
6
0
352
9
1
248

2
131
14
1
145
20
0
75

0
15
28
0
17
38
0
7
10
3
1
761
6
1
654
13
2
462
20
1
308

1
194
32
3
147
45
0
104

3
36
58
0
29
74
17
6
15
5
1
698
10
1
640
19
4
423
29
3
279

1
200
42
2
138
58
4
97

13
32
76
6
36
88
0
9
20
7
2
695
13
4
601
23
5
465
35
3
264
no en
1
193
49
7
128
65
8
98
no en
13
31
85
6
36
94
75
4
25
10
3
724
17
5
576
26
7
386
39
6
246
ties
8
128
53
11
119
70
12
83
ries
15
33
88
6
36
94
0
2
30
14
7
644
22
10
408
32
13
290
45
14
175

12
106
58
13
82
76
17
70

18
28
91
20
15
95
33
3
35
21
9
391
30
15
224
41
15
181
53
16
81

11
62
65
8
37
82
22
37

33
12
93
31
16
96
50
2
40
33
16
468
43
21
225
54
19
177
62
26
104

26
58
74
38
48
87
21
34

25
12
95
44
16
97
100
2
45
50
8
92
57
11
27
65
14
14
73
15
13

13
8
83
29
7
91
50
8

0
0
95
0
2
97
0
1
                                              5863
                                              4318
                                              2915
                                              1839
                                              1112
                                               890
                                               655
                                               208
                                               209
                                                37
                                              18046
1555  2210   2701   2552   2515   2333   1821   1043  1144    172
 129

-------
                    Table 4
25 dB-criterion years of
exposure according to
anamnesis; females.
< 80 dB
85 dB
90 dB
(AUVA 92)
95 dB
97 dB
100 dB
105 dB
(AUVA 103)
107 dB
llOdB
115dB
ISO%
AUVA%
Nr.
ISO%
AUVA%
Nr.
ISO%
AUVA%
. Nr.
1SO%
AUVA%
Nr.
ISO%
AUVA%
Nr.
ISO%
AUVA%
Nr.
1SO%
AUVA%
Nr.
ISO%
AUVA%
Nr.
ISO%
AUVA%
Nr.
ISO*
AUVA%
Nr.
PERCENTAGES
YEARS OF EXPOSURE
0
1
2
2653
1
1
1062
1
1
622
1
1
400

0
236
1
1
227
1
1
162

0
22
1
4
24
1
0
11
5
2
3
1046
3
2
859
6
2
514
9
3
354

2
203
14
3
189
20
5
150

2
41
28
3
39
38
14
7
10
3
4
764
6
2
691
13
5
499
20
3
347

3
206
32
8
142
45
4
102

5
37
58
6
31
74
0
3
15
5
6
572
10
6
611
19
8
442
29
4
273

5
173
42
9
105
58
2
64

14
44
76
8
39
88
25
8
20
7
9
394
13
6
560
23
10
431
35
12
227
no en
12
135
49
12
109
65
12
73
no en
19
33
85
13
40
94
25
4
25
10
7
300
17
12
386
26
11
300
39
12
164
tries
9
114
53
18
82
70
25
73
ries
27
22
88
13
23
94
100
2
30
14
12
94
22
20
110
32
23
77
45
21
52

9
33
58
15
26
76
33
18

50
6
91
29
7
95
0
0
35
21
8
36
30
25
28
41
8
24
53
33
12

14
7
65
44
9
82
75
8

100
1
93
60
5
96
100
1
40
33
22
9
43
22
9
54
38
8
62
25
8

50
6
74
50
4
87
0
4

50
2
95
100
1
97
100
1
45
50
0
0
57
33
3
65
0
0
73
0
2

0
0
83
0
0
91
100
1

0
0
95
0
0
97
0
0
5419  3402   2822  2331   2006  1466   423    131
52
                                                                   5868
                                                                   4319
                                                                   2917
                                                                   1839
                                                                   1113
                                                                    893
                                                                    655
                                                                    208
                                                                    209
                                                                     37
18058
                      130

-------
                   Table 5
25 dB-criterion years of
exposure according
ISO female.
< 80 dB
85 dB
90 dB
(AUVA 92)
95 dB
97 dB
100 dB
105 dB
(AUVA 103)
107 dB
llOdB
115dB
ISO%
AUVA%
Nr.
ISO%
AUVA%
Nr.
ISO%
AUVA %
Nr.
ISO%
AUVA%
Nr.
ISO%
AUVA%
Nr.
ISO%
AUVA%
Nr.
ISO%
AUVA%
Nr.
ISO%
AUVA %
Nr.
ISO%
AUVA %
Nr.
ISO%
AUVA%
Nr.
PERCENTAGES
YEARS OF EXPOSURE
0
1
0
225
1
0
234
1
0
141
1
0
104

0
48
1
0
25
1
0
30

0
9
1
0
1
1
0
0
5
1
1
313
3
0
305
6
1
149
9
0
88

0
73
14
0
22
20
2
41

0
6
28
0
0
38
0
0
10
3
0
273
6
1
267
13
1
133
20
1
88

0
53
32
3
37
45
0
40

10
10
58
0
1
74
0
0
15
5
1
244
10
0
257
19
1
145
29
1
78

2
55
42
3
44
58
4
26

0
5
76
0
0
88
0
0
20
7
1
270
13
0
277
23
1
194
35
1
108
noei
1
67
49
2
55
65
2
46
no er
33
9
85
0
1
94
0
0
25
10
3
327
17
2
300
26
2
212
39
1
114
tries
3
80
53
0
41
70
11
54
tries
10
10
88
50
2
94
0
0
30
14
4
361
22
1
342
32
3
218
45
3
115

0
67
58
2
59
76
6
52

17
6
91
0
0
95
0
1
35
21
3
183
30
4
196
41
4
142
53
4
50

0
41
65
6
34
82
21
14

0
3
93
0
0
96
0
0
40
33
4
141
43
4
131
54
8
71
62
9
43

8
38
74
14
22
87
31
13

75
4
95
0
0
97
0
0
45
50
0
10
57
33
9
65
0
6
73
0
2

33
3
83
0
2
91
0
0

0
0
95
0
0
97
0
0
                                                                  2347
                                                                  2318
                                                                  1411
                                                                   790
                                                                   525
                                                                   341
                                                                   316
                                                                    62
817    997    902   854   1027  1140   1221    663   463
32
8116
                    131

-------
                                        Table 6
25 dB-criterion years of
exposure according
anamnesis female.
< 80 dB
85 dB
90 dB
(AUVA 92)
95 dB
97 dB
100 dB
105 dB
(AUVA 103)
107 dB
UOdB
115 dB
ISO%
AUVA%
Nr.
ISO%
AUVA%
Nr.
ISO*
AUVA%
Nr.
ISO%
AUVA%
Nr.
ISO%
AUVA%
Nr.
ISO%
AUVA%
Nr.
ISO%
AUVA%
Nr.
ISO%
AUVA%
Nr.
ISO%
AUVA%
Nr.
ISO*
AUVA%
Nr.
PERCENTAGES
YEARS OF EXPOSURE
0
1
1
1344
1
0
713
1
1
306
1
0
211

0
154
1
0
77
1
0
50

0
11
1
0
2
1
0
0
5
2
1
408
3
2
529
6
1
338
9
1
189

1
123
14
0
73
20
2
89

0
15
28
0
1
38
0
0
10
3
3
222
6
1
352
13
1
265
20
2
116

3
73
32
2
59
45
2
47

10
10
58
0
0
74
0
0
15
5
5
152
10
0
266
19
3
175
29
1
91

2
57
42
2
52
58
3
31

25
8
76
0
0
88
0
0
20
7
6
144
13
3
232
23
4
163
35
4
89
no en
5
57
49
9
32
65
11
46
no en
10
10
85
50
2
94
0
1
25
10
2
53
17
4
141
26
5
110
39
5
58
tries
0
39
53
6
32
70
14
35
JKS
50
4
88
0
0
94
0
0
30
14
5
21
22
2
52
32
8
36
45
5
21

6
16
58
10
10
76
27
11

67
3
91
0
0
95
0
0
35
21
0
4
30
4
28
41
9
11
53
0
10

0
3
65
20
5
82
33
6

100
1
93
0
0
96
0
0
40
33
0
3
43
20
5
54
0
6
62
0
5

0
3
74
0
1
87
0
1

0
0
95
0
0
97
0
0
45
55
0
0
57
0
0
65
0
1
73
0
0

0
0
82
0
0
91
0
0

0
0
95
0
0
97
0
0
                                                                                  2351
                                                                                  2318
                                                                                  1411
                                                                                   790
                                                                                   525
                                                                                   341
                                                                                   316
                                                                                   62
                      2868  1765   1144   832   776    472
170
68
24
8120
     One reason for the poor correspondence between the predicted and observed number
of cases with impaired hearing could be an underrepresentation of subjects with advanced
noise-induced hearing losses in our material. An answer to this question may be given by the
distribution of hearing losses, especially at the higher frequencies, after various noise immis-
sions.
     Tables 7 to 16 show the distribution of hearing losses for the 5, 10, 25, 50, 75, 90, and
95 percentfles.
                                         132

-------
Table 7
       SPL: <80
       Years of exposure: 27-32
       Nr. of cases:  644

128 Hz
256 Hz
512 Hz
1024 Hz
1448 Hz
2048 Hz
2896 Hz
4096 Hz
5792 Hz
8192 Hz
COS
5.4
5.4
5.4
5.4
5.5
5.9
7.6
10.5
14.4
11.6
CIO
5.9
5.9
5.9
5.9
6.1
7.0
10.2
15.5
17.6
15.7
C25
7.2
7.3
7.3
7.4
7.8
10.0
16.5
25.8
26.3
24.3
C50
9.5
9.6
9.6
9.9
11.2
16.8
25.9
37.5
38.8
37.8
C75
14.5
14.4
14.0
14.3
16.7
25.3
38.5
53.7
57.1
57.7
C90
18.4
18.3
18.1
18.6
23.1
35.4
54.9
69.8
77.3
81.4
C95
19.8
19.7
19.7
21.5
28.4
43.6
63.9
79.3
89.0
95.9
Table 8
       SPL:  85
       Years of exposure: 27-32
       Nr. of cases:  408

128 Hz
256 Hz
5 12 Hz
1024 Hz
1448 Hz
2048 Hz
2896 Hz
4096 Hz
5792 Hz
8192 Hz
COS
5.4
5.4
5.4
5.5
5.5
6.1
7.9
11.5
14.0
11.8
C 10
5.9
5.9
5.9
6.0
6.1
7.3
11.0
16.3
18.2
15.9
C25
7.3
7.3
7.3
7.5
7.9
11.0
17.4
26.1
27.5
25.2
C50
9.5
9.6
9.6
10.1
11.5
17.5
26.9
39.2
40.0
38.5
C75
13.9
13.8
13.8
14.6
17.3
26.8
42.4
58.0
59.6
59.7
C90
17.9
17.9
18.1
19.1
24.2
40.1
59.6
75.0
82.1
86.2
C95
19.4
19.6
19.9
24.4
34.0
54.3
67.9
79.7
89.9
96.5
  133

-------
Tables
       SPL:  92
       Years of exposure:  27-32
       Nr. of cases: 290

128 Hz
256 Hz
5 12 Hz
1024 Hz
1448 Hz
2048 Hz
2896 Hz
4096 Hz
5792 Hz
8192 Hz
COS
5.4
5.4
5.5
5.5
5.6
6.3
8.0
13.5
15.9
12.7
CIO
5.9
5.9
5.9
6.0
6.3
7.6
11.0
17.5
19.0
17.0
C25
7.1
7.2
7.3
7.4
8.1
11.8
18.4
27.4
28.6
26.9
C50
9.3
9.3
9.6
9.9
12.3
18.5
28.8
45.3
45.5
43.4
C 75
14.8
14.5
14.4
15.0
18.5
28.8
46.9
59.8
66.1
66-9
C90
19.0
18.6
18.9
20.0
27.2
45.9
65.0
75.6
83.8
90.0
C95
22.5
20.0
22.3
24.8
31.9
55.3
71.3
87.1
105.2
97.3
Table 10
       SPL:  95
       Yean of exposure:  27-32
       Nr. of cases:  175

128 Hz
256 Hz
5 12 Hz
1024 Hz
1448 Hz
2048 Hz
2896 Hz
40% Hz
5792 Hz
8192 Hz
COS
5.5
5.5
5.4
5.4
S.6
6.1
11.3
17.5
18.8
20.0
CIO
6.0
6.0
5.8
5.9
6.2
7.1
13.8
21.1
23.9
23.4
C25
7.4
7.4
7.2
7.3
8.1
10.8
19.7
31.6
31.6
31.6
C50
9.8
9.7
9.4
9.7
12.7
20.9
36.9
52.5
54.4
56.3
C75
14.7
15.0
14.6
15.0
22.3
36.5
54.4
67.2
74.7
77.8
C90
18.8
18.7
19.2
20.4
29.4
49.2
67.3
77.4
83.6
96.0
C95
21.0
19.9
23.5
28.3
38.7
57.4
73.7
86.2
88.0
98.0
  134

-------
Table 11
       SPL:  97
       Years of exposure:  27-32
       Nr. of cases:  106

128 Hz
256 Hz
5 12 Hz
1024 Hz
1448 Hz
2048 Hz
2896 Hz
4096 Hz
5792Hz
8192 Hz
COS
5.4
5.4
5.4
5.4
5.6
7.0
11.1
15.8
14.1
12.9
CIO
5.8
5.8
5.9
5.9
6.1
9.1
15.2
19.1
21.5
17.3
C25
7.1
7.1
7.1
7.2
7.8
12.9
19.6
29.8
35.6
33.5
C50
9.1
9.2
9.3
9.3
11.3
20.0
32.9
48.5
50.0
47.7
C75
13.9
13.9
13.6
13.9
18.9
29.1
49.6
63.5
66.5
69.2
C90
18.6
18.8
18.2
18.7
28.9
47.4
59.3
78.4
87.8
88.4
C95
23.5
22.8
21.8
22.8
36.8
54.5
69.5
87.8
93.4
96.7
Table 12
        SPL:  100
        Years of exposure:  27-32
        Nr. of cases: 82

128 Hz
256 Hz
512 Hz
1024 Hz
1448 Hz
2048 Hz
2896 Hz
40% Hz
5792 Hz
8 192 Hz
COS
5.5
5.4
5.4
5.5
5.6
6.1
10.2
15.9
16.4
17.8
CIO
6.2
5.9
5.9
6.1
6.4
7.6
15.4
19.3
25.2
22.7
C2S
8.1
7.6
7.6
8.0
8.8
12.8
19.5
31.3
35.6
29.3
C50
11.5
10.5
10.5
11.5
14.4
21.7
35.7
48.3
51.0
48.8
C75
14.8
15.2
15.4
16.2
20.3
30.4
49.6
65.4
68.1
63-8
C90
18.9
19.0
18.8
19.4
29.7
47.3
64.0
76.5
76.8
77.7
C95
21.5
21.2
20.0
24.8
38.6
54.8
68.6
79.9
89.5
84.5

-------
Table 13
       SPL: 103
       Years of exposure: 27-32
       Nr. of cases:  70

128 Hz
256 Hz
5 12 Hz
1024 Hz
1448 Hz
2048 Hz
2896 Hz
4096 Hz
5 792 Hz
8 192 Hz
COS
5.4
5.6
5.4
5.6
5.8
7.2
11.5
16.9
16.3
16.3
CIO
6.0
6.1
6.1
6.3
6.7
9.4
15.0
25.0
19.2
22.5
C25
7.8
7.8
7.9
8.2
9.2
15.2
24.2
36.1
38.8
36.3
C50
11.2
11.1
11.5
11.9
15.0
22.7
34.4
50.0
55.6
55.0
C75
15.6
15.9
16.3
16.6
22.5
34.6
56.1
66.1
68.8
67.9
C90
20.0
19.6
20.0
22.5
28.9
48.8
66.7
77.5
85.0
92.5
C9S
26.9
27.5
30.6
37.5
45.8
53.8
69.6
85.6
88.5
97.1
Table 14
       SPL: 107
       Years of exposure: 27-32
       Nr. of cases: 28

128 Hz
256 Hz
5 12 Hz
1024 Hz
1448 Hz
2048 Hz
2896 Hz
4096 Hz
5 792 Hz
8192 Hz
COS
5.5
5.5
5.6
5.8
6.4
10.7
12.0
22.0
22.0
16.0
CIO
5.9
6.0
6.2
6.6
7.8
13.0
19.0
29.0
29.0
19.5
C25
7.3
7.5
7.9
8.9
12.0
18.8
27.0
37.5
38.3
32.5
C50
9.7
10.0
11.7
13.6
16.8
26.3
41.3
55.7
47.5
47.0
C75
15.0
15.0
17.1
17.8
20.0
35.0
56.7
65.0
58.3
65.0
C90
21.0
19.2
21.0
25.5
36.0
56.0
75.5
83.0
75 J
73.0
C95
28.0
28.0
28.0
29.0
43.0
68.0
79.0
93.0
79.0
93.0
  136

-------
Table 15
       SPL: 110
       Years of exposure:  27-32
       Nr. of cases: 15

128 Hz
256 Hz
5 12 Hz
1024 Hz
1448 Hz
2048 Hz
2896 Hz
4096 Hz
5792 Hz
8192 Hz
COS
5.6
5.6
5.9
6.3
6.9
11.9
13.8
18.8
23.8
18.8
CIO
6.3
6.3
6.9
7.5
8.8
13.8
16.3
22.5
27.5
21.3
C25
8.1
8.1
9.7
11.3
12.9
19.4
33.8
38.8
38.8
28.8
C50
13.8
13.8
14.4
16.3
21.3
36.3
57.5
63.8
66.3
58.8
C75
17.7
19.1
19.1
19.4
28.8
39.4
68.1
77.1
77.1
78.1
C90
19.6
26.3
28.8
36.3
46.3
57.5
77.5
82.5
82.5
88.8
C95
21.3
28.1
31.3
38.1
48.1
71.3
86.3
96.3
96.3
91.3
Table 16
        SPL: 115
        Years of exposure:  27-32
        Nr. of cases: 3

128 Hz
256 Hz
512 Hz
1024 Hz
1448 Hz
2048 Hz
2896 Hz
4096 Hz
5 792 Hz
8 192 Hz
COS
5.5
5.5
5.5
5.5
5.5
30.5
. 55.5
65.5
55.5
55.5
CIO
6.0
6.0
6.0
6.0
6.0
31.0
56.0
66.0
56.0
56.0
C25
7.5
7.5
7.5
7.5
7.5
32.5
67.5
67.5
57.5
57.5
C50
10.0
10.0
10.0
10.0
10.0
35.0
70.0
70.0
60.0
60.0
C75
17.5
17.5
17.5
17.5
27.5
57.5
77.5
77.5
82.5
82.5
C90
19.0
19.0
19.0
19.0
29.0
59.0
79.0
79.0
84.0
84.0
C95
19.5
19.5
19.5
19.5
29.5
59.5
79.5
79.5
84.5
84.5
  137

-------
     Data are shown for men with 30 years of exposure (calculated according to ISO) to
various sound pressure levels.
     Table 7, for example, demonstrates the situation in 644 men with 30 years of exposure
at sound pressure levels up to 80 dBA. Examination of the 75 percentile data shows that
subjects with severe hearing losses in the higher frequencies are rather common.
     For a second example, consider the sample of  70  men who were exposed 30 (ISO)
years to 103.5 dB(A) (table 13).
     At least one quarter of these persons show pathological audiograms of the advanced-
noise-damage type. In this group 17% meet the 25-dB criterion; according to ISO it should
be 62%.
     In countries in  which compensation is provided for ear injuries caused by occupational
noise, the frequency of persons with impairment of hearing, found at mass audiometric
investigations in noisy industries, is of practical interest, because this  frequency will to a
considerable extent determine the number of cases of compensation.
     So our  experience might also be of some practical interest to legislative and administra-
tive boards.

                                    References

(1)  Surbdck, A.R., 1971. In "Occupational Hearing  Loss." D.W. Robinson, Ed. Academic
     Press, London and New York.
(2)  International Organization for Standardization (1970). "Assessment of Noise Exposure
     during  Work for Hearing Conservation Purposes"  ISO Draft Recommendation  DR
     1999. Geneva
(3)  Committee on  Conservation  of Hearing  (1964) Guide for conservation of hearing
     impairment Trans. Amer. Acad. Ophthal. Otolaryng., 63, 236.
                                        138

-------
                       SOME EPIDEMIOLOGICAL DATA ON
                   NOISE-INDUCED HEARING LOSS IN POLAND,
                       ITS PROPHYLAXIS AND DIAGNOSIS

                                Wieslaw Sulkowski,
                          Institute of Occupational Medicine,
                                    Lodz, Poland
Introduction
     Needless to say, noise-induced  hearing damage reduces professional  efficiency and
work safety, is often a cause of disability leading to a loss of profession, and above all will
influence the sphere of private life. Glorig (1961) has rightly stressed its high potential costs
which are higher in industry than the losses brought about by olher professional diseases, if
we include in an economic reckoning the valiie of disability pensions and compensation, the
costs of absenteeism, the results of the loss of ability to work before reaching pension age,
social and health consequences of hearing impairment, etc.
     Statistical data regarding the number of persons with occupational  hearing loss, being
as a rule incomplete, are usually not  very representative and are hardly comparable due to
regional differences in the obligatory  diagnostic criteria in the particular countries, different
reference standards for calibration of audiometers, and different medico-legal regulations.
They give, however,  the  general idea of the importance  of the problem. The sources of
information  consist mainly of: records of hearing losses leading to authorization of compen-
sation, the  results of epidemiological  examination  of workers exposed to noise and, in
Poland, from the data obtained from  the central, uniformly-accepted registration of occupa-
tional hearing losses.
Epidemiological data

     Out of the 33 million population of Poland, about 4 million are employed in industry
(among them,  1.5 million are women), and out of those, roughly 15-20 per cent are exposed
to noise levels causing risk of hearing impairment.
     According to the evidence of the Sanitary-Epidemiological Department at the Ministry
of Health and  Social Welfare, based on  the individual reports of occupational diseases, the
number of occupational  hearing losses  registered during the period 1970-1972, compared
with other occupational diseases, has been presented in Fig. 1. Table 1 shows the structure
of occupational hearing losses in  the years 1970-1972, according to the industrial  branch
and age of workers. It  appears from Table I that the greatest incidence of hearing damage
occurred in the group of people of 40-49 years of age, especially in the industry of transport
means, the metal  and the textile industries. Similar conclusions come from French (Saulnier
1969), Czechoslovakian (Suntych 1970), and Austrian (Surbock 1971) statistical reports.

                                        139

-------
                          60
                          50
                       Q
                       £
                       O

                       I  <0
                       §
                       g  30
                       u.
                       I
                       U»
                       i
                          10
                                  OCCUPATIONAL DISEASES TOTAL
                                  OCCUPATIONAL rtEARJNO
                                       LOSS
                                 1910
                                            1911
                                                        1912
          Figure 1. Indices of occupational hearing loss and all occupational diseases in Poland.
Damage risk criteria

     Central accumulation of the cases of occupational hearing damage in Poland on which
the above  analysis has been made, is based on uniform criterion generally accepted for
prophylactic  and  diagnostic purposes: occupational hearing damage is a bilateral loss of
hearing in relation to the audiometric zero (ISO,  I 964) amounting to  at least 30 dB in the
better ear  after subtraction of age correction, calculated as an  arithmetic mean for fre-
quencies of 1000. 2000 and 4000 Hz. for each ear separately.
     That criterion is similar to that being used in  the USSR (Ostapkovich and Ponomareva,
1971) but differs from the  AAOO  (Davis  197 I) formula which does not accept the impor-
tance of frequencies over 2000 Hz  for perception of speech and thus limits the number of
persons receiving compensation. Without entering into the details of the controversial and
difficult  problem of defining what  degree  of hearing deterioration could be considered as a
                                          140

-------
                                       Table 1
                  OCCUPATIONAL HEARING LOSS IN POLAND IN ACCORDANCE
           TO VARIOUS INDUSTRIAL BRANCHES AND THE AGE OF WORKING POPULATION
                                 IN YEARS 1970 TO 1972
NR

1

2




3


4



S



6


7


8


9

BRANCHES
OF INDUSTRY

Transport
Means Industry
Metal

Industry


Textile

Industry

Coal-Mining
Industry

Machine

Industry

lion

Foundry
Building
Material
Industry

Other Blanches
of Industry
Great
Total
foi Industry
YEAR
1970
1971
1972
1970
1971

1972
1970

1971
1972
1970
1971
1972
1970

1971
1972
1970

1971
1972
1970
1971
1972
1970
1971
1972
1970
1971
1972
AGE GROUPS
-29
15
14
11
1
3

2
3

3
2
_
_
-
5

1
1
2

2
1
-
2
-
2
5
2
28
30
19
30-39
87
71
67
17
50

33
11

22
29
6
16
18
17

18
21
19

8
21
12
5
7
16
15
14
185
205
210
40-19
109
133
141
25
96

97
21

55
60
16
43
83
29

34
45
20

23
44
11
16
26
24
54
45
255
454
541
50-54
.30
61
78
5
26

25
2

20
23
5
25
57
10

12
18
7

10
17
3
7
11
9
15
14
71
176
243
55-59
28
55
74
8
28

24
7

14
32
6
15
28
11

6
15
12

16
24
11
14
7
6
16
15
89
164
219
60-64
17
29
61
5
11

14
10

28
81
1
24
17
6

6
11
9

8
12
9
13
19
4
13
15
61
132
230
65-
2
13
16
-
1

2
4

24
184
1
3
5
3

2
4
1

2
4
3
6
7
	
1
9
14
52
231
TOTAL
288
376
448
61
215

197
58

166
411
35
126
208
81

79
115
70

69
123
49
63
77
61
119
114
703
1213
1693
DYNAMIC
INDEX
100.0
130.6
155.6
100.0
352.5

322.3
100.0

286.2
70B.6
100.0
360.0
594.3
100.0

97.5
142.0
100.0

98.6
175.7
100.0
128.6
157.1
100.0
195.1
186.9
100.0
172.5
241.0
handicap  with regard to health impairment and/or to  the  loss or limitation of earning
ability, it seems that the criterion accepted in our country is fairer for all sufferers and is
justified both from  the  hearing-protection point of view as well as from the medico-legal
evaluation. Pensions and compensation are being granted to persons classified as disabled
due to occupational diseases by the medical board for employment and disability, taking
into account the degree of hearing loss and degree of invalidism.
    Apart from the criterion  of degree  of  hearing loss  mentioned here, the  diagnosis of
occupational hearing loss does not  take  into consideration  other etiological reasons for a
similar picture of deafness of sensorineural type, as well as other forms of hearing damage
not connected with  the working environment and, above all requires an adequate documen-
                                        141

-------
 tation of long-term noise exposure inducing the risk of hearing impairment. With regard to
 hearing protection in Poland, a noise level of not more than 90 dBA, lasting 5 hrs or more of
 continuous exposure daily is permissible.
     For continuous  noise lasting less than 5  hrs and for periodically interrupted noise.
 admissible levels are established with the help of the diagrams presented in Fig. 2. In case of
 fluctuating and irregular noise, the admissible level is specified as an equivalent sound level
 of 90  dBA. For impulse and  tonal noise  exposure, the permissible level should be 5  dBA
 lower.
                                                           
-------
Results of examinations

    1.    Textile industry

    Out of 600 randomly-selected workers in cotton-weaving, spinning and hose strand
twisting  departments exposed to continuous steady-state broad band noise, 98 subjects were
excluded who had suffered from morbid changes in the middle ear, had undergone skull
trauma or had had other diseases which would jeopardize proper evaluation of the influence
of the actual  exposure to  noise  on the  state of hearing. The mean  bilateral hearing
thresholds of the remaining 511 subjects (59 males and 452 females) (average age range 37.2
yrs ± 9.8 and with the average length of employment 12.2 ± 6.6) separately for the workers
of the particular three departments are shown in  Fig.  3.  No  statistically  significant
differences have  been found between hearing thresholds of the left and the right ear (p >
0.05). Greater  statistical significant hearing losses have  been noted in men, although the
dispersion around the mean was similar for both sexes (F < F 0.05)x.
    The significant differences of hearing thresholds among the workers of the separate
departments (F = 604.06 > F 0.01) seem to be connected with the different levels of noise,
as well  as  with  the structure  of the  examined groups, i.e. then* age and length of
employment (average age and  length of  employment  for  separate  departments differ
significantly [f=  19.38; P < 0.051)
    The influence of age and length of employment of the examined population on the rise
of hearing threshold seems to be clear. There is a significant increase of hearing loss accord-
ing to age (for men  F = 17.4; P  < 0.01 and for women F = 292.97; P < 0.01) and similar
one according to length of employment (F = 319.00 > F 0.01). The above is supported by
the regression lines (Fig. 4).
    In 46 workers the magnitude of the hearing loss (after deduction of the age correction)
reached the criterion for a compensable loss. Their distribution is shown in Table II.

    2.     Metal industry

    Out of 270 working population in three metal plants, viz. cutlery manufacture (con-
tinuous  noise with  fluctuating level), car'spring manufacturing  and press  plant  (impact
noise), 32 workers were excluded due to pathological symptoms in the middle ear or other
diseases. The mean bilateral hearing thresholds were determined in the remaining 238  sub-
jects (195 men  and 43 women) separately for  each  working group. Figure 5 shows the
oscilloscope photograms of noise in the particular plants. The results of audiometric exami-
nations of exposed populations are presented in Fig. 6.
    There  appear  significant differences  in the  mean hearing  thresholds of particular
groups, which are probably due to the different characteristics of exposure, especially as the
parameters  of age and length of employment are practically identical. In the population of
 xThe method of variance analysis according to the scheme of single classification has been used. (F denotes
 the value of F - Snedecor statistics).

                                         143

-------
                             M£M  tOUL N'59
                             HMHMliiWU
                             MEAN UVUMUNT W.&
                                                  WiA VINfr
                                                5f>L *
                                                      VI IW  l»jjj_
                                                    93 « dBA
                             MEN
                             WANA££41,t<
                             MEW
                             HOSf TWJ5TIN&  5PL  W-« *A
                             VfAN A6t 51)!
                                                                       -mean hearing threshold x; —.—.—.
Figure 3. Mean  bilateral  hearing thresholds of textile workers: —
range of dispersion ±5,	95.5 % confidence limits for the average general population (x - 25X) and (x + 2
               3 26

               g 2*

               ? 22
               tU
               T 20

               £ w
               ui
                : 10
                                      J yZ9 '0.31
                                                                                  ylf.B -09.
                          30    +0    50    60

                            AGE (YSS)
                                                    4    6    8    10   f2    14    16    19
                                                        LFNOTH Of fHfLOthCHT (V«S)
Figure 4. Regression lines of the  form y = ax  + b characterizing dependence of hearing loss on age and
length of employment.
                                                   144

-------
                                       Table 2

                      OCCUPATIONAL HEARING LOSS IN TEXTILE WORKERS
INDUSTRIAL
DEPARTMENTS
Cotton
Weaving
S*
8 c
o-S
"£
Hose Stiand
Twisting

Number of
Subjects
Occupational
Hearing
Loss
Number of
Subjects
Occupational
Hearing
Loss
Number of
Subjects
Occupational
Hearing
Loss
TOTAL
200
27
172
15
139
4
AGE (yrs)
20-29
42
-
37
-
63
-
30-39
82
10
54
5
38
-
40-49
57
11
42
4
3$
4
50-59
19
6
39
6
-
-
LENGTH OF EMPLOYMENT (yrs)
1-3
IS
-
15
-
12
-
4-6
25
1
26
1
58
-
7-10
22
3
24
4
43
1
11-15
81
11
37
4
23
3
16-
54
12
70
6
3
-
the punch press plant, consisting of men and  women  of similar distribution of age  and
length of employment, a significant rise of threshold has been noted for men (p < 0.001).
The regression lines traced on the basis of mean hearing thresholds for the separate age and
length of employment indicate the relationship between the age, length of employment and
the course of the hearing threshold curves (Figure 7).
    Table HI summarizes the cases of the stated occupational hearing losses.

    3.   Forest workers*

    Out of 12,000  woodcutters using motor-saws and therefore having been exposed to
noise of fluctuating level but without an impulsive component for less than 5 hours daily,
401 workers were chosen  for examination.  58 workers were excluded due to middle-ear
'pathology and other diseases or to exposure to noise in previous employment. In 11  of the
remaining 343, compensate hearing losses were found (Tab. IV).
     Figure  8 presents mean hearing thresholds for different  age groups, separately for left
and right ears.  Statistically significant differences were found between the left  and right
ears, which was not observed in the control group consisting of 47 tree-fellers of similar age
and length of employment, but who worked only with an axe (Table V). It seems likely that
 "The presented data are a part of the research work on: "The effects of chronic vibration and noise exposure on the health
  of woodcutters" being conducted within Polish-American Scientific Collaboration - contract No 05-005 3, Project Officer
  A. Cohen, Ph.D., Principal Investigator: H. Rafalski, M JD.                                             '
                                         145

-------
Figure 5. Oscilloscope photograms of noise in particular metal plants. A: cutlery manufacture: continuous
noise with fluctuating level, mean SPL 92-100 dBA, leq 95 dBA. B:  car spring plant: Leq 108 dBA. mean
level of acoustic background 82 dBA, impulse peak pressure level 121 dBA, rise time 0.2 sec, duration 0.6
sec, time interval 0.8 sec,  repetition of impulse 1/sec. C: punch press plant:  Leq 95 dBA,  mean level of
acoustic background  85 dBA, impulse peak pressure level 99 dBA, rise time 0.05 sec, duration 0.2 sec, time
interval 0.5 sec, repetition of impulses 2/sec.
the special posture of the chain - saw user during his work causes the left ear to be especially
exposed (Fig. 9). The calculated coefficients of both total and fractional correlation confirm
the relation between age and  length of  employment and  the  course of hearing threshold
curves (Table VI). An analogous conclusion arises from the regression lines, traced according
to  the  age and length  of employment  for evaluation  of the  average changes of hearing
thresholds (Fig. 10).

                                            146

-------
                                                    N«M  '" •-
                                                    COMTlim MfAM
                                          CJJT
                                             IfRY
                                                   r-^
                         N=109 (*EN)
                        /WAN A££.3M*14*
                        MEM EMPLOYMENT
                                        CAfl SP RIHtS _
                                             PRBH
                        MEN N-31
                        WAN Aȣ Hl>.
                                                    N-19  "
                                                    CMTimS MUM
                                                                > M
                        WOMEN N-
                        «MN A6E  ,
                        MEAN EMPLOYMENT 5,1! y
Figure 6. Mean bilateral hearing thresholds of metal workers
dispersion x ± «x).
                                                       mean hearing threshold; —.—.—. range of
     4.   Discussion

     The relation  between  noise  and irreversible changes in hearing sensitivity which arise
due  to long-term  daily exposure  has been the  subject of so many field studies that there
would not be enough room to enumerate them all here.
     In spite of standardized methods of hearing measurement  based on tonal audiometry,
the great number of variables makes the comparison and  evaluation of the results which
have been  obtained difficult-for example,  the wide variation  of individual susceptibility.
There exists, then, the problem of an adequate use of these experimental data gathered for a
better explanation of  the  effects  of the  exposure to noise on hearing. Therefore,  the
problem  is  to establish possibly  optimal damage risk  criteria (DRC) more accurate than
those now  in  use. DRC worked out on a statistical basis cannot ensure preservation of
                                           147

-------
                   	O SMMEH5
                   	* PUNCH-PIIESSfRS
                    - • CUTLERS
                            ACC (T«»»
                                                            UWTN Of ((MDYHHIT
Figure 7. Regression lines of the form y = ax + b characterizing dependence of hearing loss on age and
length of employment in metal workers.
                                             Table 3
                        OCCUPATIONAL HEARING LOSS IN METAL WORKERS
INDUSTRIAL
PLANTS
III
%h
a>*.
aJS a
M a
&r-
ws-
Q, WBL<

Number of
Subjects
Occupational
Hearing Loss
Number of
Subjects
Occupational
Heating Loss
Number of
Subjects
Occupational
Hearing Loss
TOTAL
109
16
55
32
74
5
AGE (yrs)
24-29
26
-
5
3
17
-
30-39
34
4
17
10
18
-
4049
30
8
24
13
26
2
50-
19
4
9
6
13
3
LENGTH OF EMPLOYMENT (yrs)
1-5
32
1
9
5
30
-
6-10
15
2
15
7
17
-
11-15
22
2
15
8
12
3
16-
40
11
16
12
15
2
                                               148

-------
                                                 Table 4
                           OCCUPATIONAL HEARING LOSS IN CHAIN-SAW USERS


Number of
Subjects
Occupational
Hearing Loss
Tf~lT A T


343
11
AGE (yrs)
20-29

40
1
30-39

101
1
4049

147
7
50-59

55
2
LENGTH OF EMPLOYMENT (yrs)
1 -3

57
-
4-6

106
1
7-10

137
3
11-15

36
4
16 -

7
3
                                 LEFT EAR
                                 WOOK UTTERS
                                 N«332
                                                            RIGHT  MR
101,5
                         
-------
                                      Table 5

                          EVALUATION OF MEAN HEARING LOSSES
                IN WOODCUTTERS (MEAN FOR FREQUENCIES 1000,2000.4000Hz)
EXAMINED GROUPS
Chain-Saw Users
Total (N = 343)
Chain-Saw Users
with Occupational
Hearing Loss (N = 11)

Controls (N = 47)
Significance of
Differences (Chain-Saw
Users - Controls)
LEFT EAR
22.1 ± 11.6
52.1 ± 9.1
17.1 ± 9.1
t = 2.831
p < 0.01
RIGHT EAR
18.6 ± 9.7
50.2 ± 7.1
18.0 ± 9.5
t = 0.376
p>0.5
SIGNIFICANCE
OF DIFFERENCES
t = 2.871
p < 0.01
t = 0.636
p > 0.05
t = 0.469
p > 0.05

normal hearing for the whole population exposed to noise hazard but only for the majority
— those of average susceptibility.
     Studies on a few different occupational environments which were not controlled with
regard to protection of hearing indicate a great spread of hearing thresholds of the examined
invididuals around the arithmetic mean of the whole group. A similar high standard devia-
tion was found by Riley et al. (1961) and Jatho (1966), who studied HLs as a function of
age and sex in populations not exposed to noise. In our audiograms we have not subtracted
age corrections, so the  data presented  illustrate the combined influence of noise exposure
and age-induced hearing losses which, according to Spoor (1967), increase with the logarithm
of age. However, comparison of the audiograms of non-noise-exposed subjects having the same
age as the subjects in our control groups gives an idea of the magnitude of, as it is called by
Passchier-Vermeer (1971), "the noise-induced part of the median hearing level". The results
are therefore  the key  relation between permanent  threshold  shifts for pure-tone  and
physical noise parameters, effective duration and years of exposure. They show  also, simi-
larly  to Burns and Robinson's (1970)  findings, mat the curve of hearing loss has no
tendency to a sudden flattening after 10-15 years of employment.
     Significantly lower hearing thresholds of women found in  our data as well as in many
other surveys  (Jankowski,  1952; Hassmann et al. 1970; Dieroff, 1967; Ward,  1965) are
ascribed to a higher absenteeism of women and to a lesser exposure of women  to non-
occupational noise.
     The  alarming percentage  of hearing impairment among  the workers of  the metal
industry exposed to impact noise shows the urgent need of establishing D.R.C. for this type
of noise, whose evaluation is still in the area of uncertainty of knowledge (Kryter, 1970;
Coles and Rice, 1971; Sulkowski et al. 1972).
     It should be stressed that in both the metal and textile industries, we had expected at
the outset to  find confirmation of our presumption of the considerable apparent risk of
healing loss there due to 8-hour daily exposure. But finding 11 cases of compensable hearing
losses among the woodcutters exposed  only about 3 hours daily  to fluctuating noise came as
                                        150

-------

                                                                  j
                       Figure 9. Chain-saw user engaged in tree-felling.
a surprise to us. In the entire group of 343 chain-saw users, the thresholds of the left ear
proved to be worse than those of the ritjit car. which had been reported to the examiners by
the workers themselves before the audiometric tests started.
     The risk  of damage among forest workers has also been noted  by  Holmgren  et al.
(1971) and  Passchier-Vermeer (1971). However,  they  did not give any details about  the
number of stated cases of compensable hearing losses. The results achieved by those authors
as well  as ours  seem to confirm the equal-energy  hypothesis which,  for evaluation of a
fluctuating noise as well as for an irregular one, assumes that the risk of hearing loss depends
on the total dose of acoustic energy, regardless to the distribution of the energy in time.

                                         151

-------
                                         Tables
                        EVALUATION OF DEPENDENCE OF HEARING LOSS
                       ON AGE AND LENGTH OF EMPLOYMENT BY MEANS
                    OF TOTAL AND FRACTIONAL CORRELATION COEFFICIENTS
DEPENDENCES
Age (xl) -Length
of Employment (X2)
Hearing Loss (y)
-Age(xl)
Hearing Loss (y)
-Length of
Employment (X2)
COEFFICIENTS
OF CORRELATION
TV v
X,Xj
Tyxl
TyXl.x2
^
Tyx2.x,
CHAIN-SAW
USERS TOTAL
+0.343
t = 6.03
p < 0.001
+0.299
t = 5.408
p< 0.001
+0.223
t= 3.949
p < 0.01
+0.288
t= 5.192
p < 0.01
+0.207
1=3.562
p < 0.01
CHAIN-SAW
USERS WITH
OCCUPATIONAL
HEARING LOSS
+0.394
t = 4.197
p < 0.001
+0.557
+0.471
+0.414
+0.255
t = 0.791
p > 0.05
CONTROLS
+0.685
t = 6.307
p<0.01
+0.548
t = 4.395
p < 0.01
+0.443
t= 3.318
p < 0.05
+0.360
t = 2.588
p < 0.002
-0.0154
t= 0.103
p>0.9
LEGENDS:  Coefficients of Total Correlation
                yX|, yx2
Coefficients of Fractional Correlation
        1 •
        3 n
        * M
        1:
        i:
                     l£f r CM
                     I 1M>     •
                                   MM    14   !!•<>»«
                                                    OF fmornciir 
-------
Prophylaxis and diagnosis of occupational hearing loss

    Prophylaxis of hearing damage, which starts with the physical measurement of noise
and evaluation of its harmfulness, is in the scope of the industrial health service activities in
Poland. Industrial physicians have an advisory voice in noise control which is carried out by
the industrial hygiene  staff in epidemiological centers - cities, towns, voivodeships and
districts — leading to diminution of noise levels, shortening of exposure times, and use  of
individual acoustic defenders.
    The pre-employment and follow-up audiometric examinations, which aim at assurance
of controlled, repeated and comparable evaluation of hearing of workers, are performed in
stationary audiometric laboratories at factorial, multi-factorial, district, town or voivodeship
industrial outpatient clinics. Supervision over these examinations is provided by the otolar-
yngologists, who also determine their frequency. In 1972, there were 205 otolaryngologists
employed in the Polish industrial health service. The consultant departments of the voivode-
ship outpatient  clinics verify the detected hearing damage; doubtful cases or diagnostically
difficult ones are referred for decision in conditions of clinical observation to the Institute
of Occupational Medicine. During the period of 1971 to 1972, 496 cases (356 men and 140
women) of hearing impairment among the workers of different branches of industry from
the whole country were referred for checking.  Occupational hearing loss was confirmed in
330 cases. The grounds for elimination of occupational etiology in the remaining 166 cases
are listed in the  Table VII.
                                       Table 7

                              MOTIVES FOR ELIMINATION OF
                                OCCUPATIONAL ETIOLOGY
ENUMERATION OF MOTIVES
NON-ORGANIC HEARING LOSS 	

OTOSCLEROSIS 	

TYMPANOSCLEROSIS 	

UNDERGONE CONSERVATIVE AND RADICAL OPERATIONS OF THE MIDDLE EAR 	

CHRONIC PURULENT OTITIS MEDIA . ...

CONGENITAL DEAFNESS 	

MENIERE DISEASE 	

LACK OF ESSENTIAL NOISE EXPOSURE 	

TOTAL 	

NUMBER OF
CASES
55

12

14

21

39

2

15

8

166

                                         153

-------
     It is  our experience, and Coles and Priede (1971) present a  similar opinion, that
conventional  tonal audiometry alone is  not always adequate for reliable  evaluation of
hearing, especially with regard to claimants of compensation.
     Nothing  less than the application of a large set of various audiological tests allows a
well-founded diagnosis in such cases. In our Institute these are routinely used: manual tonal
audiometry,  Bekesy automatic audiometry,  suprathreshold examinations  (mainly  tests
of Langebeck, Liischer - Zwislocki, SISI), speech audiometry, and among classic malingering
tests, mostly  Stenger's test and the delayed speech feedback test. The measurements of
stapedius muscle reflex and evoked response audiometry (Bochenek et al. 1973) are also
worth mentioning as valuable methods for evaluation of real hearing thresholds.


                                   References

 1.  Bochenek,  W., Sulkowski, W., Fialkowska,  D., Tolloczko, R., and Zmigrodzka, K.,
     Impedance and electrophysiologic (ERA) audiometry  in  diagnosis of occupational
     hearing loss. Med. Pracy (in press).
 2.  Borsuk, J., Sulkowski, W., Wendt, J., Miksza, J., and Zalewski, P., An evaluation of
     hearing status of workers of some textile plants on the  basis of audiometric examina-
     tions. Biul WAM, 13, 115-120 (1970).
 3.  Burns, W., and Robinson, D.W., Hearing and noise in industry. London: Her Majesty's
     Stationery Office (1970).
 4.  Coles, R.R.A., and Rice, C.G.,  Assessment of risk of hearing loss  due to impulse noise.
     Br. Acoust. Soc. Spec.. 1, 71-77(1971).
 5.  Coles, R.R.A, and Priede V.M., Nonorganic overlay in noise-induced hearing loss. Proc.
     roy. Soc. Med., 64, 194-199 (1971).
 6.  Davis, H., A historical introduction. Br. Acoust. Soc. Spec., 1, 7-12 (1971).
 7.  Dieroff,  H.G., Horschaden durch Industrielarm. Arbeitsmedizin, Sozialmed., Arbeit-
     shy g., 7,  256-260. (1967).
 8.  Glorig, A, cited  by Bell, A, Le bruit, Geneve, Organisation Mondiale de la Sante
     (1967).
 9.  Hassmann, W., Pietruski, J., Krochmalska, E., Filipowski, M., Kom-Rydzewska, H., and
     Kisielewska, L, Hearing impairments in  young textile workers. Proceedings of XXVII
     Conference of Polish Otolaryngologists, Warszawa, PZWL, 126-129 (1970).
10.  Holmgren, G., Johnsson, L., Kylin, B., and Linde, O., Noise and hearing of a popula-
     tion of forest workers. Br. Acoust. Soc. Spec. Vol. No. L, 35-42 (1971).
11.  Jankowski, W., Impairment of hearing in the workers of the cotton industry.  Otolar-
     yng.Pol, 6, 111-123(1952).
12.  Jatho, K., Population surveys and norms. Int. Audiol., 5, 231-239 (1966).
13.  Kryter,  K.D., Evaluations of  exposures to impulse noise. Arch. env. Health, 20,
     624-635  (1970).
14.  Ostapkovich, V.E., and Ponomareva, N.I., Determination of work capacity in workers
     with occupational hearing disorders. Gig. Truda, 15, 25-28 (1971).
15.  Passchier-Vermeer, W., Steady-state and fluctuating noise: its effects on the hearing of
     people, fir. Acoust. Soc. Spec.,  1, 15-33 (1971).

                                       154

-------
16.  Riley, E.G., Sterner, J.H,, Fassett, D.W., and Sutton, W.L., Ten years' experience with
    industrial audiometry. Am. Ind. Hyg. Ass. J., 22, 151-159 (1961).
17.  Saulnier, J., Le  bruit dans 1'industrie  et  la surdite professionnelle.  These pour le
    doctorat en pharmacie, Paris (1969).
18.  Spoor, A.,  Presbycusis values in relation to noise induced hearing loss. Int. Audiol., 6,
    48-57(1967).
19.  Sulkowski, W., and Andryszek, C., Investigation of the noise induced hearing losses in
    workers of chosen plants of light industry. Med. Pracy, 23, 296-314 (1972).
20.  Sulkowski, W., Dzwonnik, Z., Andryszek, C., and Lipowczan, A., Risk of occupational
    hearing loss due  to continuous,  intermittent and impulse noise. Med. Pracy, 23,
    465479(1972).
21.  Suntych, F., Kryze, B.,  and  Parizkova, B.,  Occupational diseases and Professional
    intoxications registered  in  Czechoslovakia in  1968. Prac. Lekarstvi, 22,  284-293
    (1970).
22.  Surbock, A., Hearing conservation  and noise control in industry organized and per-
    formed by  the Accident Branch of the Austrian Social Security Board. Br. Acoust. Soc.
    Spec., 1, 121-127(1971).
23.  Ward, W.D., The  concept of susceptibility to hearing loss.  /. Occup. Med., 1, 595-607
    (1965).
24.  Wojcieszyn, M., Adaptation and fatigue of hearing in workers  exposed to subcritical
    noise level. Proceedings  of XXVII Conference of Polish Otolaryngologists, Warszawa,
    PZWL, 39-41 (1970).
                                        155

-------
                  ON THE PROBLEM OF INDUSTRIAL NOISE AND
                       SOME HEARING LOSSES IN CERTAIN
                   PROFESSIONAL GROUPS EXPOSED TO NOISE

                                     J. Moskov
                                   Sofia, Bulgaria

     The wide development of industrial processes and extensive mechanization have greatly
facilitated physical  labor. The  modern way of life has created more comfortable  living
conditions. In spite of this, however, noise has  not diminished, but is still increasing. In
recent years, noise has proved to be one of the most widespread noxious factors of the
working  environment in  our country. Being an environmental factor,  it penetrates  all
spheres of life - both at the working places and in our ordinary living and social environment.
     To organize  and direct adequately the fight against industrial noise in our country it
was necessary to make first a complex evaluation of this factor, including both the physical
characteristics of the various noises encountered in the different industrial branches and also
the influence exerted by them on the workers.
     In view of this, the Research Institute of Labour Protection and Professional Diseases
undertook the task of studying the noise  produced at  most of the working places in the
basic branches.  We  took also into consideration the fact that not only the workers at the
machines and mechanical sources producing intensive noise are being exposed  to the noise,
but also these working in proximity to them.
     Simultaneously, we also carried out serial mass prophylactic examinations on groups of
workers exposed to noise. Different specialists took part in these examinations: profpathol-
ogists, otologists,  neurologists, internists, etc. In  instances of the presence of a noise effect
combined with other factors,  we also  carried out an investigation of these factors and of
their influence. Thus, for example,  we studied  also the vibrations  usually accompanying
industrial  noise in the timber and ore-mining industries; in the chemical industry, wood-
working, textile and other industries we conducted toxicologic studies and examinations of
the dust aerosols; in metallurgy and  in the smith-pressing departments we also studied the
overheating microclimatic conditions, etc.
     Besides all these complex studies,  we carried out further a threshold-tonal audiometric
examination. This examination was  performed in a labor-hygienic installation. In case  we
discovered substantial changes, the affected workers were taken over by the special otologic-
audiologic   departments  where  additional  audiologic  examinations,  including speech
audiometry, etc.,  were carried out. Besides an examination of hearing, in certain categories
of workers—mainly  those whose activity demanded considerable psychosensorial strain—we
have begun still other examinations: on the effect  of noise on peripheral vision, on the
processing of information,  reproduction of a dosed muscular strain, reproduction of a
spatial position of the hand, etc. In some  cases  we are  also studying the vibrational sensi-
tivity, heat  sensitivity, arterial pressure, pulse rate,  etc. We studied a  total of about 900
industrial objects of  14 different industries by measuring the noise  produced at several
thousand working places. A parallel study was also made of 6400 workers.
     The results of  these studies, as well as those of the prophylactic audiometric examina-
tions, are shown in the following figures:

                                         157

-------
                               Table 1.

     NOISE LEVELS IN SOME BRANCHES OP PRODUCTION* /in dB  (A)/

METALLUHGl                           50  60  70  8   9  1QO 110 120
  Cylinders for casts  cleaning
  Crushers
  Bell mills
  Cleaning devices
  Moulding departments

ELECTRICAL INDUSTRY
  Turbofeeders
  Turbine halls
  DEO

MACHINE-BUILDING AND METAL-
WORKING
  Tinsmith's departments
  Blacksmith's departments
  Work with riveting instruments
  Work with hand emery-wheels

WOODWORKING
  Circular saws
  Knife-grinding places
  Press departments
  Abricht-mschines

TIMBER INDUSTRY

SHOE INDUSTRY
PRODUCTION OP READY-MADE
CONSTRUCTION
  Vibromasses
  Vibroriddles
--«
     -•«
   Some most noisy sources and working processes are also  shown.
                                 158

-------
               Table 1 pg. 2 (Noise levels in some branches of production).
CHEMICAL INDUSTRY
  Soda production
  Production of nitrogen
  fertilizers
  Plastics production
  Antibiotics production
  Cement production
  Carbide and gunpowder production
  Other working places
  Production of bricks and cement

COAL NINES
  Separations

ORE-MINING
CRANES
ROPE-LINES
TEXTILE INDUSTRY
  Loons
  Spinning looms
  Ring machines
  Shuttles

POOD AND TASTE INDUSTRY
  Meat-manufacturing enterprises
  Sugar industry
  Cereals-manufacturing industry
  Production of non-alcoholic
  drinks

TELEPHONE CENTRALS AND TELETYPE
  Computer centres

LEATHER INDUSTRY

OFFICES, PLANT MANAGEMENTS AND
LABORATORIES

                                 159
0 6<










* -









MM^MI


9




















i


o e























o S










• -1





"






0 10











j
^ «•»

—4





,
"4


01'



.


4
•
,


•











0 1£























>0
























-------
250
WOO
600G
                                                              6OOO
    -/Q

     0

    JJ

    20
    40

    SO

    $0
    90

    JOO

    HO
Figure 1. Average loss of hearing in workers engaged in woodworking.

Length of work: — Less than 1 year;	1-3 years; -   - 4 - 6 years;	7-9 years,	10-15
years;	 15-20 years.
                                         I 60

-------
115     250    $00     /OOO
                                                        fa?    6000   &000
•JO
0
10
20
60
40
do
60
ro
#0
30
100
no

















	
— • —


























^w • •









JSOO



^












s*^









3000



X
\










,^
*" —
\











St
X
N>
'*" —











- — — -
	



















Figure 2. Average loss of hearing in workers engaged  in the production of ready-made construction
elements.

Length of work:	Less than 1 year; — 1-3 years; - • - Over 3 years.
                                            161

-------
         125    250    £00    '000   2000   4010   £000    40OO
-JO
  10

 20
  40

  60

  SO

  70

  00

  90

  JOO

  /10

_x
                    Figure 3. Average \oa of hearing in textile workers.

                                    —Spinners
                                   	Reel workers
                                   - • -Weavers
                                  - •- -Other
                                     162

-------
                   250    $00    1000    2000    4000   6000   80OO
   -10

     c

    10
60

40

 60

 60

 To
                                                           ,  •
    30
    /Oo

    /to
Figure 4. Average loss of hearing in workers engaged in the shoe industry.

Length of work: — Less than 1 year;	1-3 years; - • — 4 - 6 years;	 7 - 9 years;	10-15 years;
	16-20 years;	Over 20 years.
                                          163

-------
25*0
                                 JOOO   2000   4000
                                             MOO
w
20
40
 50
 fO
 ro
                                             \
 90
100
 HO
  Figure 5. Average loss of hearing in miners.
  Length of work: — 1 - 3 years; - - 4 - 6 years; -   - 7 - 9 years. -   - 10 and over 10 years.
                                      164

-------
               2SO    SDO    JO00   20DO   400G   G0OO   £000
-/O
                                   /500
jooo
20
30

So
 50
 70
 60
 90
 100
 110
       Figure 6.  Average loss of hearing in workers engaged in the timber industry.
       Length of work: — 1  • 3 years; — — 4 - 9 years; — • — 10 and over 10 years.
                                    165

-------
                SOD    fOOO   2000   30 OO   4000   SOffO  #000
'10

  0

 fO

2.0
40

 $0

 60
 no
           Figure 7. Average loss of hearing in crane and rope-line workers.

                                —Crane workers
                               	Rope-line workers


-------
         /Z5     250    500    1000   2000  4090    ***  4000
                                              3000
   20
   30
   40
   60
   60
    TO
   SO
   90
   no
Figure 8. Average data of hearing in professions unexposed to noise (administrative personnel, laboratory
workers).
                                        167

-------
     The Table indicates the average results of the noise data. The audiometric results (Figs.
1-8) are hearing levels after the age correction suggested by Hinchcliffe.
     The noise measurements are all card-indexed. A separate card is made for each  work
area studied. These  separate cards contain data about  the  kind of production and the
professions by which work is performed under the effect of noise, the category and number
of workers exposed to this noxious influence. Data are also entered concerning the duration
of exposure at the various working places.
     The purpose of our studies is the making of a contribution to elucidating the  noise
status in  our country. They are intended to draw the attention of the audiologist, industrial
health pathologist, the trade union, administrative and other bureaus, to the noise problem.
     On  the  basis of these studies, we worked out a number of prophylactic programs,
instructions for the fight against noise in various branches,  hygienic norms of industrial
noise, etc. A number of important problems still stand before us, waiting to be solved. Thus,
for example, there is still no solution to the problem of the degree of hearing needed for the
various professions; no Bulgarian standard has been worked out, as yet, for age correction in
audiometric examinations. At present,  we either do not use corrections for age changes,  or
we use those taken from foreign authors, worked  out for the population of other countries,
based on different demographic and other indices. The criterion  for  determining the work-
ing capacity of a person, in an industrial hygiene context, should depend on his profession.
Speech audiometry has to be introduced (as a rule) in establishing the  social adequacy of the
examined individual.
     Pre-employment and monitoring audiometric examinations  of workers engaged in the
so-called  noisy professions have already become standard practice. We are also having good
results in the field of preliminary sanitary control. The noise factor  is being already taken
into consideration in giving a permit and approval for different devices and machines, for
factory departments, for building a plant, etc.
     The  training programs  of medical  students and  doctors  specializing in industrial
hygiene already include a section  on noise  and vibrations.  During recent years, several
symposia and national congresses were held in our country on the problems of noise, with
participation of a number of distinguished specialists from  many foreign countries. The
development of a special law for the fight against noise and vibration is in prospect and this
will be an important step in the difficult fight against these disagreeable and noxious factors.
     We  hold  to the  concept that in establishing the noise norms in the future, both for
industrial noise and for  noises in our ordinary living and social environment, we should also
take into consideration  their extraaural effects, which in many  cases are being manifested
early, at levels that do not lead to changes in hearing.
     The fight against noise is a complex problem. The achievement of favorable results will
be possible only on the  basis of the mutual efforts of specialists from many different fields
and with the cooperation of the administrative and social organs and organizations.
                                         168

-------
               NOISE-INDUCED HEARING LOSS FROM EXPOSURE
                   TO INTERMITTENT AND VARYING NOISE

                              W. Passchier-Vermeer
                Research Institute for Public Health Engineering TNO
                              Delft, The Netherlands

INTRODUCTION

    The original intent of this review  was to consider exposure to intermittent noise only,
but exposures to varying noise have been included as well.  Since varying noise includes
intermittent noise, the  latter being nothing but noise varying between a "high" and a "low"
noise level and since it is quite unclear what "low" in this respect means, it  seems advan-
tageous to start from the more general subject of exposures to varying noise.
    At the moment no field studies are  known which give relations between exposures to
varying or intermittent noise  and  noise-induced hearing loss. In the  second part of this
review, an attempt  has been made to relate both  quantities, by using data from several
papers. Exposure to  impulsive noise, such as gunfire, is not referred to.
    In the first part of this paper, those noise exposure limits will be reviewed which refer
to exposure to intermittent and varying noises. All these  limits  are based on temporary
threshold shift by assuming that there exists a certain relation between temporary threshold
shift and permanent threshold shift.
    This review does not discuss criteria, the basic limits of safe noise exposure, concerning
the percentage of people  to be protected from so and so many decibels hearing loss at these
and those frequencies after  so and so many years of exposure, related to this or that group
of non-noise exposed people. This subject seems to be sufficiently covered by other papers
presented at the Congress (Glorig, Kry ter).

I.   TEMPORARY  THRESHOLD  SHIFT FROM  EXPOSURE TO VARYING AND IN-
    TERMITTENT NOISE

    1.  Damage risk contours prepared by NAS-NRC CHABA Working Group 46 (Kryter,
        Ward, Miller and Eldredge 1966)
        Before discussing  these contours, it should be pointed out that a tremendous
        number of  TTS  experiments  form the  basis  of these contours.  Criticism on
        extrapolations used in the derivation of these contours should be seen in the light
        of this remark. The contours are based on three postulates, (1) TTS2 (Temporary
        Threshold Shift measured 2 minutes after the  end of a noise exposure) is a
        consistent measure of a single day's exposure. This  is supported by evidence that
        TTSs maintain  their rank order  during recovery (Ward et al. 1958, 1959a) and
        that recovery does not depend on how TTS was produced (Ward et al. 1959b). (2)
        All exposures that produce a given TTS2 are equally hazardous as far as NIPTS
        (Noise-Induced Permanent Threshold Shift) is concerned (3) TTS2 is about equal
        to NIPTS  after ten years  of exposure. Noise exposures with parameters lying on

                                       169

-------
         the damage risk contours should restrict the average TTS2 at 1000 Hz and below
         to  10 dB,  at 2000 Hz to  15 dB and at  3000 Hz and above to 20 dB. These
         TTS2-values are called in the following criteria TTS2S.
             Three sets  of contours were constructed. They  were derived by using equa-
         tions, expressing the increase of TTS2 at several frequencies with exposure time,
         for constant octave band noise with  octave band  sound pressure level as the
         parameter, and  by using general frequency-independent curves for the recovery of
         TTS after exposure. These three sets are:
         a.   Damage risk contours for a single exposure daily. These contours are shown
             in  figure 1. In preparing these contours, extrapolation to the longest dura-
             tions and to the shorter ones was involved.
         b.   Damage risk contours  for  short  burst intermittent  noise (noise  bursts  2
             minutes or less in duration, alternating with effective quiet). These contours
             are  mainly  based on  the on-fraction rule (Ward 1962), which states that
             when noise alternates with quiet and is on for \% of the time, but no longer
OCTAVE  BAND  SOUND
PRESSURE  LEVEL IN  dB
         uo
          DURATION
           r!N,
          MINUTES
                           34561
34568.
           100                         '1000                          10000

                        OCTAVE  BAND  CENTER  FREQUENCY  IN  Hz
    FROM  K.D.KRYTER  ET AL

Figure 1. CHABA damage risk contours for one exposure per day to octave bands of noise. This graph can
be applied to the individual band levels present in broad-band noise (From Kryter et al, 1966).
                                      170

-------
    than 2 minutes per exposure cycle, the resulting TTS2 is x% of the TTS2
    produced by a continuous  exposure  for  the  same time. For frequencies
    below 2000 Hz,  this rule needed modification, due to  the action of the
    middle ear muscles. Again, extrapolation was involved in  applying this rule
    to higher octave band sound  pressure levels and lower on-fractions.
c.   Damage  risk contours for longer burst intermittent noise. In deriving these
    contours the "equivalent exposure rule" was used.  According to this rule,
    the residual TTS present from one noise, exposure at the start of the second
    exposure is converted to the time it would take this second noise to generate
    an amount of TTS equal to the residual TTS; this equivalent exposure time is
    added to the exposure time  of the second noise to calculate the TTS2 at the
    end of the second exposure. This rule was only shown to hold for TTS  at
    4000 Hz. Also, it was shown in the relevant paper (Ward  et al, 1959c) that
    recovery during the quiet intervals did not proceed in the expected way, but
    that recovery from higher  TTS-values was slower  than  from lower TTS-
    values. Nevertheless, the equivalent exposure rule was expanded for other
    frequencies  as well and  in deriving the damage risk contours for longer burst
    duration intermittent noise, the general recovery curves were used.
An important shortcoming  of the CHABA report concerns  the definition  of
"effective quiet": According to the report, and ignoring Ward's objections (Ward,
1966), effective  quiet is assumed to exist whenever the  noise level drops below
the octave band  sound pressure levels allowable for 8 hours a day (89 dB SPL for
the octave band with midfrequency 500 Hz, 86 dB at  1000  Hz, and  85  dB  at
2000 Hz and  400 Hz).

However,  the recovery curves used in the report were established for quiet and
they  may not be valid for 85 to 89 dB octave band SPL's. Figure 2 shows the
effect of certain  noise levels on  recovery as found  by Schwetz et al. (1970), by
Lennhardt et al. (1968) and by Ward et al. (1960). Schwetz found that recovery
from  TTS at  1000, 2000, 3000 and!4000 Hz is statistically significantly retarded
in white noise with an overall sound pressure level of 75 dB (which might be equal
to at most 70 dB SPL per octave band) and still more in white noise of 85 dB.
Results from Lehnhardt  are even more pronounced than those  from  Schwetz.
Although white  noise of 70 dB  overall  SPL (probably equal  to  at most 65 dB
octave band  SPL) allowed the same recovery of TTS at 2000, 3000,4000, 6000
and 8000 Hz as  in quiet, white noise of 80 dB SPL allowed recovery only during
the first 15 minutes or so after the noise exposure; after that time TTS increased
again! A similar effect was shown by Ward. He considered the course of TTS after
an exposure  to 105 dB SPL octave band noise, immediately followed by exposure
to 95 dB SPL octave band  noise. Ward's interpretation, which clearly emerged
from his results, is that the "excess" TTS (i.e. the difference between the TTS
produced by  the 105 dB noise level and the TTS that would have been produced
by an exposure for the same time to the 95 dB noise level) recovers in the 95 dB
noise level as in quiet, independently from  the simultaneous growth Jof TTS

                                171

-------
 TTS
                                                 Ward
                                                         Lehnhardt
                                                      75 dB
      20
      10
                               Schwetz
                                        Quiet
65 dB
   Lehnhardt
  Quiet
                  5             20       50           200
                     MINUTES  AFTER  EXPOSURE
Figure 2. Average temporary threshold shift as a function of the time after exposure, according to —
Schwetz (1970), TTS averaged over 1000, 2000, 3000 and 4000 Hz - Lehnhardt (1968), TTS averaged
over 2000, 3000. 4000. 6000 and 8000 Hz - Ward (1960), TTS averaged over 3000 and 4000 Hz Octave
band sound pressure level is parameter.
                                   172

-------
attributable to the 95 dB noise level. (By the way, by accepting the concept of
"excess" TTS, it should also be accepted that the course of ITS at a later stage
does depend on how it was produced). However, Ward's interpretation does not
hold for Lehnhardt's results, since 75 dB SPL octave band noise alone does not all
cause that much TTS during the time considered.

Klosterkotter (1971) stated that recovery from TTS in a sound level of 70 dB(A)
was slower  than recovery in 35 dB(A), with a difference in recovery of 7 dB.
Details cannot be given here, since the original paper was not at hand at the time
this paper was written.

All in all, it seems that only octave band sound pressure levels of at most 65 dB
should be considered to be "quiet", in view of recovery from TTS. This now is of
great  importance,  when  considering industrial situations.  Although CHABA
damage risk contours  for intermittent noise might be applicable, with effective
quiet at 65  dB SPL or lower, it seems that their use is limited to a few industrial
situations only.

CHABA rules are  also given for varying noise levels, at least when they do not
remain at any level for  more than two minutes. The effective level of such a
varying noise is equal  to the time-average sound pressure level of the noise over
the  exposure period.  Again,  however this is a  broad generalisation of the  one
relevant research (Ward et al, 1959a), in which it was only shown to be correct for
TTS at 4000 Hz due to exposure to noise of alternate 30 sec.-periods of 106 and
96 dB SPL. The way in which varying noises with levels remaining for more than
2 minutes at different values should be treated, is in fact not known at all. Only in
one publication, just cited (Ward 1960) was the subject touched on.

In 1970, Ward conducted  new TTS-experiments to determine,  as he states, the
degree to which the CHABA damage risk contours are in error (Ward 1970). In
general, it turned out from all experiments that the TTS2-values due to exposure,
chosen from the CHABA damage risk contours, were in the range of about 70%
up to 115% of the criteria TTS2's. Looking more closely at the results concerning
TTS2, it was shown that:
(1)   Single uninterrupted exposures up to 8 hours meet the criteria TTS2 within
     10%.
(2)   Short-burst intermittent noise does not meet the criteria TTS2- In two out
     of three experiments TTS2 after exposure was only about 70% of the criteria
     TTS2.  However,  the  other  experiment was terminated after 6 hours, al-
     though according to the CHABA contours 8 hours exposure was permitted,
     because of large values of TTS in some ears. Anyhow, even 6 hours exposure
     resulted in an average TTS2 above the criterion TTS2-
(3)   For long burst  intermittent  noise,  the criterion TTS2-values are exceeded.
     Again,  in these experiments it was demonstrated that recovery during quiet

                               173

-------
     intervals does not proceed according to the general recovery curves, but that
     recovery during successive intervals is retarded, thus not permitting sufficient
     recovery from TTS during quiet intervals.
Thus Ward's  latest published results on TTS  indicate that recovery does  not
necessarily  follow the general  recovery curves used  in  the  derivation  of  the
CHABA damage  risk  contours.  Although these curves seem to be all right for
single uninterrupted noise exposures, intermittent exposures to high frequency
(above 1500 Hz or so) high level noise, either long or short  bursts, often produced
a delayed recovery. This delayed recovery was in earlier experiments (Ward et ah,
1958,  1960)  shown to occur, after single uninterrupted  exposures, only when
TTSo was more than about 40 dB. In these experiments, however, for high level
noise it already  occurred when TTS2  was  only about  20 dB. Most alarming,
however, is the  fact  that TTS2 is not a consistent measure  of a single day's
exposure,  since TTSs do not maintain their rank order during recovery,  and
recovery from TTS does depend  on how TTS was produced. Ward, then, looking
for a practical solution suggested that TTS3Q (TTS measured  30 minutes after the
exposure) might be a useful index, since after  30 minutes the  rank order of TTS is
more or  less constant. Although this  may  be right,  it is quite unclear which
relation exists between TTS30 and NIPTS for  exposures to  intermittent noise.

Contrary to the equal-energy principle (which applied to TTS2 states that TTS2S
resulting  from exposures to noises with the  same total sound energy are equal,
irrespective of the distribution of  the energy over the exposure period), it has
been shown throughout all TTS-experiments that the distribution of the sound
energy does make a difference in the TTS2 produced.

Analogously  it has been shown  that the same TTS2  is caused by exposure to
noises with quite  different total sound energies. In figure 3 this is again shown for
Ward's latest published  results (Ward,  1970) for TTS2 at 3000 Hz, caused by
exposures to  quite different noise patterns (single exposure daily, long and short
noise bursts). In this figure, TTi>2 has been plotted against the energy-equivalent
sound  pressure level (Leq). Since the definition of Leq is given later, it is suffi-
cient to state here that Leq is nothing but a measure of the total sound energy for
an 8 hour exposure, converted into a sound pressure level.  Although Lgq in figure
3 has a range of 16 dB (from 83 to 99 dB), which corresponds to a factor of 40 in
sound energy, TTS2 is about  20 dB for all exposures. However, plotting TTS200
(200 minutes being the longest recovery period examined in all tests) against Lgq,
results in a appreciable increase of TTS2QO w^n Leq-

From this  it  is clear  that recovery from TTS is dependent on the sound energy
which  created the TTS. Since  it  is unknown at the moment which processes
underly the realization of permanent threshold shifts and  how TTS is involved in
these processes, it may be possible that recovery, and  hence, total sound energy
over a workday, plays  a more important role than TTS2 alone.

                               174

-------
         TTS  AT
         3000 Hz
          30i-
          dB
          20
           10
                            A TTS2

                            O TTS200
0
80
                                     I
                                                  I
                                    90          dB        100

                      EQUIVALENT SOUND  PRESSURE  LEVEL
Figure 3. Temporary threshold shift (TTS) at 3000 Hz, measured 2 and 200 minutes after exposure to
noise, as a function of the equivalent sound pressure level of the noise.
                                   175

-------
2.    In  1967 Botsford simplified the CHABA damage risk contours.  He recognized
     that all octave-band SPLs on the same contour for one single exposure daily (see
     figure  1), were  always  assigned practically  the  same exposure time limits for
     exposures to intermittent noise (long bursts, as well as short bursts). For instance,
     the CHABA damage risk contours for exposures to intermittent noise show that
     95  dB SPL at the octave band with midfrequency 1000 Hz and 90 dB SPL at the
     octave band with midfrequency 400 Hz (lying on the same contour of figure 1)
     both require, after an on-time of 55 minutes, an  off-time of 60 minutes; and for
     short noise  bursts, in both instances an on-fraction of  0.6 is allowed  for an
     exposure  time of 480  minutes  a  day. This fact,  that all octave  band  sound
     pressure levels on the same contour  have  the same exposure time limits for
     intermittent noise, irrespective of the exposure pattern assumed, indicates that
     the contours are general curves of equinoxious noise. The  next step of Botsford
     was to assign to each  contour an A-weighted sound level, by comparing from 580
     manufacturing noises their height  of penetration of their octave  band  sound
     pressure levels into the contours with their A-weighted sound level.
         The  results of the analysis by Botsford are shown in figure 4. The three
     curves with the highest sound levels should not be  relied on, since they come from
     an extrapolated CHABA-curve.
3.    In the "Guidelines for noise exposure control" (1970) of the Inter-society Com-
     mittee, the Botsford curves have been used; these guidelines have been included in
     the Department of Labor Standards-Walsh Healy Act-(Federal Register 34,1969)
     and are legally applicable  in the USA to industries performing work under the
     Governments Public Contracts Act.
         In the  Guidelines,  the curves given by Botsford have been modified some-
     what by shifting the 90 dB(A) curve to 480 minutes and by shifting the curves for
     the  lowest numbers of exposure cycles per day (up to 3 exp. cycles per day) to
     higher total on-time values per day. Apart  from  this figure in tabular form, the
     Guidelines also present the simple rule that exposure to 90 dB(A) is allowable for
     a full 8 hours, with an increase  of 5 dB(A)  for each halving of exposure time. As
     the  document states:  here an allowance is made for the number  of occurrences
     ordinarily found in high level noise. Referring to figure 4, this rule is overpro-
     tective when the noise  comes in short bursts,  but is highly underprotective for
     single uninterrupted exposures. E.g. calculating the TTS2,  due to an exposure for
     30 minutes to 110 dB(A)  (which is allowable  according to the rule mentioned)
     results in a TTS2 at 2000, 3000 and 4000 Hz of 25, 36 and 33 dB resp. which is
     on the average more than  10 dB above the TTS2-values from an 8 hour exposure
     to a constant sound level of 90 dB(A).
         A rule for varying noises is also given.  The ratio of the time spent at  a given
     sound level  to the allowable time at that level is calculated  and these fractions for
     all  occurring sound levels are added. If the resulting number is less than 1.0, the
     exposure is safe, and if it is more than 1.0, the exposure is unsafe.
         Here, sound levels below 90 dB(A)  do not enter into the calculations, al-
     though exposure to sound levels just below  90 dB(A) is hardly safe for 8 hours a

                                   176

-------
TOTAL  ON-TIME  PER
DAY  IN  MINUTES
          400

          200

          100
           60
           40

           20

           10
            6
             2
             1
                                        A-WE:IGHI   D
                                                   LEVEL
                          FOR  MANUFACTURING
                                   NOISES
                         4 6  10   20   40 60  100  200  UP
                        NUMBER  OF   EXPOSURE
                        CYCLES   PER  DAY
Figure 4. Total duration of a noise allowable during an 8-hour day as a function of the number of exposure
cycles per day. An exposure cycle is completed each time the sound level decreases to or below 89 dB(A).
The interruptions of potentially harmful noise are assumed to be of equal length and spacing so that a
number of identical exposure cycles are distributed uniformly throughout the day. The A-weighted sound
levels assigned to the curves were determined from manufacturing noises and may not apply to noises from
sources of other types.
                                177

-------
         day. A more realistic approach should have been to take into account sound levels
         below 90 dB(A) as well.
    4.   The American Conference of Governmental Industrial Hygienists (1968) proposed
         threshold limit values for noise, allowing 92 dB(A) for 4 to 8 hours a day, 97
         dB(A) for 2 to 4 hours, 102 dB(A) for 1 to 2 hours and 107 dB(A) for less than 1
         hour. In evaluating exposure to varying noise, the same line as in the guidelines
         was followed, exposures to sound levels of less than 92 dB(A) do not enter into
         the calculations. However,  one single  exposure for one hour  to  107 dB(A)
         (assuming a spectrum according to the CHABA equinoxious curves) results in an
         average TTS2 at 1000 and 2000 Hz of 25  dB and at 3000,4000 and 6000 Hz of
         40 dB. It seems difficult to understand why a conference of hygienists proposed a
         limit for exposure to constant noise greater than 90 dB(A)—a level which is now
         more or less generally agreed to be the maximal allowable limit—and also pro-
         posed limits for varying noise that are even less stringent than those for constant
         noise.

II.  NOISE-INDUCED HEARING LOSS FROM  EXPOSURE TO INTERMITTENT AND
    VARYING NOISE.
    Selection data
    A thorough study of the relevant literature  has been undertaken to select papers in
    which data were given that  could  enable us to relate noise exposure to noise-induced
    hearing loss for exposures  to varying  and  intermittent noise over  the workday.  In
    selecting papers the following considerations were taken into account:
    (1)  HLs  of the subjects measured a considerable time (mostly more than 12 hours)
         after their last exposure to job noise,  to permit significant recovery from tem-
         porary  threshold shift  from such noise. From the group of miners reported by
         Sataloff et al. (1969) audiograms were taken right after coming up from the
         mines, but in a pilot study it was shown that no significant TTS was at that time
         included in the hearing levels.
    (2)  Subjects selected with no previous exposure to noise at other jobs nor any prior
         ear damage or clinical abnormality. Although Sataloff reports exposure to gunfire,
         his group of miners was nevertheless included. Reasons will be given below.
    (3)  Number of subjects at least 25 per group, unless data were taken  from  a sub-group
         of a larger group. Selection  of a particular subgroup was based on number of
         subjects and exposure time.
    (4)  Total exposure time preferably more  than 10 years. When data were given for
         shorter exposures only, but for more than 4 years, these data have been included.
    (5)  Occasional wearing of ear protection at the time of the survey, reported in two of
         the papers, is included in our analysis. Both, however, report  a long service
         without ear protection and no differences were found between  the men wearing
         ear protection and those without ear protection.
    (6)  Noise exposures reported to be to several sound levels, the difference between the
         highest and lowest sound level  at  least 25 dB(A) or so. Only those surveys were
         included, for which it was sure that the noise environment did not change over

                                       178

-------
    the years. Papers dealing with exposure to noise fluctuating on a short time scale
    (impulse, impact noise) have not been included in the analysis.
(7)  Sufficient data on noise exposure to allow the calculation of a characteristic noise
    parameter of such  exposures.  Data on  the  noise  exposures were given  in the
    following ways:
    —    Overall time-distribution of  sound levels over the workday (e.g. x%  of the
         time the sound level lies between a and  a+5 dB(A), y% of the time between
         a+5 and a+10  dB(A), z%  etc.) This includes also the exposure to one par-
         ticular sound level for x% of the workday, the rest of the time being "quiet'*.
    —    time distribution over the workday of mainly one sound level (e.g. 3 minutes
         exposure to  x  dB(A), followed by 5 to 10 minutes  of "quiet")-  Unfortu-
         nately, only a few authors,  from which the data were included  in the
         analysis, give the numerical values of the sound levels  during the quiet inter-
         vals. Most authors describe the quiet intervals in a qualitative way.
Through  this stringent procedure of  data  selection,  out of about hundred papers,
(listed at  the end of this paper under the heading: further literature consulted), eleven
papers, dealing with 20 groups of subjects, could be included in the analysis.

Presentation of data

Unfortunately, insufficient data have been included in the papers about the  spread in
the hearing levels to permit any analysis of this very important subject. Therefore, the
following refers only to median and  average hearing levels and median and average
noise-induced hearing losses.
    All median (or in a few instances, average) hearing levels presented in the papers
have been converted to ISO standards, if necessary. The values given by Spoor  (Spoor
1967) and shown in figure 5 of the  age-dependent median hearing levels of non-noise-
exposed otologically normal people have been subtracted from the median and average
hearing levels of the groups, to calculate the median noise-induced hearing losses.
    Since the mean ages of the several groups are mostly around 40 years, only small
values had to be subtracted from the actual hearing levels (see Table I).
    As most noise data were presented in sound levels in dB(A), those given in octave
band sound pressure levels have been converted into sound levels in dB(A) too.
    An attempt has been made to give a classification of the noise exposures based  on
details given in the papers. Although it is realized  that intermittent noise is included in
varying noise, for the purpose of this paper a more specific definition of varying and
intermittent noise  is given.  Intermittent noise is here defined as noise with  a large
difference (at least 20 dB(A) or so) between the highest and lowest sound levels, and
where  sound levels between these  levels  are present during a negligible tune only.
Ygrying noise is here defined as noise in which several sound levels occur in the course
of time and where sound levels between the highest and lowest sound levels are  present
during a  considerable time. Looking at the times during which the sound levels remain
at a given level, the noise exposures may be grouped in exposures with short and long
times at  a  given sound  level. The limit of short-time noise exposure has been, quite

                                    179

-------
                                        FREQUENCY  IN   Hz
                           500      1000
2000     £000     8000
       3000     6000
                       ""30
                         35
MEDIAN   HEARING  LEVEL

Figure 5. Median hearing levels of otologically normal men. not exposed to noise during working hours, as a
function of frequency. Age in years is parameter.
    arbitrarily, chosen to be  5 minutes. None ot" the noise exposures to be considered here
    have exposures at a given level of less than 2 minutes (taken as the CHABA-limit for
    short noise  bursts).  The 4 resulting  classes are  illustrated in  figure 6. To give  an
    indication of the variations involved, for the varying noise exposurcs the difference is
    given between the sound level exceeded  for  27r of the  time and the sound level
    exceeded  for 5(KT of the time. For the intermittent noise exposures, the total times per
    workday at the highest sound levels are given (see Table I >.
                                     180

-------
      INTERMITTENT  NOISE
      SHORT  INTERVALS
          P.    n.  _
    INTERMITTENT  NOISE
    LONG  INTERVALS
           H	1—I Mill!	1	1—I Mill
00
                                       • 10dB(A)
H	1—I  MINI	1	1—I  I I I I II
      VARYING  NOISE
      SHORT INTERVALS
    5min
    VARYING  NOISE
    LONG  INTERVALS
           Figure 6. Illustration of classification of noise exposures into intermittent and varying noise exposures with
           long and short periods at a given sound level.

-------
         From  the noise data, for each group, the A-weighted energy equivalent sound
     level has been calculated according to

                                      !  t __LA(t')/10
Leq=101ogi/ 10            dt'
              o
     where t is the total daily exposure time (480 minutes) and LA(tr) is the momentary
     sound level at time tf.
         When  the  various levels are given in n classes, then the above formula can be
     rewritten as
                                       i= 1

     where Lj is the average sound level of class i (assuming the class widths are small) and t j
     is the total time per workday, during which  the sound levels are within class i.
          An analogous formula, although modified for a total exposure time of one week
     (40 hours) has been used in the ISO Recommendation  1999  (Assessment of noise
     exposure during work  for hearing conservation  purposes), to determine equivalent
     continuous sound levels. Essentially, the formulae given above calculate the A-weighted
     total sound energy per reference period t, converted into a sound level.
          Relevant data have been included in Table I.

Analysis of data

     Although not necessary, it seems advantageous to compare data for varying and inter-
mittent noise with data for constant noise. The data given here permit comparison with two
sets of results for more or less constant noise, namely those of Burns and Robinson (1968)
and those of Passchier-Vermeer.
     The results of the report by Passchier-Vermeer relate to continuous 8-hour exposures
to constant steady -state broadband noise for exposure times of at least 1 0 years. It is merely
a compilation of data found at that time in the relevant literature  (Bums (1964), Gallo
(1964), Subcommittee Z 24-X-2 (1954), Rosenwinkel (1957), Nixon (1961), Taylor (1961),
Kylin (1960), and Van Laar (1964).  Without going into any  detail here, the results are
shown in  Fig. 7, for  the frequencies 2000, 3000 and 4000 Hz and for an exposure time of
1 5 years.  This exposure time has been chosen  as being the average exposure time of the
groups under consideration here (see table I). To make a comparison possible, the median
noise-induced hearing losses due to the varying and intermittent noise exposures had to be
calculated for an exposure time  of 15 years  as well. Although this involves mostly slight
corrections, these corrections have been applied to the values given in the Table. It should
be noted here too that in order to take into account age-effects on hearing, the values given
by Spoor have been  applied  to both sets of data. The results are shown in Figs. 8 to 1 3 for
the frequencies 500 to 6000 Hz.

                                        182

-------
                                                            Table
Author     number    mean      average
           of                  e::p.
           subjects  age       time
                     in  years   in years
equivalent  indication of   exposure long
sound  •     intermittent or  or short times
level in    varying noise   at  the prevail-
dB( A)                       ing sound levels
average or median noise-induced
hearing loss (age-corrected
according to Spoor) at
500Hz lOOOHz 2000Hz 3000Hz 4000Hz 6000Hz
Cohen
Cohen
Taylor
Feiser
Feiser
Flach
Oo Kuiper
U> 1
Kuiper '
Ephraim
Holmgren
Holmgren
Pressel
Schneider
Schneider '
Satalof f
Jirak
Jirak
Jirak
Jirak
Jirak
18
14
40
16
16
233
144
23
97
59
26
230
69
26
26
23
25
41
40
50
37.5
30
35
45
38
57
44
41
38
33
35
45
42
39
36
38
40
40
12.5
17.5
4
7.5
12.5
15
20
40
10
10
9
4
10
20
15
17.5
12.5
15
15
15
86
86
82
102
102
82
99
99
95
95
98
94
83
83
ll'i
103
88
100
74
108
int. 120 min
int. 120 min
int. 75 min
vary. 20dB(A)
vary. 20dB(A)
vary. 25dB(A)
vary. l5dB(A)
vary. l5dB(A)
vary. l4dB(A)
vary. l4dB(A)
vary.!2.5dB(A)
int. 120 min
vary. 15dB(A)
vary. !5dB(A)
int. 180 min
int. 190 min
int. 115 min
int. 280 min
int. 250 min
int. 212 min
short
short
short
long
long
long
long
long
long
long
long
long
long
long
short
short
short
short
short
short
1
-1
-
11
12
1
-
-
2
2
-
0
12
-
-
-
-
-
2
1
-1
14
14
1
-
-
2
5
-
0
1
17
_
-
-
-
-
3
7
2.5
21
26
0
16
27
15
4
8
10
0
29.5
5
0
3
0
11
3
13
4
26
33
-
30
35
30
16
14
12
0
1
34
-
-
-
-
_
5
10
5.5
30
34
2
38
35
34
18
16
20
2
3
36.5
27
12
20
5
28
12
20
7.5
19
26
-
22
25
25
18
17
21
6
14
38.5
-
-
-
-
-

-------
     The results of Burns and Robinson do not only refer to eight-hour continuous expo-
sure to a constant sound level; noise exposures with fluctuating sound levels on a short time
scale (in the order of seconds) were also involved in their analysis. The authors relate
noise-induced hearing loss to noise immission level E^2> this quantity being equal to LA2 +
10 log T, where T is the total exposure time (e.g. in years) and L^2 is  the sound level
exceeded in 2% of the time during a workday. In their survey, L^2 ~ ^A5Q ranged from 0 to
10 dB(A), but was as much as 15 dB(A) in exceptional cases. As their report states:  "It is
more convenient to express the noise level in dB(A) in the  usual way (by meter reading)
rather than in  the form L^2> and the  average relation between the two (based  on 280
specimen noises in the survey) may be taken as L^-LA = 3.7 dB(A). Since it was found
that Lj\2 corresponds better with hearing loss than LA, using  LA instead of LA2 will be less
exact  for strongly fluctuating  noise environments". All in  all, for a constant or nearly
constant noise level, noise immission can be expressed in the form LA + 10 log T, where LA
is determined by sound level meter reading and is, according to the measuring characteristics
of sound level meters, equal to Leq.
     To permit comparison between the results of Passchier-Vermeer and those of Burns
and Robinson, for both surveys the median NIHLs at 2000, 3000 and 4000 Hz have been
plotted against Leq for an exposure time of 15 years (figure  7). Although it does not seem
the right occasion  here to go  into details about agreements and differences between the
results of both surveys, nevertheless some attention has to be paid to this subject now.
     There  is a good agreement between the median  NIHLs at 2000  Hz and lower fre-
quencies and, to a lesser extent, at 6000 Hz. However, at 3000 Hz and especially at 4000
Hz, large differences can be shownr at the highest sound levels considered. According to
Bums and Robinson, this might be mainly due to different subject-selection criteria used by
Bums and Robinson and  by the authors  of the several papers from which the values given
by Passchier-Vermeer have been derived.
     Only the fourth selection criterion used by Bums and Robinson was not used in any of
the other studies, namely that  audiograms should be compatible with clinical findings. I am
at a loss to indicate the consequences of this criterion. Also for a few of the groups used in
Passchier-Vermeer's analysis, the selection criterion  concerning military service was not as
stringent as in others,  but median results from these groups could not be distinguished from
the other median NIHLs. Moreover, subject selection does not explain that the differences
do increase with sound level. If subject-selection  were the appropriate cause, one would
expect that the  curves should not be divergent but parallel. Looking further for a possible
explanation of the  differences found (noise measurements, variance of hearing levels), none
seem to explain these differences.
     However, left  with two sets of curves for exposure to constant  noise, it seems advan-
tageous to compare the median NIHLs from the exposures to intermittent and varying noise
with both sets of curves.  As already mentioned, in figures 8 to 13 the median NIHLs from
intermittent and varying noise exposures have been plotted against Lgq,  along with  the
curves given by Passchier-Vermeer for constant noise.
     In, Figs. 14 to 19 the median NIHLs are plotted against Lgq + 10 log T, along with the
curves given by Burns and Robinson. Since in Burns and Robinson's survey, median NIHLs
have been calculated from the  actual median HLs, by subtracting median HL of non-noise

                                        184

-------
        MEDIAN
        NOISE-INDUCED
        HEARING  LOSS
            50
            dB
                 PASSCHIER
               _ ROBINSON.
        4000 Hz
                                                        4000 Hz
                                                      / 3000 Hz
                   EXPOSURE TIME
                      15 YEARS
            30
                                   90
100
SOUND
 110 dB(A)
LEVEL
Figure 7. Median noise-induced hearing losses at 2000, 3000 and 4000 Hz, caused by continuous exposure
to steady-state broadband noise for 15 years, as a function of the sound level in dB(A). Curves presented by
Burns and Robinson (1968) and Passchier-Vermeer (1968).
                                 185

-------
    MEDIAN
     NIHL
     I
 20
dB
                EXPOSURE TIME    15 YEARS
       10
            • VAR
            • INT
500 Hz
                                                     PASSCHIER
               80
                                  90          100         110 dB(A) 120
                               EQUIVALENT  SOUND  LEVEL
Figure 8. Median noise-induced hearing losses at 500 Hz from exposure to varying and intermittent noise
for 15 years, as a function of the equivalent sound level in dB(A). Curve presented by Passchier-Vermeer
(1968) for exposure for 15 years to steady-state broadband noise.

Figure 9. Median noise-induced hearing losses at 1000 Hz from exposure to varying and intermittent noise
for 15 years, as a function of the equivalent sound level in dB(A). Curve presented by Passchier-Vermeer
(1968) for exposure for 15 years to steady-state broadband noise.
                                       186

-------
  MEDIAN
  NIHL
       30
EXPOSURE TIME:  15  YEARS
                    80
         90          100         110 dB(A) 120
       EQUIVALENT  SOUND  LEVEL
Figure 10. Median noise-induced hearing losses at 2000 Hz from exposure to varying and intermittent noise
for 15 years, as a function of the equivalent sound level in dB(A). Curve presented by Passchier-Vermeer
(1968) for exposure for 15 years to steady-state broadband noise.

                                    187

-------
    MEDIAN
    NIHL
      50
EXPOSURE TIME:  15 YEARS
                               90         100         110 dB(A) 120
                          •^EQUIVALENT  SOUND   LEVEL
Figure 11. Median noise-induced hearing losses at 3000 Hz from exposure to varying and intermittent noise
for 15 yean, as a function of the equivalent sound level in dB(A). Curve presented by Passchier-Vermeer
(1968} for exposure for 15 years to steady-state broadband noise.

-------
   MEDIAN
   NIHL
     50
      EXPOSURE  TIME :   15 YEARS
      70
80
  90          100
EQUIVALENT  SOUND
110  dB(A) 120
LEVEL
Figure 12. Median noise-induced hearing losses at 4000 Hz from exposure to varying and intermittent noise
for 15 years, as a function of the equivalent sound level in dB(A). Curve presented by Passchier-Vermee
(1968) for exposure tr»r 15 years to steady-state broadband noise.
                                    189

-------
MEDIAN
NIHL
                   EXPOSURE  TIME:  15 YEARS
                                               PASSCH  ;R
                           90          100         110 dB(A) 120
                          EQUIVALENT SOUND  LEVEL
Figure 13. Median noise-induced hearing losses at 6000 Hz from exposure to varying and intermittent noise
for 15 years, as a function of the equivalent sound level in dB(A). Curve presented by Passchier-Vermeer
(1968) for exposure for 15 years to steady-state broadband noise.

                                190

-------
cq.
                                                             L°9
Figure 14. Median noise-induced hearing losses at 500 Hz from exposure to varying and intermittent noise,
as a function of the noise immission level (L^  + 10 log T, where T is the exposure time in years). Curve
presented by Burns and Robinson (1968) for continuous exposure to more or less constant noise.

-------
Figure 15. Median noise-induced hearing losses at 1000 Hz from exposure to varying and intermittent noise,
as a function of the noise immission level (l_eq + 10 log T, where T is the exposure time in years). Curve
presented by Burns and Robinson (1968) for continuous exposure to more or less constant noise.

-------
Figure 16. Median noise-induced hearing losses at 2000 Hz from exposure to varying and intermittent noise,
as a function of the noise immission level (L^ + 10 log T. where T  is the exposure time in years). Curve
presented by Burns and Robinson (1968) for continuous exposure to more or less constant noise.

-------
                                                 105           115
                                                LCq.  +  10  Log  T
Figure 17. Median noise-induced hearing losses at 3000 Hz from exposure to varying and intermittent noise.
as a function of the noise immission level (!_ + 10 log T, where T is the exposure time in years). Curve
presented by Burns and Robinson (1968) for continuous exposure to more or tess constant noise.

                                            194

-------
           MEDIAN
           NIHL
                                                            4000 Hz

                                                       ROBINSON
                                            105          115
                                            Leq. + 10 Log  T
Figure 18. Median noise-induced hearing losses at 4000 Hz from exposure to varying and intermittent noise,
as a function of the noise immission level (L™ + 10 log T, where T is the exposure time in years). Curve
presented by Burns and Robinson (1968) for continuous exposure to more or less constant noise.

                                        195

-------
             MEDIAN
             NIHL
                                               105          115
                                               Leq   +  10 log  T
125
Figure 19. Median noise-induced hearing losses at 6000 Hz from exposure to varying and intermittent noise,
as a function of the noise immission level (Lgq +  10  log T, where T is the exposure time in years). Curve
presented by Burns and Robinson (1968) for continuous exposure to more or less constant noise.

                                           196

-------
exposed people that differed from those given by Spoor, the values given in the Table have
been adjusted in order to be comparable with the curves given by Burns and Robinson. In
general, corrections of only a few decibels had to be applied.

Discussion

     Let us first consider the relation between median NIHLs at 500, 1000 and 2000 Hz
and the equivalent sound level or noise immission level. For equivalent sound  levels up to
100 dB(A) or noise imission levels up to 110, there is a reasonable agreement between the
curves for exposure to constant noise and the results for varying and intermittent noise. For
higher levels,  however,  the  median NIHLs from exposures to intermittent noise  are lower
than those from constant noise.
     For 3000, 4000 and 6000 Hz, the same applies as far as it concerns the curves given by
Passchier-Vermeer for constant noise exposure. When comparing the  median NIHLs due to
varying and intermittent noise exposures with those  for constant noise exposures given by
Burns and Robinson, it turns out  that all values are above the curves, except those from
intermittent noise exposures with noise immission levels above 110, which lie near or under
the curves at 3000 and 4000 Hz.
     Comparing the median noise-induced hearing losses from intermittent noise exposures
with those from varying noise exposures, without referring to any constant noise  exposure,
it is quite clear from  the figures that intermittent exposures below an  equivalent sound level
of 100 dB(A) agree quite  closely with varying noise exposures, but that exposures to
intermittent noise  with equivalent sound levels of at least 100 dB(A) do cause  median
NIHLs equal  to those caused by varying noise exposures  with equivalent sound levels of
about  10 dB(A) lower.  Therefore, it can be  stated  that at least some intermittent  noise
exposures are less harmful than would be expected from their equivalent sound levels, and
although not necessarily following from this, it seems reasonable to accept that intervening
quiet periods  during  work time do reduce NIHLs. Then, however, the question remains why
the exposures to intermittent noise with equivalent sound levels below 100 dB(A) do not
cause less NIHL than should be expected from their equivalent sound levels. A solution is
not found by considering the lengths of the intervening quiet periods during the  day, since
these are on  the average longer for the equivalent sound levels below 100 dB(A) than for
those above 100 dB(A). Also, the number of exposure cycles  per workday does not con-
tribute to a possible solution, since it is not  systematically different for exposures  with
equivalent sound levels below and above 100  dB(A) respectively. The only, at  the moment
unverifiable, explanation may be that the sound levels during quiet intervals were higher, or
above  a certain sound level, for the exposures with equivalent sound levels below 100 dB(A)
than those  for the equivalent sound levels above 100 dB(A). In agreement with this explana-
tion is the fact that the people exposed to  the high equivalent sound levels have  their
profession  in the mining industries, which mostly coincides with low background sound
levels.
     Rating the intermittent noise exposures according to the CHABA damage risk contours
and considering their criteria for NIPTS (10 dB at 1000 Hz and below, 15 dB at 2000 Hz
and  20 dB  at 3000 Hz and above) with exposure time equal to 15 years instead of their 10

                                        197

-------
years, it turns out that all  exposures, except one, are rated correctly safe or unsafe. The
exception  concerns the only intermittent exposure considered here with noise  bursts of
more  than  5  minutes, which is incorrectly rated safe. However, the correct rating of the
intermittent exposures considered here does not necessarily validate the CHABA contours,
since almost all noise exposures are fairly well below or above these damage risk contours.
    Returning now to the varying  noise exposures, it  seems quite  difficult to decide
whether these exposures can be rated according to their equivalent sound levels or not.
Although median NIHLs at 2000 Hz and below can be estimated with reasonable accuracy
from the equivalent sound  levels, no firm decision can  be made  for the whole frequency-
range. However, in many instances, the limit of safe noise has been fixed at a sound level of
at most 90 dB(A) for a continuous exposure for 8  hours to steady-state noise. Figures 8 to
19 show, for equivalent sound levels of  at most 90 dB(A) or noise immission levels up to
about 100, that varying noise exposures can be rated according to their equivalent sound
level in order to estimate the median noise-induced hearing losses with reasonable accuracy.
    Finally,  it should  be pointed out that all data show that the equal-energy concept is
not overprotective for varying  noise exposures. Therefore,  revision is indicated of those
varying noise exposure limits that are based on the rule that for each halving of the exposure
time an increase of 5 dB(A)  in sound level is allowed.

SUMMARY

    A review is given  of proposed limits for exposure to intermittent and varying noise,
based on temporary threshold shift (TTS) measurements. Revision is suggested of the allow-
able sound levels during quiet periods in intermittent noise.  Recent TTS-experiments show
that TTS2 (TTS measured two minutes after the end of exposure, being the basis of all noise
limits considered) is not a consistent measure of a daily noise exposure, since retarded
recovery occurred after some intermittent noise exposures. Although TT$2 itself is only to
some extent dependent upon the total sound energy of the noise, the reviewer could show
that recovery from TTS after exposure is dependent upon the total sound-energy referred to
an 8 hour noise exposure.
    By using data from several papers, median noise-induced hearing losses at 500 to 6000
Hz, from exposures to varying and intermittent noise, have been related to their A-weighted
energy-equivalent sound level (Leq). Comparisons have also been made with published rela-
tions for exposure to constant noise.
    Relations between Lgq for varying noise,  with Leq  between 80 and 102 dB(A), and
median noise-induced hearing losses at frequencies up  to 2000 Hz, agree reasonably with
those  for constant noise. The same applies for median noise-induced hearing losses at 3000,
4000 and 6000 Hz, due to exposures to equivalent sound levels up to 90 dB(A); no decision
could be made for median noise-induced hearing losses at these frequencies, due to expo-
sures to higher equivalent sound levels.
    Results  for  intermittent noise  exposures, with equivalent  sound levels below  100
dB(A), agree  closely with the results for varying noise exposures. However, exposures to
intermittent noise, with equivalent sound levels  above 100 dB(A), are considerable less
harmful to hearing that should be expected from  their  equivalent sound levels. A possible
explanation is given in terms of low sound levels during quiet  intervals.

                                        198

-------
REFERENCES

W.D. Ward, A. Glorig, D.L. Sklar. Dependence of temporary threshold shift at 4 kc on
    intensity and time. J. Ac. Soc. Am. 30 (1958) 944-953
W.D. Ward, A. Glorig, D.L. Sklar. TTS  from octave-band noise: applications to damage-risk
    criteria. J. Ac. Soc. Am. 31 (1959)  522-528
W.D. Ward, A. Glorig, D.L. Sklar. TTS  produced by intermittent exposure to noise. J. Ac.
    Soc. Am. 31 (1959) 791-794
W.D. Ward, A. Glorig, D.L. Sklar. Relation between recovery from TTS and duration of
    exposure. J. Ac. Soc. Am. 31 (1959) 600-602
W.D. Ward, A. Glorig, W. Sellers. TTS in a changing noise level. J. Ac. Soc. Am. 32 (1960)
    235-237
W.D. Ward. The  use of TTS in the derivation of damage risk criteria for noise exposure. Int.
    audiol 5 (1966) 309-313
A. Cohen,  J.R. Anticilglia; P.L. Carpenter. Temporary threshold shift in hearing from expo-
    sure to different noise spectra at equal dB(A) level. J. Ac. Soc. Am. 57 (1972, II) 503 -
    507
W.D. Ward. Studies on the aural reflex. II Reduction of TTS from intermittent noise by
    reflex activity; Implications for damage risk criteria. J. Ac. Sdc. Am. 34 (1962) 234 -
    241
W.D. Ward. Recovery from high  values of  temporary threshold shift. J. Ac. Soc. Am. 32
    (1960)497-500
K. Schroder, E. Rempt.  Untersuchungen zum Larmpausen problem. Larmbekampfung 6
    (1962) 142-143
H.H. Jones, chairman. American Conference of governmental industrial hygienists proposed
    threshold limit value for noise. Am. Ind. Hyg. Ass. J. 29 (1968) 537-540
F. Schwetz, R.  Donner, G.  Langer, M.  Haider. Experimentelle Horermudung und ihre
    Riickbildung  unter  Ruhe- und  Larmbedingungen.  M.  Schrift Ohrenheilk.  Laryngo-
    Rhinol 104 (1970) 162-167
E. Lehnhardt, J. Bucking. Larmpausen  - eine Moglichkeit zur Prophylaxe der Larmschwer-
    gehorigkeit. Int. Arch. Gewerbepath, Gewerbehyg. 25 (1968) 65-74
W.D. Ward. TTS and damage-risk criteria for intermittent noise exposures. J. Ac. Soc. Am.
    48 (1970) 2 (part 2) 561-574
W. Klosterkotter.  Vorsorge- und Ueberwachungsuntersuchungen  bei Larmarbeitern,  nach
    der VDI-Richtlinie 2058 Blatt 2. Kampf dem Larm 18 (1971) 29-33
"Guidelines for noise  exposure control" prepared by Intersociety Committee on Guidelines
    for noise exposure control, Intersociety  Committee Report. Sound  and Vibr.  1970,
    21-24
G. Holmgren, L. Johnsson, B. Kylin, O. Linde. Noise and hearing of a Population of Forest
    Workers. Br. Ac. Soc. Spec. Vol. no. 1 (1971)
W. Passchier-Vermeer. Steady-state and  fluctuating noise, its effect on the hearing of people.
    Br. Ac. Soc. Spec. Vol. no.  1 (1971)
J. Sataloff, L. Vassalo, H. Menduke. Hearing loss from exposure to interrupted noise. Arch.
    Environ. Health 18 (1969) 972-981
M. Flach,  E.  Aschoff. Zur  Frage berufsbedingter  Schwerhorigkeit beim Musiker. Laryn-
    gologie 4.? (1966) 595-605
                                       199

-------
W. Feiser, R. Hauf, U. Heuft. Larmmessungen und audiometrische Untersuchungen in der
     holzverarbeitenden Industrie. A.S.A. 2 (1968) 38-43
W. Taylor, J. Pearson, A. Mair. The hearing threshold levels of dental practitioners exposed
     to air turbine drill noise. Br. Dent. Journ. 1965, 206-210
D.W. Robinson. The  relationships  between hearing  loss and noise exposure. NPL Aero
     Report Ac 32, 1968
G. Pressel. Horschaden durch Larm bei Ladearbeitern eines grossen zivilen Flughafens. Int.
     Arch. Arbeitsmed. 26 (1970) 231-249
K.D. Kryter, W.D. Ward, J.D. Miller, D.H. Eldredge. Hazardous exposure to intermittent and
     steady-state noise. Journ. Ac. Soc. Am. 39 (1966) (3) 451-464
J.H.  Botsford.  A new method for rating noise exposures.  Am. Ind. Hyg. Ass. J. (1967)
     431-446
E.J.  Schneider, I.E.  Mutchler, H.R.  Hoyle,  E.H.  Ode, B.B. Holder.  The progression of
     hearing loss from industrial noise exposure. Am. Ind. Hyg. Ass. J. 31 (1970) 368-376
A. Cohen, J.R.  Anticaglia, H.H. Jones. Noise-induced  hearing loss. Arch. Environ. Health 20
     (1970)614-623
W. Burns, R. Hinchcliffe, T.S.  Littler.  An exploratory  study of hearing loss and noise
     exposure in textile workers. The Ann. of Occ. Hyg. 7(1964) 323-333
R. Gallo, A. Glorig. P.T.S. changes produced by noise exposure and ageing. Am. Ind. Hyg.
     Ass. Journal 25 (1964) 237-245
The  relations of hearing loss to noise  exposure. A report by subcommittee Z 24-X-2 (1954)
     34
N.E.  Rosenwinkel, K.C.  Stewart. The  relationship  of hearing loss to steady-state noise
     exposure. Am. Ind. Hyg. Ass. Quart. 18 (1957) 227-230
J. Nixon, A. Glorig. Noise induced P.T.S. at 2000 and 4000 Hz. J.A.S.A. 33 (1961) 904-913
W. Taylor, J. Pearson, A.Mair, W. Burns. Study  on noise and hearing in Jute Weaving.
     J.A.S.A. 37 (1964) 113-120
B. Kylin.  TTS  and auditory trauma  following exposure to steady-state noise.  Acta Oto-
     Laryng. Suppl. 152(1960)
F. v.  Laar. Results of audiometric research at some hundreds of persons, working in differ-
     ent Dutch factories. Publication:  A.G./S.A. C 23  of N.I.P.G. TNO (1964)
A. Spoor. Presbyacusis values in relation to noise-induced hearing loss. Int. Aud. 6 (1967)
     48-57
W. Burns, D.W. Robinson. Hearing and noise in industry. London: Her  Majesty's Stationery
     Office 1970
J.P.  Kuiper.  Gehoorbeschadiging bij  machinale houtbewerkers.  Report from "Arbeidsin-
     spectie" 1968
A.C. Ephraim. Discontinu lawaai.  Report from "Arbeidsinspectie" 1970
ISO  Recommendation  1999. Assessment of noise  exposure during  work for hearing con-
     servation purposes 1971
W. Passchier-Vermeer. Hearing loss due to continuous exposure to steady-state broadband
     noise from "Institute of public health engineering" 1968
Z. Jirak, B. Mautner, J.  Kostal,  A. Andel,  C.  Losert.  Larnihorschaden bei Bergleute des
     Ostrau-Karwiner Kohlenreviers. Int. Arch. Arb. med. 28 (1971) 49-61

                                       200

-------
              EVALUATION OF THE HEARING DAMAGE RISK FROM
                 INTERMITTENT NOISE ACCORDING TO THE ISO
                               RECOMMENDATIONS

                       B. Johansson, B. Kylin and S. Reopstorff
                         Stockholm and Sandviken, Sweden

     Comprehensive international investigations, carried out during the last 15-20 years,
have finally  resulted in the publication of an international recommendation (ISO Recom-
mendation R 1999), which lays down the guiding principles for the risk of hearing damage
from noise. On the basis of this recommendation a Swedish standard, SEN 590111, has been
developed, and has been in force since February  1972.
     The assessment of  irregular noise is  in accordance  with  that specified in the ISO
Recommendation, and the damage risk limit  of 85 dB(A) has been taken as a satisfactory
.degree of safety.
     With regard  to regular  continuous noise, this recommendation is  based on reliable
empirical data. However, regarding the effect of irregular noise on hearing, from the point of
view of damage, sufficient data are still not available.
     In an earlier  Swedish investigation (Holmgren et al,  1971), which, comprised studies on
the hearing status of forestry workers who were exposed to very irregular noise from power
saws, it was  found that the conversion method  recommended apparently overestimates the
risk of hearing damage.
     The purpose  of the present study was to investigate in industry the irregular noise that
Usually occurs, to estimate this in accordance with  the recommendations, and to relate the
results to the hearing status found in the persons exposed. The study also includes experi-
ments on a  laboratory scale  in  which a comparison was made  between the temporary
threshold shifts which occur after  exposure to  both regular and irregular noise.

A.   FIELD  STUDIES

     Materials and Methods.  About 170  employees in  the  engineering industry were
selected. The subjects,  between the ages of 20 and 35 years, had to be exposed, without ear
defenders, to intermittent noise,  from which, however,  extreme impulsive noises were ex-
cluded. Exposure must have been  for not less than 2 years. Moreover, their working environ-
ment must, to a great extent, have remained unchanged since the beginning of the exposure
period.
     Routine otoscopy preceded the hearing examinations. Persons with a case history of or
with objective signs of hearing damage due to disease or accidents were excluded.
     The employees selected had  to  wear ear  muffs with an attenuation of about 25-30 dB
from the beginning of the working day up to the time the audiogram was made.
     The recording of the relevant noise for  the employee was made at  his place of work
with regard to his various jobs. The recording of the noise was continued for not less than
one  typical  working cycle. The  most usual  types of work were  sheet-metal production,
turning, welding and chiselling.

                                       201

-------
     The recorded noise was analyzed, in the laboratory, for spectral frequency and the
time distribution of the sound level  during a typical working cycle. The results of the
distribution analysis were subsequently used to calculate, in dB(A), the equivalent con-
tinuous sound level, according to the given ISO Recommendations.
     Results: The investigated employees  were distributed according to exposure level, in
the form of an equivalent continuous sound level in dB(A), ECNL, and the exposure time,
which is shown in table 1.
     On the basis of the composition of the material, the employees were divided into two
groups with regard to exposure time, namely less than and more than 5 years exposure time.
The mean exposure time was 3 and 9 years respectively.
     The mean age of the persons investigated was rather similar both for the short and the
long exposure time, with the exception of the group in the lowest noise level, less than 85
dB(A), (table 2.)
                                     Table 1.
      EMPLOYEES, NUMBER OF YEARS EXPOSED TO NOISE. WITHOUT EAR DEFENDERS,
                            AT DIFFERENT EXPOSURE LEVELS.
dB(A)
^85
85-90

90-95
>95
Years
2-3
7
11

12
5
4-5
7
19

6
4
6-10
17
30

19
8
11-15
4
12
Total
35
72
i
10 j 47
7
24
                                      Table 2.
AGE DISTRIBUTION FOR GROUPS WITH DIFFERENT EXPOSURE LEVELS AND EXPOSURE TIMES.
' dB(A) Exposure time
i
i >&
! 85-90 <'5
j
i '
\ 90-95 <5
i ->5
!>95 <5
: :5
n
14
21
29
42
18
29
9
15
Age, mean, years
24.7
29.2
29,0
29,9
29.4
31.2
30,4
29.7
                                      202

-------
     The employees' hearing loss was positively correlated both with increasing noise level
for the same exposure  time,  and with increasing exposure time at the  same noise level
(figure 1.). For the exposure level below 85 dB(A), there was slight deterioration in the
hearing status in relation to increasing exposure time.
     The mean  age of the persons in the group with less than 5 years exposure time was,
however, about 6 years lower than that of the corresponding group where the exposure time
exceeded 5 years.
     As regards the exposure level between 85 and 90 dB(A) the average hearing was normal
from a clinical point of view, i.e., hearing loss did not exceed 20 dB for the group that was
exposed for less than 5 years. For the group with more  than  5 years exposure time the
hearing status was slightly worse. This average hearing loss, however, does not interfere with
the so-called speech zone of the Swedish language. The mean audiogram for the noise level
90-95 dB(A) showed a further slight deterioration of the hearing, and the hearing loss now
affects the speech zone of the  Swedish language. Thus, the results show that an equivalent,
continuous noise  level between  90-95  dB(A) is a hearing-damage risk level, if the same
definition of hearing damage is applied as in the Swedish standard, SEN 590111.
     Table 3 lists  detailed data  on the composition of the material with regard to age,
exposure level, exposure time, and hearing status.
     A comparison between the results of the present investigation and those of a previous
investigation (Kylin,  1960), which dealt with the relation between the exposure level and
the hearing status of employees who were exposed to continuous noise, shows that exposure
to a continuous noise at a level  between  85-90  dB(A)  causes a hearing loss which corre-
sponds to that on  exposure to irregular noise with an equivalent, continuous noise level
between  90-95  dB(A). Thus the results support a previous observation (Holmgren  et al.
1971), that the  recommended conversion method for irregular noise causes a shift in the
damage risk limit corresponding to about 5 dB.

B.  LABORATORY STUDIES

    Material and Methods. Twenty male normal-hearing subjects between the ages of 18
and 27 years (mean age 23 years) were subjected to a series of noise-exposure experiments
through earphones. The temporary threshold shifts that occurred after exposure was meas-
ured by means of a Bekesy audiometer.
    The noise exposure consisted of both regular and irregular noise. The following condi-
tions were observed.
    a) The irregular noise corresponded  to the  type  of noise  that usually occurs in the
engineering industry.
    b) In the irregular noise the relation between the lowest sound levels and the highest
sound levels was such, that a certain restitution effect could be expected after the individual
peaks (figure 2.).
    c) The frequency spectrum for the regular noise was similar to that which occurred in
the high sound levels in the irregular noise (figs. 3 and 4). As suitable noise of the irregular
type, a two-minute  interval of a recording from a shipyard (fig. 2) was played. The peak
noise was produced by welders in the immediate vicinity who, during short periods, ground

                                        203

-------
                -M


                  10


                  :
             o

             f    1
             c
             z    w
                                         tr«gutncy H:


                           in     i»o     too     1000     noo    ax     MOO



, — >
X
^
\
V
\
s
\
s
\



<• »«.
85-90
00-95
> 95




,
^


X
X
N
^

HA IA\ i
dB (A) •
dB (A)
dB (At


>•>

i

**» ^


"-^


	
! Hi


ytors 01

	 ,
££^=







-



noiM

ir*-'











t

.
"-<










kpot


fc
\



_ • —





—

s

V
,3
\
\










^
— Jf3
*^

^^
rf*
^









/
s
r~

)
r*










v
\
1
/
/
S
s
S






                 -JO


                 -10


                   :
             r.
                  K


                  H
                 100
Tt»t (rtautncv Hi
in rso MO A


*«•
**






•*
/•
^^
?


,'








N
\
\
1
;
/
s
s
*





Figure 1. Average hearing loss (best ear) for group exposed to different levels of irregular noise. The area

marked with broken  lines covers  the  so-called  speech-zone of the Swedish language (G. Fant,  1948 and

1949).
                                                204

-------
                                  Table 3.
AVERAGE HEARING  LOSS (BEST EAR) FOR GROUPS WITH DIFFERENT EXPOSURE LEVELS
AND DIFFERENT EXPOSURE TIMES (Qlf Q2, Q3 INDICATE QUART!LES).

dB(A) ECKL
(mean)
^85
(83,5)



*85
(84.0)



85-90
(88,6)



85-90
(88.5)


90-95
(93.5)



90-95
(92.4)


>95
(98.9)


>95
(96.9)


Kumbers
Years of of
noise exp, subj.
<5 14 raean

Q!
Q2
Q3
>5 21 mean

Q!
Q2
Q3
<5 29 mean

Q!
Q2
Q3
>-5 42 mean
Q1
Q2
Q3
<5 18 mean

Qj
Q2
Q3
>5 29 mean
Ql
Q2
Q3
<5 9 mean
Ql
Q2
Q3
>5 15 mean
Q1
Q2
*3
Frequency

0.5
7.9

5.0
10,0
10,0
6.3

0.0
5.0
10.0
10.0

5.0
10,0
10,0
9.3
5,0
10,0
10.0
11,4

5,0
10.0
20,0
8.4
5,0
10,0
10,0
10,0
5,0
10.0
15,0
3.7
10.0
10,0
10.0

1
5.4

0,0
5.0
10.0
14,0

0,0
5.0
5.0
5.9

2,5
5,0
10.0
6,6
5.0
5.0
10. 0
7.5

5,0
5.0
10,0
6.6
0.0
5.0
5.0
5.6
3.8
5.0
10.0
8.3
5.0
10.0
10.0

2
5.

5.
5.
10.
4.

2.
5.
10,
7.

5.
s.
10,
9.
5.
10.
10.
8,

5,
5.
10.
a.
3.
5.
10.
10.
5.
10.
11.
13.
5.
10.
20.


0

0
0
0
5

5
0
0
1

0
0
0
2
0
0
0
1

0
0
0
3
8
0
0
0
0
0
3
3
0
0
0


7

5
5
10
5

0
5
10
10

5
10
15
15
5
10
20
11

5
15
18
20
5
15
32
15
5
10
25
30
10
30
45
in

3
.5

.0
,0
.0
,3

,0
.0
.0
.4

.0
,o
.0
,3
.0
.0
.0
.7

,o
.0
.8
.0
.0
,0
.5
,6
.0
.0
.0
.7
.0
,0
,0
kHz

4
12.1

10.0
10,0
15.0
10.8

5.0
10.0
15.0
14,8

10,0
15,0
17,5
23.2
10.0
20,0
30.0
19.7

10.0
17,5
25.0
28.3
10,0
30,0
40,0
27,8
10.0
30.0
41,3
46,0
27,5
55,0
60,0


6
11.

5.
10.
20.
15.

7.
10.
25.
16.

10.
15.
25.
22.
15.
20.
30.
25.

10.
25.
35,
26.
15.
25.
35.
21.
13.
25.
31,
51.
37.
50,
63,



n

0
0
0
5

5
0
0
3

0
0
0
4
0
0
0
8

0
0
0
1
0
0
0
7
8
0
3
0
5
0
8


8
10.8

5.0
7.5
15.0
9.7

5.0
10.0
15.0
10.8

6,3
10.0
15.0
16.1
10.0
10.0
25.0
21.4

10.0
15.0
27.5
17.9
5.0
15.0
35.0
17.5
12.5
15.0
22.5
39.4
22,5
35.0
56.3
                                   205

-------
  Noise level dB(A)
  no -
  100 i
  80 -

                           .

                                                             .
                        Figure 2. Example of the irregular noise used.

down the irregular joint between  two steel plates that were welded together. The back-
ground noise was produced by the activities occurring in the other parts of the large welding
hall. For the regular noise a suitable section of the recording of the continuous grinding
noise was played. The above-mentioned recording was subsequently played several times on
to a tape recorder, so that a exposure time of one hour was obtained for each type of noise.
     For the regular noise an exposure level was decided on  which produced a maximal
threshold shift  of 20 dB measured 15 minutes after exposure.  For the  irregular noise the
highest sound level chosen  was that compatible with the subjective tolerance threshold for
some of the more sensitive subjects. In addition to this, two lower levels, in 5 dB steps, were
used for each type of noise in the experimental  noise exposures.
     Each subject had to undergo  noise exposure tests in the  following  order: first, the
lowest level  of the regular noise, followed by the intermediate and highest level respectively:
then the  irregular noise  in the same order. If the threshold shift was greater than 10 dB, the
next experimental-noise exposure test was carried out only after an interval  of at least 48
hours. Otherwise, the  intervals were usually  24 hours (minimum). For audiometry, the
frequencies were  tested  in the following order: First,  2.5 kHz to 8kHz, then from 2.5 kll/.
to 0.5 kHz, the  right  ear  first. Each subject  was trained in  Bekesy audiometry until he
reacted regularly  and reproducibly. For each  ear the hearing  test took 8 minutes. Audi-
ometry was started 15 minutes after the end of the exposure, and was thus  completed 31
minutes after the end of the exposure. A control  audiogram was made before each experi-
mental noise exposure.
                                         206

-------
               08 >•! J.IO  * N-m?
              31 S    U     ITS    350     MO   1000    ?OOO    40OO   ROOD   '6000
               31 i    63     IJ4    740     MO    1000   MOO    «OOO    8OOO    HOOO  M,
Figure 3-4. Frequency spectrum for regular noise (above) and irregular noise (below).
                                          207

-------
     Results: The mean hearing-threshold shifts are shown in table 4.
     A significant (p<0.05) effect of the noise exposure (indicated by underlining) occurred
mainly at the test frequencies 3, 4, and 6 kHz. Although the size of the threshold shifts was
small throughout,  nevertheless there was an evident  trend for  these to increase with in-
creasing noise level.  On exposure to regular noise, a  sound level of 95 dB(A) produced a
definite effect  at the  three  frequencies,  whereas for the irregular  noise,  this effect was
obtained already at  92 dB(A) equivalent continuous  noise level, calculated in accordance
with the ISO Recommendation. However, this threshold  shift was, throughout, less than
that for the 95 dB(A) continuous noise level.
     Moreover, the results indicate that at first the highest noise level for the irregular noise
caused a temporary  threshold shift corresponding to that obtained at 95  dB(A) for con-
tinuous noise.  According to  the conversion method used,  the equivalent level  for the
irregular noise  was  calculated to be 97 dB(A). Consequently,  the  investigation has con-
firmed,  to a  certain  extent,  the results of the field study,  namely, that the risk of hearing
damage is somewhat  overestimated by the conversion method.

DISCUSSION

     In  all retrospective investigations  whose purpose is to look for a correlation between
the degree of influence and exposure to the relevant substance, there are always difficulties
in obtaining  representative and well-defined material. During recent years  propaganda for
the use of ear defenders has been increasingly intensified.  In  the present investigation, this
situation may have led to some uncertainty in assessing  the noise exposure to  which the
-persons investigated  were exposed to. On the whole, however, this uncertainty is the same
for all the groups investigated, and consequently the results for these are comparable.
     The  relatively high correlation between increasing deterioration in hearing status and
increasing noise  level, which agrees also with relevant facts  previously known (Kylin, 1960),
supports the  view that  the estimation of the exposure conditions for the present selection of
the population seems to be pertinent.
     The results of the field study show, however, that the method employed in  evaluating
the exposure level for  irregular noise apparently causes a shift in the risk limit for hearing
damage by about 5 dB compared with the corresponding value for continuous noise. Similar
results were  obtained  in an earlier  investigation on the hearing status of forestry workers
who were exposed to very  irregular noise  which occurs  when  working with  power saws
(Holmgren etal., 1971).
     In the present investigation, the experimental study on the temporary  threshold shifts
after exposure to continuous and irregular noise has also shown indications that the con-
version method  used overestimate the damage risk limit when it concerns the irregular noise.

SUMMARY

     About 170 employees, between the ages of 20 and 35 years, who were exposed in their
work to the  irregular noise that usually ocurs in the engineering industry, were subjected to
a  hearing  examination. The aim of the examination was to  test a proposed method for
assessing the  risk of hearing damage from irregular noise.
                                        208

-------
                                      Table 4.

AVERAGE TEMPORARY THRESHOLD SHIFT IN 20 PERSONS 15-31 MINUTES AFTER EXPOSURE
TO DIFFERENT LEVELS OF STEADY STATE, AND VARYING NOISE. (SIGNIFICANT, PH 0.05
EFFECT OF THE NOISE EXPOSURE IS INDICATED BY UNDERLINING.).




Regular
Noise





Irregular
Koise



dB(A)
90 right
left
95 right
left
100 right
left
d3(A)
87 right
left
92 right
left
97 right
left

0.5
0.5
1.2
0
1.4
1.1
1.0

kHz
1 2
0
1
0
1
MM
1
MB!
0
Equivalent
0
0.3
0
0,2
0.2
1.0
0
0
0
0
0
0
0
.3 1.3
0,8
O. i.o
iJS 1.1
.7 2,2

3
1.2
3^5
2,3
2.1
lal
5.2
continuous
0
0
.3 0
.5 0,5
.7 0,5
.1 0.8
0.9
0,7
1,9
MMMMM
1.5
1,6
•••M^HMI
1.8

*
1.3
0.6
3,1
3,0
4.3
5.8
noise
0,5
1.1
1,7
2.5
2,0
3,1

6
1.5
1.1
2,6
5,0
7.4
8,7
level
2^2
1.1
2.8
2.8
5.0
_!—
    The proposed method  enabled the conversion of an irregular noise to an equivalent
continuous  noise level. The persons investigated were divided into groups with different
exposure levels, in 5 dB steps, between 80 and 100 dB(A).
    The mean hearing loss obtained showed a high correlation with increasing noise level.
The results  also indicated that the conversion method now used for irregular noise seems to
cause a shift in the damage risk limit of about 5 dB.
    In  the experimental study 20  persons with  normal hearing were exposed to both
regular and irregular occupational noise.  The temporary  threshold shift shift which  still
affected the subjects 15 minutes after the completion of the exposure was recorded.  The
irregular noise was converted to an equivalent continuous noise level. The results showed
that the temporary threshold shift after exposure to irregular noise was less than that after

                                       209

-------
exposure to the corresponding noise level for continuous noise. This supports the conclusion
that the conversion method employed  for irregular noise overestimates the damage risk
limit.

REFERENCES

ISO Recommendation R 1999, Acoustics. Assessment of Occupational Noise Exposure for
     Hearing Conservation Purposes. May, 1971.
SEN 590111,  Acoustics. Estimation  of risk of hearing damage from noise.  Measuring
     methods and acceptable values. Swedish Electrical Commission, 1972.
Fant, G.: Analys av de svenska vokalljuden.  LME Report H/P-1035, 1948.
Fant, G.: Analys av de svenska konsonantljuden. LME Report H/P-1064, 1949.
Holmgren, G., Johnsson, L, Kylin, B.  and  Linde, O.: Noise and Hearing of a Population of
     Forest Workers, in Robinson, D.W.  (ed) Occupational Hearing Loss. British Acoustical
     Soc. Spec. Vol. No 1, p 35^2, 1971.
Kylin, B.: Temporary Threshold Shift and Auditory Trauma Following Exposure to Steady-
     State Noise. An  experimental  and field study.  Acta  Oto-Laryngologica,  vol. 51:6,
     suppl. 152, 1960.
                                       210

-------
                      NOISE-INDUCED HEARING LOSS FROM
                        IMPULSE NOISE: PRESENT STATUS

                       R.R.A. Coles, C.G. Rice and A.M. Martin
                     {Institute of Sound and Vibration Research,
                        University of Southampton, England)

    Although  a large number of damage risk criteria have been published and there are
considerable variations in the legislation and codes of practice put forward by national and
state governments  and by numerous occupational authorities, it is evident that the relation-
ships between steady-state noise exposure and hearing loss are becoming a matter of general
agreement in all but  detail. Perhaps the main uncertainty revolves round the question of
intermittent exposures and fluctuating levels of noise,  and whether the energy principle
provides an accurate basis or acceptable approximation on which assessments of hazard can
be made.
    The divergence of opinion on  such exposures is illustrated well by comparing British
and American  attitudes. In the U.K.,  the Department of Employment's Code of Practice
(1972), based on the  research of Burns and Robinson (1970), utilizes the energy principle.
In the U.S.A., the criteria applicable under the Walsh-Healey Act (1969) and the Occupa-
tional Safety and Health Act (1970), together with those of several occupational groups and
those recommended in 1972 by the National Institute for Occupational Safety and Health,
favor a 5-dB correction per halving of daily exposure duration. It is interesting, though, that
the  American National Standards Institute, which of course benefits from the most com-
petent academic advice available and presumably represents the consensus of opinion of
American  industry, is a signatory to  the International Standard (1971) on occupational
noise exposure which utilizes the  energy principle. It is also  noteworthy that the U.K.
evidence is based mainly on Burns and Robinson's PTS (permanent threshold shift) studies,
whereas the U.S.  evidence appears  to be based on  TTS2 (temporary  threshold shift 2
minutes after cessation of noise exposure): this is probably a significant procedural  differ-
ence, and further calls into question the  relevance of TTS studies for derivation of correc-
tion factors for interimttency of exposure or fluctuation of noise level.
    In comparison to the situation concerning steady-state noise, there is remarkably little
legislation or governmental guidance on impulse noise hazards. Indeed, in most instances the
laws, codes and criteria specifically  exclude impulse noise, except perhaps where it consists
of rapidly repeated impacts.
    In looking at the historical aspects  of noise-induced hearing loss, impulse noises have
been prominent causes and have led to such descriptive terms as "gunfire deafness" and the
"boilermaker's notch" (referring to  the  4-kHz dip  in the audiogram). At  the same time,
these noises have  been the most difficult to quantify.  An almost traditional awe has sur-
rounded them, such that the writers of earlier criteria have spoken guardedly of additional
hazards  where a noise contains impulsive components, or they have even suggested addi-
tional arbitrary hazard factors equivalent to an increase in sound level of up to 10 or 15 dB
where such components are present.
    In keeping with the term  "gunfire deafness", the  first attempts to quantify the rela-
tionships between  impulse noise and hearing  loss have come from military fields. The
                                        211

-------
pioneer work was that of Murray and Reid (1946 a, 1946b) and Reid (1946). Almost two
decades passed  before Pfander (1965)  and Rice and  Coles (1965) made further moves
towards establishment of measurement and assessment criteria,  although a number of TTS
studies on various aspects of impulse noise had been published and in one case (Ward, 1962)
led to inclusion in the CHABA recommendations (Kryter et al., 1966) of a 140-dB limit for
impulse noise.
     The first  relatively comprehensive set of recommendations on impulse noise were
published by Coles, Garinther, Hodge and Rice in 1968, after a pooling of research data and
thought derived from studies mainly of gunfire-induced TTS with the British Royal Marines
and the U.S. Army. With minor modifications, these were adopted by CHABA (1968) for its
recommendations on gunfire noise exposure.
     Figure 1  illustrates the Coles et al criterion. A-duration refers to simple (Friedlander)
waveforms and refers to the duration of the positive pressure wave. It is comparatively rare,
and much more commonly there is a succession of pressure fluctuations when B_-duration is
the relevant parameter.  This  refers to  the  total time that the envelope of the pressure
fluctuations, positive  and negative, is within 20 dB of the peak pressure level. Exposure to
100 impulses, whose  physical characteristics fall on the relevant criterion line, will lead to
CHABA (Kryter et al. 1966) levels of TTS2 in 25% of  those exposed. The criterion should
be lowered by 5 or 10 dB if only 10% or 5% respectively are permitted to be affected to this
degree.  By the criterion, as published, the level could be raised by 10 dB for exposures to
one impulse per occasion.
            175
            170
        °   160
         K
            155
         E
         I
ISO
            145
            140
                                                      ....1
                         100
                       1ms       10ms      100 m s

                         Duration
             Figure 1. Damage risk criterion for impulse noise. (After Coles et al, 1968).
                                       212

-------
     In  an attempt to extend the criterion to embrace industrial  impact types of noise,
where the peak levels tend to be lower, the B-durations longer, and the numbers of impulses
per day very much greater, Coles and Rice (1970) proposed a revised and extended series of
corrections for numbers of impulses per exposure occasion. This is  shown in Figure 2. The
arguments for it are given in the reference, but it is worth noting that a major factor in its
construction was the need to  take into account the TTS studies of  Cohen et al. (1966) and
of Walker (1969).
     In  1970 Burns  and  Robinson published the  results of their 10-year government-
sponsored study of the relationship  between  noise  exposure and hearing loss  in British
industry. An important conclusion of this work was that their data  were consistent with an
energy concept;  that is,  the  total A-weighted sound energy received is a representative
measure of noise exposure with respect to injury to hearing. At the conference at which this
work was first presented,  Coles and Rice also gave a general resume of the current situation
on impulse noise hazards. A significant point to future trends came in the discussion, when
Martin (1970) linked the two together and suggested that  the energy  concept might be
extended to include impact noises up to peak pressures of 145 dB. This assertion  was based
on PTS studies in industry published later by Atherley and Martin (1971) and their method
of assessment of impact noise (Martin and Atherley,  1973).
     Recently, Rice and Martin (1973) have made a series of calculations which  show that
an extension of the energy concept gives results which are close to but rather more con-
servative than the Coles et  al (1968) criterion or its modification by CHABA (1968), and
 10

 5

 0

 -5

-10

-15

-20

-25



-35

-40
              - Nixon (1969)
               Tentative estimate by
               Coles et al (196S)
                                  Coles at al (1966) reference point
                                / (DRC expressed for about 100 impulses >

                             ^    
-------
very similar to the modifications by Coles and Rice (1970) and an earlier revision suggested
at its draft stage by Forrest (1967). The results of this analysis are shown in Figure 3. Two
practical points concerning the good agreement between the earlier impulse criterion and
the equivalent continuous noise level (ECNL) extension should be mentioned.
     First,  the  method of  assessment  of  industrial impact noises used  by Martin and
Atherley (1973) needs validation with respect to its application to gunfire noises. Indeed,
research is being carried out to find an easily  practical way to do this, and to adapt a noise
dosimeter for this purpose.  The instrumentation needs have been discussed generally by
Martin (1973).
     Second,  the 90 dBA  ECNL criterion based on an energy concept relates to daily
workday exposures, whilst  Coles  et al. one referred  to  10 or 20 exposure occasions per
annum. In fact, this discrepancy  is more apparent than real. On the  one hand, the ISO
recommendation  and the British government's Code of  Practice reflect  most scientific
opinion in making no allowances for non-habitual exposures. On  the other, the Coles et al.
criterion was derived mainly from TTS  studies and rests on the  generalization that TTS2
indicates the  likelihood of  PTS from habitual exposure. The criterion  was expected to be
applied to military situations, with perhaps  only 10 tor 20 exposure occasions per annum.
This was accepted as the 'norm', and the stated auditory effects were expected in practice to
arise from as  few as 10-20 exposures in many instances. This expectation was based mainly
on  the variable nature of actual gunfire-noise exposure in military situations, which some-
times leads to a more rapid accumulation of hearing loss than expected. On the other hand,
the total amount of hearing loss  accumulating  over a large number of exposure occasions
would not necessarily be markedly greater than that predicted by the TTS measurements on
which the criterion was  largely based, if the physical characteristics of the noise exposures
were strictly controlled. As this cannot be done in practice, and also because unpredictable
variations in susceptibility seem to occur (Reid, 1946), the criterion was considered to apply
to ordinary military conditions of 10 to 20 exposure occasions a year; warning was, how-
ever, given against taking a chance with even  a single unprotected occasion.
     For those individuals and countries who  accept the energy principle, it would seem on
the basis of the work described that we  are now on the threshold of being able to measure
and evaluate the auditory hazard of all types  of noises (from steady-state and intermittent
and fluctuating noises, and also industrial impact noises and high-intensity explosive im-
pulses) with one general principle and possibly with one single instrument The advantages
of this are so obvious, however, that one has to be careful not to lose one's critical appraisal
in face of the claims of expediency: certain further studies are desirable, in addition to the
two items already mentioned.
     Further  PTS  studies in a range  of populations, covering high-intensity impulses of
explosive type to moderate-intensity ones of impact type, are needed in order to validate
further  me agreement  illustrated in Figure 3  and  strengthen the evidence provided by
Atherley and  Martin (1971). These studies will not be easy though, because of the difficulty
in finding suitable populations. In industry, noise levels change with development of manu-
facturing processes, workers are less static in respect of their place of employment than they
used to be, and  hearing conservation  measures are becoming more effective or at least
complicate the  problem of selecting  an unbiased  and fully exposed sample. In military

                                       214

-------
       183
       175
    3
    I
    O
•;  I5S
CSI
S
"  145
    J
    e
       135
       125
                 -(R&M, 1973)
                                               All curvm corrected for 25% affected up lo CHAIA or
                                               equivalent amounts of threshold shift.
                                     Coles et al (1968),modified by Cotei and Rice ( 1970)for 1-100 impuliet
                                                           Rice and Martin (!97i),(ECNL = 90 dBA)
                                                                       CHAM (1968)
                              Coles et al (1968), modified by
                              Coles and Rice (1970) for
                              1000 impulses
          10"1         1          10        10'2        103

           'S'-DURATION » NO. OF IMPULSES, MILLISECONDS.
                                                          10
                                                                    10
 Figure 3. Comparison of damage risk criteria for impulse noise with an equivalent continuous noise level of
 90 dB(A), calculated according to the method of Martin and  Atherley (1973). (After Rice and Martin,
 1973). (Note:  the sound  levels should be lowered by 5 dB where the ears are at normal incidence to the
 impulse sound waves).
situations,  the same  complicating factors apply and there are still greater difficulties in
defining the amount of noise exposure (in terms of both level and numbers of impulses).
     It may be necessary to have recourse to TTS studies. The trouble here is that there are
great doubts as to the validity of the supposed arithmetic equivalence of TTS2 and  ex-
pectancy of PTS. Martin's studies (1970)  suggest that with impact-type noise TTS2 is a poor
quantitative predictor of PTS and, unlike his and other PTS ones, do not agree with  the
energy principle. Ward (1970) has shown that recovery from a given amount of TTS is  not
independent of how the TTS arose, as was originally postulated by Ward, Glorig and Sklar in
1959. Moreover, the rate of recovery from TTS, which must surely be related to risk  of PTS,
is not related in a simple manner to  the  amount of TTS  even from a given pattern of noise
exposure; where the  TTS2 is  over 40 dB, a much slower rate of recovery applies (Ward,
Glorig and Sklar, 1958,  1960). Ward (1970) himself suggested that TTS^Q may be a more
relevant index of risk of PTS,  but the relationship  between this and PTS is not yet  known.
As Passchier-Vermeer (1973) has stated  earlier in  this Congress,  "it may be possible that
recovery  and, hence,  total sound energy  over a workday  plays a more  important role than
can be expected from TTS2 alone".
     There is also a theoretical argument. If the energy principle is postulated for a  damage
risk  criterion, it is evident that TTS2 measurement applied to  a TTS2/PTS relationship
                                         215

-------
cannot be expected to validate the criterion. Consider the time patterns of the two types of
exposure illustrated in Figure 4. Both have four hours of exposure per  day, and  on the
energy principle the risk of PTS must be the same. The TTS2 measurement resulting from
exposure pattern B will be considerably greater than from exposure pattern A however. The
same sort of arguments could be applied to hypothetical patterns of impulse noise exposure.
    On the other  hand, hi spite of all the doubts and uncertainties about TTS2 studies, the
Coles  et al (1968)  criterion for impulse noises, together with Forrest's extension of it, and
the evidence of Cohen et al. (1966) and Walker (1968), were all based on TTS2 measure-
ments. It is these which appear to agree so well with the energy principle and the extrapola-
tions of ECNL 90  dBA into impulse noise fields, both of which were based on PTS studies.
    Finally, there is  the question of interactions  between impulse and steady-state noise
when  combined. On the energy principle, there is of course no problem: the energies from
the two sources are simply summed. However, this would not appear to be the complete
story. Cohen et al. and Walker have both demonstrated protective effects (for TTS) when
the two types of noise  were  added, and attributed these to enhancement of the acoustic
reflex. It should be noted, though, that  these were TTS studies from relatively short-dura-
tion exposures and with subjects who were not habitually exposed to high-level noise. It
seems quite likely that  there would be greater adaptation of the reflex  with longer and
habitual exposures.
    The contrary  and more traditional view is that the presence of impulsive components
hi a steady-state noise involves a disproportionately  greater auditory hazard. Such a view has
found support recently from the work of Hamernik and his colleagues (1972) who observed
the effects of impulse and steady-state noise on the hearing (PTS) and hair-cells (histology)
of chinchillas.
    In turn, these and other animal experiments, together with some electrophysiological
studies, open up much wider questions as to whether  TTS (or even PTS) is in fact a
satisfactory  measure  of  auditory damage.  However, until we have something that  is both
substantially  better and at least as easy to measure (and verifiable objectively when it comes

 NOISE EXPOSURE PATTERN A
                     noise
                   // S
             /
       /no
       ss
                                4
                               hours
             8
 NOISE EXPOSURE PATTERN B
                                             /—q°'
liet-
                                                                TTS measured 2 minutes
                                                                after last noise exposure
                                                                of an eight-hour workday.
Y///S
/noise '
///A
quiet
y///
'- noise S
'//y/

'////.
/noise

-------
to compensation assessment), pure-tone threshold shifts must remain our principal tool for
assessment of auditory damage resulting from noise exposures of all types.
    Acknowledgement: For much of their own material included in this general review, the
authors are indebted to the Medical Research Council for financial support.

                                     Summary

    Until fairly recently, governmental and occupational noise  hazard criteria have either
omitted or specifically excluded reference to impulse noise situations, or merely mentioned
them in guarded and largely non-quantitative  terms. The criteria put forward by Coles et al
in 1968, modified and extended by CHABA and Forrest, have provided some help but refer
mainly to gunfire-type noises. Coles and  Rice  (1970)  offered extensions for  industrial
impact noises,  and Atherley and Martin (1971) studied the problem further showing that
their PTS data gave  support to the equivalent A-weighted sound energy concept as a meas-
ure  of hazard  to  hearing. Recent analyses by Rice and Martin (1973) suggest  that such
equal-energy extrapolations may possibly be extended to cover the high-intensity explosive
type of noise. The highly  desirable prospect of one comprehensive method of measurement
and auditory hazard evaluation for all types of noise is thus revealed. Problems relating to
validation, instrumentation, and other uncertainties are discussed  briefly.
REFERENCES

Atherley, G.R.C. and Martin, A.M. Equivalent-continuous noise level as a measure of injury
     from impact and impulse noise. Annals of Occupational Hygiene, 14, 11-28 (1971).
Burns,  W. and  Robinson, D.W. Hearing and noise in Industry.  Her Majesty's Stationery
     Office, London (1970).
CHABA.  Proposed damage-risk criterion for impulse noise (gunfire). National Academy of
     Sciences - National Research Council's Committee on Hearing, Bio-acoustics and Bio-
     mechanics, Report of Working Group 57 (ed. W.D. Ward), Washington, D.C. (1968).
Cohen, A., Kylin, B. and LaBenz, P.J. Temporary threshold shifts in hearing from exposure
     to combined impact/steady-state noise conditions. Journal of the Acoustical Society of
     Americano, 1371-1380(1966).
Coles, R.R.A., Garinther, G.R., Hodges, D.C. and Rice, C.G. Hazardous exposure to impulse
     noise. Journal of the Acoustical Society of America, 43, 336-343 (1968).
Coles, R.R.A. and Rice, C.G. Towards a criterion for impulse noise in industry. Annals of
     Occupational Hygiene, 13, 43-50 (1970).
Department  of  Employment.  Code of Practice for reducing the exposure of employed
     persons to Noise. Her Majesty's Stationery Office, London (1972).
Forrest, M.R. The effects of high intensity impulsive noise on hearing. Master of Science
     dissertation, University of Southampton (1967).
Hamernik, R.P., Crossley, J.J., Henderson,  D. and Salvi, R.J. The interaction of continuous
     and impulse noise. Personal communication, State University of New York, Syracuse,
     N.Y. (1972).

                                        217

-------
International Organisation for Standardization. Assessment of occupational noise exposure
    for hearing conservation purposes. ISO Recommendation R1999 (1971).
Kryter, K.D., Ward, W.D., Miller, J.D. and Eldredge, D. Hazardous exposure to intermittent
    and steady-state noise. Journal of the Acoustical Society of America, 39, 451-464
    (1966), and Report of CHABA Working Group 46.
Martin, A.M. Industrial  impact noise and hearing. Doctor of Philosophy thesis, University of
    Salford(1970).
Martin, A.M. Discussion on papers in Section 1. Pages 89 and 90, in Occupational Hearing
    Loss, British  Acoustical Society Special  Volume  No. 1, Academic Press,  London
    (1971).
Martin,  A.M. and  Atherley, G.R.C. A  method for the assessment of impact noise with
    respect to injury to hearing. Annals of Occupational Hygiene, 16 19-26 (1973),
Martin, A.M. The measurement of impulse noise. Journal of Sound and Vibration, 342-344
    (1973).
Murray, N.E. and  Reid, G. Experimental observations on the aural effects of gun blast.
    Medical Journal of Australia, 1, 611-617 (1946,a).
Murray, N.E. and Reid, G. Temporary deafness due to gunfire. Journal of Laryngology and
    Otology, 60, 92-130 (1946, b).
National Institute for Occupational Safety and Health. Criteria for a recommended standard
    ... Occupational  Exposure to  Noise. U.S. Department of Health,  Education, and
    Welfare; NIOSH Report HSM 73-11001 (1972).
Passchier-Vermeer,  W. Noise-induced hearing loss from exposure to intermittent and varying
    noise. Proceedings of the International Congress on Noise as Public Health Problem,
    Dubrovnik(1973).
Pfander, von F. Uber die Toleranzgrenze bie Akustischen Einwirkungen. Hals-, Nasen- und
    Ohrenarzte, 13, 27-28 (1965 ).
Reid,  G. Further observations on temporary deafness following exposure to gunfire. Journal
    of Laryngology and Otology, 60, 608-633, (1946).
Rice, C. G. and Coles, R. R. A. Impulsive noise studies and temporary threshold shift. Paper
    B67, in Proceedings of Fifth International Congress of Acoustics, Liege (1965).
Rice,  C. G. and Martin, A.M. Impulse noise damage risk criteria.  Journal of Sound and
    Vibration, 359-367 (1973).
Walker,  J.G. Temporary  threshold  shift from impulse noise. Annals  of Occupational
    Hygiene, 13, 51-58(1970).
Ward, W.D., Glorig, A. and Sklar, D.L. Dependence of temporary threshold shift at 4 kc on
    intensity and time. Journal of the Acoustical Society of America, 30,944-954 (1958).
Ward, W.D., Glorig, A. and Sklar, D.L. Relation between recovery from temporary threshold
    shift and duration of exposure. Journal of the Acoustical Society of America, 31,
    600-602(1959).
Ward, W.D., Glorig, A.  and Sklar, D.L. Recovery from high values of temporary threshold
    shift Journal of the Acoustical Society of America, 32,497-500 (1960).
Ward, W.D. Effect of temporal spacing on  temporary threshold shift from impulses. Journal
    of the Acoustical Society of America, 34, 1230-1232 (1962).
Ward, W.D. Temporary threshold shift and damage risk criteria for intermittent noise ex-
    posures. Journal of the Acoustical Society of America, 48(2), 561-574 (1970).

                                        218

-------
                    HEARING LOSS DUE TO IMPULSE NOISE
                                A FIELD STUDY.

                         Tadeusz Ceypek, Jerzy J. Kuzniarz

                            Otolaryngology Department
                             Silesian Medical  Academy
                                  Francuska 20,
                             40-027 Katowice, Poland

                                 Adam Lipowczan
                              Central Mining Institute
                         P1. Gwarkow 1, Katowice, Poland

    Present  knowledge on consequences of exposure to industrial impulse noise is still
rather scanty (Acton, 1967; Coles, 1970). That was the reason we have started a study on
hearing loss in drop-forge operators, exposed to several thousand impulses each day. For
that purpose we have selected a factory where both the local tradition and a great distance
from another possible place of work made the staff pretty stable over the years. In that way
213, drop-forge operators (426 ears), working up to 30 years at the same place and exposed
to the same kind of noise, were tested.

Characteristics of noise exposure.

    The impulses  are generated by iron drop-forge hammers, weighing 1  to 5 tons, falling
from a height of approximately 1.5 m onto an iron base. Owing to the distance between the
forges (about 5 m) each operator was exposed mainly to impulses from his own forge, from
a distance of less than 1 m. As a rule they did not use ear protectors.
    The impulses  were measured with  a Bruel and Kjaer Impulse Sound  Level Meter type
2204 and 1/2-in. microphone, and recorded on a Kudelski tape-recorder Nagra IV L (speed
19.05 cm/sec., frequency range of 20  - 18 000 Hz, dynamic range 50 dB over the back-
ground level). Using the frequency modulation technique the lower limiting frequency equal
DC was obtained.  The recorded impulses were  measured with the  use of storage oscillo-
scope. (Fig. 1). The following values have been determined:
    (a) Peak pressure level: 127 to 134 dB, independent of the weight of the hammer,
    (b) Rise time: almost instantaneous (a few microseconds),
    (c) Impulse duration (from peak to the ambient level): 100 to 200 msec,
    (d) Level of background noise: 110 dB,
    (e) Repetition rate (during on-time): 0.5 to 2 per sec.
    (f) As drop-forges work in an on-off manner the total number  of impulses have been
calculated from the known  number of strokes necessary to produce each item, and from the
number of items produced  by each drop-forge during the work-day.  Depending on the item
produced there were 3000 to 10,000 impulses a day to which every drop-forge operator was
directly exposed.

                                       219

-------
: -
                      Figure 1. Oscilloscopic picture of impulses produced by drop-forge. Lp = peak level. Lb = background level,
                      Tr • time of repetition, Tj = duration time.

-------
    (g) Frequency spectrum has not been analyzed because of known difficulties with that
procedure (Coles, 1970).

Hearing Tests.

    Hearing examinations were conducted before  work in order to measure the PTS (at
least 16 hours after the last exposure to noise), and after work in order to measure TTS.
Otological  examinations were done beforehand to exclude other possible causes of hearing
loss. Audiometers were calibrated according to ISO standards. Ambient noise did not exceed
the allowable levels.

Results.

    All the workers were divided into 8 groups, according to exposure time in years (Tab.
I). After correction of hearing losses for presbyacusis (Glorig,  1962) a statistical analysis of
the results  was performed.  The scatter of the individual results  was quite large, but dis-
tributed in a near-normal manner (medians and means were equal), so means and standard
deviations  were determined. The results discussed here cover PTS only, as the study on TTS
has not been finished yet (Fig. 2-5)
     Some  characteristic  features of the hearing  losses in that  population may be  sum-
marized as follows:
1.   The most prominent hearing loss during early  exposure (e.g. in groups "under 1 year"
     and "1-2 years of exposure") occurs at 6000 Hz (Fig. 2). It was a "leading frequency"
     in our series.
2.   Hearing loss at 4000 Hz was next in frequency of occurrence and magnitude and after
     5 years of exposure becomes as large as that at 6000 Hz.
3.   The greatest drop in hearing threshold at 6000  Hz and 4000 Hz appears during the first
     two years of exposure (the average rate was 20  dB/year).
4.   Fully-developed hearing loss at 4000 and  6000 Hz appears during the first 5 years of
     exposure  (average rate 10 dB/year). A further drop in hearing threshold occurs slowly
     (on the average, 5 dB over a period of 10 years).
5.   The growth of hearing loss at the lower frequencies was much smaller in the first two
     years (at  2000 Hz about 10 dB/year on the average); a steady slow progression during
     the following years was observed, at an average rate of 5 dB in 10 years. (Fig. 5).
6.   After  2 years  of exposure not a single worker with normal threshold of hearing could
     be found, but individual variations in hearing loss were quite large.

Comment.

     The pattern of noise exposure which has  been found in the present study is different
from a laboratory  type of impulse noise, but typical for industry: the impulses are super-
imposed on a rather high noise background which, by itself, might cause damage to hearing.
But it was probably  the impulses that have  induced a quicker development of PTS in
comparison to  steady-state noise exposure:  fully-developed PTS occurred after 5 years of

                                        221

-------
                                                       Table 1.
                     NOISE-INDUCED HEARING-LOSS IN dB VS. YEARS OF EXPOSURE (MEAN AND STANDARD
                                                DEVIATION - S.D.)
Exposure
time
years
<1
1-2
3-5
6-10
11-15
16-20
21-25
26-30
<1
1
2
3
4
5
srrsirsrrssriEEjscsssErsiz:!:
Age
/meantS.D./
24 ±5
26 ±6
27 ±5
30 ±5
39 ±6
46 ±6
47 ±6
49 ±5






N
/ears/
34
84
98
68
60
22
42
18
34
40
44
34
34
30
Hearing level / dB ISO / - mean and S.D.
500
Mean S.D.
10.3 8.2
12.1 10.3
12.4 9.0
13.2 11.6
14.3 10.9
17.8 15.5
17.8 11.2
18.2 12.3






1000
Mean S.D.
12.42 8.7
17.3 10.1
18.0 8.-
24.9 14.8
26.8 14.8
30.9 20.0
28.8 13.9
31.9 18.6
0 - 5 years






2000
Mean S.D.
17.4 8.2
25.0 15.7
27.4 12.8
34.9 17.6
33.0 19.0
47.2 15.8
36.7 14.9
42.7 13.9
17.4 8.2
18.6 10.5
25.0 16.8
25.6 13.0
24.5 8.-
31.8 15.4
4000
Mean S.D.
24.8 13.3
37.6 23.6
44.6 22.3
49.7 23.2
49.1 17.9
56.3 12.4
48.9 16.5
55.2 17.1
24.8 13.3
27.1 15.3
46.2 26.1
45.5 19.6
41.0 23.0
49.1 23.0
6000
Mean S.D.
28.1 15.7
43.4 23.6
44.6 22.1
50.5 20.2
48.3 16.3
57.0 16.1
50.2 15.2
61.6 24.6
28.1 15.7
31.8 19.0
51.2 25.5
45.7 22.5
40.7 23.0
45.6 20.7
8000
Mean S.D.
24.8 14.2
40.8 22.5
42.4 22.7
48.9 21.1
48.3 17.6
57.9 18.0
50.1 17.8
60.2 26.5
24.8 14.2
31.6 16.1




K)

-------
                tup. hmt < 1y*or.    N* 34 ears.
                     Exp. fimf 1- 2 years.  N * 84 ears.
• i
• -
,.
                                        WOO   6000 Hz
                 Exp.  timt  3-5ytor3.  N* 98 ears.
                     599
9000 Hi
                         500    WO   HOO    WO   MOO  Hz
                     Exp. time  6-10 years,  tf- 68 tars.
WOO  6000  HZ
                  Figure 2. Noise-induced hearing loss (Hearing Levels corrected for presbyacusis) during the first 10 years of
                  exposure. Mean: thick line; S.D.: thin line.

-------
! <

' '

i-
             Cxp. time  11-15 years.   N'Meors
                                              Exp. time n-20 years.     H*2? ears
                              fOOO  tOOO  9000 Hz
txp. time #-?5 years.    N* tf ears
                              two   wo  oooo H:
    500    1000  ZOOO  ¥300  6000 Hi

ftp. time 26-30 years.   N* 18 ears.
                              —i
                                                                           9000  Hz
                 Figure 3. Noise-induced hearing loss (HL corrected for presbyacusis) after 10 years of exposure. Mean:
                 thick line; S.D.. thin line.

-------
to
K)
            10.

            20.

           so
Exposure  time in years
                    A—A   1-2
                    m—•   3-5
                    x—x   6-10
                   500
          WOO
                    Exposure  time in

                       •—•   11 -  15
                       A	A  16 -  2(1
                       m—•   ^J-^5
                       X—x   ffi-30
WOO 6000 6000  Hz    500
1000
tooo
4000 6000 9000
                    Figure 4.  Mean noise-induced hearing loss  (KL adjusted for  presbyacusis)  years of exposure as the
                    parameter.

-------
           10
to
to
co

-------
exposure while it takes approximately 10 years in steady-state noise exposure (Glorig et al.,
1961; Nixon et al., 1961). Similar observations were made by Sulkowski et al. (1972).
    The occurrence of the greatest impulse-noise-produced hearing loss at 6000 Hz has also
been reported by other authors (Loeb and Fletcher, 1965; Salmivalli, 1967; Gravendeel et
al., 1959; Zalin, 1971).
    As far as DRC are concerned, the criteria proposed for gunfire impulses (Coles et al.,
1968) cannot be applied to industrial noise, as Coles and Rice (1970), Walker (1970), and
Martin  et al. (1970) supposed. The impulses of peak level of 125 to 135 dB and duration
100 to  200 msec were clearly harmful when repeated several thousands times  a day for
several  years. The protective action of intraaural muscles seems to be  questionable during
such long exposure, although Cohen et al (1966) reported less harmful effect of impulses
when superimposed on steady state noise. The new concept of the evaluation of risk caused
by impulses reported here by Coles seems to be very promising.
    The large variations in the extent of hearing loss caused  by impulse noise may be
attributed to individual susceptibility: some workers in our series show 50 dB of hearing loss
after 1  or 2 years of exposure, while others have only 20 dB or less after 25 years of work.
The cause  is still obscure,  so routine audiometric tests should be advocated in order to
exclude persons with "tender ears" from  further exposure  or to introduce appropriate
preventive measures (ear protection, enclosures, etc.).
Conclusions.

1.   Fully-developed PTS occurs after 5 years of exposure to impulse noise, which implies
     that rapid transition from TTS to PTS in that type of exposure is quite probable.
2.   The fact that the earliest and greatest change in threshold occurs at 6000 Hz indicates a
     slightly different pattern of development of hearing loss in  impulse noise than in
     steady-state noise exposure.
3.   A very slow increase in hearing loss after 5 years of exposure indicates that PTS caused
     by impulse noise stabilizes  with time, as has been found in  exposure to steady-state
     noise.

                                      Summary

     Results are presented  of hearing examinations in 213 drop-forge operators who were
exposed, for  1  to 30 years, to 3000 to 10000 impulses a day, the impulses having a peak
level of 127-134 dB and duration of 100 to 200 msce. The impulses are superimposed on a
background of 110 dB of steady-state noise.
     The maximum PTS during early exposure (1-2 years)  occurred at 6000 Hz, at a rate
of 20 dB a year. Hearing loss at 4000  Hz was smaller, but after 5 years of exposure  they
become equal. PTS at 6000 Hz and 4000 Hz reached maximum in first 5 years of exposure
(about 50 dB on the average); further increase was rather slow. At 2000 Hz a smaller PTS
during early exposure is observed: about 10 dB per year during first 1 - 2 years, later about
SdBin 10 years.

                                        227

-------
     Fairly large individual variations in threshold shift were observed, which resulted in a
20-dB standard deviation from mean values of PTS (the scatter of values was close to
"normal curve").

                                  REFERENCES

Acton, W.I., A review of hearing damage risk criteria. Ann.  Occup. Hyg.,  10,  143 - 153
     (1967).
Cohen, A., Kylin, B., La Benz, P. J., Temporary threshold shifts in hearing from exposure to
     combined impact/steady-state noise  conditions. /. Acoust. Soc. Am., 40, 1371-1380,
     (1966).
Coles, R.R.A., Garinther, G.R., Hodge, D.C., Rice, C.G., Hazardous exposure to impulse
     noise./. Acoust. Soc. Am.,43, 336-343 (1968).
Coles, R.R.A., Rice, C.G., Towards a criterion for impulse noise in industry. Ann. Occup.
  .   Hyg., 13,43-50(1970).
Glorig, A., Nixon, J., Hearing loss  as a  function of age. Laryngoscope, 72, 1596-1611
     (1962).
Glorig, A., Ward, W.D., Nixon, J., Damage risk criteria and noise-induced hearing loss. Arch.
     Otolaryng., 74, 413-423  (1961).
Gravendeel, D.W., Plomp, R., The relation between permanent and temporary noise dips.
     Arch. Otolaryng., 69, 714-719 (1959).
Loeb, M., Fletcher J.L., Benson R.W., Some preliminary studies of temporary threshold
     shift with an arc discharge impulse noise./. Acoust. Soc. Am., 37, 313 (1965).
Martin, A.M.,  Atherley, G.R.C., Hempstock, T.I., Recurrent impact noise from pneumatic
     hammers. Ann. Occup. Hyg.,  13, 59, 67 (1970).
Nixon, J.C., Glorig, A., Noise-induced permanent threshold shift at 2000 cps and 4000 cps.
     /. Acoust. Soc. Am., 33, 964-968 (1961).
Salmivalli, A., Acoustic trauma in regular army personnel. A eta Otolar.,  (Stockh.), Suppl.
     222(1967).
Sulkowski, W., Dzwonnik, Z.,  Andryszek, Cz.,  Kipowczan,  A., Ryzyko Sawodowych
     uszkodzen sluchu w halasie ciaglym, przerywanym i impulsowym. Med. Pracy, (Pol.),
     23,465-481 (1972).
Walker,  J.G.,  Temporary threshold shift from impulse noise. Ann.  Occup. Hyg.,  13, 51-58
     (1970).
Zalin, H., Noise induced hearing loss. Unmasking other pathology. Proc. Roy.  Soc. Med., 64,
     187-190(1971).
                                       228

-------
                  HEARING DAMAGE CAUSED BY VERY SHORT,
                        HIGH-INTENSITY IMPULSE NOISE

                                   H.G. Dieroff
                          HNO-Klinik, University of Jena
                         Jena, German Democratic Republic
Introduction
    Much has  been reported in recent years about the harmful effect of impulse noise-
especially of very short impulses-on hearing (DIEROFF, GARINTHER and MORELAND,
KRYTER, POCHE, RICE and  COLES, STOCKWELL and ADES). Findings  of damage
caused by ultra-short impulses such as produced by certain toys have also become more
frequent (GJAEVENES, HODGE and McCOMMONS, etc.).
    Rather striking cases of hearing loss are often found in  persons occupied in metal-
working trades  and working in places where continuous sound pressure levels hardly exceed
the crucial intensity of about 90 dB (A). Again and again, individuals are found to suffer
from hearing loss that seems to bear no relation to the sound pressure level (SPL) prevailing
at their work place. The frequently-encountered hearing impairment  in  welders is repre-
sentative of the situation to be described here, as Fig. 1 demonstrates. There is no doubt
that impulse noise, even only a few high-intensity impulses, produced when burrs and slag
are removed from welds  or when the acetylene flame is ignited, largely account for the
extent of the damage. Measurements of such noise reveal very irregularly scattered impulses
of varying quality. Level frequency counts permit a considerably better  assessment of
auditory stress, but even they are not sufficient to establish a relationship between the
actual noise stress  and  the  hearing  loss  detected.  Therefore, today's still-inadequate
techniques of measuring the actual stress involved in densely pulsed industrial noise has to
be regarded as  a major reason for the apparent discrepancy between noise levels and the
degree of hearing impairment. Moreover, the question arises whether impulse noise involves
a damage mechanism different from that caused by continuous noise, which in the end leads
to the familiar severe hearing damages.

Theoretical considerations

    The author's investigations, especially in the metal-working industry, have  shown that
the clashing of two metal parts frequently produces SPL peaks between  150 and 160 dB.
Peaks of that intensity may cause hair-cell damage by way of a heavy electrical discharge of
the hair-cells when the hairs  touch the  tectorial  membrane,  which  leaves a  scar in the
hair-cell's microstructure. A hair-cell thus damaged may sooner or later suffer degeneration.
    Damage after steady-state acoustic stimulation has been described as a deficiency in the
supply  of nutrients to hair-cells; but  assumptions  primarily relying on  the same type of
metabolic disturbance to account for impulse-noise-induced  hearing damage  carry little
conviction.
    A  research  team  consisting  of   BIEDERMANN,  GEYER,   GUTTMACHER,
KASCHOWITZ, MEIER and  QUADE, in addition  to the author, investigated the issue of
                                       229

-------
                                      £r$chweisscr
                      '.".  'Vt IikX  >.',." *«'* r'
                                                                    Ql,r
Figure 1. Single audiograms (air conduction) of welders, each curve with age and number of working years
in noise {after DIEROFF, 1962).

-------
whether the exposure of guinea pigs' ears to very short sound impulses causes functional
hearing changes as well as structural deformations of and metabolic damage to the organ of
Corti. We  hoped  to determine  to  what  extent  there is  any correlation between  the
functional behavior and the microscopical and histochemical findings.

Methods

     1. Functional behavior test series, carried out by BIEDERMAN and KASCHOWITZ:
Guinea pigs weighing 250 to 450 g and showing normal PREYER reflexes were exposed to
acoustic stimulation in  a low-reflection sound chamber of about  1 x 1 x 2.2 m in size. A
spark-discharge sound generator was used as the source; sound pressure peaked around  162
dB with a standard  deviation of ± 1 dB. Impulse widths varied from 200 to 400 jus. The
functional behavior  of  the guinea pigs' ears was  measured in terms of the microphonic
potential  (MP) measured at the round  window. MP was measured exclusively during the
experiment.
     2. Metabolic behavior test series, carried out by Geyer, Quade und Guttmacher, Meier:
Experimental  subjects were guinea pigs weighing  250 to 450 g and exhibiting positive
PREYER reflexes. The  animals were placed  two at a time in a sound chamber containing
very narrow  cages. Acoustic measurements  demonstrated a  uniform sound  pressure
distribution in the area of the animals' heads. Sound pressure levels of  135 or 158 dB ± 1 dB
were  used. Investigation  included  the microstructure and  the behavior  of succinic-
dehydrogenase (SDH) activity in the cochlea.  The same spark-discharge sound generator was
used for  both test series. In either series, some control animals  served  as a  basis of
comparison for the pathological changes to be detected.

Results

     In the functional  behavior experiments, the  guinea  pigs were exposed to 1, 3 or 5
impulses at intervals of about 3.5 seconds, which was sufficient to allow the middle-ear
muscle reflex aroused by the preceding impulse to relax. The first  MP measurement was
made 5 minutes after the start of each  experiment. The MP amplitude, which was recorded
for 120 minutes after stimulation (Fig. 2), showed a noticeable  attenuation after a single
impulse, decreased more rapidly after  3 impulses, and continued to fade after 5  impulses
with a tendency toward an asymptote. Throughout the observation  period, no increase in
MP was detected. Extending the period any  longer was not practicable,  since the MP
measurements were carried out with anaesthesized animals.
     In contrast to the functional behavior described,  the second test-series required a much
longer period of acoustic stimulation before a distinct decrease in SDH could be found. The
first SDH changes were observed only after 8 days, with a daily sound exposure time  of 9
hours alternating  with  15 hours of rest, and a pulse rate of 16 per minute. A marked
increase was detected after the period of sound exposure had been extended to 14 days. On
the other hand, acoustic stimulation of these animals did not produce any MP or any Preyer
reflexes prior to sacrificing that would be indicative of a residual function.
     The decrease in SDH activity extended over  the entire cochlea  and also included the
nerve endings. SDH activity decreased  more in the outer than in the inner hair-cells. No

                                        231

-------
micro-structural changes could be detected. Other histochemical details are not considered
here.

Discussion

    The two series of experiments showed a large discrepancy between the behavior of MP
and that of SDH.  A change in MP can be noticed after the very first impulse, whereas a
decrease in SDH activity will occur only after extensive exposure to impulse sound. How-
ever, histochemical changes after  stimulation by impulse sound are identical to those
observed after continuous stimulation, since the results are well in agreement with observa-
tions reported by VOSTEEN (1958, 1960, 1961), VINNIKOV and TITOVA (1958, 1963)
and QUADE and GEYER (1972). As the only divergence, VOSTEEN in his  experiments
found no decrease of enzyme in the inner hair-cells and nerve endings after minor sonic
stress. According to VOSTEEN (1958), the decrease of SDH activity is to be taken as a stage
preceding the disintegration of hair-cells, with the decrease in SDH activity corresponding to
a state of exhaustion after functional over-exertion.
     Summarizing  the results of the  two test series, we find  that sound impulses of high-
peaked SPL cause  a reduction of MP  that was irreversible for the  duration of the measure-
ment, whereas the impulse  sound stress required for histochemical changes is much longer
than the continuous sound stress that would produce the same changes. The reduction of
MP suggests the very rapid occurrence of a functional over-exertion or  damage due to a few
very short pulses  of some hundreds of microseconds  duration. These can  probably be
explained  by  the  electron-microscopic  observations  made  by SPOENDLIN. After
                                                           70S (min) IX
 Figure 2. The behavior of microphonpotentials of guinea pigs after 1,3 and 5 sparks depending from time
 (after BIEDERMANN and KASCHEWITZ, 1973).

                                       232

-------
continuous  acoustic stimulation  with wide-band  noises  ranging from  125 dB upward,
SPOENDLIN  found changes in the ultrastructure of the outer, and later of the inner,
hair-cells and of the nerve endings, which changes progressed with growing sound intensity.
These changes were of a partially mechanical and partially metabolic nature. The influence
exerted on MP by a few single impulses might be due primarily to purely mechanical damage
to the ultrastructure of the organ of Corti, caused by direct contact between the tectorial
membrane and the outer hair-cells, with a possibly stronger affection of the sensory hairs.

Summary

    A  few  single impulses (1, 3 or 5) having  sound pressure peaks of  162 dB  and pulse
widths  of up  to 400 /its induce a permanent reduction of  microphonic potential in guinea
pigs.  A considerably longer  period  of acoustic  impulse stimulation (Le. 8 to 14 days) is
required to  detect histochemically  the same SDH decrease as that found after continuous
sound stimulation, a decrease assumed to be a stage preceding hair-cell degeneration. What
changes in the organ of Corti are responsible for the observed microphonic potential reduc-
tion remain to be identified.

                                    References

Biedermann and Kaschowitz: Personal communication 1973
Dieroff, H.G.: Zur Problematik der Schlagimpulse  in Industrielarm.  Arch. Ohr.,-Nas.-
    u.Kehlk.-Heilk. 179, 409 (1962)
Dieroff, H.G.: Schlagimpulse und Larmschwerhorigkeit. Intern. Audiology 2 (1963)
Dieroff, H.G.: Horschaden durch impulsreichen  Larm und dessen Erfassung. Kongre/3ber-4.
    Akustiche Konferenz Budapest 1967, Teil I
Dieroff, H.G.: Der gehorschadigende Impulslarm in  der  Industrie  und seine Erfassungs-
    probleme. Kampf dem Larm, H. 2 (1969)
Dieroff, H.G.: Zur besonderen Larmsituation bei Schlossern,  Schwei/Jern. Pressern sowie
    Stanzern und den  zu erwartenden Larmhorschaden. Praxis der Larmbekampfung, AICB
    KongrejS Baden-Baden 1966 Verlag fur Medizin - Technik, K.H. Walter, Baden-Baden.
Dieroff, H.G.: Einige  spezielle Fragen der Gehorschadigung durch Impulslarm in  der
    Industrie. Proc.  Vol.  Ill  S.  37. 3.  Konferenz Kampf gegen Larm und  Vibration
    Bukarest 1969
Dieroff, H.G.: Der Einflu(3 von Schlagimpulsen  auf das AusmajS der Horschadigung eines
    Larmarbeiters. Z. ges. Hyg. H. 2 (1973)
Garinther, G.R., and J.B. Moreland: Transducer  Techniques of Measurin the Effect of Small
    Arms Noise on Hearing. U.S. Army Human Eng. Labs., Aberdeen Proving Ground, Md.,
    Tech. Mem No. TM 11-65
Geyer,  G.,  R. Quade und  H.  Guttmacher: Histochemischer  Nachweis von Succinode-
    hydrogenase-Aktivitat   im   Cortischen  Organ   vom  normalen   und  beschallten
    Meerschweinchen. Z. ges. Hyg.  19, 25 (1973)
Gjaevenes, K.: Measurements on the  Impulsive Noise from Crackers and Toy Firearms. J.
    Acoust. Soc.  Am. 39, 403 (1966). Damage-Risk Criterion for the Impulsive Noise of
    Toys. J. Acoust. Soc. Am. 42, 268 (1967)

                                       233

-------
Hodge, D.C. and R.B. McCommons: Acoustical Hazards of Children's Toys. J. Acoust. Soc.
    Am. 40, 911 (1966)
Kryter, K.D.: The Effects of Noise on Man. Academic Press New York and London 1970
Meier, Ch.:  Personal communication 1973
Poche, L.B., Ch.W. Stockwell and H.W. Ades: Cochlear Hair-Cell Damage in Guinea Pigs
    after Exposure to Impulse Noise. J. Acoust. Soc. Am. 46, 947 (1969)
Quade, R.  und G. Geyer: Der Succinatdehydrogenase-Nachweis mit Hilfe der Perfusion-
    stechnik an der Cochlea des Meerschweinchens  unter Normalbedingungen  und nach
    Dauerlarmeinwirkung. Acta Otolaryng. (Stockh.) 75, 45 (1973)
Rice, C.G. and R.R.A. Coles: Impulsive Noise Studies and Temporary Threshold Shift. Proc.
    Intern. Congr.  Acoust. 5th, paper B 67 (1965)
Spoendlin, H.: Primary Structural Changes in the Organ of Corti after Acoustic Overstimula-
    tion. Acta Otolaryng. (Stockh.)  71, 166  (1971)
Vinnikov, J.A. and L.K. Titova: Intern.  Rev. Cyt 14, 157 (1963)
Vosteen,  K.H.:  Die Erschopfung der Phonorezeptoren nach funktioneller Belastung. Arch.
    Ohr.,-Nas,-und Kehlk.-Heilk. 172,  489 (1958). The Histochemistry of the Enzymes of
    the Oxygen Metabolism in the Inner Ear. Laryngoscope 70, 351 (1960). Neue Aspekte
    zur  Biologic und Pathologic des Innenohres. Arch. Ohr.,-Nas.-u.Kehlk.-Heilk. 178, 1
    (1961)
                                       234

-------
                SESSION 3





 NOISE-INDUCED HEARING LOSS-MECHANISM



Chairmen: H. G. Dieroff (DDR), R. Hinchcliffe (UK)
                   235

-------
          BEHAVIORAL, PHYSIOLOGICAL AND ANATOMICAL STUDIES
                      OF THRESHOLD SHIFTS IN ANIMALS1

                  Donald H. Eldredge, James D. Miller, John H. Mills
                              and Barbara A. Bonne2
                  Research Department, Central Institute for the Deaf
                             St. Louis, Missouri 63110

    In a Congress on  Noise  as a Public  Health Problem we  are really very much more
concerned with man  than  we are with animals.  So what is it that we would like to know
about noise and man  that we may learn from animals? As representatives from an Institute
for the Deaf we would like to know how  to evaluate hazards for hearing and the ear from
information about exposures to noise. Exposures to noise are commonly specified in terms
of the  level  and spectrum of the noise, and the temporal pattern  and total duration  of
exposures. For man, the effect on the ear is usually specified in terms of temporary and/or
permanent shifts of the  auditory thresholds for pure tones. In animal  experiments we would
like to learn as much as possible about the above relations and in addition to learn some-
thing of the pathological physiology and pathological anatomy.
    From animal experiments we  have long known  that exposures to sound can  lead  to
degeneration of the hair cells of the organ of Corti with associated loss of neurons. Most
such experiments  have involved  short  exposures  at high  levels  and  have  served  to
demonstrate  the susceptibility of the ear to sudden acoustic trauma. We are all familiar with
these experiments, and they will not be of much concern to us today because these studies
have not been so  clearly  oriented toward the  human situation. The exposures that are
important for man  do not seem to injure with a single event. Instead, loss of hearing follows
only after repeated exposures over relatively long periods. Accordingly we wish to review
some of our  recent work that emphasizes prolonged exposures at lower levels and exposures
that are not  immediately associated with permanent loss  of threshold  sensitivity for tones.
On the basis of this work we can already state several important relations between exposure
to noise  and loss of hearing in the chinchilla. The measured relations for  the chinchilla
correspond so well to  similar relations for man that we believe common principles apply.
    Advantages of Animal Experiments'. Animal experiments can have certain rather clear
advantages. First, and possibly foremost, it is possible to restrict activities and exposures in
such a way  as to reduce individual variability in responses to exposures to a minimum.
Secondly, it is  possible to look systematically at some  limiting conditions that are not
always  reasonable for man. Thirdly, it is reasonable and possible to  examine the  ears of
animals  under  the  microscope  at appropriate times  and with  good  preservation  of
anatomical detail.
    Advantages  of Chinchillas: The choice of an animal is another important step. We
chose the chinchilla because:
     1)  it can  be  easily  trained,  using  standard  shock avoidance  techniques, to give
behavioral responses to tones;
     2) its threshold of audibility has a sensitivity and a frequency range similar to those of
man;

                                        237

-------
     3) it is relatively healthy and free of diseases of the middle ear;
     4) it has a long life, at least 10 years and up to 20 years; and
     5) three turns of its cochlea are surgically accessible to electrodes for the recording of
cochlear potentials (Miller, 1970).
                                 Behavioral Studies

     Acquisition of Threshold Shifts: Let us begin by asking what happens to the behavioral
auditory thresholds when exposures to noise are continuous. Carder and Miller (1971, 1972)
showed that the threshold for a 715-Hz  tone increased with duration  of exposure to an
octave band of noise centered at 500 Hz for only about 24 hours (1440 min.) and then
remained at a plateau or asymptotic value as exposures were continued for 7 days and even
for 21 days. Four  examples of this kind of growth of threshold shift to asymptotic values
are shown in the left panel of Fig. 1. Here we see that the effect of changing the level of the
band of noise from 75 to 85, 95, and  105 dB SPL is to shift the asymptotic threshold shift
from about  17 dB to 31 dB, 49 dB, and 63 dB, respectively. Note that for each exposure the
asymptote is reached in about the same time and that rate of growth of shift has increased
correspondingly with level.
     At this point  a prominent footnote is required. One of our most important findings is
that the terms "temporary threshold shift" (TTS) and, "permanent threshold shift" (PTS)
are not enough to distinguish among the  operationally important features of all of the
various threshold shifts (TS) that we have encountered. At a minimum we  will need a new
term similar to "asymptotic threshold shift" (ATS). All of the evidence is not yet available
and  we are not yet prepared to recommend a definitive set of terms. The notations we have
used in the  past reflect the fact that an animal must be removed from the noise to measure
his TS. Some recovery from the level of TS present in the noise begins as soon as the animal
has been placed in the quiet, but  this is usually small and the animal is returned to the noise
as soon as  thresholds have been measured. Measurements of  threshold twice at a single
frequency can be made at an average time out of noise  of 4  minutes and this time may be
given as a subscript; e.g., TS4, TTS4. Measurement of thresholds for an audiogram of five or
six frequencies can be made at an average time out of noise of 11 minutes; e.g., TTSj j. In
addition we have earlier  expressed ATS4 as TTS4°° or the TS4 measured with the TS has
stopped growing.
     Part of the data in the left panel of Fig. 1 are replotted in Fig. 2A to show the relation
of ATS4 at  715 Hz to the level of the band of noise. Note that the relation is well described
by the equation

                               ATS4=1.6(OBL-65)

That is, for  every decibel that the band level exceeds the subtractive constant 65 there will
be a 1.6-dB increase in ATS.

                                        238

-------
to
OJ
so
          dB

          70
       (/>
       
-------
     TTS4   =1.6(081-65)
                                                          70
                                                       N  60
                                                       I
                                                          40
                                                     V)
                                                     g
                                                     X
                                                       2 30
                                                     tr
                                                     <
                                                     CL
                                                     S
                                                     UJ
                                                          20
                                                          10
                                                                      B.
                 TTS4fl0= 1.6 (OBL-47)
50      60     70      80      90     100     110
  SPL OF OCTAVE BAND  CENTERED  AT 0.5kHz
          30     40      50     60      70      80     90
           SPL OF OCTAVE BAND CENTERED  AT 4.0 kHz
 Figure 2A. The relation of threshold shift at asymptote (ATS^ for a test-tone of 715 Hz as a function of
 level of noise in an octave band centered at 500 Hz. Filled circles are means of the data in Fig. 1. The open
 square is from Miller, Rothenberg and Eldredge (1971) and the open circle from unpublished data of Miller,
 Eldredge and Bredberg (after Carder and Miller, 1972).
 Figure 2B. The relation of threshold shift at asymptote (ATS^ for a test-tone of 5.7 kHz as a function of
 level of noise in an octave band centered at 4 kHz. The open circles are from the data in Fig. 4A. The filled
 .squares are from other experiments (after Mills and Talo, 1972).
      The curvilinear extrapolations in both  Fig. 2A and Fig. 2B are based on the equation
                   ATS4 = 1.6
10log10l-£j—£)
where le is the square of the
                                           •c
 sound pressure of the noise and lc is a constant such that 10 log^n. lc = 65 for Fig. 4 A and 10 log-jg
 for Fig. 4B.

-------
    The threshold shifts (ATS j ]) as a function of frequency for the above exposures are
shown as audiograms with exposure level as the parameter in Fig. 3A. Here we see that the
largest shifts were at 715  Hz and that there was significant spread of ATS^ to higher
frequencies. Parenthetically we find  the shifts at 715  Hz to be easily replicated. There is
always some spread of ATS to higher frequencies, but it is often not so much as shown here.
    Von Bismarck (1967)  has measured the ratio of sound pressure at the ear drum (P^) of
the chinchilla to that in the free field (Pf) as functions of frequency, angle of incidence, and
static pressure in the middle ear. For the frequencies included in an  octave band centered
around 500 Hz he found the ratio P^/Pf to be about 3 dB. For the frequencies included in
an octave band centered around 4 kHz he found the ratio Pd/Pf to be 15-20 dB. These
differences are similar to those reported for man by Wiener and Ross (1946) and for the cat
by Wiener, Pfeiffer and Backus  (1966),  and represent sound pressure transformations
produced by combinations of acoustic diffractions about the head and pinna and resonances
in the outer ear  canal. Carder and Miller reasoned that the physiologically important
measure of exposure level will be P
-------
to
          dfi
            o
           10
           20
           30
40
           50
           60
                  OCTAVE  BAND NOISE
                  cf= 500 Hz
                        85 dB
                                  (A)
                        AFTER CARDER 8
                        MILLER
                                                                         I   I
OCTAVE BAND  NOISE
cf = 4.0 kHz

AFTER MILLS & TALO
                                                                                   5.7   11.4
                                                                                    I   I    I   I
               0.25     0.715      2.0   4.0   8.0   0.25  0.5    1.0   2.0   4.0   8.0   16.0

                                       FREQUENCY IN  KILOHERTZ

            Figure 3A. Mean audiograms at asymptote (ATS-ji) across ears and days at asymptote for the ears and
            conditions described in Fig. 1. (after Carder and Miller, 1972).
            Figure 3B. Mean audiograms at asymptote (ATS^) across ears and days at asymptote for the ears and
            conditions described in Fig. 4A. (after Mills and Talo, 1972).

-------
to
       60
    1	T-	J	T
       50 -
     Kl
     X
     JC
       30
    m
20
     (ft
        10
                   GROWTH
        *..^*—+J

     f    57 dB	*
                             65 dB-
                                                        1	P
                                               80dB
                                        m ^•^•^'•~*

                                        **'
                                          72 HB	••
                                              EXPOSURE:
                                              OCT-BAND NOISE
                                              cf - 4000 Hz
                                                                           T	p
      DECAY
                                                                                                                           T-f>-
        10
                                                                            60
                                                                                           t  t  t 1  1    t  f  1  t  t  t  1   t
                                                                                      12 1824 36
                         I
                                                                     MINUTES
                                                                                    HOURS
347  8 12 19  29  89
        'DAVS
                 6           12
                    DAYS OF  EXPOSURE
                                                 IS
24   4       50   100         500  IK         5K  10K  20K
                   TIME  AFTER NOISE  !N  MINUTES
                50K"lOOK
                    Figure 4A. Growth of TTS4 at 5.7 kHz. The levels of noise were 57,65, 72, and 80 dB SPL. For each level
                    of noise, the duration  is six days; the total duration is 24 days. Each point is the mean lor four animals
                    {after Mills and Talo, 1972).

                    Figure 4B. Decay of TTS at 5.7 kHz. The filled circles show the delayed recovery measured following the
                    exposure in Fig. 4A. The other data points are from other experiments  in which the more characteristic
                    slow recovery was observed (after Mills and Talo, 1972).

-------
     Up to this point in our review we can conclude:
     1) Continuous exposures to noise lead to losses of threshold sensitivity for pure tones
that increase for about 24 hours to an asymptotic threshold shift that may be maintained
for many days.
     2) The apparent stress on the  ear as indicated by the size of the ATS grows by 1.6 to
1.7 dB for each decibel increase in level above some critical level.3
     3) The critical level, if measured at the eardrum, is probably  only weakly related to
center frequency of the exposure-band over the range from about 0.4 kHz to about 6.0 kHz.
When measured in the field, the critical level is strongly related to the center frequency of
the exposure-band because of the enhancement of sound pressure level produced by
acoustic diffractions around the head and resonances of the ear canal.
     Recovery from ATS: Once the state we have called ATS4 has been reached, exposures
may be terminated and we  can study the recovery of sensitivity for pure tones. When this is
done we learn that ATS is different from some other forms  of temporary threshold shift in
that when ATS4 exceeds about 10-15 dB and the exposure has been nearly Continuous, then
recovery to normal will take several days. Under these circumstances recovery is never rapid
or even prompt. In the right half of Fig.  1 we see that recovery to normal threshold at 715
Hz required 2 to 6 days following exposures to the octave band centered at 500 Hz. There is
a slight trend such that with greater  ATS more time is required for recovery.
     Two  trends for recovery of sensitivity at 5.7 kHz following exposure to octave-band
noise centered at  4 kHz and at 80 dB SPL  are shown in Fig. 4B. The open symbols,
collected from several experiments  in which miscellaneous parameters were being explored,
show the same slow recovery observed at 715 Hz in Fig. 1. The filled symbols show an even
slower or delayed recovery  that was observed following the exposures and ATSs shown in
Fig.  4A.  Recovery to thresholds that were not significantly different from  pre-exposure
thresholds required at least  12 to  28 days and, as we shall see, these exposures produced
small changes in cochlear potentials  and small cochlear injuries.
    Permanent Threshold  Shifts:  The preceding paragraphs and  figures summarize  the
important trends in those relations  of behavioral threshold shifts to noise exposures that we
have already published. From the  point  of view of environmental safety these exposures
have already gone far enough. The  results of subsequent  studies now in press or in progress
show that increases in either dimension of exposure spell environmental hazard.  For example,
continuous exposures for 9  days  to an octave band of noise centered at 4 kHz  and at 80 dB
SPL have not produced permanent  shifts  in behavioral thresholds and are accompanied by
only minor changes in  cochlear potentials  and minor injuries to the inner ear. However,
when the level of exposure is increased to  86 dB or when the exposure to 80 dB is extended
for 90 days, we  see permanent threshold shifts. The recovery of  behavioral  threshold
sensitivity is delayed or even prolonged to more than 60 days and is not complete. When
there are permanent threshold shifts of this kind we have always found reduced cochlear
potentials and injuries to the inner ear.

                                Physiological Studies

     In our descriptions  of the behavioral  studies we have  alluded to changes in cochlear
potentials and to injuries to the inner ear. In the context  of the  present conference the
                                        244

-------
physiological studies are important because we find physiological changes that are quantita-
tively large enough to account for all of the behavioral loss of sensitivity. Thus we need not
invoke any central nervous system components to account for the threshold shifts.
     In a small rodent such as the chinchilla the three scalae of the individual turns of the
cochlea are readily  accessible for the insertion of electrodes. Cochlear microphonic (CM)
potentials can be measured differentially in each of three turns by recording the electrical
differences between scala  vestibuli and  scala tympani in the manner of Tasaki, Davis and
Legouix (1952). The  classic N^-N2 waveform of the whole-nerve action potential (AP)
response  can be recorded  by taking the average potential difference between the pair of
electrodes in the basal turn and a ground reference in the tissues of the neck wound. This
response appears most clearly for clicks  or  the onsets of tones and can be grossly analyzed
into smaller components  by observing  the changes  produced by masking noises with a
computerized version of the method reported by Teas, Eldredge and Davis (1962). It is also
possible to measure the endocochlear DC potential with pipette electrodes inserted in scala
media of each turn.
     Cochlear Potentials Early in Recovery: Benitez et al. (1972) exposed chinchillas to an
octave band of noise centered at 500 Hz and at 95 dB SPL for two to three days. This is
long enough to  assure that the state of ATS had been reached. Then DC, CM and AP were
measured at the  times indicated on the right half of Fig.  1. Two hours after the end of the
exposures the DC potentials were within our range of normal values.
     The  changes in CM responses to tones at 200 Hz are shown in Fig. 5. Panel A shows the
normal growth of CM voltage with increasing sound pressure level at the eardrum for each of
the three cochlear turns. At low levels the CM measured in the third, or apical, turn (CM3) is
larger than CM2 and CM j  in a manner consistent with the envelope of the Bekesy traveling
wave at this frequency. However, the three functions become nonlinear at different levels
and the  CM arising more basally  continues to  increase at higher levels. At maximum
response  the rank  order is reversed so that CMj gives the  largest voltage. In panel B the
average of CMj functions after  5  hours of recovery is compared to the mean control
function  for CMj.  There is 12 dB loss of sensitivity, 6 dB loss of maximum voltage, and the
SPL required to  produce maximum  CMj is shifted about 6 dB higher. Panels C and D show
similar comparisons for CM2 and CM3 after recovery  for  5 hours.  This low-frequency
exposure has produced a  clear gradient of increasing loss from less near the base to more
near the apex.
     The  recovery  of loss of sensitivity for CM as measured 5, 24,  and 48 hours after
termination  of exposure is shown in  Fig.  6 along with  the recovery of  mean behavioral
threshold  sensitivity at 715 Hz.  The physiological data are the means of three different
groups of five ears each at the three different times.  The trends for recovery are the  same for
all measures.
     For  more than 6  hours after the termination of exposure at 95 dB SPL to the band of
noise centered at 500  Hz it was not possible to elicit  an AP response to a wide-band click
Jeyen at levels 90 dB above normal visual detection levels  for this AP. A brain-stem evoked
response  could be  elicited at levels consistent with the behavioral and CM losses of sensi-
tivity. These two observations imply changes in degree of synchrony as well as changes in
sensitivity of neural responses. We  are not now prepared to interpret these changes any
further.
                                         245

-------
         3K
         IK
              CONTROLS
            -  200 Hz
    a.
     i
    a.
       300
   UJ
   CD
   
-------
       N
      X
      o
      o
      CM
      00
to
      00
      2
      LU
      00
           dB
           50
40
           30
           20
10
                           CM
           BEHAVIORAL
           ITS       x,
           (715  Hz)
                CM
                                  1
                                                          1
1
                                  5              10                  24

                           RECOVERY  TIME AFTER  NOISE   (HOURS)

             Figure 6. Recovery of CM sensitivity as a function of time after termination of exposure. The shifts in
             sensitivity at the lOO-^V level on the input-output functions for 200-Hz tones are shown by the ordinates.
             The parameter is the cochlear turn. Recovery from behavioral TTS for tones at 715 Hz is shown by the
             dashed line for comparison (from Benitez et al., 1972).
                                                                         48   HRS

-------
     Cochlear Potentials Late in Recovery. Even  when behavioral thresholds recover to a
sensitivity that cannot be statistically distinguished from normal pre-exposure values (s.e.m.
about 4 dB) we often find residual changes in cochlear potentials (Eldredge et al., in press).
For example, there were four ears in the group used for the four levels in Fig. 4. The CMj
responses were measured in each of these ears after the course of exposures and the recovery
shown. In each ear, CMj was found to be about two standard deviations below mean normal
values throughout most of the  dynamic range. The whole-nerve AP responses had normal
thresholds and in one ear grew quite normally; however, the other three ears gave responses
at suprathreshold levels that were up to more than two standard deviations below normal
means.

                                Anatomical Studies

     Early in these studies involving ATS we were  surprised to  find evidence for some
injuries to the organ of Corti and loss of hair cells with little or  no change in behavioral
thresholds. The first examples were briefly noted by Miller et al. (1971) and by Carder and
Miller (1972). These early indications of injury have now been confirmed repeatedly.
     Changes Seen Early in Recovery: Benitez et al. (1972)had restricted their exposures to
only two or three days in an attempt to achieve ATS without the minor loss of hair cells
reported by Carder and Miller (1972). Nevertheless, 23 of 30 ears so exposed and examined
by conventional histological methods showed unmistakably-missing hair cells and in  13 of
these ears the number of missing cells clearly exceeded  those we had seen in unexposed
control ears. The samples provided by only every fifth section did not allow more precise
evaluation of the injuries.  Also, the fixation of cell structure in  decalcified, celloidin-
embedded sections was not adequate for evaluation of more subtle cytological changes. For
this reason we adopted the  osmium-fixed, araldite-embedded, flat preparation described by
Bohne (1972) for our subsequent studies.
     We have  already been handsomely rewarded by this extra  care in our anatomical
preparations. Bohne (in preparation) has reviewed ears exposed continuously for 48  hours
to an octave band centered at 500 Hz with levels  at 75, 85, or 95 dB SPL and to an octave
band centered at 4  kHz at 80  dB  SPL.  Ears that are fixed one to two  hours after the
termination  of exposure show, by phase  contrast  microscopy, outer  hair cells  with
thickened walls and with intracellular accumulations of homogeneous-appearing material.
These cells are seen primarily in the lower part of the third  turn after exposures to the
octave band centered at 500 Hz and in the lower part of the first turn after exposures to the
octave band centered at 4  kHz. The incidence of outer hair cells with the above changes
decreases with level of exposure.
     When the outer hair cells are examined  by electron microscopy, it becomes clear that
there has been a marked proliferation of the cisternae of the smooth endoplasmic reticulum
which comprises the peripheral  membrane  system of the cells. The top of Fig. 7 shows part
of a normal outer hair cell sectioned at a horizontal angle. Eight rows of flattened cisternae
are present just inside the plasma membrane (at the arrow) of the cell. There are 30 rows of
cisternae hi the peripheral membrane system in the outer hair cell from the third turn which
is shown in the bottom part of Fig. 7. This cell is from an animal exposed to octave-band

                                        248

-------
                                         ."           .

         "

Figure 7. Electron micrographs of portions of two  outer  hair cells showing  the  smooth endoplasmic
reticulum which forms the peripheral membrane system of the cells. Top shows eight  rows of cistemac in a
normal cell. The plasma membrane is at arrow. Bottom shows proliferation of the cisternae to form 30 rows
in a cell from a noise-exposed chinchilla with an asymptotic threshold shift.

-------
noise centered at 500 Hz at 95 dB SPL; the specimen was collected 1-2 hours after the end
of the exposure. The homogeneous-appearing material seen in the outer hair cells by phase
contrast microscopy was found to consist of extensions of the cisternae of smooth endo-
plasmic reticulum into the central portions of the cells. These changes are slowly reversible.
After 7 days of recovery the membrane systems in the outer hair cells show a return toward
average proportions. However, some scattered outer hair cells still contained small areas of
the proliferated cisternae even in ears allowed to recover for 70 days.
     Changes Seen Late in Recovery: Both  early and late,  scattered loss of up to a few
hundred hair cells has been a  regular finding. On closer  examination some of these ears are
showing changes that  persist more than three months after  exposure, at a time when
recovery to stable behavioral  thresholds is complete.  For example,  some of the outer hair
cells in places where outer hair cells have been shown to develop extra rows of cisternae in
the peripheral membrane systems may have membrane systems of average dimensions, but
the shape of the cells may be irregular and the peripherally-arranged mitochondria reduced
in number (Bonne et al., in press). Also, the afferent  nerve fibers under the inner hair cells
and  in  the  outer spiral bundles may show increased numbers of vesicles, vacuoles, and
strands of endoplasmic reticulum.

                         Parallels Between Man and  Chinchilla

     Behavioral measures of  thresholds are the only ways we can compare man to  the
chinchilla. For the chinchilla we have demonstrated the phenomenon of asymptotic thres-
hold shift. Mills et al.  (1970) and Mosko  et al.  (1970)  have demonstrated asymptotic
threshold shifts in man. For both man and chinchilla, slow recovery of-threshold sensitivity
after the end of exposure has been associated with the state of ATS. In chinchilla we have
also seen recoveries delayed even longer than those  we have  characterized as slow. This
finding invites closer comparisons along the dimension of rate of recovery. We are not  yet
ready to specify quantitatively the different rates of recovery that we believe are opera-
tionally important. Nevertheless, we believe that it may be necessary to distinguish among
different rates which may be qualitatively  described as: a) rapid,  b) prompt, 
-------
                                  Table 1.
CITATIONS SUPPORTING DIFFERENT RATES OF RECOVERY FROM AUDITORY THRESHOLD SHIFTS
 Rate of

 Recovery
Chinchilla
 A.  rapid         Peters (1965)

    ^  3 hours*


 B.  prompt        Saunders, et al.

    ^  16 hours*   (in preparation)
 C.  slow

    > 1 day*
Carder & Miller  (1971)

Miller et al.  (1971)
 D.  delayed       Mills & Talo  (1972)

     > 1 week*


 E.  prolonged     Mills  (in press)

     > 2-3 weeks*
Man
                          Ward, Glorig & Sklar (1958)
                          Ward, Glorig & Sklar (1958)
Davis et al.  (1950)

Ward (1960)

Mills et al.  (1970)

Ward (1970)
 * times are not yet well  established but are similar to examples shown.
                                   251

-------
perspective to the data we have acquired at lower levels and reviewed here. Table II sum-
marizes for the octave bands centered at 500 Hz and 4 kHz a) the exposure levels implied by
the subtractive constants in Fig.  2A and 2B, b) the exposure levels required to produce an
ATS4 of about 50 dB, and c) the exposure levels that have been shown to produce in 3.5
hours injuries that progress to total loss of organ of Corti over some distance along the
basilar membrane.
     The subtractive  constants,  65 and 47, of Fig.  2A and 2B imply that exposures to
corresponding levels could be prolonged indefinitely and that the ATS4 would be only 0-5
dB. The functions in this figure state  that above these  levels the stress on  the ear, as
measured by ATS, grows rapidly at about 1.6 dB per decibel increase in level. For exposures
that are about 30 dB above the levels implied by the subtractive constant, ATS4 is about 50
dB and behavioral thresholds can recover completely after exposures for a few days. But
there will usually be  some loss of hair cells after these exposures. We know that when these
exposures are continued beyond 9  days to 90 days (Mills, in preparation) permanent loss
appears. We know that an increase of only 6 dB can lead to permanent threshold shift (Mills,
in press) after exposure for only 9 days. Experiments in progress suggest that exposures for
only 6 hours daily with 18 hours in quiet will only delay the acquisition of injury and loss.
The dynamic range between levels that appear to be entirely safe and those that are clearly
injurious is only 30 dB.
     When  exposure levels are increased by an additional 30 dB, severe cochlear injuries, loss
of cochlear potentials, and permanent threshold shifts follow exposures of only 3.5 hours.
dearly such exposures are excessive and their equivalents are to be avoided by man at all
costs.
     It is tempting to make a final extrapolation to  man.  The data  collected on one man
(Mills et a/., 1970) suggest that the subtractive constant for  the octave band centered at 500
Hz is about 75 and that continuous exposures at 75 dB SPL should produce asymptotic
threshold shifts  less  than 5 dB. The same data  and  some unpublished data provided by
Melnick indicate that for man ATS4 also grows by about  1.6 dB for each decibel increase in
exposure level.

                                Concluding Remarks

     We believe  the results of our studies will support the following assertions concerning
hearing and noise exposure for both man and chinchilla:
     1) Sound Levels: Above some critical level, manifestations of stress on the ear increase
by about 1.6 dB3 for each decibel increase in level of exposure.
     2) Frequency Spectrum:  We can account for most of the differential hazard related to
frequency  spectrum by the differences in the ratios of sound pressures at the tympanic
membrane to the sound pressures in the sound field.
     3) Stress on Hearing:  There are combinations of level and durations of exposure to
noise that  produce temporary threshold shifts characterized by slow recovery to normal
thresholds.
     There is a fourth assertion that we know is true for chinchilla and that we suspect may
also be true for man.

                                        252

-------
                              Table 2.
              CRITICAL EXPOSURE LEVELS FOR CHINCHILLA
Completely Safe Levels,

     ATS4 = 0-5 dB no matter

     the duration
                                           Center Frequency of

                                               Octave Band

                                        0.5 kHz          4.0 kHz
  65 dB SPL
 47  dB SPL
Borderline Levels,

     ATS4 ^ 50 dB and is

     temporary
  95 dB
 77-80 dB
Level for Severe Injury after

     3.5 hour-exposure
120-128 dB
108 dB
 * All values are approximate,  i.e. "t 3 dB
                               253

-------
    4) Cochlear Injury: The pathological anatomy and physiology observed in cochleas of
chinchillas following exposures that were characterized by slow or delayed recovery from
ITS have in all or nearly all instances shown destruction of hair cells and permanent loss of
cochlear potentials.

1 Supported in part by Grant No. NS-03856 from the National Institute of Neurological Diseases and Stroke
 to the Central Institute for the Deaf and in part by Grant No. NS-01791 from the National Institute of
 Neurological Diseases and Stroke to the Department of Otolaryngology, Washington University School of
 Medicine.
2 also Department of Otolaryngology, Washington University School of Medicine.
3 As more data have been accumulated in unpublished repetitions and extensions of measurements of ATS,
 a slope of 1.7 dB per decibel tends to describe the relations better than 1.6 dB per decibel.

                                    References

BENITEZ, L.D., ELDREDGE, D.H., and TEMPLER, J. W., Temporary threshold shifts in
     chinchilla: Electrophysiological correlates. /. acoust. Soc. Amer., 52 1115-1123 (1972)
BOHNE, B.A., Location of small cochlear lesions by phase contrast microscopy prior to thin
     sectioning. Laryngoscope LXXXII, 1-16 (1972).
BOHNE, B.A., ELDREDGE, D.H.,  and MILLS, J.H.,  Electrophysiology and electron-
     microscopy for the study of small cochlear lesions. Ann. Otol  Rhinol Laryngol. (in
     press).
BOHNE, B.A. (in preparation) Anatomical correlates of asymptotic temporary shift of the
     threshold of hearing.
CARDER, H.M. and MILLER, J.D., Temporary threshold shifts produced by noise exposure
     of long duration. Trans. Am. Acad. Ophthal Otolaryng., 75, 1346-1354 (1971).
CARDER, H.M. and MILLER, J.D., Temporary threshold shifts from prolonged exposure to
     noise./. Sp. Hear. Res., 15, 603-623 (1972).
DAVIS, H., MORGAN, C.T.,  HAWKINS,  I.E., Jr., GALAMBOS, R., and  SMITH, F.W.,
     Temporary deafness following exposure to loud tones and noise. Acta otolaryngol.,
     Supp.88, (1950).
ELDREDGE, D.H., MILLS, J.H.,  and BOHNE, B.A., Anatomical, behavioral and electro-
     physiological observations on chinchillas after long exposures to noise.  Proceedings of
     the International Symposium on Otophysiology, Ann Arbor, Mich., Univ. of Michigan
     Medical Center, May 20-22, 1971, (in press).
MELNICK, W., Personal communication.
MILLER, J.D.,  Audibility curve of the chinchilla. /. acoust.  Soc. Amer., 48, 513-523
     (1970).
MILLER, J.D., ROTHENBERG, S J., and ELDREDGE, D.H., Preliminary observations on
     the effects of exposure to noise for seven days on the hearing and inner  ear of the
     chinchilla./, acoust. Soc. Amer., 50, 1199-1203 (1971).
MILLS, J.H., Temporary and permanent threshold shifts produced by nine-day exposures to
     noise. /. Sp. Hear. Res. (in press).
MILLS, J.H. (in preparation) Threshold shifts produced by a 90-day exposure to noise.


                                        254

-------
MILLS, J.H., GENGEL, R.W., WATSON, C.S., and MILLER, J.D., Temporary changes of
    the auditory system due to exposure  to noise for one or two days. /. acoust. Soc.
    Amer., 48, 524-530(1970).
MILLS, J.H. and TALO, S.A., Temporary threshold shifts produced by exposure to high-
    frequency noise./. Sp. Hear. Res.,  15, 624-631  (1972).
MOSKO, J.D., FLETCHER, LTC J.L., and LUZ, CPT G.A., (1970). "Growth and Recovery
    of  Temporary  Threshold Shifts Following  Extended Exposure  to ,High-Level,
    Continuous Noise," Rept. No. 911, (Exp. Psych. Div., U.S. Army Med. Res. Lab., Ft.
    Knox, Kentucky).
PETERS, E.N.,  Temporary shifts in auditory thresholds of chinchilla after exposure to
    noise./, acoust. Soc. Amer., 37, 831-833 (1965).
SAUNDERS, J.C., MILLS, J.H. and MILLER, J.D. (in preparation) Threshold shift in the
    chinchilla from daily exposures to noise for six hours.
TASAKI, I., DAVIS,  H., and LEGOUIX,  J.P.,  The space-time patterns of the  cochlear
    micro phonics (guinea pig) as recorded by differential electrodes. /. acoust. Soc. Amer.,
    24,502-519(1952).
TEAS, D.C., ELDREDGE, D.H., and DAVIS, H., Cochlear responses to acoustic transients;
    an  interpretation of whole-nerve  action potentials.  /. acoust. Soc. Amer.,  34,
    1438-1459(1962).
VON BISMARCK, G. (1967), The sound pressure transformation function from free-field to
    the  eardrum of  chinchilla.  M.S.  thesis, Massachusetts Institute of Technology,
    Cambridge, Mass.
WARD, W.D., GLORIG, A. and SKLAR, D.L., Temporary threshold shift from octave-band
    noise: Applications to damage risk criteria./, acoust.  Soc. Amer., 31, 522-528 (1959).
WARD,  W.D., Recovery from high values of temporary threshold shift. /. acoust.  Soc.
    Amer., 32,497-500(1960).
WIENER,  P.M.  and  ROSS,  D.A.,  The pressure distribution in the auditory canal  in  a
    progressive sound field. /. acoust. Soc. Amer., 18, 401-408 (1946).
WIENER,  F.M.,  PFEIFFER,  R.R., and  BACKUS, Ann S.N., On the sound  pressure
    transformation by the head  and auditory  meatus of  the act. Acta otolarying., 61,
    255-269 (1965).
                                      255

-------
       PRESBYACUSIS IN RELATION TO NOISE-INDUCED HEARING-LOSS

                                     A. Spoor
                        ENT Department, University Hospital
                                Leiden, Netherlands

Introduction

     The well-known phenomenon that hearing deteriorates with age is commonly called
presbyacusis and has in this sense a very broad meaning. On this point we must be more
precise. Presbyacusis is  the hearing loss caused by  the pure process of aging itself, but aging
can happen under more or less favorable conditions. Unfavorable conditions can arise from
genetics,  drugs, noise,  nutrition,  stress, illness, climate and  maybe even  from  unknown
circumstances. These conditions differ widely  and therefore it  can be expected that the
hearing acuity is different in different populations and even may change in time. So we have
to choose as a definition of presbyacusis the process of deterioration of hearing under
circumstances that are  normal for the group under consideration. Perhaps for this kind of
hearing loss a better name would be presbya-socia-cusis. However, it is impossible to have
many different values for the hearing  loss caused by presbyacusis and therefore we  will
present here data based on several field surveys.  The reason why we  want  to know the
hearing level at different ages is that we want to have a basis on which we can  judge the
influence of some special factor, which is for this congress the factor "noise". The influence
of noise on hearing-level is measured  as the amount of noise-induced hearing  loss (NIHL).
With regard to the interaction between presbyacusis and NIHL there are a few questions to
be considered:
1.   Is the interaction between presbyacusis and NIHL additive or  nonadditive and  is
     sensitivity to NIHL dependent on age?
2.   Is the influence of noise on hearing level comparable with that of other factors?
3.   Is it possible to arrive at standard values for presbyacusis, i.e. is it possible to  give
     audiometric zeros for different ages?
4.   If the latter is possible, what is the spread of these values?

1.   Interaction between presbyacusis and noise-induced hearing-loss.

     The pathology of the influence of noise on the organ of hearing is pretty well known.
The destruction confines itself to the hair-cells  of  the basilar membrane and in severe cases
the ganglion  cells are  affected also.  According  to Gacek and Schuknecht (1969) the
pathology of presbyacusis is very complex. The hearing-loss is of the sensorineural type and
may involve one or more of four types: a) sensory presbyacusis, i.e.  degeneration of the
organ of Corti; b) neural presbyacusis with auditory neuron degeneration;  c) metabolic
presbyacusis with atrophy of the stria vascularis and d) mechanical presbyacusis by restric-
tions in the mobility of the basilar membrane.
     One can expect that the interaction between  presbyacusis and NIHL is dependent on
the type  of presbyacusis and simple addition can only be expected for the fourth type of
presbyacusis,  but in general it is not. Some authors, however, report summing without

                                        257

-------
influencing each  other, e.g. Mollica (1969). The  reason  should be that in presbyacusis
mostly ganglion cells are affected but it is not well understandable that this is sufficient to
explain addition.  Mollica observed in groups of people with the same noise exposure that
after deduction for age the hearing loss  is the same at different ages, which can be explained
by addition. He concludes that older people have more hearing loss not because of greater
sensitivity to noise but because of old age phenomena. On  the other hand, Gallo and Glorig
(1964)  observed that for a noise-exposed group, the hearing level at 4000 Hz is constant
after 15 years of  exposure. In a non-noise-exposed group, the hearing level continued to fall
and this suggests  that there is  no simple addition. Although the answer to question one is
not clear in general, the conclusion  must be that the interaction between presbyacusis and
noise-induced hearing loss is not purely  additive.

2.  Influence of noise on hearing compared with other factors.

    There are strong suggestions that the influence on hearing from noise is accompanied
by precapillary vasoconstriction. Precapillary vasoconstruction in man in response to various
types of noise was demonstrated by Jansen et al.  (1964), while Lawrence et al. (1967)
demonstrated  vasoconstriction histologically in  animal experiments  for the spiral vessels
underlying the basilar membrane  of  the cochlea. Friedman  et al.  (1967) showed that
atherosclerosis was greater in noise-exposed animals. Rosen (1969) concluded from hearing
surveys in  the Mabaans, in Finland, in Crete and the Bahamas that accelerated loss of
hearing with age  is strongly correlated with atherosclerosis and coronary diseases and not
with vascular  hypertension and  cerebro-vascular incidents. From these findings  one may
conclude that the influence of noise  on  hearing runs parallel with a factor like diet by means
of the blood supply even to the cochlea.
    A  subquestion might be: Is there an essential difference in hearing-level  between
different populations? The findings of Rosen (1962) in the Mabaan-tribe in the  Sudan at
first suggested that  there were  people with essentially better hearing. Bergman (1966) how-
ever demonstrated in a critical  analysis  of the Rosen data that the hearing of the very young
Mabaans was the  same as that  of the very young people from  cities in other countries and
also that the 10% best hearers  in the Mabaans were equal  in HL with the 10%  best hearers
from  other populations. These findings indicate that the Mabaans preserve their hearing
better, especially for the high frequencies.

3.   Presbyacusis values.

     In order to  evaluate the  influence of noise on hearing of people we must know the
normal hearing levels as a function of age. The latter will  be called presbyacusis values. We
already saw that  there will be  many differences but we must see how far we can get. In a
working group on noise influences of the Organization for  Health Research of the Organiza-
tion for Applied  Scientific Research in the Netherlands (T.N.O.), studies have  been under-
taken to analyse the results of several hearing surveys in terms of both hearing levels and the
spread in hearing levels (Spoor, 1967; Spoor and Passchier-Vermeer, 1969). For determina-
tion of the presbyacusis values, eight hearing surveys in the literature have been compared

                                        258

-------
and analyzed: Hinchcliffe (1959), Corso (1963), Jatho and Heck (1959), Johanson (1943),
Beasley (1938), A.S.A. Report (1954), Glorig (1957) and Glorig (1962).
    Table I gives some details of these surveys: The number of people involved, selection or
not, the kind of population  and the kind  of values given for the hearing-loss (median or
mean).

                                       Table 1.
                         INVESTIGATIONS USED FOR ANALYSIS
author
Hinchcliffe, 1959
Corso, 1963
Jatho and Heck, 1959
Johansen, 19^3
Beasley, 1938
A.S.A. Report, 195^
Glorig, 1962
Qlorig (WSF), 195^
number
men
326
493
399
155
2002
—
2518CR)
1724
women
319
75^
361
155
2660
—
—
17^1
selected +
non-select. -
+
+
+
+
•f
-
-
-
kind of
population
rural
-
-
-
random
-
professional
industrial
value
given
median
median
mean
mean
mean
mean
median
median
     At first glance, there are great differences in the values for the hearing losses, but it was
proven that they can easily be compared by taking the age group around 25 years in each
survey as a reference. A further analysis starts with the assumption that hearing levels for
the age group of 25 years may be equated in order to evaluate the influence of age.
     The  data  from the survey mentioned were brought together in this way  for each
frequency as a function of age. Figure 1 gives an example: the frequency is 4000 Hz and the
values concern men. It is clear that the differences are small except for the Wisconsin State
Fair data of Glorig (1957) and it is easily possible to draw a best-fitting curve. This has been
done for several frequencies for both men and women. In the article mentioned (Spoor,
1967) it is proposed that these best fitting curves can be given by an equation: log (HL+ c)
= b log  (age) - a in which a, b and c are  frequency- and sex-dependent constants. With the
aid of these equations, the hearing levels can be calculated for any age at different frequencies
(Fig. 2). From these data, the normal audiograms for different ages can be drawn for men
and for women. Fig. 3 shows these audiograms for men. It was also discussed that these data
can be considered  as median values; they have been brought together in table 2 and table 3.
It can be repeated that  these data have been calculated with respect to levels for the 25 age
group, but the standard zero level for audiometers is also based on this group and therefore
the data given may be considered as median hearing levels for different age groups of
non-noise-exposed  people.  It may be mentioned that  the hearing  level for men at the
frequency of 4000 Hz can be given with the following rule of thumb:
HL =  inn  - 6; for women this value has to be multiplied by 2/3.
                                        259

-------
Figure 1. Data points derived from 8 investigations giving hearing level as a function of age with respect to
the hearing level in the 25- or 21 Vz-year-old age group in the same investigation. Frequency 4000 Hz. Solid
curve according to the equation (see text).
     After completion of our work two other surveys came to our attention: Riley (1961)
and Glorig and Roberts (1965). The data of Riley for selected people hardly differed from
our data. The Glorig and  Roberts data are  from the  1960-1962 United States National
Health Survey and differ more, but the data for the better ear come pretty close to our data.
In  any  case it can be concluded  that our calculated data would not have been essentially
different when these other data had been included. Naturally, our data show the well-known
age, frequency and sex dependency.

4.   Spread of presbyacusis values of non-noise exposed people.

     From the surveys already mentioned, three could not be used for calculation of the
variability: A.S.A. Report (1954),  Glorig (1957) and Glorig (1962). From the remaining five
surveys the values M-QU and Qj-M have been calculated for each frequency and age group,
where M is the median hearing level and Qu and Qj are the upper and lower quartile values,

                                         260

-------
          -t—
Figure 2.  Curves giving  the relation between hearing level  and age for different frequencies for men
according to the calculated data.
i.e. the hearing-levels respectively not exceeded in 25% and 75% of the people in the age
group. Fig.  4 gives values of M-QU and Qj-M in the age groups of 25, 35, 45, 55 and 65
years in men for different frequencies. Here also the well-known fact  is found  that the
spread increases with frequency and age.  There is little difference for the sexes  and also
there is little difference  between M-QU and Qj-M.  The values of Qu  and Qj  are given  in
table 2 and  3 together with the M values. At last we can compare the quartile values and
median values for the different age groups and then it can approximately  be concluded that
the lower quartile value of one age group coincides with the median value  of the next higher
age group of ten years while the upper quartile value coincides with the median value of the
next lower age group. This can also be formulated as: from a certain age  group of ten years
50% of the people have hearing-levels in between the median hearing-levels of the next lower
and the next higher age group. This is illustrated in fig. 5.

                                          261

-------
                • 0,75-  1    • n.-s
Fig. 3.  Curves giving the relation between hearing level and frequency at different ages for men according
to the calculated data.
Summary:

     In this review article the following points are stressed. Presbyacusis is the deterioration
of hearing caused by the process of ageing. The circumstances for this ageing process can be
more or less favourable. Unfavourable circumstances can arise from genetics, noise, drugs.
nutrition, stress, illness and climate.  In  general, the interaction between presbyacusis and
NIHL  is not  additive. There is no essential  difference in hearing-level between different
populations. The sensitivity for NIHL  is not  clearly dependent  on age. The  influence of
noise to arrive at standard values of presbyacusis: data are proposed. The variability of these
median hearing-levels is also given.

                                          262

-------
                               Table 2.
  MEDIAN HEARING LEVELS AND UPPER AND LOWER QUARTILE VALUES FOR MALES
                       IN AGE GROUP OF 10 YEARS.
age
frequency
HZ
250
500
1000
2000
3000
4000
6000
8000
20-29
«,,

-4
-4
-4
-4

-6
-6
-6
M

0
0
0
0

-------
 dB

 10-
M-0
                                                                        M-
(),'_»5
  ]() -I
 dB
                                                                                 35

                                                                                 45
                                                                                65
                                                                                age
    Fig. 4. Curves giving M-Q,, and Q-j-M as a function of frequency for male groups with age the parameter.
 References:

  1. A.S.A. Report Z-24-X-2, The relation of hearing loss to noise exposure, 1954.
  2. Beasley, W.C.:  National Health Survey, 1938; data quoted from: Sunderman, F.W.
     and Boerner, F.: Normal Values in Clinical Medicine, Philadelphia, 1949.
  3. Bergman, M.: Hearing in the Mabaans. Arch. Otolaryng. 84: 411-415(1966).
  4. Corso, J.F.:  Age and sex  differences in pure-tone  thresholds. Arch, Otolaryng., 77:
     385-405(1963).
  5. Friedman, M.; Byers, S.O. and Brown, A.E.: Plasma lipid resppnses of rats and rabbits
     to an auditory stimulus. Amer. J. Physiol. 272: 1174-1178 (1967).
  6. Gacek,  R.R.  and Schuknecht, H.F.: Pathology of presbyacusis. Int. Audiol. 8: 199-209
     (1969).
                                         264

-------
                       0,5
4
I
6
I
                                                                            M ( 35
   •4O —i
                                                                            M (75
 Fig. 5. Comparison of Qu (45) with M (35), Q1 (45) and Qu (65) with M (55) and Q1 (65) with M (75) at
 various frequencies for male groups.
 7. Gallo, R.  and Glorig, A.: Permanent  threshold  shift changes produced by noise
    exposure and aging. Amer. Industr. Hyg. Assoc. J. 25: 237-245 (1964).
 8. Glorig, A.  e.a.: Wisconsin State Fair Hearing Survey, American Academy of Ophthal-
    mology and Otolaryngology, 1957.
 9. Glorig, A.  and Nixon, J.:  Hearing-loss as a function of age. Laryngoscope 72:
    1596-1610(1962).
10. Glorig, A.  and Roberts,  J.: Hearing levels of adults by age and sex, United States
    1960-1962. National Center for Health Statistics: Series 11, number 11  (1963).
11. Hinchcliffe, R.: Threshold of hearing as a function of age. Acustica 9, 303-308 (1959).

12.  Jansen, G. e.a.: Vegetative reactions to auditory stimuli. Trans. Amer. Acad. Ophthal.
    Otolaryng.: 68: 445-455 (1964).
                                         265

-------
13.  Jatho,  K. und  Heck, K.H.:  Schwellenaudiometrische  Untersuchungen uber  die
    Progredienz  und  Characteristik , der  Alterschwerhorigkeit  in  den  verschiedenen
    Lebensabschnitten. Zeitschr. Laryng., 38: 72-88 (1959).
14.  Johansen, h.: Den Aldersbetingede Tunghorhed, Munksgaard, Kobenhavn (1943).
15.  Lawrence, M.; Gonzales, G. and Hawkins, I.E.: Some physiological factors in noise-
    induced hearing-loss. Amer. Industr. Hyg. Assoc. J., 28: 425-430 (1967).
16.  Mollica, V.: Acoustic trauma and presbyacusis. Int. Audiol., 8: 305-311 (1969).
17.  Riley, S.C. e.a.:  Ten years' experience with Industrial Audiometry. Amer.  Industr.
    Hyg. Assoc. J., 22: 151 (1961).
18.  Rosen,  S.; Bergman, M.; Plester, D.; El-Mofty, A. and Satti, M.H.: Presbyacusis study
    of  a relatively  noise-free  population in the Sundan.  Ann. Otolaryng., 77: 727-743
    (1962).
19.  Rosen, S.: Epidemiology of hearing loss. Int. Audiol. 8: 260-277 (1969).
20.  Spoor, A.: Presbyacusis values in relation to noise-induced hearing-loss. Int. Audiol. ^:
    48-57 (1967).
21.  Spoor,  A. and Passchier-Vermeer, W.:  Spread  in hearing-levels of non-noise  exposed
    people at various ages. Int. Audiol. 5:328-336 (1969).
                                        266

-------
                   NOISE EXPOSURE, ATHEROSCLEROSIS AND
                         ACCELERATED PRESBYACUSIS

                              Z. Bochenek, W. Bochenek
                      Otolaryngological Dept., Medical Academy
                                   Warsaw, Poland


    In  1968, we presented the results of studies, undertaken in the Audiological Labora-
tory of  the Central Research Institute of the Polish State Railroad Health Service, for a
group  of  engine-drivers  concerning  noise  exposure,  atherosclerosis  and  accelerated
presbyacusis.
    The investigations concerning this group were multidimensional and included, besides
the examinations of the ear and  hearing, ophthalmological, neurological, psychological,
electrocardiographic and biochemical examinations. The aim of these investigations was to
evaluate the ability of an engineer to do further work in his profession.
    It  is generally  understood that  the  work  of a railroad engineer involves chronic
exposure to noise, nervous tension,  and irregular eating and resting schedules.
    The results of the ophthalmological, neurological and psychological examinations are
not included  here, since they are not directly connected with the subject of hearing loss. We
shall give only the results of the audiological examinations and tests bearing on the existence
or absence of atherosclerosis.
    The state of hearing was evaluated by means of pure-tone thresholds.
    The evidence for atherosclerosis was classified as either "definite" or "probable." As
definite  symptoms, the  following were  included:  I. heart infarction, 2.  coronary disease,
and 3.  intermittent  claudication  with decreased  oscillometric  deviations in the lower
extremities.
    As  "probable"  symptoms, the following  items of evidence were accepted:  arterial
hypertension, hypertrophy of the left ventricle, accentuation of the second aortic sound in
patients with normal blood pressure, systolic  murmur  heard at  the base of the heart,
diminished elasticity  of the radial arteries, asymmetric or absent pulse in the dorsal pedis or
posterior tibial  arteries, asymmetric oscillations  in the lower  extremities,  non-specific
changes  in the distal part of the ventricular ECG complex and cardiac rhythm disorders in
the form of atrial fibrillations, or multiple premature extrasystoles in an individual without
clinical evidence for heart disease. When the examined person showed two or more probable
signs of atherosclerosis, these  individuals were then assumed to have "definite"  athero-
sclerosis. The results of tonal audiometry were then compared with curves proposed by
Aubry (BBL) as typical for the age group. Aubry also  differentiates  two forms of pres-
byacusis: i.e.  "pure" or "physiological" and "accelerated".
    The group of 110 engineers, aged 51-60 years, with noise exposure in their professional
Hfe of about  30 years—where the intensity of noise in the cabin reached, in certain periods,
values up to  112 dB  SPL—was divided into two subgroups, one with signs of atherosclerosis
and the  other without such signs. All individuals with otoscopic abnormalities or history of
chronic  inflammation of the ears, and persons with a history of skull trauma or who had
been treated with ototoxic drugs were excluded.

                                        267

-------
     The  results are demonstrated in Table I. These results indicate that the accelerated
form of presbyacusis is more common in the group of persons with symptoms of athero-
sclerosis than in the group without any detectable signs of this disease.
Sroup

 I
II
 Age

51-60
                                        Table I
Atherosclerosis present /+/
                absent /-/
Atherosclerosis /+/
Atherosclerosis /-/
                                  The number of
                                  observations
                                      59
                                      51
The form of presbyacusis
pure
22 /37.V
25 /47.V
accerelated
37 /€3.v/
2? /53.V
                               according to Aubry s curves
     For comparison, the examination of hearing was performed on a group of 45 engine-
 drivers aged 41-50 years. The results are shown in Table II. The comparison indicates that
 in this  age group the accelerated  form of presbyacusis was encountered in a relatively
-smaller number of individuals than in the age group 51-60 years. However, it should be
 mentioned that in  the group of engine-drivers aged 41-50 years, presented in Table II, no
 systematic examination for atherosclerosis was performed. It may be only speculated that
 definite signs of atherosclerosis appeared less frequently in this group of individuals than in
 the former group, that is, in the persons aged 51-60 years.
    Group

     I
    II
 Age

41-50
51-60
                                 Table II

           The  number of observations

                      45
                     110
                                    The form  of presbyacusis
                                        pure
                                    27 /60,-V
                                    46 /42jS/
accerelated
13
64
                                   according to Aubry'o curves
     In order to establish  the function of hearing in  persons who are not professionally
 exposed to noise, but demonstrate definite signs of atherosclerosis, studies were recently
 undertaken in the Department of Otolaryngology of Warsaw Medical Academy.
     Preliminary results include the following: The studies concerned 32 men, aged 41-60
 years, who were under  the care  of  the Institute of Cardiology of the Warsaw Medical
 Academy because of at least one previous heart infarction. None of the 32 men examined
 was professionally exposed to noise. The remaining criteria of selection were the same as in
 the previous group. Patients with diabetes, diseases of kidneys, etc. were eliminated.
     Table  III presents the results of tonal audiometry in this group. A slightly different
 system of classification was used  than in the  two former tables, since the norms given by
 Aubry for so-called "pure" presbyacusis seem to be rather elevated, at least according to the
 norms given by other authors, such as Glorig, Hinchcliffe, Jatho and Heck, Leisti, Spoor and
 van Laar.  It  appears  from this Table III that,  particularly in the age group 51-60, an
 accelerated form of presbyacusis was encountered. A comparison of the hearing levels of
                                         268

-------
                                       Table III
Group

 I
II
  Age

41-50
51-60
The number of
observations
     16
     16
Audiometric
better than "pure"
2
3
"pure"
0
2
A. **. A -~ *~ 	
threshold
worse than "pure"
       7
       3
accerelated
     4
     8
                               in comparison to Aubry  a  curves of pure and
                               accerelated form of  presbyacuais
this group (Table III 51-60 years) with the hearing levels for the same age group, but
exposed to noise (Table I) shows that the  accelerated form of presbyacusis appears more
frequently in the group exposed to noise.
                                         269

-------
               HIGH-FREQUENCY HEARING AND NOISE EXPOSURE*

                                  John L. Fletcher
                              Department of Psychology
                    Memphis State University, Memphis, Tennessee

    Little or no research or clinical concern had been shown toward the measurement of
hearing for frequencies above 800 Hz until fairly recently. There were many reasons for this
disregard for quantification and study  of high-frequency hearing. A very real and practical
reason was that early efforts to test high-frequency hearing were highly unreliable, at least
partly because of the problems arising in the coupling of the ear with the transducer used to
deliver the acoustic stimulus. With  higher-frequency signals and the shorter wave lengths
that  are  associated with  higher  frequencies,  the placement  of  typical  over-the-ear
transducers, earphones, is highly  critical and  the thresholds might reflect more upon the
exact placement of the earphone than upon the actual ability of the Subjects (5) to hear the
signal. For that and other technical reasons, then, little was known .until recently about high
frequency hearing in humans, nor were norms or standards available to allow us to evaluate
or compare high-frequency hearing among various S's. The breakthrough in this area came
when Rudmose, after pondering upon the problem, came  up with a simple but effective
solution. He utilized a small microphone with a conical probe tip as an earphone, inserting
the tip of the probe into the ear.  This provided a reliable and effective coupling of the ear
and the transducer. He then proceeded to examine  the hearing of several young healthy
non-noise exposed  male and female high school students, pooling their data to provide an
interim biological baseline that could  be used with caution to evaluate hearing data from
future persons tested. Fletcher (1965), using an early model of the Rudmose high frequency
audiometer, determined that reliability  of the technique compared favorably with that of
conventional audiometry (Fig. 1). Zislis and Fletcher (1966) then tested male and female
non-noise-exposed  6th, 9th, and  12th grade students to establish whether high-frequency
hearing varied within such age limits.  Essentially, they found that it did not, that females
were better than males, and that probably Rudmose's original data were from too select a
population and people  do not hear quite as well as he had supposed. Northern et al. (1972)
in a later standardization study coordinated with earlier studies, emerged with proposed
high-frequency hearing standards. It is now apparent that a technique is available for reliably
testing hearing for frequencies above 8000 Hz. With a technique available for valid and
reliable testing of high-frequency  hearing and with tentative norms or standards proposed,
efforts have begun to determine practical implications of high-frequency hearing.
    In an early study, Fletcher et al.  (1967) found that meningitis  patients who had been
categorized as seriously ill during  the course of the disease had significant losses of hearing
above 8000 Hz compared to those who  had not been seriously ill, while neither group had
noticeable losses of hearing for the conventional frequencies. These results suggested that
high-frequency hearing might possibly be  a sensitive index of possible trauma within the
cochlea. This hypothesis was supported by research conducted by Jacobson et al. (1969)
into the effects of ototoxic drugs on  high frequency hearing. In  a study begun before
chemotherapy on tuberculosis patients,  they  found that ototoxic drug effects upon high-

                                        271

-------
UJ
2  40


«  30
ca

55  20
UJ

~  10


s   o
   -20
	 Ist SESSION
	7n« SESSION

	3" SESSION
                                                            I
                                 9      10     11      12     13
                                 FREQUENCY  IN  KILOCYCLES
                                      14
15
16     II
                 Fig. 1.  Repeatability of high frequency thresholds over three sessions.


 frequency hearing were detected from  41-76 days earlier than effects could be detected
 within the conventional frequency range (Fig. 2). Hearing losses were most apparent first in
 the frequency range from 9-13  kHz. Kanamycin was the drug found most likely to have
 caused the hearing loss, although the patients were also receiving other drugs.
      The indications of the usefulness of high-frequency hearing for the early detection of
 ototoxic drug reactions, and of the sensitivity of high-frequency hearing as an indicator of
 damage from disease,  suggested that high-frequency hearing might also provide an early
 warning system of noise-induced hearing loss, or might differentiate  hearing levels at an
 early stage between populations exposed to various  non-occupational noises such as rock
 music, sport shooting, drag racing, etc.
      Conventional and high-frequency hearing of rock band members, rock spectators, sport
 shooters, drag racers, and motorcyclists was tested for a population of  18-21-year-old males
 and females. A normal or control population of 18-21-year-old males and females was also
 tested to  provide  a  basis for comparison (Fletcher,  1972). In this study (Figs.  3  and 4)
 motorcyclists were found to have incurred the greatest loss of hearing, followed by drag
 racers. Sport shooters had much less loss than expected, probably because most shooters are
                                          272

-------
 0
_C
"c
 *
 O
 v
 O
M
•o
 c
 c
 9
          O •
          10-
         70.
         30-
         4O-
         50-
         60-
         70-
 0 -
 10-
20-
30-
40-
50-
60-
7O-
                                                           C.E.
                                                   8-16-67
                    I      I      i      I      I       I
 0-
 10-
 20-
 30-
 40-
 50-
 6O-
 70-
                                                                     J.G.
                  1-2-68
                                 12-19  67
                                                               — 10-3-67
                           I      I
                                     I      I
                                                                     F.J.
                                                                     8.8.67
       - 1 - 1 - \
        46      8      9
                                               i - 1 - 1 - 1 - 1 - 1 - T
                                           10     11      12     13     M     15     16
                                        Frequency  (kHx)
 Figure 2. Changes over time in high frequency hearing of ototoxic drug patients (subjects C.E.. J.G., and
 F.J.).
                                            273

-------
                                      DRAG RACERS
   ic:

   ec

   t r


   7C

  . 6'
c
z
AGE
 ie •
 19
 20»
 21
                                ROCK BAND MEMBERS
   5C
          AGE
            13 ••
            19 -
            2C-
            21 •
                   2    3   4    6    8   9    10   II   12   13   14   15   16   18
                                     FREQUENCY IN KHZ

                       Figure 3. Percent responses by exposure category.
Acutely  aware of the hazard to hearing and utilize some form of ear protection while
engaged in shooting. Hearing losses among musicians in rock bands were surprisingly small
but not totally unexpected according to the author. He attributed the small losses exhibited
by the musicians to a sampling flaw, saying, "Many times, in trying to schedule known rock
band members or drag racers, we were told they were out of town playing an engagement,
or driving at some track, and repeated calls received the same answer. It would appear that
those at the upper level of experience and skill, if they desire, can spend a great deal of time
at this activity and  make reasonably good money. Therefore,  it could well be that our
sample is missing many,  if not most of those at this level.. .those  who would be expected to
have suffered the greatest exposure and therefore the largest losses".
     Another study presently underway to determine whether high-frequency hearing is a
useful early detector of noise-induced hearing loss involves aviators  (Fletcher, 1973).  The
                                        274

-------
                                      MOTORCYCLISTS
   ICC


   9C


   sc


   7C

o
Z.  6C
O
Z
o  50
V
a:
>-
Z
Ul
           AGE
            !8 	
            !<>	
            20 —-
            21 	
c
u
ICC




e:

7C

6C

50
                                     SPORTS SHOOTERS
           AGE
            18  	
            19  '—
            20
            21  —
                         3     4     6    8    9    10   II    I?    13   U   IS   16   18
                                         FREQUENCY  IN  KHZ


                       Figure 4.  Percent responses by exposure category.
 conventional and high-frequency hearing of a large number of professional pilots was tested.
 Pilots were found who, predominantly, flew jet aircraft, propeller-driven airplanes, or rotary
 wing (helicopter) craft. Also, a sufficiently large sample of each type of pilot was obtained
 to study hearing from entry into aviation training through up to about 9,000 hours flight
 time. Results of the  study to  date (Figs. 5-9,  inclusive) show earliest losses at  the higher
 frequencies, as expected, with  a gradual erosion of hearing with continued exposure, with
 lower and lower frequencies progressively becoming involved with continued exposure and a
 larger percent of the  S's  not hearing the higher  frequencies. In fact, percent of S's hearing a
 frequency appeared a more sensitive index of exposure than average hearing level.
     Propeller-driven  planes  appeared  to  cause the greatest loss of  hearing, followed by
 helicopters,  with jets the least  hazardous (Fig.  9). The results for  percent response, by
 comparison, showed a drop beginning at 4 kHz for the prop  pilots (Fig. 5), whereas for jet
                                          275

-------
   too

   »5

   90

   Bi

   80


   7S
o
5;  /o
w

-------
u
w
_l
u
VI
a
M
O
2  40
W
K
Z
                                               10   11    12    13   U   IS    16   18
                                   FREQUENCY IN KHZ
                      Figure 6.  Mean hearing levels by flight time—prop.
both sides  of the shooter randomly.  The  non-noise-exposed male population  used for
comparison had hearing levels an average of about 26 dB better at the frequencies 10, 12,
14, 16, and 18 kHz than  those for the boys' rifle team members. For the females, the girls'
rifle team member average hearing levels were some 17 dB higher (less sensitive) than those
of the levels found for the non-exposed females.
    These  data  do demonstrate rather  clearly the usefulness of tests of high-frequency
hearing for the early detection of at least some types of noise-induced hearing loss.
    In summary,  data have been  presented suggesting the usefulness of high-frequency
hearing testing in early detection of not only noise-induced hearing losses, but also ototoxic
responses, and losses attendant upon certain types of illness. Tests of high-frequency hearing
have been shown to be reliable, equipment for  such testing is now commercially available,
techniques for testing are no more complex than those of conventional tests, nor is the
required testing environment as quiet as that necessary for conventional testing. Therefore,
it is recommended that serious consideration be given to use of high-frequency hearing tests
in the early detection of noise-induced hearing loss.
                                         277

-------
    100
    90
    85
    80
o
z

S
z
o

Si   70
UJ
tE
    65
vj
«E
60


55


50


45
                AVG. MRS.
    35
                                                       10     II    12    13    U    15    16    IB
                      Figure 7.  Percent responses by frequency and hours—Jet
                                              278

-------
  ul
  z
  VI
  <
  a
  a)
  •a
  1/1
  a
  VI
  hi
                                 6    6     9    10    II   12   U
                                   FREQUENCY IN KHZ
                                                                  U
16
                      Figure 8. Mean hearing levels by flight time—jet.
                                    References

1.   Fletcher, John  L. Reliability  of high-frequency  thresholds. /. Aud. Res., 1965, 5,
    133-137.
2.   Zislis, T. and Fletcher, John L. Relation of high frequency thresholds to age and sex. /
    Aud. Res., 1966,6, 189-198.
3.   Northern, Jerry L., Downs, Marion P., Rudmose, Wayne,  Glorig, Aram, and Fletcher,
    John  L. Recommended high-frequency audiometric threshold levels (8,000-18,000
    Hz)./. Acoust. Soc. Amer., 1972,52, 585-595.
4.   Fletcher, John  L., Cairns,  Adrian B.,  Collins,  Fred G., and Endicott, James. High
    frequency hearing following meningitis. /. Aud, Res., 1967, 7, 223-227.
5.   Jacobson, Edward J., Downs, Marion P., and Fletcher, John L. Clinical findings in high
    frequency  thresholds during known ototoxic drug usage. US Army Med.  Res. Lab.,
    Rep. 820, Mar. 25, 1969.
                                        279

-------
Ul
<
at
    -5


     0




    10




    JO
 m  30
 o"
 X
 in
 ui
 IE
    40
    50
    60
                     HOUttS  N
               JET VI   U?2  \7
               HELO VI  Jlftt  22
               PROP VI  4!7I  29
          .5    1    2    3    t    6    8    9    10   II    12   13   14   15   16   18
                                   FREQUENCY  IN KHZ


          Figure 9. Selected comparisons of mean hearing levels by aircraft type & flight time.
6.   Fletcher, John L. Effects of non-occupational noise exposure on a young adult popula-
     tion. Final Rep., Contract HSM-099-71-52, NIOSH, Dept. HEW, October 13, 1972.
7.   Fletcher, John L. Conventional and high frequency hearing of naval aviators. Office of
     Naval Res., Contract N00014-71-C-0354, Progress Rep., February, 1973.
8.   Gasaway, Don. Personal communication, 1972.
9.   Simonton, Jane K. Results of high frequency tests of members of Denver public
     schools high school rifle teams during Dec., 1972. Personal communication,  1972.
                                        280

-------
                        SUSCEPTIBILITY TO TTS AND PTS*

                                   W. Dixon Ward

        Department of Otolaryngology; Department of Communication Disorders
                               University of Minnesota
                               Minneapolis, Minnesota

    More  than 140  years  ago,  John  Fosbroke  noted  that  people differed in their
susceptibility to hearing loss,  as they did in so many other ways, and speculated about the
underlying constitutional factors responsible for these differences. About a century later, it
occurred to Jakob Temkin that it might not be necessary to determine individual differences
in how well sound was conducted to the inner ear, in the elasticity of cochlear structures, in
blood circulation, etc., or to institute monitoring audiometry in noisy industries (which he
did favor, however), in order  to forestall hearing loss. Instead, one could merely expose the
ears of all the workers to a moderate  sound, and measure the auditory fatigue produced.
The ear showing the greatest effect  surely should be the most susceptible  to permanent
hearing loss. This  idea was developed by Alfred Peyser, who proposed the first formal
susceptibility test: the change in the  amount of time a 250-Hz tuning fork could be heard
immediately after half an  hour of exposure to the noise from a  "Klappermaschine aus
Metall"  (1930). In  the intervening period, literally millions of man-hours have been
expended in a test  of Temkin's hypothesis, as exemplified in at least 20 different proposed
susceptibility tests  (reviewed  by Ward,  1965). I am indeed sorry that Dr. Temkin, who is
still active in the field of noise in Moscow, is unable to attend the Congress and participate
in a discussion of the present status of his idea.
    As we  all know, the relation between auditory fatigue, or temporary threshold shift
(TTS), and permanent threshold shift  (PTS) did  not turn out to be  as simple  as Temkin
hoped. Already 24 years ago, when Walter Rosenblith got Ira Hirsh and myself interested in
TTS, Theilgaard (1949) and Greisen (1951) had shown that since there was little correlation
between TTSs produced by pure tones of different  frequencies, susceptibility could hardly
be a unitary function. The same independence of susceptibilities was implied by Theilgaard's
(1951) finding that there was no  consistent correlation between the TTS produced by a
1500-Hz pure tone and the magnitude of the hearing loss at 4 kHz in a group  of 59 weavers.
And of course Flugel had shown in 1920 that the two ears of a given individual differed in
fatigability.
    However,  there  was  and is  no  question  that there  were large  differences in
susceptibility to PTS,  as Borge Larsen (1952) pointed out while discussing the implications
of Theilgaard and Greisen's work. He  cited two persons tested by him and  found to have
normal hearing, despite employment histories of 15 years as a boilermaker or  14 as a riveter,
respectively.  The same point was made a few years later by Shapiro (1956), who found a
normal-hearing drop-forge operator with 28 years of experience. Clearly, such workers are
unusually resistant.
 •Preparation of this manuscript was supported by Grant NS-04403 from the Public Health Service, U.S.
 Dept. of Health, Education and Welfare.

                                         281

-------
     The reverse, however, does not follow—i.e., that an individual who has a hearing loss is
 necessarily more susceptible than the average man, although this assumption is sometimes
 made (e.g. Harris, 1965). He might be more susceptible, but he might also have been merely
 more unlucky in  being over-exposed  on a  particular unusual  day-that  is, in having
 experienced  a  noise dose that would give anyone a large PTS. In any actual situation, the
 group of workers with elevated Hearing Levels will include both the susceptible and the
 unlucky.
     This fact causes all sorts of trouble when one tries to validate a susceptibility test based
 on ITS  with a cross-sectional study of a group of  men who have been working in noise for
 many years and who have a wide range of hearing losses. One can, of course, give these men
 the same susceptibility test, measure the TTSs produced, and then determine the correlation
 between these  TTSs and the existing Hearing Level (HL). However, a source of bias exists if
 this procedure is followed. Men with hearing loss at the frequency whose shift is taken as a
 susceptibility index will, on the average, show less  ITS than normals, and so there will be a
 significant negative correlation even  if  the susceptibility test per se is worthless. In extreme
 cases, this is intuitively obvious. When  the loss is due to conductive factors, then of course
 the effective level of the sound reaching the cochlea will be reduced and so less effect would
 be expected. On the other hand, when the loss represents a sensorineural deficit, then there is
 less shift possible.
     Figure  1 shows the curve relating average TTS at 4  kHz to the resting HL of some
 seasoned workers in a planet with uniform levels throughout the working area of about 100
 dB(A).  It can be seen that the average TTS decreases with HL in a linear fashion, with an
 intercept at about 80 dB HL.
     The individual results of three workers are also shown. Workers A and B have resting
 HLs of  10 and 50 dB, respectively, but  both show TTSs of 20 dB." Worker C, with a
 threshold of 40 dB HL shows a TTS of 10 dB.
     It is clear that despite the equivalence of TTSs, worker B should be considered to  show
 a greater "effect" than A; he displays considerably more TTS than the average man with a
 50-db loss, while A shows less than the average of his group. What is not so clear is whether
 C  is more susceptible than A, or vice versa. Both show a TTS that is  5 dB less than the
 average for their respective  HL groups. However,  one might say,  "Yes, but C shows only
 67% as much TTS as the average (10  vs.  15), while A shows 80% as  much (20 vs. 25).
 Therefore A is the  more susceptible." This is supported by considering also the variability
 involved. The variance of TTSs for HLs of 40 dB will be smaller than for HLs of 10 dB, so
 the 5-dB departure  from the mean does indeed imply that worker C is "farther"-i.e.,  more
 standard deviations-below the average line than is A.
     This is the procedure that was used by Bums et al. (1970) in their recent study of 218
workers, one of the few recent results that offers  much encouragement for susceptibility
tests. Not only were all individual TTSs-in their case, the  TTSs produced by the workers'
own normal working day—converted to standard scores based on the use of average TTSs and
their respective variances, but HLs themselves were also normalized to scores reflecting how
the individual's HL compared to  the HLs of men  of the  same  age and cumulative noise
exposure; that  is, corrected  for both average noise-induced PTS and for presbyacusis plus
sociacusis. With this procedure, they were at  least  able  to'show a correlation of 0.34

                                        282

-------
     LJ —
     1-0
     Lu Z
        LJ
        o:
2
    HO:
           '°
                                                                 I
                   0         20         40
                    HEARING  LEVEL  IN
                                             60          80
                                             DB  (ISO)
Figure 1. Dependence of ITS on the resting threshold. The noise had a sound level of about 100 dBA
Crosses indicate hypothetical results from three workers (see text).
between the normalized TTS averaged for 1 and 2 kHz and the normalized HL averaged over
3,4 and 6 kHz.
    Correlation coefficients of about one-third or smaller seem to be the rule, being about
the same as that found by Jerger and Carhart  (1955), in  their study  of the changes in
hearing suffered by  178  airmen at a  school for jet mechanics during a 10-week training
course. The time to recover to within 20  or 10 dB of original threshold at 4500 Hz,
respectively, after a 1-min exposure  to a 3000-Hz tone at 100 dB SPL, was correlated
against the  average shift in HI at the end of the course (presumably PTS). The correlation
coefficient was 0.36 for the recovery-to-20-dB-TTS criterion, though only 0.23 for recovery
to 10 dB.
    These results are about what one would expect on the  basis of the extensive study of
the intercorrelation among different types of susceptibility indices done in our laboratory
several years ago (Ward, 1965, 1968). In these experiments, 49 college students with normal
hearing, 24 men  and 25 women, were exposed  to a host of different short susceptibility
tests: TTSs from exposure to pure tones and octave-band noises for various times and at
various levels, measurement of the intensity of noises and of simulated gunfire  required to
produce a given TTS, perstimulatory adaptation, and contralateral remote masking were
measured, using earphones or free-field exposure, in one ear or two.
                                       283

-------
     The test-retest correlation was about 0.65, even when there was a 6-month intervening
period; this increased to 0.77 if the average of TTSs at the three frequencies most affected
were used instead of only the frequency at which the maximum TTS occurred. Correlations
among  TTSs  from different tests involving exposure in the same frequency range had a
median value of about 0.55. The  correlation between TTSs produced by different ranges of
frequency was smaller,  being statistically insignificant (less than  0.3) even for two 3-min
exposures to 1000-Hz octave-band noise at 120 dB SPLand to a 2000-Hz octave band noise at
116 dB, for example. Between the former and the TTS from a 15-min exposure to 500-Hz noise
at  120 dB SPL, however, the correlation was 0.5. Factor analysis  of the correlation matrix
implied that there is a common thread of "general susceptibility" to auditory stimulation,
but that this  would account  for only about a third of the communality  in  the matrix.
Varimax analysis indicated that one should really speak of susceptibilities to TTS-to low-,
medium- or high-frequency noises or tones,  or to impulse noise.  However, exposure to
broad-band  noise did tend to produce TTSs that agreed well with  those produced by the
appropriate single octave band presented alone, so it appeared that the best course of action
would  be  to  use broad-band noise as the fatiguer, but  measuring  TTS  at  the  various
relatively-independent frequency ranges.
     It appears, then, that a quick test that will reliably measure all aspects of susceptibility
even to TTS is not  at hand, much less one that produces an effect-either the TTS or the
recovery time—that is a valid predictor of eventual PTS. However, it is clear that some of the
validation studies so far attempted, such as those of Jerger and Carhart (1955) and Burns et
al. (1970),  suffer from the fact that  the true noise exposure  was  only estimated or was
assumed to  be the same for all workers. Other attempts at validation have another problem
as well—namely, that PTSs may be produced in so few ears that the correlation between the
results of a  susceptibility test given at the beginning of employment and the change in HL is
meaningless. For example,  Sataloff et  al. (1965) in 1951 gave a 2-kHz 95-dB-HL (ASA) test
to 105 Ss,  and then in J 962  examined the hearing of the 33 still at this jet-engine test
facility. Unfortunately,  no important PTSs had been produced in this time, so the absence
of a significant correlation proves nothing about the test per se.
     There  is  clearly only  one socially-acceptable solution to the lack of control over the
noise exposure of the test subjects in  all tests of PTS involving humans, and that is to use
experimental animals whose auditory history is completely known. Toward this end, we
have for the past few years been conducting studies using chinchillas. They are trained, using
a conditioned  shock-avoidance technique, to jump across a barrier when they hear a sound.
Temporary  and permanent threshold shifts thereby can be measured. However, We find that
if our animals all come from  breeders who maintain a quiet environment,  the differences
among them in resting thresholds are so small that they barely reach statistical significance
even after weeks of  testing. Furthermore, when the exposures are done, as ours have been,
by restraining them  in a head-holding device in a fixed position in front of a loudspeaker,
which gives  good control of the exact noise dose they receive, the amount of TTSs and PTSs
produced are  also nearly  the same.  While occasionally there is an animal who shows a
significantly greater TTS  than the rest; a repetition of the exposure usually will  fail to
substantiate his indicated higher susceptibility. Similarly, the occasional animal who shows
less PTS than the average has always showed a normal (in this case, average) TTS.

                                        284

-------
    The results just cited are of course subject to the limitation that we are at the moment
still not sure that we did not have an artifact of some sort in our testing situation. Animals
who had been given TTSs of 70 to 80 dB apparently recovered completely in about two
weeks, yet histological examination showed extensive, even complete destruction of the hair
cells in the basal turn of the cochlea (Ward and Duvall,  1971). Furthermore, the degree of
destruction could  be  quite different for two animals that showed the same TTS. These
results are difficult to explain. We are only now beginning a set of experiments to confirm
or deny the accuracy of these observations, our laboratory having been inoperative for the
last 6 months because it was necessary to move it.
    In view of the fact that the exposure that produced equal TTS and PTS in all animals
did not produce the same histological damage, we rather expect to find that individual
differences in TTS and PTS are not as small as thus far indicated, so that we can proceed to
study individual susceptibility. Even if we do not, however,-if, for example, the histological
differences are artifactual instead-this  will not  be particularly disheartening. If these
animals, kept free of sociacusic influences such as avocational noise, blows to the head, and
middle-ear infection, really all do have the same susceptibility, at least it  will not require a
very large number of animals in order to establish the group relations between TTS and PTS,
especially in, regard to intermittent noise, which is our other chief area  of activity at the
moment.
    Other laboratories have not had much success with different animals either. Herman
and Clack (1963) got a zero correlation between TTS and PTS from white noise in the rat,
and Luz et al.  (1971)  had a similar outcome exposing monkeys to high-intensity impulse
stimuli. However, it must be noted that in the Luz et al. study, test-retest correlation for
TTS was significantly negative, so one  can hardly expect any correlation with anything else
to be significant.
    In man, only a few recent papers  claim to have established a positive relation between
differences in TTS and PTS. Sulkowski  (1969) claims  that in a*sj;udy of 127 beginning
workers in a textile mill, a combination of a Peyser-type test (4 min  of 4000 Hz at 90 dB
HL) and a tone-decay test predicted the degree of hearing losses developing in the next two
years,  Strubinski (1970)  also  reports success with  such a  combination in forecasting
hearing-loss development  in 33 diabetics. Pfander (1968) exposed  100  recruits to three
susceptibility tests, two involving white noise,  the third being exposure to five shots from an
ordinary military weapon (161 dB peak). He  indicates that five soldiers who showed TTSs
from the rifle shots that required 3 to  6 days for full recovery had permanent losses at the
end of their training. However, it is not made clear in the article how many of the men
originally had high values of HL to begin with, how many dB constituted a "significant"
loss, or even whether or not these 5 were the only cases showing permanent shift, so some
uncertainty still exists.
    No one  seems to have followed up a study of telephone operators by Kuroyanagj
(1960). Despite the fact that there are consistent differences in susceptibility between the
two ears, the median correlation between ears for TTS is 0.63 (.Ward, i 968), which indicates
that the two ears of a given observer are generally quite similar. If, then,  one ear is always
used for telephone listening, and that ear is exposed to intense noises (.or clicks and buzzes,
in the case at hand), then the more susceptible will end up with a hearing loss, but in only

                                        285

-------
one ear. Kuroyanogi found that in 914 telephone operators, 18% had "some" loss in the ear
used. It is indicated that those with such loss showed more TTS than the average on the
normal side.  Unfortunately  for science,  though not for operators, hearing losses are no
longer being produced by telephone noises (Glorig et al., 1969), so this source of possible
validation of susceptibility tests may no longer exist.
    If the relation between susceptibility to TTS and the susceptibility to PTS seems
uncertain, it is no more so that the question of what it is that determines either one. Many
articles have been  written  about the relation of mastoid pneumatization  and PTS in
highly-exposed  workers, for example. Link  and Handl (1955) and Ceypek et al. (1956)
found a statistically significant though not large indication that well-pneumatized ears were
less susceptible  than poorly-pneumatized ones. Kosa and Lampe (1967), however, found no
correlation whatever, and Kubo, who also concluded that pneumatization was a factor of no
importance,  actually got data that show conclusively that men with the most  pneumatiza-
tion had  more loss than those with the least. I tend to believe that Kosa and Lampe have hit
upon the truth.
    Another equivocal area, although this always seems odd, is in the area of middle-ear
problems. The  natural inclination is to believe  that anything that interferes with the
conduction of sound must act as a protection and thus reduce susceptibility to both  TTS
and PTS. Although this does seem to be true of ears with ordinary otosclerosis (Gerth,
1966), even after stapedectomy (Fletcher and King, 1963), other middle-ear problems such
as otitis  media do not seem to reduce TTS or PTS to a degree commensurate with the
conductive deficit that can be measured. Results reported by Paparella et al. (1,970) on 279
ears with otitis media actually led them to the conclusion that otitis media can cause
sensorineural loss by somehow invading the cochlea; however, it may merely be that this
condition has its effect by changing the susceptibility to noise damage rather than by direct
action. The  question of the effect of middle-ear pathology on susceptibility  is still open.
However, I   am  glad tto  report that no  one has yet  challenged Kristensen's (1946)
demonstration that there is no relation between hearing loss and body type.
    The role of the middle-ear muscles  has also received considerable study; inoperative
middle-ear muscles would be expected to produce an ear that was unusually susceptible to
low-frequency stimulation, particularly to intermittent sounds. Such ears are occasionally
found; in; one  that displayed the symptoms just cited, impedance measurements indeed
implied that the reflex was inoperative (Ward, 1962). In a study of 40 Marines who had just
completed training, Coles and Knight (1965) found that men with "poor high-tone hearing"
had a higher reflex threshold than those with "good" or "fair". So in this field the evidence
seems to  point at least in the  same—and, happily, the expected—direction.
    Theoretically, individual differences  in susceptibility to acoustic stimulation having a
particular frequency can be ascribed to a near-infinite number of parameters, not only those
dealing with transmission of sound to the cochlea but also with inferred characteristics of
the cochlea itself such as blood circulation, thickness of the various membranes, etc. A most
unusual correlation has recently been reported by Tota and Bocci (1967)j for example, who
report that bhie-eyed persons showed twice as much TTS as brown-eyed ones (27 dB vs. 13
dB) following a 3-min exposure to a  1000-Hz tone at  100 dB  (HL, presumably), and so
conclude that melanin plays an important role in protecting the ear from hypoxia.

                                       286

-------
     There seems to be a consistent relation between individual differences in the degree of
vasoeonstriction  caused by  a noise and  the resultant TTS, persons with the  smaller
vasoeonstriction  showing  the greater  TTS,  provided  that  one  uses as  the  index of
vasoeonstriction the value observed 20 sec after onset of the noise, not at the end of a long
exposure (Jansen, 1970). This may account for the fact that Oppliger et al. (196) found no
consistent relation. It must  be pointed  out,  however, that this correlation  does not
necessarily mean that vasoeonstriction is causing a greater TTS; if the vasoconstrictive effect
depends on the loudness of the sound producing it,  then in those persons for whom more
energy is reaching  the inner ear,  the sound  will appear louder, hence produce  more
vasoeonstriction, and will also produce more TTS, whether or not the vasoeonstriction has
anything to do with, say, the accumulation of fatigue products.
     Both eye color and vasoeonstriction should be investigated further. There continues to
be no convincing evidence that ears ever get "tough"-more resistant to damage-because of
habitual exposure, or that younger or older persons are more susceptible than young adults,
or even that any sort  of medication can decrease susceptibility to TTS and  PTS, although
there is no doubt that it can be increased by the administration of certain ototoxic drugs,
even though the doses of drug are by themselves subtoxic. An unreported study by John
Park in our laboratory found no differences in TTSs from noise in the chinchilla to be
caused by injection of hydergin or adenosine triphosphate (cf. Plester,  1953; Faltynekand
Vesely, 1964). Dextran, which is reported by Kellerhals (1972) to reduce TTS, also was
without  effect;  however, in  this  case  the  dosage  turned  out  to be, less than  that
recommended by Kellernals, so additional experiments are planned.
     The concept of a "critical intensity" for a given ear seems to have  died a natural death
as it became clear than an intensity  level that was "critical" for one duration of noise was
not "critical" for a different duration. The notion of a "critical energy"—a measure of noise
exposure, not just  of noise—may have some merit, but if it merely denotes the  "breaking
point" above which permanent damage will result, it is not usually worth looking for.
     On  the other  hand, there are characteristics of noise as such whose perception varies
from individual to individual.  Perhaps determination  of the LDL (loudness discomfort level)
is worth  intensive study hi regard to susceptibility (Hood, 1968)—that  is, provided we can
agree on  the exact instructions to the listener,  in this rather instruction-sensitive task.
Perhaps someone will be stimulated to  extend some results reported by van Dishoeck and
Spoor  (1958)  15 years ago, who claim to have gotten a "good criterion of the  individual
sensitivity of noise" by asking subjects  to "indicate  at which intensity a pure tone acquires
an impure and sharp character".
     Nowadays,  the possibility must be kept in mind  that perhaps PTS may not be the
correct validating  index for susceptibility in  the  first place.  It  may be that auditory
sensitivity is not much affected  until all of the hair cells in  a certain area of the basilar
membrane are destroyed (Eldredge  and Miller,  1969; Ward and  Duvall, 1971). If, then, a
given exposure destroys only  a few hair cells, then the full recovery of the TTS that ensues
will incorrectly imply that the exposure was innocuous. It may be that in man a much lower
TTS should be permitted than we now  deem to be safe; however, in the chinchilla a 40-dB
shift—at least if caused by a short exposure—is known to be safe. Figure 2 shows the course
of recovery from TTS in two groups of chinchillas. One group of 5 animals was exposed to

                                        287

-------
 114 dB SPL of 700-2800-Hz noise for 10 min, the other for 20 min. As can be seen, full
 audiometric recovery occurred after one day in the 10-min group, but required 4 days in the
 other. Inspection of the cochleas showed no missing hair cells in the 10-min group, but a few
 (3 to 7) hi each ear of the 20-rnin ones. We have not yet run the experiments to answer the
 question of whether or not  a limit of 40  dB of TTS is also safe if it was produced by a
 longer or an intermittent exposure. In view of the longer persistence of such TTSs (Ward,
 1970), it is certainly not to be taken for granted that this will be the case.
     On the other hand, perhaps hair-cell destruction is itself no better a validating criterion.
 Both Elliott (1961) and Hunter-Duvar (1971), among others, have produced  tonal gaps in
 their experimental animals (cats and monkeys, respectively), yet upon examination the hair
 cells all appeared normal.
     We are truly in a difficult position at the moment. We cannot rule out the possibility
 that what are  clearly innocuous exposures to noise, in the sense that recovery from any
 measurable effects is complete within a  few hours and that no cumulative effects can be
.seen over  a period of many weeks, may nevertheless be producing latent damage—changes
 that are unobservable in the intact organism.  It is not enough, in this case, to argue that a
 difference that makes no difference is not a difference. If noise exposures that cause no
 change in  auditory function are nevertheless gnawing away at nature's safety factor, then
 the  fact that no functional deficit is observed is not completely relevant. There are many
 who support such a cautious viewpoint.                                       >
     However, if this were the case, then one would expect that years of work  in a noise
 that is slowly destroying hair cells one at a time—the essence of the "microtrauma" theory
 of Gravendeel  and Plomp(I960)—would suddenly produce severe impairment of sensitivity,
 as the last few hair cells in a given area finally succumbed. I know of no evidence for such
 sudden growth of impairment in men working for a long time in uniform noise; on the other
 hand, I  doubt that anyone has looked very  carefully for it. Nevertheless, the burden of
 proof,  in  my opinion, still rests  on those  who  assume such  latent  (or  residual but
 immeasurable) effects  to  be occurring. Otherwise we may be forced to adopt absurd
 damage-risk  criteria, as for example that proposed, apparently in  all seriousness,  by
 GePtischcheva and Ponomarenko (1968) in Russia: a measurable TTS at any frequency!
 Because a  1-hr exposure to a 500-Hz octave band of noise at 75  dB (SPL, presumably), to a
 1-kHz  octave  band  at 65  dB, or  to  a  4-kHz  octave band  at 60  dB  produced in
 15-to-16-year-olds  a  3-dB TTS measured  within the first minute after exposure, they
 propose that adolescents be protected from anything higher than these levels.
     At the  moment, therefore, I shall continue to conduct my research as if it were true
 that if recovery  from TTS induced by a daily exposure is complete before the next day's
 exposure begins, no hazard to hearing exists. I hope this is correct.
                                        288

-------
    70
    60
    50
 g  40
 ^  30

 £20
 H
    10


    0
       EXPOSURE:
700-2800-HZ NOISE
                                                    0 M1N
                             1
           I
I
I    I   I   I   I
               IH           4H     8H         ID      2D
                     TIME   SINCE   EXPOSURE
                                     4D    7D
Figure 2. Recovery of chinchillas exposed only once to noise: either a 10- or a 20-minute exposure to
700-2800-Hz noise. The 20-minute group showed a few missing hair cells; the 10-minute group did not
                                References

Burns,  W.,  Stead, J.  C.  and  Penney,  H.  W., The  relations between temporary
    threshold shift and occupational hearing loss. In: Hearing and Noise in Industry,
    W. Burns  and D. W. Robinson,  Eds., Her Majesty's  Stationery Office,  London,
    183-210 (1970).
Ceypek, T., Lepkowski, A. and Szymczyk, K., Sensitivity  to  acoustic traumas and
    pneumatisation of  the mastoid, Otolaryng. Pol 10, 329-334 (1956) ref. ZHNO
    57, 219 (1957).
                                    289

-------
Coles,  R.  R.  A.  and  Knight,  J.  J., The  problem of  noise  in  the Royal  Navy  and
     Royal Marines.,/. Laryng. Otol,  79,  131-147 (1965).
Dishoeck, H. A. E.  van  and  Spoor, A., Auditory fatigue and occupational deafness.,
     Laryngoscope, 68, 645-659 (1958).
Eldredge, D. H. and Miller, J. D.,  Acceptable noise  exposures—damage risk  criteria.
     In:  Noise  as  a Public Health Hazard; ASHA Reports  No. 4,  W. D.  Ward  and
     Fricke,  J. E.,  Eds. Am.  Speech Hearing Assoc., Washington, D.C.
Elliot,  D. N.,  The  effect of  sensorineural  lesions on pitch discrimination  in  cats,
     Ann. Otol. Rhinol. Laryngol., 70, 582-598 (1961).
Faltynek, L.,  and  Vesely,  C.,  Zur Restitution  der  Mikrophonpotentiale  des  Meer-
     schweinchens nach Kurzfristiger Larmbelastung, Arch.  Ohren— Nasen— u. Kehlk-
     opfheilk, ver. Z. Hals-,  Nasen- u.  Ohrenheilk, 184,  109-114 (1964).
Fletcher,  J.  K. and  King,  W. P.,  Susceptibility of stapedectomized patients  to noise
     induced  temporary threshold  shifts, Ann.  Otol. Rhinol. Laryngol, 72,  900-907
     (1963).
Flugel, J. C., On  local fatigue in  the auditory system., Brit.  J.  Psychol., 11,  103-134
     (1920).
Fosbroke, J., Practical observations the pathology and  treatment of deafness. Lancet,
     1,  645-648 (1830, 1831).
Gel'tishcheva, E. A. and Ponomarenko, I. I., Noise standards for adolescents, Hygiene and
     Sanitation, 33, 199-203(1968).
Gerth, B. Otosklerose bei  Larmarbeitern. HNO (Berlin), 14,  205-208 (1966).
Glorig, A.,  Hearing studies of telephone operating personnel, /. Speech Hearing Res.,
     12,  169-178 (1969).
Gravendeel,   D.  W. and Plomp, R., Micro-noise trama? AMA Arch. Otolaryngol,  71,
     656^663 (1960).
Greisen,   L.   Comparative   investigations  of  different  auditory  fatigue  tests.  Acta
     OtO'Laryngol, 39, 132-135 (1951).
Harris,  J.  D.,  Hearing-loss  trend  curves  and  the  damage-risk  criterion  in diesel-
     engineeroom personnel, /. Acoust. Soc. Am., 37, 444-452  (1965).
Herman,  P.  N. and Clack, T. D.,  Post-stimulatory shift of the Preyer  reflex threshold
     in the  white  rat, /. Aud. Res.,  3,  189-204 (1963).
Hood,  J. D.,  Observations  upon  the  relationship of  loudness  discomfort level  and
     auditory fatigue to sound-pressure level and  sensation level, /. Acoust. Soc. Am.
     44,  959-964 (1968).                                               r
Hunter-Duvar, I. (Henry Ford), unpublished thesis, Wayne State U., (1971).
Jansen,   G.,  Relation  between temporary  threshold shift and  peripheral  circulatory
     effects   of  sound, In: Physiological Effects  of Noise,   B.  L. Welch  and  A.  S.
     Welch,  Eds., Plenum,  New York,  67-74 (1970).
Jerger,  J.  F.  and  Carhart,  R;i  Temporary  threshold shift as  an  index of noise
     susceptibility, /. Acoust. Soc.  Am.,  28, 611-613 (1956).
Kellerhals,  B.,   Acoustic   trauma   and  cochlear microcirculation, Adv.  Oto-Rhino
     Laryng.,  18, 91-168  (1972).
                                        290

-------
Kristensen, H.  K.,  On the  relation  between  the hearing of  weavers  and the body
    type, Acta Oto-Laryngol,  34,  82-94 (1946).
Kosa,  D.  and   Lampe,   I.,  Uber   die   Pneumatisation  des  Warzenfortsatzes  bei
    Larmarbeitern,  HNO  (Berlin)  15,  324-325 (1967).
Kubo,  M.,  Acoustic  trauma-its  development  and  prevention  in  transport workers,
    Proc.  16th   General  Assembly of the Japan  Medical  Congress,  Osaka,  (April
     1963).
Kuroyanagi, S., A contribution of  the  so-called telephone-deafness, J. Oto-rhinolaryng.
    Soc. Jap.. 63,  1438-1448  (1960),  ref ZNHO 69, 242  (1960/61).
Larsen,  B., Occupational deafness, Acta  Oto-Laryngol, 41,  139-157  (1952).
Link,  R.  and  Handl,  K.,  Die  Pneumatisation  des  Schafenbeins,  ein Schutz  gegen
    Larmschwerhorigkeit,  Arch. Ohr. Nas. Kehlkopfheilk,  167,  610-613 (1955).
Luz, G. A.,  Fletcher, J.  L.,  Fravel,  W.  J,  and  Mosko,  J. D.,  The  relationship
    between  temporary  threshold shift  and  permanent  threshold shift  in  rhesus
    monkeys  exposed  to  impulse noise,  US  AMRL  Report  No.  928,  U.S.  Army
    Med. Res. Lab.,  Fort Knox, Ky.  April  (1971).
Oppliger, G. C., Schulthess, G. V. and Grandjean, E., Uber die Beziehung der Gehorermu-
    dung zur bleibenden traumatischen Schwerhorigkeit, Acta Oto-Laryngol, 52,  415-428
    (1960).
Paparella,, M. M.,  Brady, D.  R.  and  Hoel, R., Sensori-neural  hearing loss in chronic otitis
    media and mastoiditis, Trans. Am. Ophthalm. OtolaryngoL, 74, 108-115 (1970).
Peyser, A., Gesundheitswesen u. Krankenfursorge. Theoretische und experimentelle  grund-
    lagen des personlichen schallschutzes. Deutsche  med.  Wochenschrift 56,  150-151
    (1930).
Pfander, F., Hone der  temporaren Schwellenabwanderung (TTS) in Audiogram und "Ruck-
    wanderungszeit"  gerausch- und  knallbelasteter  Ohren als Test  knallgefahrdeter
    Hororgane.^rc/i.  Klin. Exp. Ohren-, Nasen- Kehlkopfhlk. 191, 586-590 (1968).
Hester,  D., Der  Einfluss vegetativ wirksamer  Pharmaka  auf  die  Adaptation  bezw.
    Horermudung. Arch. Ohren-, Nasen- Kehlkopfheilk, 162, 473-487 (1953).
Sataloff, J., Vassallo,  L.,  and Menduke, H.,  Temporary  and permanent  hearing loss. A
    ten-year follow-up, Arch. Env. Health 10, 67-70 (1965).
Shapiro, S. L., Deafness following short-term exposure to industrial noise, Ann. Otol.  etc.
    68, 1170-1181 (1959).
Strubinski,  A., The  significance of the  tests of hearing adaptation in diabetes, Otolaryng.
    Pol 24, 311-318(1970).
Sulkowski, W., Badania nad przydatnoscia wybranych prob zmeczenia i adaptacji sluchu w
    profilaktyce  przemyslowych urazow akustycznych. Research on the value of certain
    hearing fatigue and hearing adaptation tests in the prevention of occupational deafness,
    Medycyna pracy (Warsaw),  20, 354-368 (1969).
Temkin, J., Die Schadigung des Onres durch Larm und Erschutterung, Mschr. Ohrenheilk. u.
    Laryngo-Rhinologie 67,  257-299, 450-457, 527-553,705-736, 823-834 (1933).
Theilgaard, E., Testing of the organ  of hearing with special reference to noise prophylaxis,
    Acta OtolaryngoL 37,  347-354 (1949).
                                        291

-------
Theilgaard,  E., Investigations in auditory fatigue in individuals with normal hearing and in
     noise workers (weavers)., Acta Oto-Laryngol. 39, 527-537 (1951).
Tota, G.  and Bocci, G.,  The importance  of the  color of the iris in the evaluation  of
     resistance to auditory fatigue (in Italian), Rivista Oto-Neuro-Oftalmologica (Bologna,
     Italy) 42, 183-193(1967).
Ward,  W. D.,  Studies on  the aural reflex, HI. Reflex latency as inferred from reductions  of
     temporary threshold shift from impulses, /. Acoust. Soc. Am. 34,  1132-1137 (1962).
Ward,  W. D., The  concept of susceptibility to hearing loss, /. Occup. Med 7, 595-607
     (1965).
Ward, W.  D., Susceptibility to auditory  fatigue. In: Contributions to Sensory Physiology, 3,
     W. D. Neff, Ed., Academic Press, New York (1968).
Ward,  W. D., Temporary threshold  shift and damage-risk criteria for intermittent noise
     exposures, /. Acoust, Soc. Am. 48, 561-574 (1970).
Ward,  W. D. and Duvall, A. J. HI, Behavioral and ultrastructural correlates of acoustic
     trauma, Ann. Otol. Rhinol Laryngol. 80, 881-896 (1971).
                                        292

-------
     GROWTH OF ITS AND COURSE OF RECOVERY FOR DIFFERENT NOISES;
                       IMPLICATIONS FOR GROWTH OF PTS

                                 Wolfgang KRAAK
              Technical University of Dresden, Sektion Informationstechn.
                          8027 Dresden, Germ. Dem. Rep.
    Temporary threshold shift (TTS) is often taken as a measure of noise effects that are
detrimental to hearing. Here it is mainly the TTS2 - i.e., the temporary threshold shift 2 min
after noise  exposure - which serves as criterion for designating the accumulating stress on
hearing. But there are various severe  objections to such a procedure. Noise exposure has
shown that after prolonged periods the TTS2 approaches a limiting value. The reduction of
the TT$2, the course  of recovery, depends on how long exposure is continued after this
limiting value has been reached.
    Fig. 1 shows the growth and course of TTS recovery for wideband noise having a sound
pressure level of L^ = 100 dB at test frequencies of 0.5,  1, 2, 4 and 8 kHz as measured in 20
young people (40 ears) of normal hearing [ 1 ]. This clearly  demonstrates the approaching
of a limiting value and the strongly delayed course of recovery after prolonged retention at
this asymptotic value. A similar asymptotic behavior  and delay in recovery time  at the
retention on the limit  value has also been established for intermittent noise [2] and pulse
sequences [ 3 ].
    Because there is no unambiguous connection between  the  time needed for TTS re-
covery  and TTS2 the  latter alone proves to be inadequate for characterizing  the stress on
hearing.
    Knowledge of the biochemical processes in the inner ear, especially in the hair cells [4]
and the course of TTS at noise exposure suggests the introduction of
                                      S= /(TTS) dt                          (1)
                                          •X

as a measure  of physiological stress where the integral  over TTS has to cover the period
during and after noise exposure as well.
    For further investigation, the TTS at 4 kHz has been selected to characterize the stress
on hearing.
    G. FUDER and L. KRACHT [2]  have studied the physiological stress at different noise
exposure according to Fig. 2. Each series of measurements has been made on 15 to  25
young people of normal  hearing (students and apprentices). It can be shown that for steady
noise  (types a and  b in Fig. 2), there is a simple relation between cause (noise) and effect
(physiological stress S) if noise exposure is expressed by

                                             (t)| dt.                          (2)

Here the integral of the magnitude of the A-weighted sound pressure should cover the entire
period of noise exposure.

                                        293

-------
dB

25

20-

 15-

 V

 5
        773P
              10
 dB
 25\  irr*2

 20

  15
 iO

 5*
 dB
 25
11
 20

  V

  10

  5"
       rrs
                      100
                                  30min
                            •fOOOmin
                              tE*12Omin
                                 -I-
                                10OO rrtin
                       W
                                 •+•
                               •COOmin
Figure 1. Growth of ITS and 'course of recovery as a function of time of noise exposure. Sound level L
100 dBA. Mean ITS of 20 persons.
                                        294

-------
     Steady noise
     L - 90... 105
     t=Bs... 8h
 Steady noise
 intermittent exposure
L = $3.  105 dBA
ft = 8s...40 min
if 8s... 40min
                           Impulse noise
                           t=KO... «5  dB
                           tz = 0,1ms ...3ms
                                    .  00
Figure 2. Types of noises used for noise exposure during the laboratory tests. The values below the figures
indicate the range in which the tests take place. Not all combinations could be realized.
     Figure 3 shows the averages S§ ^/(TTS)  dt = f (B§) obtained from 54 different noises
of types a to b. This correlation can be expressed by the linear equation
where
                                                (3)

                                                (4)
                                    PS"1"J(/ubar)^

For single pulses and pulse trains (type c) the studies of H. ERTEL [3]  - performed with
peak levels of L ^ 140 dB and pulse spacings of tp ^ 3 sec. - resulted in
                                     Si=0iB!                                     (5)
where

and
=Jp2 (t) dt

   1     dB
                                         15  (/Ubar)2-
                                                (6)

                                                (7)
The time  integral  of the squared sound pressure  in Eq. (6) again covers the whole noise
exposure time. With a low pulse spacing of tp ^ 0.5 sec. as obtained, for example, with
automatic small arms,  the physiological stress will reduce. If the pulse rate is sufficiently
high, the effective value ft reduces to approx. 10%, an effect that, in all probability, can be
attributed to the acoustic reflex.
     But a measure of the noise-dependent physiological stress on hearing  proves to be
useful only in the event that a connection can be established to the noise-induced hearing
                                          295

-------
»•
• ~-
I
«
1
I
V
1
•
4
2
»'
I
f
4
I
«'
1
t
t
t
•0'









	

















































» V











































y
y[
o
B
oy
/
*
















»
>
'


















/



















/
1















.
*
7.
^

















y

i


















^







































/
/
















0
jf
V










































t
»


































y
^
















.^
f



















/















































Sj -/(Trs; vo*M a/am IfH-urm tntirmtntnt t*f>oiurr to «*od> noat
             TTS • tut AvfMOT} 4 000 Mi e-20penorLi pur tut itntt
                  •/ type o andtt n F,g 2
                     Figure 3. TTS - integral as a function of noise exposure.
loss. Because of the linear relationship of Eq. (3) it was merely the expression PTS = f (85)
for steady  noise that has been searched for (PTS = permanent threshold shift). Apart from
some other data, the material collected by PASCHIER-VERMEER (5]  was used for evalua-
tion based  on the following hypothesis:
     The loss of hearing with increasing age (presbyacusis)  shall be expressed by a load
variable that is equivalent to Eq. (2), viz: -
                                          Padt=Pa(tL-tE)
(8)
                                           296

-------
where
                             tg = Overall duration of noise exposure as covered by Eq.(2)

                             pa = 0.4/ubar (= La = 66 dB).

As a rule, work published in the relevant literature about noise-induced hearing loss gives the
PTS after presbyausis correction. Here the threshold shift measured in subjects suffering
from loss of hearing has been reduced by the probable age-dependent threshold shift. In our
evaluating for the literature about noise-induced hearing loss, this  correction had  to be
revoked.
    The result of evaluating the data  collected  by PASCHIER-VERMEER [5]  and other
researchers is shown in Fig. 4 where the ordinate contains the prospective actual threshold
shift at 4 kHz (caused by noise  and age) plotted against the function
                                = BS+Ba
In close approximation this yields: -

                              PTS=551g- dB
with
                                         B
                                           o
                              B0 = 2.5 • 108/ubar • s.
                                               (9)



                                               (10)

                                               (11)
If it holds true that S of Eq. (1) is equal to the physiciological stress, a hypothesis can be
worked out on the loss of hearing, both for steady and impulsive noises
with
where
   PTS = 55 lg4-dB
               PC
                                            f(p)
                                                                           (12)
                     dt                        (13)

         = instantaneous value of A-weighted sound pressure
In a range of 100/ubar <
been determined yet.
   pc   = 22.5 ubar

   |3C   =  1.1-107

   f(p) -  1 for (PA(0  < 100/ubar

   f(p) = 2 for (PA(0  > 2 • 103/ubar

(t)| < 2 • 103/ubar, f(p) follows a transfer function that has not
                                         297

-------
PTS
       •c
      H
      M
      >C
      :c
  Sound level L in

 •>• -m   105
 *--^   101
 •   •   100
                                                           .-»   9V
                                                                         *- -4   9Z
                        H
                        IS
                        71
r-3 . f(P).     ! • ', » ra
(nee tq.C>). (0) and (?)).
Fvery point represents the "
at 40CC lit of i« to rtfjrly ?
ployeee «ho rere tipo-e^1 to
noiien (after
                                                                          i.-ls due
                                                        hypothetic .1 oound IcTel of
                                                        (PTS . t(*ft) after
 tr .
-------
                                    References

1.   Funder, G., and Kracht, L., Der Einfluss von Grenzwerten der TTS auf die Beurteilung
    der gehorschadigenden Wirkung stationaren Larms.
    VII. AICB-Kongrej3 Dresden Proc., 87-91, 1972.
2.   Fuder,  G., and  Kracht,  L.,  Ph.D.-THESES  at  Techn.  Univ. Dresden. Paper in
    preparation.
3.   Ertel, H., Ph.D.—THESES at Techn. Univ. Dresden. Paper in preparation.
4.   Vosteen, K. H., Neue Aspekte zur Biologie und Pathologic des Innenohres.
    Arch. Ohr.-Nas.-und Kehlk.-Heilk. 178, 1-104, 1961.
5.   Passchier-Vermeer, W., Hearing loss due to exposure to steady-state broadband noise.
    Research Institute for Public Health Engineering. Delft, The Netherlands,  Report 35,
    April 1968.
6.   Coles, R. R. A., and Rice, C. G., Assessment of Risk of Hearing Loss due to Impulse
    Noise.
    In Kryter, K. D., Effects of Noise on men. Academic Press, 71-94, 1970.
                                        299

-------
      EXPERIMENTS ON ANIMALS SUBJECT TO ACUTE ACOUSTIC TRAUMA

                                  Wiktor Jankowski
                       Clinic of Otolaryngology of the Medical
                            Academy in Wroclaw, Poland.

    It is well known that the  development of temporary hearing loss in operators working
under bad acoustic conditions during the  8-hour shift is not linear with respect to the time
of work. The greatest amount of that loss occurs during the first hour or hour and a half of
the initial time period.
    Thus in our experiments  made on animals our particular attention has been paid to
that period of acoustic exposure.
    The development of hearing loss in our animals has been observed by us on the basis of
the measurement of the loss of cochlear microphonics (CM) and action potential (AP). The
concept of using the decrease of  both these potentials, occurring during the acoustic expo-
sure, as  the measure of the drop  in hearing  sensitivity is not a new one. Other authors have
long been making such experiments; still, they were applying mostly only short (lasting only
a few minutes) acoustic exposures.
    In  our experiments we have been using white noise at levels of 80,85, 90,95 and 100
dB, with the exposure times of 5,15, 30, 60 and 90 minutes.
    In  one of our experiments we have conducted 25 test series, each of them embodying
four to  five individual tests - thus making jointly in that part of work more  than 100
individual tests.
    The values of CM have been measured for six pure tones and the mean value of the CM
voltage for all six tones has been calculated. In our experiments, the values of both potentials
have been expressed as a percentage of the potential's initial value, i.e. the value obtained
prior to exposure. The obtained results are illustrated on the diagram. The exposure time
periods are marked off on abscissa, while  the loss values  of both potentials are marked off
on the ordinate. The solid curves  represent the behavior of the CM and the dashed lines that
of the AP. On the right hand side of the diagram the exposure level is marked off.
    I should like to direct your attention only to the relation between the CM loss and the
AP loss. As may be seen on the diagram, the respective curves are not parallel  to each other.
After  short exposures, the loss of CM  exceeds that of AP. That can  be best seen when
employing high exposure levels. I should like to bring your attention to the variation of the
AP with these high intensities.  With  95 dB, the AP value falls below the lowest value that
could be measured after ninety minutes of exposure. At the 100-dB level, the  AP disappears
already after 60 minutes of exposure. That never has been observed in the case of the CM. It
is known that after an animal dies, the AP disappears at once, while the CM endures-to be
sure, at a lower voltage—still for several minutes.
    We have been also observing the recovery of the losses of both potentials, after termi-
nation of the exposure, during the  next  90 minutes.  In general, the smaller the loss, the
quicker  the recovery of the potential loss. After terminating the 90 minute exposure to the
100 dB level, during the next 90 minutes, the reappearance of AP has not been observed.
     The above losses of both potentials constitute a result of metabolic unbalance. We
have tried to prove that assumption by administering cytochrone C (which promotes the

                                        301

-------
                                  Figure 1. Jankowskj
oxidation processes) before exposure, as well as ATP (which promotes the energetic processes)
or both of these substances simultaneously, and then observing their effect on the formation
of losses appearing during exposure.
     As the most prominent differences between  the shape or both potentials have been
occurring in response to 100 dB, we have repeated these experiments, except that the above
mentioned substances have been given before applying the load. The prophylactic giving of
cytochrome C, ATP or both had a most beneficial effect on the reduction of the loss of CM
following exposures lasting 5,  15 and 30 minutes. However, for the longer exposures, the
beneficial effect of pre-administration of the above substances has not been so distinct: the
prophylactic administration of the above preparations had not so advantageous effect on
reduction of the AP loss; for  durations exceeding 60 to 90  minutes the  effect of these
preparations on the AP-curve shape was negligible.
     How are the results of these experiments to be explained? It is known that, as the
sensory cells of the organ of Corti operate under rather unfavorable oxygen supply condi-
                                        302

-------
tions, even a weak disturbance of the exygen or energetic economy provokes a distinct drop
of CM, and the administration of cytochrome C or ATP distinctly improves the efficiency of
the  sensory cells of the organ of cort. The drop  of AP occurs after the longer and stronger
exposures because its supply with metabolites necessary for keeping up its efficiency is, so
to say, better, because they are drawn directly from the blood circuit system. In that way
could be explained in some measure the differences between the behavior of both poten-
tials.
     It might  be asked if the behavior of both  potentials  is  characteristic only for the
acoustic trauma?
     Doc. dr  Zb. Ziemski, my  associate, has noticed a  similar behavior of both potentials
when he  was poisoning  guinea  pigs with  some organic solvents (polyethylene glycol,
propylene glycol,  dimethyl formamide, dioxane,  four-hydrofuran) by way of inhalation.
During the initial period of poisoning the drop of CM exceeded that of the AP. As the time
of poisoning was extended, the loss of AP exceeded the loss of CM.
     The same results have been obtained by Ziemski and myself when poisoning animals
with sodium salicylate. During the initial periods  of poisoning the losses of CM were greater,
but with extension of the time of poisoning, the losses of AP exceeded those of CM.
     Similar results have been obtained by Preibisch-Effenberger and Ziemski in the com-
bined work in which they have been exposing guinea pigs to ultra sonic noise (800 kHz at an
intensity of 7 watts/cm2). After 5 min of exposure, the  loss of CM greatly exceeded that of
AP, while after 30 min the values of the loss of AP approach those of CM.
     Similar differences in the  behavior of  these two potentials have been observed by
Deutsch, who, using a series of hypoxia periods,  found a greater loss of AP than of CM.
Simmons et al.,  after a strong and long-lasting  exposure, have found that the loss of AP
exceeds greatly that of CM.  Silverstein  has obtained, in  case of the salicylate poisoning,
results much like ours.  Spoendlin has noticed, in his ultramicroscopic studies after severe
exposures,  greater changes  in the mitochondria  of the  nerve endings  than in the
mitochondria of the hair cells.
     All above cited studies stress the difference in behavior of these potentials ascribed to
disturbance in the metabolism of the hair cells of  the organ of Corti and in the metabolism
of the aural nerve.
     If we  were perhaps to propose a general law, it would be that "any disturbance of the
metabolism of the organ of Corti or of nerve metabolism  will at first produce the strongest
effect on the CM, while more prolonged exposure to  the disturbing factor will lead to a
greater diminution of AP."
     At any rate it seems that, in view of the experimental results, it would be worthwhile
to continue investigations regarding the different behavior of both morphologic elements.

                                     References

 1.   Aubry, M.,  Pialoux, P.,  Burgeat, M.: Influence d'une stimulation acquistique intense
     sur la response de la cochlee. Acta Otolarying.  (Stock 60, 191-196, (1965).
 2.   Burgeat, M.,  Burgeat, C.: Study of changes  in the  cochlear microphonic potentials
     arising after intense auditory stimulation. J. Physiol  (Paris) 56, 225-232, (1964).

                                        303

-------
 3.  Deutsch, E.: Cumulative effects of oxygen lack on the electrical phenomene of the
     cochlea. Ann Otol Rhinol Laryngol. 73,348-357, (1964).
 4.  Faltynek,  L.,  Vesely,  C.:  Zur  Restitution  der Mikrophonpotentiale  des Meer-
     schweinches nach Kurzfristiger Larmbelastung.  Arch. Ohr.-Nas.-KehlkJHeilk.  184,
     109-115. (1964).
 5.  Gerhardt, H. J.,  Wagner, H.: Die Wirkung dosierter Ger&usch-belastung auf Mikrophon-
     potentiale der Meerschweinchenschnecke.  Arch. Ohr.-Nas.-Kehlk.Heilk. 179, 458-472,
     (1962).
 6.  Gisselsson, L., Soerensen, H.: Auditory adaptation and fatigue in cochlear potentials.
     Acta Otolaryng.  Stockh. 50,39M05, (1959).
 7.  Jankowski, W., Ziemski, Z., Cyrulewska-Orowska, J.: Pharmacological tests of Preven-
    tion  the acoustic trauma in  electrophysiologic  experiments. Archium Akustyki 5,
     129-143(1970).
 8.  Jankowski, W., Ziemski, Z., Birecki, W., Cyrulewska-Orowska, J., Praga, J., Kowalew-
  .   ska,  M.: The effects of acute acoustic trauma and urethan anesthesia on the patterns of
     biopotentials of internal ear in guinea pigs. Otolaryng. Pol., 25,253-265, (1971 A).
 9.  Jankowski, W., Ziemski, Z.: Differences in Performance of Cortis Organ sensual cells
     and  of the Aural Nerve at Experimental  Disturbances of their Efficiency.  Archiwum
     Akustyki. 3, 233-231, (1971B).
10.  Kumagai, K.: Studies on  the  microphonic action of the cohchlea in auditory
     disturbance. Excerpta Med. sect. XL, vol. 13, 904, (1960).
11.  Meyer zum  Gottesberge, H.: Rauch, S., Koburg, E.: Unterschiede in Metabolism us der
     eincelnen Schneckenwindungen. Acta Otolaryng. (Stockholm) 59, 116-123,  (1965).
12.  Preibisch-Effenberger R., Freigang  B., Seidel, P., Ziemski Z., Reczek-Krauss H.: The
     effect of ultrasounds and cytochrome C  on functional  disturbances of Corti's organ
     and  cochlear nerve in animals. Otolaryng. Pol. 24, 21-26, (1970).
13.  Sekula, J.,  Trabka, J., Miodonski, A.: Noise Influence on Biopotentials  of Hearing
     Organ. Pamietnik XXVII Zjazdu Otolaryngologow Polskich. Katowice  1968, Warszawa,
     PZWL,(1970)24.
14.  Silverstein, H., Bomstein, J. M., Davies,  J.: Salicylate ototoxicity a  biochemical and
     electrophysiological study. Ann of Otol Rhinol Laryngol, 76, 118-129, (1967).
15.  Simmons, F. B.,: Middle Ear Muscle Protection from the Acoustic  Trauma of loud
     continuous Sound. Ann Otol Rhinol Laryngol. 69,  1063-1071, (1960).
16.  Spoendlin,  H.:  Submikroskopische Veranderungen an Cortischen Organ des Meer-
     schweichens nach  acustischen  Belastung.  Practica Otolaryngologica.  20, 197-201,
     (1958).
17.  Wagner, H.,  Gerhardt, H. J.: Die Wirkung  dosierter Reintonbelastung auf die Mikro-
     phonpotentiale  der Meerschweinchen-schnecke.  Arch. Ohr.-Nas.—Kehlk. Heilk.,  181,
     82-106, (1963).
18.  Ziemski, Z., Jankowski,  W.: Ototoxidity of sodium salicylate. Otolaryng. Pol. 26,
     391-395. (1972).
                                        304

-------
                         SESSION 4 A
INTERACTION OF NOISE WITH OTHER NOXIOUS AGENTS IN PRODUCTION OF
                        HEARING LOSS

                   Chairman: E. Lehnhardt, BRD
                             305

-------
             INFLUENCES OF CHEMICAL AGENTS ON HEARING LOSS

                                 M. Haider, Vienna

    The ototoxic effects of chemical substances have been known for a long time. There
are otologic reports on carbon  monoxide  poisoning that were written in the seventeenth
century (van Helmont,  1667).  Other known ototoxic industrial chemicals  include lead,
phosphorus, halogenated hydrocarbons, mercury, carbon disulfide, etc. An extensive review
on the industrial health aspects of these substances is given by Lehnhardt (1965).
     Besides industrial products there are  many drugs for which ototoxic side-effects have
been described. At  the  moment the best  known are  some antibiotics like streptomycin,
dihydrostreptomycin, kanamycin, and neomycin. Some other ototoxic drugs are salicylates,
quinine and substances with similar effects (e.g. chloroquine), arsenic, alcaloids (strychnine,
morphine, scopolamine), oleum chenopoii, etc.  Finally, the ototoxic effects  of stimulants
(nicotine, alcohol), narcotics and endogenous intoxications (e.g. during some infections and
other diseases) are worth mentioning. Huizing (1966) gives a summary of these problems.
    This paper will mainly review the combined effects of ototoxic substances and noise.
On principle, the combination of chemical agents and noise can be:
    (a) indifferent (combination does not  differ from its most effective component),
    (b) additive (combination corresponds roughly to sum of both factors),
    (c) synergistic (effect of combination  is higher than sum of individual components).
    (d) antagonistic (effect of combination is less than most effective component), or
    (e) protective (substances make the ear less susceptible to hearing loss).
    Data  on such combined effects may  be derived  from  different sources, e.g.: clinical
case studies, systematic field studies, experimental research with animals  and  experimental
research in connection with temporary threshold shift.
    There are many clinical observations on combined effects of intoxication (e.g.: under
CO, C$2,  Nitrobenzol)  and noise.  For twelve  years, the working conditions of a man
described by Wagemann (1960) contained  noise  of 80-90 Phon,  but the clinical symptoms
(audiogram, vestibular signs, etc.) seemed to indicate a CO-intoxication.  It  is possible that
both noxes had a combined  effect. One  case, described by Lehnhardt  (1965) showed a
similar combined effect of trichlorethylene and noise. Some of the experimental studies on
drugs are based on clinical  observations. Darrouzet (1967) mentions a case of hearing loss
after streptomycin treatment and surgical intervention on one ear with noise-stress through
a milling-machine. Dayal et al. (1971) designed their experiments according  to the common
clinical situation of premature babies in incubators (generating noise of about 68-72 dB)
receiving kanamycin treatment.
    Field studies show the frequency of hearing loss that has to be expected under the
influence of certain  chemical substances. Examples are given in the Scandinavian reports on
hearing loss caused  by "chronic" carbon monoxide poisoning.  Lumio (1948 a, b) found
hearing deficiencies  in 78% of his patients. 44% of them he assumed as typical for carbon
monoxide poisoning.
  With financial support of the Austrian "Fonds zur Forderung der wissenschaftlichen Forschung" and
   the "Allgem Unfallversicherungsanstalt".

                                        307

-------
     Some other examples are given in reports on hearing loss caused by carbon disulfide.
Zenk (1970,1971) found hearing losses under this condition to have a higher incidence than
control groups of the same age. In such cases it seems difficult to rule out some industrial
noise effect. Unfortunately none of the reports I  found in the literature tried to single out
the influence of hearing loss in connection with chemical agents versus hearing loss due to
industrial noise. It could be interesting to compare equivalent groups of individuals with and
without  noise exposure to groups with or without the influence of chemical  agents.
     Animal experiments have shown that the combined effect of noise and chemical agents
may be  synergistic. Darrouzet (1962)  demonstrated that antibiotics (kanamycin) might
sensitize the cochlea to the damaging influence of noise. If noise was given before the drug,
no synergistic effect occurred. Quante et al. (1970) reported a potentiating effect of noise
(90, 100, 110 dB) with  kanamycin treatment. Jauhiainen et al. (1972) examined harmful
effects of noise  (115 dB) and neomycin both electrophysiologically and microsopically. In
the guinea pig they found a synergistic effect in hair cell damage as well as in amplitude
reduction of cochlea microphonic potentials as demonstrated  in Fig. 1. The authors con-
cluded that there is greater susceptibility to noise-induced hearing loss in persons treated
with these antibiotics.
     Many authors have shown  that ototoxic effects damage  mainly the  outer hair  cells,
maximally at  the basal end. Combination with noise seems to extend the damage further
towards the apical end. Dayal et al. (1971) found that even a low-level noise (68 to 72 dB at
125 Hz) combined with low dosage of kanamycin may have produced a synergistic effect.
                             toe,
                             10
u
x
o
a
u
»  M
                          •4
                          a
                                    • 10
                                                 -10
                                                        -4.0
                                                           dB
                                 DECREASE  OF  VOLT ASK

                                    OF  CM -PO IE N T I A I
Fig. 1.  Relation between percentage outer hair  cell damage and average loss in cochfear microphonic
potentials. Point A refers to animals exposed to neomycin, point B to those exposed to noise alone, and
point C to those exposed to both noise and kanamycin. (From Jauhianinen et al., 1972).
                                         308

-------
Two factors which are ineffective (subthreshold) by themselves may in such way give rise to
manifest damage. In this case the changes of the outer hair cells were seen primarily in the
apical turn of the cochlea. This is demonstrated in Fig. 2.
     The authors assume that the hair cells of the  apical turn were sensitized to damage by
the low frequency noise.
     A possible cumulative effect for carbon monoxide and noise has been assumed by Zorn
(1968), who found a delaying effect on carbon monoxide elimination with 65 dB noise.
Klosterkotter (1972), however, could not verify these results.
     Some of the possibilities of protective influences by various chemical agents will only
be mentioned here briefly. A combination effect of neomycin and noise of 120 dB together
with a protective effect of adenosintriphosphate (ATP) has been described by Faltynek and
Vesely (1969). The course of normalization of cochlear microphonic potentials after com-
bined neomycin and noise influence was favorably affected by ATP. Protective effects of
vitamin B and aminoacids have been described amongst others  by Darrouzet (1962, 1963,
1967). The possible protective effects of  vitamin  A have been under discussion since the
                 ooooooooooooooooooooooooooooooooooooo
           3*  o<
JOOOOOOOOOOOOOOOOOC
'^OOOOOOOOOOOC
 3OOOOOOO»OO«
3OOOOOOOOOOOOOOC
3OOOOOOOOOOOOC
X3«OOOOOOOOOOC
                 ooooooooooooooooooooooooooooooooooooo
               oooooooooooooooooooooooooooooooooooooooooc
                OOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOC
               oooooooooooooooooooooooooo«o*oooooooo*«
                                      JOOOOO
                 ooooooooooooooooooooooooooooooooooooo
                     JOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOO
                     3OOOOOQOOOOOOOOOOOOOOQOOOOOOOOOOOOOOOOOOOOOOO
                     JOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOO
                 ooooooooooooooooooooooooooooooooooooo
            2 OOOOOOOOOQOOOOOOOOOOOOOOOOOOOC
               ooooooooooooooooooooooooooooc
              oooooooooooooooooooooooooooooc

Fig. 2. Cochleogram of guinea pig exposed to incubator noise and receiving kanamycin 15 mgm/kgm body
wt, per day for 5 weeks. Dark circles indicate damaged hair cells. The damage is predominant in the OHC of
the apical turn. (From Dayal et al., 1971)
                                      309

-------
early reports of Willemse  (1952) and Ruedi (1954). A positive effect of nicotinic acid is
described in Sheehy (1960) and Nowak et al. (1971). As one example of the combination
effects of chemical agents and  noise on temporary threshold shift I  will mention some
preliminary  results of our own experiments. Eighteen normal-hearing students were tested
under two conditions. In the experimental condition they were exposed to 200 ppm CO for
4 hours. Before and after  the exposure, auditory thresholds were measured with a Bekesy
audiometer. After half an hour and after two and a half hours, 100 clicks were given and the
evoked potentials were computer-analyzed out of the EEG (Vertex-Mastoid). In the control
condition, the subjects performed the same tests without the influence of CO. The situa-
tions were rotated systematically in  a double-blind  design. The threshold measurements  at
2000, 3000 and 4000 c/s showed no systematic change and no statistical significant dif-
ferences.
     For the auditory evoked potentials no significant latency changes but amplitude reduc-
tion under  CO-exposure of the  main negative-positive  peak to peak amplitude could be
demonstrated. This is shown in Fig. 3.
    - There was also a diminution of amplitudes from the 05-hour exposure to the 2.5-hour
exposure. This diminution occurs in both situations but it is significant only under the
CO-condition. Some authors have shown similar results for the late part of the visual evoked
potentials for animals under the CO-condition (Xintaras et  al, 1966). In an earlier experi-
ment, we demonstrated that the slow electric brain potentials (expectancy waves) show a
marked reduction under the influence of even  lower CO-concentrations in the air (Groll-
Knapp et al. 1972). So it seems that even very low concentrations of CO in the inhaled air
may change the brain reactions evoked by acoustic signals.
    €0

    50-

    W-

    30-

    ao-
          of  E?
             ew  crz
            Control
EM Etl
ioorr,co
Latenet)
msec.
410-
410-
400-
90-
*>•
10-


b








































EM EfZ EM ffl
CoKtrol iOOfrCO
Figure 3.  Amplitudes (relative values) and latencies (msec) of auditory evoked potentials before and after 4
hours CO-exposure (200 ppm) compared to the control condition.
                                        310

-------
     To get some information on possible combination effects of CO and noise, the subjects
were  exposed to a 105-dB octave-band noise with a middle frequency of 2000 Hz for 15
min in both situations. One result of TTS-measurements at 4, 8, 16, 32, and 64 minutes
after exposure is shown in Fig. 4.
     The TTS values, measured with  a test-tone of 3000 c/s are slightly higher immediately
after  the CO-exposition than under the  control-situation. But  the late TTS values are
                                Control
CO	
                                                         33t
.Figure 4. TTS at 3000 HZ after  15-min exposure to octaveband noise (center frequency  2000 Hz)
following a 4-hour CO exposure (200 ppm) compared to control condition.

                                         311

-------
practically identical. There is no statistical significant difference between both situations. It
must be concluded that ««hrer the circumstances described no synergistic effect of CO- and
noise exposure could be demonstrated.
                                  REFERENCES

(1) Darrouzet, J. and E. de Lima-Sobrinho: Oreille interne kanamycine et Traumatisme
     acoustique. Etude experimentelle: Rev. Laryng.Otol. 83, 781-786, (1962)
( 2) Darrouzet,  J.: Essais  de protection  de 1'organ  decorti centre 1'otoxicite. Rev.—
     Laryng. (Bordeaux) 84,459-478 (1962).
( 3) Darrouzet, J.: Essais  de  protection de 1'organ decorti centre 1'ototoxicite des anti-
     biotiques. Table Ronde: Rev. Laryng. (bordeaux), 84,478-488 (1963).
( 4) Darrouzet, J.: Essais de protection de 1'organe de  corti centre 1'ototoxicite  anti-
     biotiques. Etude experimentelle: Acta otolarying. (Stockholm) 63, 49-64 (1967).
( 5) Dayal, S., A. Kokkainen and P. Mitchell: Combined effects of Noise and Kanamycin.
     Ann.Otol. 80,897-902 (1971).
( 6) Fairy nek,  K. and C. Vesely: Einflu/3 des Neomycin auf  die  Funktion  der Meer-
     schweinchenschnecke. Mschr.f. OHrhk. 103, 545-547 (1969).
( 7) Groll-Knapp, E., H. Wagner, H. Hauck and M. Haider: Effect of low Carbon Monoxide
     Concentrations on  Vigilance and Computer-Analysed Brain Potentials. In: Carbon
     Monoxide-Origin,  Measurement and Air Quality Criteria, 116-119, Dusseld. (1972).
( 8) Huizing, E.  H.:  Toxische Schaden des Hororgans.  In:  Berendes:  HNO-Heilkunde-
     Lehrbuch, HI/3 (1966).
( 9) Jauhiainen, T., A. Kohonen and M. Jauhiainen: Combined effect of noise and neomy-
     cin on the cochlea. Acta otolaryng. 73, 387-390 (1972)
(10) KlosterkStter, W.:  Der Umweltfaktor Larm als Komponente kumulativer Umwelt-
     wirkungen. Arb.-Med., Soz.-Med., Arb.-Hyg. 10,281-286 (1972).
(11) Lumino, J. S.: Otoneurological Studies of Chronic  Carbon Monoxide Poisoning in
     Finland. Acta oto-Iaryng. (Stockholm), Suppl.67,65-112 (1948 a)
(12) Lumino, J. S.: Hearing deficiencies caused by carbon monoxide (generator-gas).  Acta
     oto-laryng. (Stockholm), Suppl. 71, 1-112 (1948 b).
(13) Lehnhardt, E.: Die Berufsscnaden des Ohres. Arch.f.Ohr.-Nas.us. Kehlk.-Heilk. 185,
     11-242(1965).
(14) Nowak,  R.,  B.  Kleinfeld und D.  Dahl:  Tierexperimentelle Untersuchungen  zum
     Verhalten  der  Mikrophonpotentiale  nach  Nikotinsaure-gabe am  durch  Schall
     vorgeschadigten Ohr. Mschr. Ohr.Hk. Wien, 11, 495-498, (1971).
(15)Quante,  M.,  H.   Stupp  und J.  P. Brun:  Ototoxikosen  unter Larmbelastung
     Arch.klin.exp. Ohr.-Nas.-Kehlk.Heilk. 1960, 233-237. (1970).
(16) Ruedi,  L.: Wirkungen des Vitamin A im menschlichen  und tierischen Gehorgang.
     Schweiz.med.Wschr. 84,1411-1414 (1954)
(17) Sheehy, J. L.: Vasodilator therapy in sensorineural  hearing loss. Laryngoscope 70,
     885-914. (1960).

                                       312

-------
(18) Wagemann, W.: Das otologische Bild der Kohlenoxidvergiftung. Zschr. f.Lar.Rhin.Oto.
    38,691-702(1960).
(19) Willemse, C.:  Protection against occupational deafness: Role played by Vitamin A.
    Acta oto-laryng. belg. 6, 319-324, (1952).
(20) Xintaras, C., B. L. Johnson, C. E. Terrill and M. F. Sobecki: Application of the Evoked
    Response Technique in Air Pollution Toxicology. Toxikol. and AppLPharm. 8, 77-87
    (1966).
(21) Zenk, H.: Schwefelkohlenstoffeinwirkungen auf die rhino-otologischen Funktionen der
    Beschaftigten in der Kunstfaserindustrie. Int.Arch.ArbMed. 27, 210-220 (1960).
(22) Zenk,   H.:  Zum   rhino-otologischen   Bilde   chronischer  Schwefelkohlenstoffein-
    wirkungen. Z.Laryng.-Rhinol. 50, 170-175 (1971)
(23) Zorn, H.: Die kombinierte Wirkung physikahscher und chemischer Noxen, aufgezeigt
    am  Beispiel  der Schadigung durch Larm und CO. Schriftenr.  Arb.Med.Soz.Med.
    Arb.Hyg. Stuttgart, Bd.27, (1968)
                                        313

-------
       HEARING LOSS OF FOREST WORKERS AND OF TRACTOR OPERATORS
                    (INTERACTION OF NOISE WITH VIBRATION)

                                    Istvan Pinter
               State Institute of Occupational Health, Budapest, Hungary

    In our investigations to be presented an answer was sought to the question, whether
simultaneous exposure to noise and vibration has any influence on the dynamics of the
development of hearing loss.
    The starting point of the investigations is, on one hand, the equal energy principle, i.-e.
that exposure of the same magnitude results in the same degree of hearing loss and, on the
other hand, the interrelationship between the PTS and TT$2, namely that the value of ITS
of healthy young people measured 2 minutes after the end of a daily noise exposure equals
the average PTS caused by a ten-year exposure.
    In order to make a judgement of the question possible,  the HLs found in tractor
operators and forest workers were compared to those of control persons, who were exposed
to noise only, and for whom the basic principle mentioned above could be proven to hold,
on the basis of data  in the  literature and of the results of investigations of our own.
Accordingly, the persons investigated were grouped as follows:
    1.) Tractor drivers with noise exposure and exposed simultaneously to vibration of a
frequency around 10 Hz; workers of the furniture industry served as their controls.
    2.) Contra - Stihl power  saw operators of the forest industry with noise exposure and
with a  simultaneous exposure to vibration of frequency in the range of 125-350 Hz; their
controls were workers in the textile industry, exposed to the same level of noise.
    3.) Forest workers with clinically verified vibration damage (alterations in the locomo-
tion and vascular systems Raynaud-syndrome); "healthy" forest workers served as their
controls.
    The tractor drivers and the workers of the forest and furniture industry are the em-
ployees of the  same forest company at the same geographical location near the capital; the
textile industry workers are employees of a metropolitan plant.
Noise analysis

     Noise exposure was determined according to the R 1999 ISO noise-measuring recom-
mendations on the basis of the equal-energy principle. For this, the noise in the different
working cycles was picked up for a period of time according to the regulations at the height
of the ears of the persons working on the working places, and recorded on magnetic tape.
For the recording on magnetic tape, a precision noise level meter (Bruel & Kjaer type 2204)
and  a portable measuring tape recorder (Nagra III) were used. The noise recorded was
analyzed  in the laboratory (Bruel & Kjaer Real-time 1/3 Octave  Analyzer type 33-47, Level
Recorder type 2305) and was evaluated with mathematical methods and the Lgq value was
determined with a dosimeter (Bruel & Kjaer type  4423).
     The results  of the noise analysis  in case of several typical occupational activities are
shown in Fig. 1 and 2.

                                        315

-------
                100
                •'.
                s:


                II


                70
              g  M
                «o
              K
              tf 30
M


II

           WOOD WOfiK
     ;  !  TRACTOR
     i  •
                                                 r
                                                          110 dBA
                H
               IM
                                                                        ••— » »»-
                                                         ,..,,,...,.
Figure 1. Percent distribution of sound level, Lgq and 1/3 octave spectra measured for tractor driving and
at working places in the furniture industry during appropriate working cycles.
     The upper parts of Figures 1  and 2 demonstrate the  percent distribution of the noise
levels and the corresponding Leq -values in the period of time investigated, in case of forest
workers (Contra-Stihl power saw operators), in relation to work done on the tractor, in the
                                           316

-------
100
 '.
 K

 U

 70
              —
              U
                40
                30
                10
n
                          TEXTILE WORK
                        i  STIHL-SAW
                                  >-r
                                     80
                           *•
                             100     ifa  dBA
                            ...... 1f •.,,.,,
                                       I
                                       ST^-S*.                  :::|  B. I
Figure 2. Percent distribution of sound levels, L^ and 1/3 octave spectra measured at the power saw and
at working places in the textile industry during appropriate working cycles.
furniture factory and in the  textile plant; the lower parts of the figures show the l/3-octave
spectra of these same noises.
                                          31 7

-------
     The equivalent sound levels of the different workers vary, depending upon the working
phase and upon the location of the measurement, as follows:

                   Forest workers:     96-100   dB(A)
                   Tractor drivers:     90-98    dB(A)
                   Workers in the furniture industry:    90-98  dB(A)
                   Workers in the textile industry: .      97-101 dB(A)


Audiometric investigations

     The  hearing investigations were performed with a Peters Type AP-6 clinical audio-
meter, standardized according to  the  1964 ISO recommendation /R 389/. The measure-
ments were done 16 hours after the last noise exposure in an anechoic chamber correspond-
ing to the ANSI SI. 3-1960 standard.
     By the evaluation of the audiograms the ISO R 1999 and the AAOO 1970 recommen-
dations were taken into account.                             .         -•••••
     The  de facto noise-induced hearing loss was determined from the audiograms taking
into consideration the sociacusis according to sex (Spoor 1967, Passchier-Vermeer 1968),
thus the audiograms can be compared on ground of the exposure  times and levels only,
without the necessity of taking the variations due to age and sex into account.
     From the audiograms corrected according to the procedure above, the mean values of
the hearing curves (059) were calculated in case of all the 4 groups, at  1-4, 5-14, 15-24 and
finally,  at precisely 10 years  of  exposure. Exposures of less than one year were not con-
sidered. At the mean  value curves the average width of the field of scatter was calculated
according  to  frequencies; from this the relative deviation  was computed,  which varies
between 0.53 and 0.20 in the frequency range of 3000-6000 Hz. Accordingly, they are
within the values evaluable mathematically.
     The audiograms (right ear) of only such workers are entered into the evaluation, whose
hearing loss can be ascribed with high probability  on the  basis of the anamnesis and of
otological investigation, to noise exposure of occupational origin only.
Results of the investigations
     On  the  basis of the criteria outlined above, in the first  approximation the average
hearing thresholds (059) were determined as the function of the type of noise exposure only.
     This is shown by Figure 3.
     No. 1 shows the D$Q audiograms of forest workers (ContraStihl power saw operators),
No. 2 those of the corresponding control textile industry 'workers, No. 3 demonstrates the
curves of tractor1 operators working in the forest and No. 4 those of their controls, workers
in a furniture factory.
     It is to  be seen on the figure that, on  the one  hand,  the hearing loss in the high-
frequency range of both the tractor drivers and forest workers is greater than that of the
controls and, oil the other hand, that the hearing threshold in the low frequency range (250
Hz -1 kHz) of tractor drivers and of forest workers is higher than usual.

                                        318

-------
                                                                       1 FELLERS
                                                                       2 TEXTILE
                                                                        WORKERS
                                                                       3 TRACTOR
                                                                        DRIVERS
                                                                       4 WOOD
                                                                        WORKERS
               2SO
1000
2000    3000  tOOC    6000  8000   Hz
             Figure 3.  Average hearing loss due to noise as a function of noise exposure.
    In order to follow the development of  this type of audiogram, the D$Q values were
compared also in terms of exposure time. The audiograrns corresponding to  1-4, 5-14 and
15-24 years of exposure, of tractor drivers and of furniture industry workers are seen in Fig.
4.  In case of tractor drivers the group with 1-4 years of exposure is not depicted, because
there were only two workers in this category.
    Beyond the course of the increase in the hearing loss, the  figure clearly  demonstrates
the different character of  the curves that is correlated with doubling the exposure time, and
within this, especially the difference at the low frequencies.
    The same comparison in case of forest- and of textile-industry workers is demonstrated
in  Fig. 5.
    Here an even  more marked difference between the two groups can be seen. In case of
the textile-industry workers the  equal-energy principle is valid, and according to this the
hearing threshold increases with exposure time.
    This is, however, not  observed in forest workers; on the contrary, there  is hardly any
change in  the already-developed hearing loss between 2 kHz and 8 kHz  with  an increase in
the exposure time. It can, however,  be seen also in case of forest workers that the hearing
thresholds are higher between 250 Hz and  1 kHz.

                                         319

-------
••
-









:?













































•*













ta

9!.





^
- - " '












T.,f
-
S-H






-*;



'»«;
-







! ' N 1
_J
•
•.





^



s 1
a • K
. c
- - m
1








	 	





,,
>
N










Vvv,
S.
S
^







k
— ""
l»^
N
s
\








k
A
\
k
N
1





\
\
\
\


ft
4








X
•N
> •*
\
\
\
,



"~ — ' — _

1 	



* «M

IHACTOfi




r
i
1 c
t *
i

i 1
y

WOOD
WORKERS


....
,— - t.

, '

<
,'

                         a
                                 •ODD
                                          i
                                                  *OOD
Figure 4. Average hearing loss due to none of trader drivers and of workers in the furniture industry for
exposures of 1-4, 5-14 and 15-24 years, respectively.
     The object of our further investigations will he the nowadays accepted interrelation-
ship between the TTS2 and PTS on 4000 Hz in case of a 10 year exposure. The upper part
of  Figure 6  shows the D5Q-curves which developed in the various groups in case of an
exposure of  IO years.  The lower part demonstrates  the mean values measured  at 4000  Hz
and  the TTS2  values expected 2 minutes after a one-day  exposure and calculated on the
basis of the Ward et al. (1958) and Nakamura (1967) equation.
     It can be seen from the curves that the equal-energy principle is acceptable in relation
to the controls (furniture- and textile-industry workers), but it does not seem valid in case
of tractor drivers  and  forest workers. The  situation is similar at  an equal noise exposure
when comparing the controls and the groups corresponding to them. In the equation  for
calculating the  TTS2 for the investigation of the interrelationship between the PTS and the
                                         320

-------



                                                                FEU c»s
                                            .^".
                                             - V
                                                *



                                                      :
                                                           (COD WOO  M;
Figure 5. Average hearing loss due to noise of forest workers and those of textile industry for exposures of
1-4, 5-14 and 15-24 years, respectively.
T]$2 tMC values of K  the on-traction  was calculated from the daily exposure times found
on the basis of field investigations of several year duration. Comparing the PTS and TTS2
values we - as many others - found good agreement in case of the controls. In case of forest
workers and  tractor drivers, however,  the  measured  values exceeded the calculated TTS->
values. Correcting all the values on the basis of an equal value for R, it turns out even more
pregnant that, in  case of forest workers and tractor drivers the found PTS values exceed the
expected values.  Hence, if we take the daily exposure  times into consideration, then the
discrepancy having displayed itself by the curves is solved on 4000 Hz.
     Finally,  the  audiograrns of power  saw  operators suffering from clinically verified vibra-
tion disease (alterations in the locomotor or in the vascular system) were compared to those
of healthy power  saw operators exposed to noise  tor the same length of time.
                                          321

-------
             -••
              :
1 FEUERS
1 TEXTILE
 WORKERS
                                                               3 TRACTOR
                                                                DRIVERS
                                                               X WOOD
                                                                WORKERS
                                             N.
                                             ••x 's
                                             V-^
             U
                                 nco
                                           MOO
                                                          tooo eon HI
                            TTS,-  1.06R(S-85) log
&TOUP
I
2
1
4
U
4104
58
55
T.,,
•*
s
D
•
V
o
R
0.7
1.0
03
0.9
4000 Hz
PIS
30.0
35,6
28.8
24.0
TTS1
22.0
34.0
24.G
24.0
n
ncofr
0.7
0,7
0.7
0.7
400QHz«*r
PTS
J0.0
24.9
20.2
16.8
TTS,
22.0
23,8
16.8
16,8
                              I  2  )  i
      Figure 6. Average hearing Ion in case of 10 yean exposure and the TTS calculated at 4000 Hz.
     The average hearing threshold and the range of deviation for all the forest workers are
Jepicted by  the curves in  the figures. It can be seen  that the  hearing losses of patients
oiffering from the damage of the locomotor organs only (the points in Fig,  7) correspond to
hose of the "healthy persons". In contrast to this, the hearing loss of those  with vascular
iamage (Raynaud  disease) at 3000-6000 Hz is always greater than that of the healthy
                                          322

-------
                                              GROUP I
                         	
                   '-•
                   •
                   -

                                            ODD
                                                      WOO    3000 4030    6000 MOD Hi
Figure 7. A Hearing  loss in the range of 2000-6000 Hz due to noise in patients with  vibration damage of
the locomotor system.
                           B. X-ray photograph of typical vibration damage.

                                                323

-------
persons.  Figure  9 shows the audiogram and X-ray angiogram  of a patient  with  Raynaud-
syndrome.
     The  D5Q-curves found in  the various groups correspond  quantitatively as  well  as
qualitatively  to  the  results of investigations upon workers of similar occupation  published
by Passchier-Vermeer(1968), Kylin(1971), Burns(!964). Robinson (1970), Dieroff (1963)
and others.
     On  the  basis of the data presented we  feel justified in concluding that the simultaneous
effects of vibration and noise result  in an influencing of the dynamics of hearing  loss,
differing from that elicited by noise alone.
                                         iROUl •
                                       BOO
                                                2000   xoo ton    no torn HI
   Figure 8. Hearing loss in the range of 2000-8000 Hi due to noise in patients with Raynaud syndrome.


                                         324

-------

                                                       XOO     JOCC OOD    tOOO tOOOnr
Figure 9.  A. Hearing loss of patient D.M. suffering from Reynaud-syndrome following a 10-year exposure.
                                    B. The angioram of O.M.




                                               325

-------
Discussion

     Acoustic stimuli reach the inner ear normally by air conduction and thus, according to
the general opinion accepted at present, vibration of the whole body plays only a secondary
role in the development of hearing losses.
     The measurements of Bekesy  (1960) proved that hearing loss develops by way of bone
conduction, too.
     Wittmaack (1928, 1934), Popov (1928), Jokoyama (1963) and Morita (1958) proved
on the basis of animal experiments that the simultaneous effect of body vibrations and noise
is more marked on the upper turns of the cochlea corresponding in the low tones, contrary
to the anatomical effect of noise alone, which manifests itself at the base.
     Our  investigations verify that the results of experiments with animals are  valid for
humans too.
     From the investigations  demonstrated it can undoubtedly be concluded that vibration,
acting together with noise, has a potentiating effect which, depending on the frequency
range of vibration, exerts different influences upon the dynamics of hearing  losses. Vibra-
tion in the subacoustic frequency range (tractor drivers) damages  the hearing at the low
frequencies (250 Hz -  1 kHz) and,  at the high frequencies (3-6  kHz) it  shows an increase of
the threshold, increasing with the increase of the exposure time and anyhow exceeds the
values of the controls. Vibration in the audible range (125-350 Hz)  similarly causes damage
in the range of the low tones, but also, the loss brought about in the range of the high tones
is greater than that in case of the controls, with the limitation that here the damage develops
during an exposure of 1-4 years  duration, and hardly changes with an increase  of the
exposure  time. The potentiating effect is even more  marked in patients with Raynaud-
syndrome.
     The data demonstrated prove  that the change in the dynamics of hearing loss in case of
tractor drivers and forest workers is affected  by the simultaneous effect of  vibration and
noise, but they do not give any hints as for the course of pathomechanism. The clarification
of this needs further investigation.

                                    References

American Standard Criteria for Background Noise in Audiometer Rooms S -1.  3. (1960).
Bekesy, G., Experiments in hearing. McGraw Hill, New-York (1960).
Burns, W., Hinchcliffe, R., Littler, T.  S.,  An exploratory study of hearing loss and noise
     exposure in textile workers. Ann. Occ.Hyg. 7, 323-333 (1964).
Burns, W. and Robinson, D. W., Hearing and Noise in Industry. H.M.S.O. London (1970).
Dieroff, H. G., Die Larmschwerhdrigkeit in der Industrie, Johann Ambrosius Barth, Leipzig,
     (1963).
Guide for Conservation of hearing  in Noise. Supplement to the Transactions  of the Ameri-
     can Academy of Ophtalmology and Otolaryngology, (1970).
Holmgren, G., Johansson, L., Kylin, B. and Linden, O., Noise and Hearing of a Population
     of Forest Workers. Cit. Robinson, D. W.: Occupational Hearing Loss, Academic Press
     London and New-York (1971).

                                       326

-------
International Organization for Standardization. "Standard Reference 200 for the Calibra-
    tion of Pure-tone Audiometers"; ISO Recommendation R 389, Geneve, (1964).
International  Organization  for  Standardization.  "Assessment of  Noise  Exposure during
    Work  for  Hearing  Conversation  Porposes"; ISO Draft Recommendation R.  1999,
    Geneve, (1970).
Lehnhardt, E.,  Die  Berufschaden  des Ohres.  Arch.Ohn-Nas.-Kehlk.-Heilk., 185, 11-242
    (1965).
Kryter, K. D., The effect of noise on man. Academic Press, New York and London, (1970).
Morita,  M., Experimental  studies  on  the  acoustic  influence of  vibration  and  noise.
    OtoLFukuoka, 4, Suppl. 5, 327-332 (1958).
Nakamura, S. and Katano, Y., Relationship between temporary threshold shift and duration
    of noise exposure J.Aud.Res. 7,401-411 (1967).
Nitschkoff, St., Kriwizkaja,  G., Larmbelastung, akustischer  Reiz  und neurovegetative
    Stdrungen. VEB Georg Thieme Leipzig, (1968).
Nixon, J., Glorig, A., Noise Induced P.T.S. at 2000 and 4000 Hz., J. Acoust.Soc.Amer., 33,
    904-913(1961).
Passchier-Vermeer, W., Hearing loss due to exposure to steady-state broadband noise. Insti-
    tute for Public Health Engineering TNO, Netherlands, (1968).
Pinter, I., Gehorschaden  der Arbeiter verschiedenartiger Betriebe. International Congress of
    the Hungarian Society of Occupational Health, Budapest, (1971).
Pinter, I., A textilipar zajproblemainak munkaegeszsegugyi vetulete. Magyar Textiltechnika,
    24, 329-332, (1972).
Popov, N. F., cit by Nitschkoff and Kriwizkaja (1968).
Spoor, A.,  Presbyacusis  values in relation to noise-induced hearing loss. Int.Aud. 6,  48-57
    (1967).
Ward,  W. D., Glorig, A. and Sklar, D. L., Dependence of Temporary Threshold Shift at 4000
    cps on Intensity and Time. J.Acoust. Soc.Amer., 30, 944-954 (1958).
Wittmaack, K.,  Vergleichende Untersuchungen uber die Luftschall  Luftleitung und Boden-
    schwingung und Korperleitungschadigungen des akustischen Apparates. Arch-Ohr.-Nas.
    u. Kehlk.-Heilk. 102, 96-100 (1928).
Wittmaack, K.,  Uber die Wege der Knochenleitung mit besonderer  Beriicksichtigung der
    Schallschadigung durch  Kochenleitung Acta oto-larying., Stockholm, 19, 105-11
    (1934).
Yokoyama, T.,  Studies  on the occupational deafness.IV. J.Oto-rhino-laryng.Soc.Jap., 66,
    777-783(1963).
                                        327

-------
                           INFRASOUND AND HEARING

                        Charles W. Nixon and Daniel L. Johnson
                        Aerospace Medical Research Laboratory
                      Wright-Patterson Air Force Base. Ohio USA

                                  INTRODUCTION

     Airborne acoustic energy in the frequency region below 20 Hz is arbitrarily described
as infrasound. Human hearing is insensitive  to infrasound  except at exceedingly intense
levels. Technical knowledge on the effects of infrasound on  man is rather sparse; however,
limited information obtained from  a few real life experiences(2 > *7) and experimental in-
quiries clearly suggests that  infrasound  exposures  at high  levels might adversely  affect
man(3'9). The extent to which  infrasound experienced during routine living and occupa-
tional activities might influence human performance, health  and wellbeing is an open ques-
tion. Infrasound is generated by various events in nature as well as numerous  man-made
systems and activities and is experienced by all of us to varying degrees (Table 1).
     Infrasound occurs in nature at  relatively low levels as a result of actions such as winds,
air turbulence, thunder, volcanic activity, storms, large waterfalls and even the impact of
waves on beaches (4>14). Natural activities such as  walking, jogging, and swimming must
theoretically produce  the same low  frequency  pressure fluctuations  as infrasound on the
auditory system. For example, walking or jogging in a way which causes the head  to vary 15
cm in altitude at each step is equivalent to  approximately 90 dB.  Swimming in such a way
that  the ear becomes submerged in  7.5 cm of water during part of the stroke (and  not
submerged otherwise) is equivalent to 141 dB.
     A variety of adverse effects of  naturally  occurring infrasound on human behavior has
been speculated, however essentially no objective  data relating  human response to the
infrasound exposure have been generated. A single study does report^5) a correlation (0.5)
of infrasound exposure with activities such  as automobile accidents, absenteeism in school
children and in unskilled workers during a period of high infrasonic exposure  in a metropoli-
tan area. Although these data are not conclusive, they do sustain the possibility that such
relationship and effects might exist for low level exposures.
     The incidence of  infrasound from man-made sources appears to  be growing both in
terms of intensity and in number of exposures.  Infrasonic energy is found in a wide variety
of sources including air heating and cooling systems, occupational environs  in which com-
pressors, pneumatic devices, air turbulence,  and the like, are  found, in essentially all forms
of transportation systems including the high powered propulsion systems for space vehicles,
and many more(ls'16l  Man-made  infrasound  generally occurs at much higher intensity
levels than  that found  from natural causes, consequently the threat of potential adverse
effects on people is also much greater. Subjective reports of  effects of infrasonic exposure
from other than natural sources have included disorientation, nausea and general unpleasant-
ness as well as a variety of other symptoms  (2>! 5). A  comprehensive study  by Mohr et
al,(9) which examined intense infrasound and  low  frequency effects'on humans  demon-
strated clearcut adverse symptoms which are summarized in Figure 1 (! °). The nature of the
observed behavior indicates that human subjective tolerance  limits for these short duration

                                        329

-------
                                       Table 1
           REPRESENTING SOURCES OF INFRASOUND FOUND IN NATURE AND IN
                               MAN-MADE ACTIVITIES

                               SOURCES OF INFRASOUND
Nature
Source Est Freq Est Max SPL
Thunder
Earthquake
Ocean Waves < 1
Wind: lOOKm/hr 135 dB
25Km/hr 110 dB
Atmospheric Pressure

Fluctuations < 1 100 dB

Volcano







Man -Made
Source
Free Field
Jet Engines
Helicopters
Large Rockets
Diesel Engines

Activities

Running

Swimming
Riding in:
Aircraft
Submarines
Rockets
Automobiles
Helicopters
Est Freq

1-20
1-20
1-20
10-20



<2

<2

<10
5-20
1-20
'' 1-20 '
5-20
Est Max SPL

135
115
150
110



90

140

120
140
145
120
130
(~2 min) exposures may have been very close. The extent to which these symptoms may
disappear as the levels of the exposures decrease has not been defined.
    Auditory system response has long been a criterion for measuring the acceptability of
noise  exposures in terms of both subjective and objective effects. Although infrasound is
inaudible, except at high intensity levels, its potential effect on the human auditory system
and function must  be fully  examined.  Due to such factors as the harmonic contents of
overtones of infrasound, the fact that infrasound rarely if ever occurs free from higher
frequency sounds and the fact that the ear itself causes distortion at high pressure levels, it is
                                        330

-------
                                          EXPOSURE
                                    OBSERVED BEHAVIOR
                                   HOUR
       FREQUENCY
    20
HERTZ
                                            TEST
                                             (10)

                                             (n)
                                Middle ear discomfort, pressure
                                build up; Tickle in ear;
                                One subject had mild
                                Abdominal wall vibration

                                No TTS  1 hour later
                                                          All tolerable; But subjective
                                                          sensations rose rapidly
                                                          for exposures above 145 dB
                                                                No Protection
                                                          One hour exposures;
                                                          No protection; No subjective
                                                          effects observed
                                                            Adapted From
                                                            Mohr et al
                                                            Reference 9
Figure 1. Representative Low Frequency and Infrasound Noise Environs and Human Subjective Responses.
desirable that maximum permissible exposure conditions for infrasound be defined relative
to human auditory function. Knowledge of infrasound and hearing, based on the few studies
reported in the technical literature and on results of investigative efforts recently completed
or underway in our own laboratory, is discussed herein.
               GENERATION AND MEASUREMENT OF INFRASOUND

     Infrasound lies below the range of frequency operation of many high quality items of
instrumentation typically  used in the study of acoustics and psychoacoustics. In order to
accurately measure and analyze infrasound some  additional  instrumentation performance
characteristics must be employed and certain precautions must be exercised (8>16). The
transducer must be capable of responding to DC to  insure acquisition of the total signal. The
remainder of the system must reflect an equivalent low frequency response, especially tape
recorders which must  operate in the FM range. Frequency analysis should be accomplished
on a "per cycle" basis or a small percentage bandwidth filtering of about 5% or less. Sound
level meters are not appropriate for assessing infrasound. It is extremely important that
technically acceptable instrumentation be  used and the frequency response be accurately
                                         331

-------
described. The higher harmonics generated with high level infrasonic signals will likely spill
over into the lower audio frequency regions where they are well above threshold of hearing
(Figure 2). It is clear that this energy is much louder than the fundamental infrasound and
may, in fact, determine the response of the experimental subject. Unless the experimenter is
fully aware of all of the energy present in his test conditions, as described by technically
appropriate  measurement instrumentation, human  responses to these  higher frequency
energies may erroneously be attributed solely to infrasound.
     This matter becomes more important as one considers the design and fabrication of
generators of infrasound for experimental investigations in  which humans will serve as
subjects. On the basis of threshold of hearing values for infrasound shown in Figure 3, a 10
Hz signal at 120 dB would be approximately 18-20 dB above threshold, on average. In order
for this signal to be inaudible at 20 Hz it must be more than 40 dBdown and at 100 Hz it
must be greater than 80 dB down. The very great difficulty in generating an infrasonic signal
CM
   140
    130
 0120
 0)
 miio
 UJ
LJ
en
en
o
I
en
    90
    80
    60
REPRESENTATIVE
   INFRASOUND
THRESHOLD  OF HEARING
"ACTUAL"SOUND
 SOURCE SHOWING
   " SPILLOVER"
                            "IDEALIZED"
                            SOUND SOURCE
                                4   5  6  7  8 9 10
                                 FREQUENCY  IN HERTZ
                                                    20
             30   40 50
 Figure 2. A Graphic Display of the "Spillover" or Upward Spread of Energy Into the Audiofrequency
 Region Frequently Encountered When An  Infrasound Source is Used for Experimental Purposes With
 Humans.
                                       332

-------
    140
"E  130
 3-120
 CM
 S NO
uJ 100
tu
uj 90
a:
V)
to
LJ
or
o.

170
    60
                                            I   I  I  I
                  tx.
            —  BEKESY, MAP
             O   YEOWART, MAP
             O   YEOWART, MAP
             *   YEOWART, TONE THRESHOLD
             A   YEOWART, NOISE THRESHOLD
                                                 I	1
                           3    4   56789K)
                                FREQUENCY   IN  HERTZ
                                                               20
30   40  50
Figure 3. Hearing Threshold Levels for Maximum Audible Pressure (MAP), Minimum Audible Field (MAF)
and for Bands of Noise.

with harmonics  below such values is rather obvious. Consequently, it is not only essential
that infrasound be accurately measured and analyzed, but in addition that interpretation of
human responses be made with knowledge of the total frequency response of the specific
test signal to which the observers were exposed. This problem of "spillover" or upward
spread  of the signal is not as great for measurement of hearing threshold levels as for the
high level energy required for studies of temporary threshold shift (TTS) due to infrasound.
    One of the reasons why so little research has been accomplished at these frequencies
may well be the difficulties encountered in the generation of infrasound signals.  Although
the  number of infrasound investigators is small, a variety of systems have been used for
infrasound generation, ranging  from small pistonphones coupled to an ear to large special
purpose pressure chambers which enclose the whole body. The characteristics of a number
of systems are summarized in Table 2. Data collected from the use of these generators are
sufficient to allow research on  infrasound  to be summarized  in a technical discussion as is
done in this paper.
    The purpose of this paper is to present a review of technical literature and experience
which represents the state of the  scientific understanding of the influence of infrasound on
                                       333

-------
                                                 Table 2

                 TYPICAL EXPERIMENTAL SOURCES USED IN INVESTIGATIONS OF
                                     INFRASOUND AND HEARING
 INVESTIGATOR

Bekesy  (1)
     FACILITY

Thermophone
BENOX, von Gierke
(18)
Mohr. etal  19)
Nixon (lit
Leventhall and
Hood (8)
Yeowart (22)
Pistonphone and
Mercury Manometer
Whole Body
Enclosures; and
Free Field:  Jet
Engine; High
Pressure Air Source;
Low Frequency Test
Chamber

Pistonphones
Whole Body Pressure
Chamber 3x4x6
feet
Monaural'Binaural
Headphones From
0.3m Diameter
Loudspeakers

Whole Body 1200
Litre Cabinet
    GENERATOR

Loudspeaker
Coupled Via
Manometer
Motor Driven
Pistonphone; Manual
Control Manometer

Hydraulic Loud-
speakers; High
Velocity Air Flow;
Low Frequency
Siren
Coupled to Ear via
Closed Tube
Four 15 meter Diam.
Loudspeakers: 300 W
Amplifier
Earmuffs; Drivers
Worked Into Volume
of 1 Litre
                                          Six 0.46m Diameter
                                          Loudspeakers on
                                          Sides of Chamber
  OPERATION

Beating of Two AC
Inside Thermophone
Capsule; Frequency
Response Down to
IHz

Alternating
Pressures Up to
50 Hz

Normal Operating
Modes for Various
Devices and
Facilities
Motor Driven
Alternating
Pressures

Operates as a
Helmholtz
Resonator Tunable
Over Range of 3 Hz
to 18 Hz

Response is Flat up
to 200 Hz
                       Electrodynamic
   PERFORMANCE

104 - 105 dynes/cm2
Sound Pressure
165 dB Alternating
Pressure;  180 dB
Static Pressure

Discrete Tones and
Bands of Noise at
Levels of 150
154 dB
165 dB + Alternating
Pressure
145 dB for Single
Frequency; 126 dB
For Noise Band
Maximum SPL is
ISOdBat IHz
                        Maximum SPL is
                        140 dB
Johnson  (7)
Whole Body Chamber
55 Cu. Ft.
Hydraulic Driven
6 Ft. Piston and
1.5 Ft. Piston
Alternating
Pressures, 0.5-
30 Hz
6 Ft. Piston: 172 dB
(0.5-10Hz). Falling
to!58dB($30Hz

1.5 Ft. Piston:
144 dB d- 10 Hz)
Falling to 135 dB
(?)20 Hz
                                                   334

-------
the human auditory system. The body of the report, which comprises the basic data review,
is  organized into four functional areas, (1) hearing threshold  levels  for infrasound, (2)
temporary hearing loss (TTS) due to infrasound, (3) additional effects of infrasound on the
auditory mechanism, and (4) infrasound effects on the  speech  reception aspect of voice
communication. These data are discussed  in terms of exposure guidelines, and tentative
limiting noise levels for infrasound exposures are recommended.

                          AUDITORY SYSTEM RESPONSE

Hearing Threshold Levels

     In order to evaluate effects of infrasound on the human auditory system and function,
the nominal  response  of the system must first be determined.  Perhaps one of the most
long-standing descriptions of hearing threshold levels (MAP) for acoustic energy below 20
Hz is that  of Bekesy('X A number of other investigators, at different times and using a
variety  of  instrumentation have  independently measured infrasound  hearing thresholds.
Most recently this has been accomplished by WhittleO9) and YeowartC20). The hearing
threshold levels from a series of studies by Yeowart are  compared in Figure 3. The agree-
ment among  the  various values is very good and  it provides confidence that the general
sensitivity curve for the human ear for this frequency region has been described.
     Measured MAP  values for infrasound are  also  contained in Figure 3. The classical
MAP-MAP  difference of 3 dB for audio frequencies is also present for infrasound (22).
Hearing threshold level values for  noise bands in this frequency region are also shown (2').
It  is evident  that no significant  difference between tone and  noise  threshold data are
observed between about 30 Hz and 100 Hz. The noise thresholds are significantly lower, by
about 4 dB, for frequencies below about  16 Hz. The greater sensitivity of the ear for the
bands of noise is attributed to detection of the peak factors present in the noise signal.

Temporary Threshold Shift (TTS)

     A quantitative relationship between human exposure to infrasound and hearing loss is
not well established.  Very few investigations of this phenomenon are found in the technical
literature, partly  because of a general low level of interest in this frequency region  and
partly because of the problems associated with the measurement of hearing thresholds for
infrasound  as well as the general inability to produce infrasound exposures free of audible
overtones. A  few investigators who have ventured into infrasound research with cognizance
of the latter problems are identified in Table 3.
     For threshold  determinations the general  approach has  been to measure effects of
infrasound  on  the standard audiometric test frequencies instead of for lower frequency
signals.  The question of audible overtones is not at all clear because most reports do not
contain spectral representations of the test signals. It cannot be determined from these
reports if the stimulus was truly infrasound  or if it was a multiple-component signal with the
fundamental  an infrasonic frequency. Nevertheless, it is a reasonable assumption that the
adverse effects of the infrasound alone would be no worse than those of the infrasound plus

                                         335

-------
overtones. This review considers the data as it is reported and does not attempt to critically
analyze the acoustic exposure while recognizing that actual observed effects may have been
highly influenced by energy above 20 Hz.
     Some early observations of possible infrasound effects on hearing were not described in
terms of audiometric test frequencies. Tonndorf C1 7) reports on the effects of infrasound in
the diesel rooms of submarines on the hearing of crew members. Depression of the upper
limits of hearing were demonstrated by decreased time  periods during which tuning fork
tests were audible. Recovery occurred after various intervals of time outside the diesel room.
Mohr et al(9) exposed subjects to infrasonic signals, both pure tones and noise bands, for 2
minutes or less at levels of 150-154 dB. Audiometry was not performed immediately follow-
ing the  exposures; however, measures taken about one hour later showed no ITS. In the
latter work an exposure signal was experienced only  once, while on-board submarine expo-
sures were experienced daily.
     Jerger (6) exposed 19 males to repeated three-minute signals of from 2-12 Hz at levels
of 119 dB to 144 dB SPL. ITS in the range of 3000-8000 Hz was observed in 11 of the 19
subjects for exposures of  137 to 141 dB. All ITS values were small, ranging from 10-22 dB.
The author indicates  that  the 7-12 Hz signals at 120-144 dB did produce considerable
masking over  the 100-4000 Hz range. It seems likely that some of the measured TTS was
caused by the masking signal.
     In  our laboratory a  number of studies of TTS and infrasound have been conducted
using  pistonphones and a large pressure chamber as signal generators. Using a pistonphone
coupled tightly  to the ear via an earmuff, Nixon 0 1) investigated effects of 14 Hz at 140 dB
and 18 Hz at 135 dB for 30 minute exposure durations on hearing threshold levels. Some
subjects experienced no changes in hearing due to the exposures  while others showed
various amounts of TTS, with one subject showing 20-25 dB at one test frequency.
    In another series of investigations, Johnson (7) measured the effects of auditory expo-
sures of 135 dB to 171 dB  at 0.5 Hz to 12 Hz and whole body exposures of 135 dB to 144
dB  at 1 Hz to 20 Hz. A pressure chamber which provides whole body exposures to infra-
sound at levels as high as 172 dB was used to generate  the stimuli. Exposure durations varied
from 26 sec of 7 Hz at 171 dB to 30 min of 4, 7 and 12 Hz at 140 dB. The various exposure
parameters, effects on hearing, if any, and recovery are itemized in Table 3.
    It is clear from the data contained in  Table  3 that TTS has been measured following
infrasound exposures at moderately intense levels. The observed changes in hearing thresh-
old  levels have  been small and recovery of pre-exposure hearing levels has been rapid for
the  few situations in which TTS did occur.
    Susceptibility,  Susceptibility of ears to infrasound-induced TTS appears to be gen-
erally the same  as TTS induced by higher-frequency energy. Amount of TTS induced by a
specific  exposure or whether or not TTS occurs,  both within and between subjects, show
about the same variability  as for audio frequency  exposures. Data are not available to
determine if susceptibility to infrasound due to age or to sex is different from that due to
exposures to audio frequency energy.
    Middle Ear Ventilation. As will be discussed later, infrasound exposure at levels suffi-
cient  to induce TTS also produces retraction of eardrum  membrane. The  efficiency of
middle ear transmission of energy to the inner ear is reduced when this system is retracted.
                           j
                                        336

-------
As  an investigator, one  must consider the advisability of having experimental subjects
periodically  ventilate  the middle ear system during studies of infrasound since different
effects would be expected from exposure of a retracted vs non-retracted drum-membrane
system.  Regardless of the exposure, it is critical that the middle ear system be adequately
ventilated prior  to measurement of post-exposure hearing threshold levels. A retracted
middle ear system will show reduced sensitivity which may be attributed to sensorineural
effects.


Other Effects on the Auditory System

     Infrasound  may  stimulate the auditory system at rather low levels  so as to be un-
detected or  at levels of sufficient  magnitude to cause aural pain. During whole body and
aural exposures to infrasound, particularly below about 5 Hz, at levels of 120 dB and above,
subjects may report a sensation that the eardrum membrane is being mechanically massaged.
At the lower intensity levels, perception of the sensation is not unpleasant and it becomes
less noticeable after a little time. At  higher intensity levels, subjects may report the sensa-
tion as being quite unpleasant and disliked; however,  this too appears to dissipate during
continuation of the  exposure period. At one time, pneumatic massage of the eardrum-
middle ear system was a common otological practice. It has been reported that mild massage
under properly controlled conditions is beneficial to the ear. However,  excessive mechanical
massage could be  detrimental,  presumably  because the mechanical  displacement due to
intense infrasound is so much larger than during typical listening situations. Massage of the
drum membrane system by infrasound, at rather high levels and/or for long duration expo-
sures, has produced effects clearly recognized at the drum membrane by investigators^' ' l).
    Pressure Build-up. Experimental subjects  almost  universally  describe a sensation of
pressure buildup in the ear shortly after initiation of infrasound exposure. This sensation is
reported by  many subjects at 126 dB and by virtually all persons at 132 dB. This fullness is
experienced  for  both  aural and whole body exposures. The sensation remains throughout
the exposure and persists for some time afterwards in many subjects. Ventilation of the ear
during exposure  may relieve the sensation of fullness; however, it is only temporary, for the
feeling of pressure quickly returns. This phenomenon appears  to occur a little earlier than
injection of the drum membrane is observed.
     Vascular Infection. A vascular injection  of the eardrum  membrane may be observed
during and following exposure. This injection is  similar to  that produced by therapeutic
massage  of the  drum membrane. The degree of injection may be slight to severe in which
case congestion appears all along the  handle of the malleus and in the folds. Although slight
injection may be caused by many different factors, it is not considered "abnormal". Severe
congestion must be recognized as a positive indication of overexposure.
     Drum Membrane Retraction.  The  cyclical displacement  of the  drum membrane, in-
ward  phase, during infrasound exposure appears to force gases from the middle ear cavity
out through the collapsed Eustachian tube. The negative pressure created by this action is
not automatically equalized  on the alternate phase of the cycle and drum membrane retrac-
tion will likely  occur. The effect of retraction on hearing is to reduce transmission of
acoustic energy to the inner ear and  in this mode is likely beneficial during exposure if not

                                        337

-------
allowed to become severe. There is some evidence^1'»'7) that drum membrane retraction
acquired during infrasound exposures may not be due entirely to a negative pressure in the
middle ear. During exposures to 14 Hz at  140 dB, all subjects  exhibited mild retraction.
After termination of the exposure each individual ventilated the middle ear system via
Valsalva, under the observation and guidance of an  otologist. Following confirmed ventila-
tion or pressure equalization, the retracted drum condition remained and persisted for some
time. An unexplainable function of the middle ear muscle system is believed to account for
the persistent retraction; however, this assumption  remains  to be investigated. One conse-
quence of this retraction phenomenon is that even with positive ventilation following expo-
sure, prior to audiometric testing, some retraction  may remain and indicate a slightly
elevated threshold.
     Drum Membrane Scar Tissue. The potential adverse effects of long duration exposures
of this nature have not been examined in the laboratory. Tonndorf (! 7),  in describing the
condition of drum membranes  of submariners exposed daily to diesel  room infrasound
found the formation  of cicatrized  tissue and loss of elastic fibers. Although a direct causal
relationship was not established, the incidence of these conditions among  the exposed crew
members clearly exceeded what  would normally be  expected in that  population from non-
infrasound and noise factors.
     Pain. Pain in the ear is easily recognized by individuals and is related to mechanical
displacement of the middle ear system beyond its limits of normal operation. Aural pain is
not related to sensitivity, as evidenced by the fact that normal and hard-of-hearing persons
have the same average aural pain thresholds.  Likewise,  it is not associated with sensorineural
hearing loss which can become  very  severe without any experience of pain. At infrasonic
signals, pain may be experienced at levels which pose no risk to the hearing function.
     Thresholds for aural pain are summarized in Figure 4. The data of Bekesy (* )and from
the BENOX report (13) are highly consistent. One datum on the figure represents a condi-
tion where one subject reported pain from a pistonphone exposure of around  140 dB at 14
Hz, which disappeared after a few minutes exposure at that level. It  appears that the pain
threshold might be elevated a few dB, 2-3 dB, with high-level exposure experience of the
subject.  The pain threshold appears to be about 140 dB around 20 Hz rapidly increasing to
about 162 dB at 2  Hz. Pain produced by static  pressure  on the ear, either positive or
negative, appears between 175 dB  and 180 dB.  Any form of aural pain during infrasound
exposure must be considered  an indicator that the  tolerance  limits of the mechanical
systems  have been reached and the exposure  should  be terminated and avoided in the
future.

                  VOICE COMMUNICATION-SPEECH RECEPTION

     Effects of infrasound exposure on voice communication are  generally considered to be
those affecting the talker. High-intensity  infrasound may influence various organisms and
functions involved in speech production. Amplitude modulation during speech production is
obvious  at very low frequencies as a result of the respiratory cage or chest being driven by
the infrasound, and is reported by essentially all exposed persons. Choking, coughing, gag
sensations,  chest wall vibration and modulation of respiratory rhytm have been reported for

                                        338

-------

-------
Hearing Protection

     Human exposure to intense infrasound may occur at hearing tolerance limits where
potential hazards exist or at lower levels which pose no  hearing risk but are subjectively
disagreeable. Effective hearing protection  is highly desirable in each of the situations
described. Classically, insert hearing protection  has provided good performance across the
audio  frequency range whereas earmuff protector performance decreases with  decreasing
frequency.
     Subjective reports  of ear protector effectiveness in intense  infrasound indicate that
good insert-type earplugs provide appreciable attenuation of the acoustic energy. Earmuff
type protectors appeared  to provide  negligible  protection  and on occasion appeared to
amplify  the noise under the muff. Earmuffs, which are  suspended from lightweight spring
tension headbands,  were noticed to visibly vibrate against the sides of the subject's head
during infrasound  exposure.  When worn over insert earplugs, earmuffs appeared to  add
attenuation obtained by the wearer.
     An experimental investigation of earmuff effectiveness in infrasound, using both a
subjective and a physical method, confirms the subjective observations reported above(* 2).
Good earmuff protectors provide about 10 dB of sound  protection between 20 Hz and  100
Hz and very little protection in the infrasound region.  For optimum protection in sound
fields below  20 Hz, good  insert earplugs are  recommended for intense exposures of long
duration.


Limiting Levels of infrasound

     Limiting levels of infrasound exposure effects on the auditory system must consider in
addition to potential hearing loss,  mechanical effects on the middle ear system-including
pain, speech reception and discomfort. The available knowledge from which limiting levels
may be  formulated  comes from  experience in intense infrasound and  from  laboratory
investigations.
     The  purpose  of defining relationships of infrasound  exposure  to  auditory system
effects, is  to allow potential risk  to  be determined on  the basis of descriptions of the
physical stimulus, There is somewhat of a problem in depicting the three stimulus variables
of importarice-SPL, duration and  frequehcy-in a simple fashion. Consequently, we have
adopted  a method  of representing exposures in terms  of level and  of number of cycles
(frequency times time) as parameters. Although the utilization  of this procedure does not
extend to the extremely low frequencies or to high frequencies it does appear to be a very
good approximation for representing exposure for the range 1 Hz to 20 Hz.
     The proposed equal risk formulation based upon adoption of this method is:

                         SPL = 10 log t+10 log f+base SPL

     A number  of experimental subjects have experienced exposures to 10 Hz at 144 dB for
durations of 8  minutes, via auditory  only or whole body presentation of the stimulus.
Although not necessarily enjoyable, no adverse effects  have been observed which would

                                       340

-------
indicate that these exposure conditions are threatening or harmful.  Accepting this set of
conditions as a base acceptable exposure, the formulation becomes:
                                           t           _f_
                    Limiting SPL = 10 log 8 min + 10 Log 10+144

     Various infrasound exposures conducted in  our laboratory are displayed in Figure 5
along with the curve which  represents the Limiting-SPL formulation. The Limiting-SPL
curve shows reasonable agreement with the experimental data collected to date, as well as
the proposed criterion. The same data are presented in a more conventional form in Figure
6. It is clear that exposure durations of 8 minutes for levels up to 150 dB caused essentially
no TTS. Actually, over 100 ear-exposures are shown for this duration range and  only two
experienced a mild TTS of 8 dB, with immediate recovery.
     Limiting levels for frequencies of 0.5 Hz to 20 Hz and  exposure durations of 0.5 min
to 1440 minutes in terms of "Limiting  SPL" are displayed in Table 3. On the basis of
experience  to date and lack of more complete data, it is essential that the table  values be
qualified. It is clear that non-auditory, whole body effects of infrasound occur at levels of
UJ

UJ
3j 140
v>
V)
UJ
SOUND PR
g
   120
                                        t   ^
                                   CHINCHILLA
                                        ( DRUM MEMBRANE
                                         AND MIDDLE EAR
                                           DAMAGE)
                                                          O WHOLE BODY EXPOSURES
                                                          A AURAL EXPOSURES
                                                          O SOME  TTS OBSERVED
                                                                     SUBMARINERS
                                                                    (TONNDORF)
        EXPOSURE LEVEL     ._
        LIMITED BY           <->
        NON-AUDITORY EFFECTS
                                 A

                                 AO

                                 A
                                      i
                      10
                                    100
1000
                                           111
                                           60
10,000
 Figure 5 Various Laboratory Infrasound Exposures in Terms of Level and Number of Cycles (frequency X
 time) and a Limiting Sound Pressure Level Curve Based on the Formulation:
                     t         _f_
 Limiting SPL = 10 log 8 min + 10 log 10 + 144
                                         341

-------
    30
 CO
 LJ
rr
D
Q
iu
o:
z>
CO
p
Q_
X
LJ
    25
    20
     15
     10
     0
             EXPOSURE
                      LEVELS
          A GREATER THAN ISOdB
          O 140-149 dB
          Q !30-l39dB
          O 120- 129 dB
         FILLED SYMBOLS INDICATE
         THAT SOME TTS OCCURRED
       D
CSD
  mn
                                                                        QD
                               3     4    5678910
                           FREQUENCY  IN  HERTZ
                                                                  15   20
                 30
Figure 6 A  Conventional Display  of Individual Exposures Recorded in Our Laboratory in Terms of
Frequency and Duration With Level as the Parameter. Solid symbols indicate that some TTS was observed;
the vast majority of exposures show no TTS.
150 dB and above. Consequently, whole body effects impose limitations at levels considered
to be safe for the auditory system.
    The proposed limiting noise levels, in more general terms, which may be considered as
acceptable are 150 dB at 1 Hz - 7 Hz,  145 dB at 8 Hz - 11 Hz and  HOdBat 12 Hz- 20 Hz.
These levels apply to discrete  frequencies or octave bands centered about the stated fre-
quencies. Maximum exposure duration is eight minutes with 16 hours rest between expo-
sures.  The use of good insert earplugs may increase the permissible  levels by 5 dB for the
same exposure times by reducing the aural contribution to the overall response. Earplugs are
strongly recommended for all  intense  infrasound exposures to minimize subjective  sensa-
tions.  Levels above 150 dB should be avoided even with maximum hearing protection until
additional technical data are accumulated.
    The normal levels at which aural pain is induced  by infrasound correspond  closely to
the limiting  values shown above for the 10 Hz  and  20 Hz frequency regions. At  2 Hz the
value is much higher at about 162 dB. Consequently, the threshold  regions for aural pain are
compatible with the proposed values and do not  impose any additional limitations.
                                        342

-------
                                             TABLE 3

       A SUMMARY OF STUDIES OF TEMPORARY HEARING LOSS FOLLOWING EXPOSURE
                                        TO INFRASOUND
INVESTIGATOR
 Tonndorf (17)
Mohr. et al  (9)
Jerger, et al  (6)
Nixon (11)
Nixon (11)
Johnson (7)
    EXPOSURE
Submarine Diesel Room
10 Hz -20 Hz.  No Level Given
Discrete Tones; Narrow Band
Noise in 10 Hz - 20 Hz Region.
150 - 154 dB Exposures of
About 2 Minutes

Successive 3 Minute Whole
Body Exposures. 7-12 Hz;
119- 144 dB

Pistonphone Coupled to Ear
viaEarmuff.  18 Hz at 135 dB.
Series of 6,  5 Minute
Exposures Rapid in Succession

Pistonphone Coupled to Ear
viaEarmuff.  14 Hz@140dB.
Six Individual Exposures of
5,  10,  15, 20, 25 and 30
Minutes

Ear Only:   Pressure Chamber
  Coupled to Ear via Tuned
  Hose  and Muff
171 dB  (1 - 10 Hz) 26 sec.  Is
168 dB  (7 Hz) 1 min, Is
155 dB  (7 Hz) 5 min, 2s
140 dB  (4,  7, 12 Hz)
  30 min, Is
140 dB  (4,  7, 12 Hz) 5 min.
  8s
135 dB  (.6.  1.6, 2.9 Hz)
  5 min. 12s
126 dB  (.6.  1.6, 2.9 Hz)
  16 min, 11s
  HEARING RESPONSE
Depression of Upper Limits of
Hearing as Measured by Number
of Seconds a Tuning Fork was
Heard - No Conversion to MAP

No Change in Hearing Sensi-
tivity Reported by Subjects;
No TTS Measured About One Hour
Post Exposure

TTS in 3000-6000 Hz Range
For 11 of 19 Subjects (TTS of
10dB -22dB)

Average TTS of 0 - 15 dB After
30 Minute Exposures
Three Experienced Subjects
NOTTS in One; Slight TTS in
One; 20-25dBTTS in One
    RECOVERY
Recovery in Few Hours
Outside of Diesel Room
Recovery Within Hours
Recovery Within 30
Minutes
Recovery With in 30
Minutes
                                               NOTTS
                                               NOTTS
                                               NOTTS
                                               14- 17dBTTS

                                               8dB TTS for 1 Subject

                                               NoTTS

                                               NOTTS
                               Recovery Within 30 min

                               Recovery Within 30 min
                 Whole Body:  All Exposures. 2s:
                   8 min at 8 Hz at SPL'sof        NoTTS
                    120. 126, 132. 138
                   8 min at 1, 2, 4, 6, 8, 10       NoTTS
                   .Hz at 144 dB
                   8 min at 12.  16. 20 Hz at        No TTS
                    135 dB to 142 dB
                                               343

-------
     Injection and retraction of the eardrum membrane may occur at values well below the
 limiting levels shown above. No potential risk to the auditory system is expected to develop
 because of the brief durations of exposure permitted by  the  limits. No change in the
 proposed values is indicated with respect to injection and retraction.
     General face-to-face speech reception in  experimental noise exposures at  the  same
 levels  as the limiting values was  considered acceptable by participants in those studies.
 Headphone reception of speech during laboratory studies has been observed to involve some
 slight  difficulties at levels of about 145 dB. Data  are insufficient at this time to justify
 lowering the proposed  levels at exposures of  7 Hz and below  on  the basis of potential
 interference to speech reception with headphone listening.
     The limiting values proposed  in Table 4 are estimates based upon the various kinds of
 responses made by the auditory system to infrasound. Individuals observing these limita-
 tions may be expected  to display no symptoms of overexposure or abuse of the human
 auditory system.


                            FUTURE  CONSIDERATIONS

     The few studies-reported in the literature  and  a series of studies accomplished in our
 laboratory have been used to formulate tentative exposure criteria for infrasound. In order
 to  more  firmly establish  limiting levels additional information is required  on specific
 responses of the auditory system in infrasound.
     The  parameter of exposure duration appears to be a significant one  requiring better
 definition. The majority of work completed on TTS has been limited to exposures of 8
 minutes and less duration. Essentially no adverse effects have been observed. Conversely a
 small number of exposures of 30 minutes show a reasonably high incidence of TTS. The role
 of duration between  10 minutes and 30 minutes certainly requires better definition for
exposures up to 150 dB.
     Voice communication in infrasound, both speech reception and production, requires
 investigation. Speech reception via headphone listening has already been identified as a
 potential  problem.  Face-to-face communication  in exposures  longer than  one  to  two
 minutes  have not  been  considered.  An important  aspect is the degree to which  the ear
 distorts during intense  infrasound  exposure and its effect on  normal speech  reception
 (300-2000 Hz range).
     Of particular interest to the authors is the  observation that drum membrane retraction
 which  was incurred  during infrasound  exposure  did not disappear following confirmed
inflation of the retracted middle ear system. This implies retraction due to some mechanism
other than negative middle ear pressure, possibly the middle ear  muscle system. Identifica-
 tion of the mechanism sustaining retraction is in  order.
     Additional information is desirable on relative effects of aural vs whole body exposures
 to specific stimuli. Although it is asserted that  middle ear pressure equalization will occur
 for whole body exposures  but not for  aural exposures, this appears unlikely due to the
 relatively  slow action of the eustachian tube in its automatic mode.  It is conceivable that
 individuals who experience  relatively long duration exposures will learn to equalize pressure
 buildup as it occurs in essentially an unconscious manner.

                                        344

-------
                                         Table 4

     LIMITING VALUES FOR INFRASOUND EXPOSURE AS A FUNCTION OF FREQUENCY
                                    AND DURATION
                                     MAXIMUM PERMISSIBLE EXPOSURES
                                                  FREQUENCY in hertz

.5
1
2
4
8
10
20
30
Ihr. 60
120
480
1 day 1440
0.5
169
166
163
160
157
156
153
151
148
*>S-^ «,
|s$
i
166
163
160
157
154
153
150
148
145
t *. •.
•«£ ff r * *f '••'<
'^131'*"
*>! Si
2
163
160
157
154
151
150
147
145
142
'•/' - ?>
* 139
^ 133
V*128./
4
160
157
154
151
148
147
144
142
139
*- 136''
130,
- v?%
8 10
159 156
154 153
151 150
148 147
145 144
144 143
141 140
139 138
136 135
^ 133 132
:" 127, -126,
-, 122 / I?, &!
12 16 20
155 154 153
152 151 150
149 148 147
146 145 144
143 1.42 141
142 141 140
139 138 137
137 136 135
134 133 132
- 131 v 130 * 129
- 125, * }' 124 ^ 123 ,v ,
;;'/i2o;^ ,"ii9 j)iiy ^
  Table of Recommended Maximum Permissible Exposures Based on SPL    = 10 log   t      +10log_f_  + 144
                                                      max      Trtim"         10

   No whole body exposure is recommended over 150 dB for frequencies greater than 0.5 Hi.  Shaded area is an
   extrapolation and should be used with care.
     Improvements in instrumentation and test facilities for evaluating all infrasound effects
on man are also needed, particularly for infrasound exposure signals which typically have
higher frequency energy at levels well above  threshold. It is equally as important to accu-
rately and conveniently  measure hearing at these very low frequencies in order that possible
changes due  to  exposure, which will be missed by testing only audio frequencies, may be
identified.
                                          345

-------
                                    SUMMARY

     Infrasonic energy, both natural and man-made, is present at varying levels in a wide
variety of environments occupied by man. Efforts to describe natural and potential adverse
effects of these exposures are just beginning. The small amount of data available have been
reviewed.
     (1) Nominal infrasound hearing threshold levels are reasonably well defined.
     (2) Hearing function for infrasound appears equivalent to hearing function for audio-
frequency.
     (3) Tentative limiting levels of infrasound exposure are recommended on the basis of
measured  effects on hearing threshold and on other characteristics of the auditory system.
     (4) Improvement in the  tentative  criteria and expansion of its scope,  will require
additional research on factors such as TTS, voice communication and instrumentation.
     This  report has been limited to acoustic energy below 20 Hz; however,  most of the
questions  may relate equally as well to energy from 20 Hz to 50 Hz and even 20 Hz to 100
Hz. Also,  the review has been restricted  to effects only on the auditory system. The overall
technical area of infrasound effects on man is equally or even more in need of information
than is the specific auditory system effects area.
                                  REFERENCES

 1. von Bekesy, Georg, Experiments in Hearing. McGraw-Hill, 1960.
 2. Bryan, M.  E., Annoyance Effects Due to Low Frequency Sound. Proceedings of Fall
    Meeting of British Acoustical Society, 71.109, November 1971.
 3. Evans, Margaret J., Infrasonic Effects on the Human Organs of Equilibrium, Proceed-
    ings of Fall Meeting of the British Acoustical Society, 71.104, November 1971.
 4. Fehr, Vri, Measurements of Infrasound from Artificial and Natural Sources. Journal of
    Geophysical Research, Vol. 12, No. 9, pp. 2403-2417, May 1967.
 5. Green, J. E. and F. Dunn, Correlation of Naturally Occurring Infrasonics and Selected
    Human Behavior. Journal of ASA, 44(5), 1456, 1968.
 6. Jerger, J., B. Alford, A. Coats and B. French. Effects of Very Low Frequency Tones on
    Auditory Thresholds. Journal of Speech and Hearing Research, 9, 150-160, 1966.
 7. Johnson, D. L (Unpublished Data.)
 8. Leventhall, H. G. and R. A. Hood. Instrumentation For Infrasound, Proceedings of Fall
    Meeting of British Acoustical Society, 71.101, November 1971.
 9. Mohr, G. C, J. N. Cole, E. Guild and H. E. von Gierke, Effects of Low Frequency and
    Infrasonic Noise on Man. Aerospace Medicine, 36, 817-824, 1965.
10. Nixon Charles W., Some  Effects of Noise on Man. Proceedings of 1971 Intersociety
    Energy Conversion Engineering Conference, Boston, Mass, August 1971.
11. Nixon, C. W. (Unpublished Data.)

                                       346

-------
12.  Nixon, C. W., H. K. Hille and L. K. Kettler, Attenuation Characteristics of Earmuffs at
    Low Audio and  Infrasonic Frequencies,  Aerospace Medical Research Laboratory
    Technical Report No. 67-27, May 1967.
1-3.  Pickett, Ji  M.,t Low Frequency Noise and Methods  for Calculating Speech Intelligi-
    bility. JASA, 31.9, 1259, Sept 1959.
14.  Stephens, R. W. B., Natural Sources of Low Frequency Sound. Proceedings of Fall
    Meeting of British Acoustical Society, 71.105, November 1971.
15.  Stephens, R. W. B., Very Low Frequency Vibrations and Their Mechanical and Biologi-
    cal Effects,  Seventh International Congress on Acoustics, 26G1, Budapest, 1971.
16.  Tempest, W., Low Frequency Noise  in Road Vehicles. Proceedings of Fall Meeting of
    British Acoustical Society 71.106, November 1971.
17.  Tonndorf, M. D., The Influence of  Service on Submarines On The Auditory Organ.
    (Personal Notes.)
18.  von Gierke, H. E., H. Davis, D. H. Eldredge and J. D. Hardy., Aural Pain Produced by
    Sound, Benox  Report, Contract N6  ori-020, Task Order 44, ONR Project Nr. 144079,
    University of Chicago, December 1953.
19.  Whittle, L.  S.,  The Audibility of Low Frequency Sounds, Proceedings of Spring Meet-
    ing of the British Acoustical Society,  April 1971.
20.  Yeowart, N. S., M. E. Bryan  and W. Tempest, The Monaural M.A.P.  Threshold of
    Hearing at  Frequencies From 1.5 to 100 c/s. J. Sound and Vibration, 6(3), 335-342,
     1967.
21.   Yeowart, N.  S., M. E.  Bryan and  W. Tempest.  Low Frequency  Noise Thresholds.
     Journal of Sound and Vibration, 9(3), 447-453,  1969.
22.   Yeowart, N.  S.,  Low Frequency Threshold Effects. Proceedings of Fall Meeting of
     British Acoustical Society, 71.103, November 1971.
                                        347

-------
      THE EFFECTS OF AIRBORNE ULTRASOUND AND NEAR-ULTRASOUND

                                  W. I. ACTON,
                    Wolfson Unit for Noise and Vibration Control,
                      Institute of Sound and Vibration Research,
                       The University, Southampton, England.

INTRODUCTION

     Ultrasonic devices are now widely used in production industries for a variety of pro-
cesses, including drilling, dicing, soldering, cleaning, welding plastics, emulsification, mixing
liquids, initiating free-radical chemical reactions and so on.  Relatively low ultrasonic fre-
quencies in the range 20 to 40 kHz are generally employed for mechanical reasons, although
small apparatus has been encountered operating at a frequency as low as 16 kHz. Measured
sound pressure levels  at the operator's working position rarely exceed  110 to 120 dB
(ACTON,  1968, GRIGOR'EVA, 1966a, KNIGHT,  1968).
     These sources invariably emit air-borne noise, not only at the operating frequency and
its harmonics, but also at sub-harmonics which may be audible. Furthermore, processes
involving liquids, e.g.'washing, mixing and using a liquid-suspension of abrasive powder, are
accompanied by the phenomenon of "cavitation". This is thought to involve the formation
of bubbles of gas previously held in solution around nuclei such as the abrasive particles in
suspension or dirt on objects being cleaned. The bubbles grow  until they reach  a resonant
size, when they oscillate with an increasing amplitude until they implode. Non-linear radial
and  surface oscillations  of the  gas-filled bubbles may be responsible for more tonal noise,
and  the violent collapse of cavities is responsible for the generation of high levels of random
noise at frequencies of approximately 3 kHz upwards (WEBSTER, 1963).
     Ultrasonic frequencies used in medicine for cell destruction are generally in the range 1
to 3 MHz and for  diagnosis in the range  1 to 20 MHz. Diagnostic exposures were not
considered likely to be  potentially harmful by HILL (1970). As these frequencies do not
appear to have found widespread industrial application yet, they will not be  considered
further.


HISTORICAL REVIEW

     When jet aircraft  were introduced, the term "ultrasonic  sickness" was coined (DAVIS,
1948, PARRACK,  1952)  to  cover a complex  of  symptoms which included excessive
fatigue,  headache, nausea, vomiting, etc., exhibited by  personnel working in their vicinity.
ALLEN,  FRINGS and RUDNICK (1948) observed a loss of the sense of equilibrium  or
slight dizziness on exposure to  intense (160 to  165 dB) high  frequency, audible sound, and
unsteadiness and dizziness have been reported in personnel exposed without ear defenders
and  at close range to the noise  from the air intake of jet engines (DICKSON and WATSON,
1949, DICKSON and CHADWICK, 1951). The latter authors suggest that this  might be due
to vestibular  disturbances  caused  by intense acoustic stimulation. In any case published
analyses of jet engine noise show that radiated airborne ultrasound is not present at signifi-

                                       349

-------
 cant intensities. As ultrasonic frequencies are rapidly absorbed by air, intense ultrasound
 would only be encountered in  regions where approach was  normally barred by safety
 considerations (DICKSON, 1953, GUIGNARD, 1965). Finally, PARRACK (1966) stated
 that "ultrasonic sickness" was "largely psychosomatic in origin", although, of course, the
 other effects had been real enough.
     Then  followed a period when  the possibility of effects  from exposure to airborne
 ultrasound was dismissed. DAVIS, PARRACK and ELDREDGE (1949) stated that there
 was no evidence that airborne ultrasonics themselves constituted a hazard to the hearing,
 and, in general, high-intensity audible noise was potentially more hazardous. PARRACK
 (1952) concluded that there was no hazard from laboratory sources of airborne frequencies.
     A note of caution was introduced in the mid-1950's. CRAWFORD (1955) reported
 that some laboratory workers had suffered unusual fatigue,  loss of equilibrium, nausea and
 headaches which persisted after the exposure had ceased, and "some loss of hearing in the
 upper audible frequencies", although this was not substantiated by audiometry and was
 probably based on purely subjective observations. Systematic  research into the biological
 effects of ultrasound  was started in Russia in the late 1950's (GORSLIKOV, GORBUNOV
 and ANTROPOV, 1965),  but some of the translations and reviews available in the West
 should be viewed  critically,  as effects observed with liquid- or solid-coupling to the ultra-
 sonic source have apparently been attributed to airborne ultrasound.
     There have been a number of audiometric temporary threshold shift investigations
 involving  laboratory  (DOBROSERDOV,  1967),  PARRACK,  1966, SMITH,  1967) and
 industrial (ACTON and CARSON, 1967) exposure, and at least one retrospective permanent
 threshold shift investigation (KNIGHT,  1968). Subjective effects have been correlated with
 measured exposure levels by SKILLERN (1965) and ACTON and CARSON (1967). Finally,
 a number of exposure criteria for the prevention of both  auditory and subjective effects
 have been proposed, and these do not differ widely (ACTON, 1968, GRIGOR'EVA, 1966a,
 1966b, GORSLIKOV et al, 1965, ROSCIN et al, 1967).

PHYSIOLOGICAL EFFECTS

     One difficulty in reviewing the physiological effects is  to be certain that the exposure
was to airborne ultrasound. Another difficulty is that many,  often bizarre, effects have been
reported without  the exposure  level being quantified. Consequently, this section of the
review has been limited to references which specifically stated exposure conditions, and
these have been summarized in Fig. 1.
     In the case of airborne ultrasound, the acoustic mismatch between the air and tissue
leads to a  very poor transfer of energy. The effects on small fur-covered afnimals are more
dramatic because the fur acts as an impedance-matching device, they have a greater surface
area to mass  ratio, and they have a much lower total body  mass to  dissipate the  heat
generated than man.  Furthermore, the lower ultrasonic frequencies may well be audible to
these animals,  and the exposures have  been to high sound pressure levels. Therefore, the
effects on small laboratory animals cannot be extrapolated directly to the human species.
     Mild biological changes  have been observed in rats and rabbits as a result of prolonged
exposure to sound pressure levels in the range 95 to 130 dB at frequencies of from 10 to 54

                                       350

-------
  HUMAN

 Death (calculated)
Loss of equilibrium

Dizziness  —^—.
Mild warming
     (body surface)
Mild heating     _
     (skin clefts)
No physiological
     changes
Industrial exposure
no hearing loss
            SMALL  ANIMALS
                       180
160
..I-
                        120
                        100
     dB
           Death  (rabbits)
           Body  temperature rise
           (hairless mice)

           Death (mice, rats,
           guinea-pigs)


           Body  temperature rise
           (haired mice)
        — Mild biological changes
           (rats, rabbits)
             Figure 1 Physiological effects of ultrasound

                          351

-------
kHz (ANTHONY and ACKERMAN, 1955, BUGARD,  1960, GORSLIKOV et al, 1965,
GORSKOV et al, 1964, and others). Where the sound was audible to the animals, these
represented relatively high sensation levels and the biological changes were typical of any
stress condition in many cases. Actual body heating in mice was not measured until a level
of 144 dB at 18 to 20 kHz was reached. With hairless mice, the corresponding level was 155
dB,  indicating  the role  of the fur  in absorbing energy (BANNER,  ACKERMAN and
FRINGS, 1954). The deaths of mice and guinea pigs as a result of exposure to a level of 150
to 155 dB at 30 kHz (DICKSON, 1953), of rats and guinea pigs to 144 to 157 dB at 1 to
18.5 kHz (ELDREDGE and PARRACK, 1948), and of rabbits to 160 to 165 dB at 22.5 and
25 kHz (BUGARD, 1960, ROMANI and BUGARD, 1960) have been reported also.
     In man, there are reports of both a drop  (ASBEL, 1965) and an increase in the blood
sugar level (BYALKO et al, 1963) and electrolyte balance  changes in the nervous tissues
(ANGELUSCHEFF, 1957) as a result  of exposure to ultrasound, although neither sound
levels or frequencies were reported. However, BATOLSKA et al (1969) rightly pointed out
that many  of the effects attributed to ultrasound are also typical of exposure to other
physical and toxic conditions at their places of work, and conclusions should not be drawn
without comparison of results with a control group. GRIGOR'EVA (1966a) failed to find
any significant physiological changes as a result of one hour exposure to 110 to 115 dB at
20 kHz in a comparison with control subjects.
     Slight heating of skin clefts was observed by PARRACK  and  PERRET  (1962) as a
result of exposure to ultrasound at levels of 140 to 150 dB. At 159 dB there may be a mild
warming of  the body surface (PARRACK, 1951). Loss of equilibrium and dizziness oc-
curred at levels of 160 to 165 dB at 20 kHz (ALLEN et al, 1948). The calculated lethal dose
for man is at least 180 dB (PARRACK, 1966).
AUDITORY EFFECTS

    The ear constitutes an efficient impedance matching device for high frequency airborne
sound, and it seems likely  that any hazard from airborne ultrasound will manifest itself
initially as a hearing loss or an associated psychological effect.
    An investigation to determine if the noise from  industrial ultrasonic devices caused
auditory effects was described by ACTON and CARSON (1967). The hearing threshold
levels  of 16 subjects (31 ears) were measured in the frequency range 2 to 12 kHz before and
after exposure to the noise over a working day. No significant temporary threshold shifts
were detected (Fig. 2). On the assumption that if a noise exposure is not severe enough to
cause  a temporary threshold shift, then it cannot produce permanent damage, it was con-
cluded that hearing damage due to exposure to the  noise from industrial ultrasonic devices is
unlikely. A parallel retrospective investigation by KNIGHT (1968) on a group of 18 young
normal subjects using ultrasonic devices showed a  median hearing level within 5 dB of that
of a matched control group of hospital staff except at 4 kHz where the departure was 7 dB
(Figure 3). It was concluded that it would have  been difficult to attribute this exposure
solely to ultrasonic radiation. In addition, no abnormal vestibular function  test (caloric test)
results were found.

                                      352

-------
        -2

        -1

          0
        dB
          1

          2

          3

          4

          5
                     I	1	1	
                     248

                                    Frequency  KHz

                 Figure 2. Temporary threshold shift due to industrial exposure
12
    Some temporary threshold shifts have been reported as a result of exposures to ultra-
sound under laboratory conditions, and the exposure conditions have  been summarized in
Figure 4. (PARRACK, 1966, DOBROSERDOV, 1967, SMITH,  1967).
    The exposures used  by Dobroserdov were at high audible frequencies, and those by
Smith contained high-audible-frequency noise. The results due to Parrack are interesting in
that he exposed subjects to discrete frequencies mainly in the  ultrasonic region, and meas-
ured temporary  threshold  shifts  at subharmonics of one half of the fundamental and
occasionally at lower subharmonic frequencies after 5-minute  exposures to discrete fre-
quencies in the range 17 to 37 kHz  at levels  of 148 to 154  dB. Subharmonic distortion
products have  been reported  in the cochlear-microphonic  potentials  of guinea pigs
(DALLOS and LINNEL, 1966a) and have also been monitored in the sound field in front of
the eardrum using a probe-tube microphone (DALLOS and LINNEL, 1966b). They were
believed to result from nonlinear amplitude distortion of the eardrum, and they appeared at
a magnitude of the same order as that of the  fundamental. This observation may help to
explain Parrack's findings.
    Many sources of ultrasound, and particularly processes involving cavitation, produce
substantial levels of noise in the high  audible range. Reported auditory effects can often be
explained  in  terms of the  audible noise only, and  these references have been omitted
deliberately.
                                        353

-------
   2

dB3


   4


   5

   6

   7
              I
            0.25
0.5
                                     Frequency  KHz
        Figure 3 Permanent threshold shift due to industrial exposure relative to control group
SUBJECTIVE EFFECTS

     It has been mentioned already that early laboratory workers reported suffering unusual
fatigue, loss of equilibrium, nausea, and headaches which persisted after the stimulation had
ceased,  as a result of their exposure  to airborne ultrasound. Complaints of fatigue, head-
aches,  nausea and tinnitus are frequently made by the operators of industrial ultrasonic
devices, but their exposure does not seem to be sufficiently intense to cause loss of equilib-
rium. Observers entering the sound field for shorter periods often experience an unpleasant
sensation of "fullness" or pressure in the ears.
     It has been  shown that  these subjective effects are due to the high levels of high-
frequency audible noise usually produced as a by-product of industrial ultrasonic processes,
and especially those involving cavitation (ACTON and CARSON, 1967). SKILLERN (1965)
attempted to correlate  these effects with frequency, and erroneously concluded that the ear
was sensitive to a narrow band of frequencies centered on 25 kHz. Examination of his data
shows that all frequency spectra he quoted as producing effects contained high levels of high
frequency audible noise as well as an ultrasonic component at about 25 kHz.
                                        354

-------
   160 _
   140 -
 PQ
 •o
 > 120
 V
 3
 o>
 •a
 c
 o
 CO
100 -i
    80  -
                                                 Parrack  1966
                                                     Y/// ////,
                                               Uobroserdov 1967
                                  Smith 1967
                                       I
                                      10

                              Stimulus frequency KHz
                                                     20
30
50
              Figure 4 Laboratory exposures producing temporary threshold shifts
    The effects  are often only reported by young females in exposed populations, but
ACTON and CARSON (1967) showed that this was a function of auditory threshold. Males
employed  industrially  often have high-frequency  hearing losses, probably due to noise-
induced hearing loss and presbyacusis, or "sociacusis". The effects were not considered to
be psychosomatic in origin or due to hysteria.

EXPOSURE CRITERIA

    Unlike audible noise, the area of ultrasonic noise is characterized by a lack of published
exposure criteria. Early Russian workers (GORSLIKOV et al, 1965, ROSCIN et al, 1967)
proposed an overall limit to ultrasonic exposure of 100 dB, regardless of frequency. This
was probably a cautious move made in the absence of data to quantify some of the physio-
logical  effects  being reported at that time. Only two frequency-dependent criteria are
known. The first was due to GRIGOR'EVA (1966a) and is shown in Figure 5. The levels at
and below a frequency  of 16  kHz are based  on the  experimental results of temporary
threshold shift measurements, and at 20 kHz and above as a result of experiments to detect
physiological changes. However, later in the same year,  GRIGOR'EVA (1966b) proposed a
                                       355

-------
level of 110 dB in the 20 to 100 kHz frequency range, and mentioned that this had been
incorporated into a (USSR) Ministry of Health memorandum.
     The criterion due to ACTON (1968), also shown in  Figure 5, was based on experi-
mental evidence  to prevent both auditory and subjective effects  in the greater part of
population exposed over a working day. The author did not feel justified in extrapolating
this  criterion  beyond the one-third-octave band  centered on  31.5 kHz on the  basis of
available experimental evidence.

CONCLUSION

     Many of the reports of effects  due to exposure to ultrasound must be  regarded as
anecdotal rather  than factual. Further confusion has undoubtedly arisen because results
obtained with small, fur-covered  animals have been transposed directly to man, and because
airborne  exposure has not been sufficiently  differentiated from liquid- or solid-coupled
exposure. Nevertheless,  there is  ample evidence to show that  exposure  to high levels of
ultrasound can have some effects  on man.
     In industry,  the  exposure to the high levels  of high-frequency audible sound which
accompanies many ultrasonic processes is more likely  to prove troublesome than the ultra-
   120 -
«  110 _
   100 -
M
2   90
1
.a
v
o
o
    70  -1
                                  Grigor'eva 1966a
                                                              Grigor'eva 1966b
                       Acton 1968
        I
        3
         I
       10

Frequency  KHz
20
 f

30
50
                         Figure 5 Exposure criteria for ultrasound

                                        356

-------
sonic frequencies themselves. Subjective effects include headaches, nausea, tinnitus, possibly
fatigue and so on, and some temporary threshold shifts in hearing have been observed as a
result  of experimental laboratory exposures to ultrasound. Two similar exposure criteria
have been published, both in the 1960's.

ACKNOWLEDGEMENT

    The author wishes to thank Mr. A. E. Crawford for bringing some of the references to
his notice, and Dr. R. R. A. Coles for reading and commenting on the draft.
                                 REFERENCES

ACTON, W. I. (1968) A criterion for the prediction of auditory and subjective effects due
     to  air-borne noise from ultrasonic sources, Annals of Occupational Hygiene, 11, 227.
ACTON, W. I. and CARSON, M. B. (1967) Auditory and subjective effects of airborne noise
     from industrial ultrasonic sources, British Journal of Industrial Medicine, 24, 297.
ALLEN, C. H.,  FRINGS, H. and RUDNICK, I. (1948) Some biological effects of intense
     high frequency airborne sound, Journal of the Acoustical Society of America, 20, 62.
ANGELUSCHEFF, Z. D. (1957) Ultrasonics, resonance and  deafness,  Revue de Laryn-
     gologie, d'otologie  et de Rhinologie, July-August 1957, abstract in CORDELL (1969).
ANTHONY, A. and ACKERMAN, E. (1955) Effects of noise on the blood eosinophil levels
     and adrenals of mice, Journal of the  Acoustical Society of America, 27, 1144.
ASBEL, Z. Z. (1965) The effect of ultrasound and high-frequency noise on blood sugar
     level, Gigena truda i professional  'Nye  Zabolevanija (Moscow),  9, 29.,  abstract in
     Occupational Safety and Health Abstracts 4, 104, (1966).
BATOLSKA, A. et al (1969) Occupational disorders due to ultrasound, Work of the Scien-
     tific Research Institute of Labour  Protection and Occupational Diseases (Sofia), No.
     19, pp 63-69, abstract in Noise and Vibration Bulletin March 1971, p. 97.
BUGARD, P. (1960) The  effects of noise on the organism, the importance of non specific
     effects Revue des  Corps de  Sante des Armees (Paris),  1, 58,  abstract in CORDELL
     (1968).
BYALKO, N. et al (1963) Certain biochemical abnormalities in workers exposed to high
     frequency noise, Doklady Vsesoyvznogo Nauchno - Prakticheskogo Soveshchaniya Po
     Izucheniyu Deistriya Shuma  Na Organizm (Moscow), pp. 87-89, abstract No. 2659 in
     Excerpta Medica 17, 570 (1964).
CORDELL, J. (1968) Physiological effects of airborne ultrasound; a bibligraphy with ab-
     stracts, Labrary  Report No.  4, Commonwealth Acoustic  Laboratories, Sydney, Aus-
     tralia.
CRAWFORD, A. E. (1955) Ultrasonic engineering, Butterworth, London.
DALLOS,  P.  J. and  LINNEL,  C. O.  (1966a)  Subharmonic  components in  cochlear-
     microphonic potentials, Journal of the Acoustical Society of America, 40, 4.

                                       357

-------
 DALLOS, P. J. and LINNEL, C. O. (19665) Even-order subharmonics in the peripheral
     auditory system, Journal of the Acoustical Society of America, 40, 561.
 BANNER, P. A., ACKERMAN, E. and FRINGS, H.  W. (1954) Heating of haired and
     hairless mice in high intensity sound fields from 6 to 22kc.
 DAVIS, H. (1948) Biological and psychological effects of ultrasonics, Journal of the Acous-
     tical Society of America, 20, 605.
 DAVIS, H., PARRACK, H. O. and ELDREDGE, D. H. (1949) Hazards of intense sound and
     ultrasound, Annals of Otology (St. Louis), 58, 732.
 DICKSON, E. D. D. (1953) Some  effects of intense  sound and ultrasound on the ear,
     Proceedings of the Royal Society of Medicine, 46, 139.
 DICKSON, E. D. D. and CHADWICK, D. L. (1951) Observations on disturbances of equilib-
     rium and other symptons induced by jet engine noise, Journal of Laryngology and
     Otology, 65, 154.
 DICKSON, E.  D.  D. and  WATSON, N. P. (1949) A  clinical survey into the effects of
     turbo-jet engine noise on service personnel, Journal of Laryngology and Otology, 63,
     276.
 DOBROSERDOV, V. K. (1967) The effect of low  frequency ultrasound and high frequency
     sound on exposed workers, Gigiena i sanitarija (Moscow), 32, 17, abstract in Occupa-
     tional Safety and Health Abstracts, 5,658 (1967)
 ELDREDGE, D. H., and PARRACK, H. O. (1948)  Biological effects of intense sound, Paper
     presented at 36th meeting of the Acoustical Society of America, Cleveland, Ohio,
     November, 1948, abstract in Journal of Acoustical Society of America, 21,55.
 GORSKOV, S. I. e't al (1964) The biological effects of ultrasonics in relation to their use in
     industry, Gigiena i sanitarija (Moscow), 29,  37, abstract in Occupational Safety and
     Health Abstracts, 3,39(1965).
 GORSJJKOV, S. L, GORBUNOV, O. N. and ANTROPOV, S. A. (1965) Biological effects
     of ultrasound, Medicina, Moscow, review in Ultrasonics, 4, 211 (1966).
 GRIGOR'EVA, V. M. (1966a) Effect of ultrasonic vibrations on personnel working with
     ultrasonic equipment, Soviet Physics - Acoustics 11,426.
GRIGOR'EVA, V.  M.  (1966b) Ultrasound  and  the question  of occupational hazards,
     Maschinstreochiya No. 8, p. 32, abstract hi Ultrasonics, 4, 214.
GUIGNARD, J. C. (1965) Noise, Chapter 30 in  GILLIES, J. A. (Editor) A textbook of
     aviation physiology, Pergamon, Oxford.
HILL, C. R. (1970) Paper presented to conference on "Clinical aspects of ultrasonic diag-
     nosis", London, 10 December 1970, review in Ultrasonics, 9, 172 (1971).
KNIGHT, J. Jl (1968) Effects of airborne ultrasound on man, Ultrasonics, 6, 39.
PARRACK, H. O. (1951) Physiological and psychological effects of noise, Proceedings 2nd
     Annual National Noise  Abatement symposium, Chicago, Illinois, 1951, abstract in
     CORDELL(1968).
PARRACK, H. O. (1952) Ultrasound and industrial medicine, Industrial Medicine and Sur-
     gery, 21, 156.
PARRACK, H. O. (1966)  Effect of airborne ultrasound on humans, International Audio-
     logy, 5, 294.
                                      358

-------
PARRACK, H. O. and FERRET (1962) Effect on man of low frequency ultrasonics pro-
    duced by aircraft, Report presented at meeting of group of experts on struggle against
    noise caused by aircraft, Organisation de  Co-operation et de Development Econom-
    iques, Paris, abstract in CORDELL, (1968)
ROMANI, J. D. and BUGARD, P. (1957) Further experiments on the effect of noise on the
    endocrine system, Acustica, 7, 91.
ROSCIN, I. V. et al (1967) Occupational health hazards of technical applications of ultra-
    sound, Gigiena truda i professional 'Nye Zabolevanija (Moscow), 11, 5., abstract in
    occupational Safety  and Health Abstracts, 5, 657 (1967).
SKILLERN, C. P. (1965) Human  response to  measured sound pressure levels from ultra-
    sonic devices, American Industrial Hygiene  Association Journal, 26, 132.
SMITH, P. E. (1967) Temporary threshold shift  produced by exposure to high-frequency
    noise, American Industrial Hygiene Association Journal, 28, 447.
WEBSTER, E. (1963) Cavitation, Ultrasonics, 1, 39.
                                       359

-------
         SESSION 4 B





PERFORMANCE AND BEHAVIOR




  Chairman: D. E. Broadbent, UK
             361

-------
                 PSYCHOLOGICAL CONSEQUENCES OF EXPOSURE
                       TO NOISE, FACTS AND EXPLANATIONS

                                     Edith Gufian
                                 Institute of Psychology
                                  Bucharest, Romania

     I hate beginning this paper with a pessimistic statement, but I must say that systematic
investigations carried out on the psychological effects of noise since  1950*) are in a rather
controversial state and therefore no firm conclusions can be drawn.
     There is, however, a sound reason for the rather ambiguous nature of the psychological
results, and this will be evident from Figure  1. This reason is the extraordinary complexity
of the various factors which intervene in the effect of noise on man.
     In considering the effects of noise on behavior, one usually assumes  that noise is a
more or less undifferentiated  variable, defined mainly by its intensity. However, the  facts
are that the different possible  interactions between the several intrinsic parameters of noise
x)This paper deals primarily with experimental results since 1960, as there are excellent reviews of former data by Krytei
(1950) and Broadbent (1957).
                   VAIUAIU>:M
•
IS 1 KNSI 1 V <
- I'ltK- I K.SCY^r-
1
i
DI u.uiiiN

- I'YPK-^S^^^^
I-IIMPI.KMTY
XAITKK ^— —
^

; CONSTANT
VAH1AIU.K
- CONST AN f
^ VAIUAIILE


- CON riNl'OTS
- INTERMITTENT <

-TIIAKFIC
•MMilSITUAL
" AIIICKAFT ETC.
                                          IVfKHMEUAItY VAIUAHLES
                                          llnlra-h.d tnlvriiMlividual iliirc'risu-L's>
                                                                  DEPENDENT VAUIAHI.KK
                          PACED vs. I SPACED
                               :s HATE
                          KA.VDOM vsHEGl-LAn
                               :VEUIIAL
                               NON-VERBAL
                      visi AI. t:rc



M



A









0
H
A
X
]
S
M
C

P
S
Y
C
II
O
L
O
G
I
C
A'
L
- UENERAL HEALTH STATE 1
- AUDITORY SYSTEM ,
- OTHER SENSORY SYSTEMS (
y SELECTIVE 1
- A HOUS A U LEVEL^pjppj.gj, ^
- ACCUMULATED PHYSIO LOGICAL
EFFECTS (FATIGUE ETC. I ;

- SENSORY INTERACTION
- AUDITORY PEHCEPTION
- AllOl'SAL LEVEL^ TASK '
'DURING
- PltEVIOUS EXPERIENCE !
. NOISE
WITH < TASK '
- MOTIVATION >
- ATTITUDE *
1
S
-
ACCU.ACY c<
C
3 - SPEK1I
I
I - TOY
! -DISCIIIMlSAilliN
: -OV'I'Pl'T


^
* - IMHUW Al.
)
'
I
C - OIUU'PKKA



    Figure 1. A schematic representation of the ways and modalities through which noise acts on man

                                          363

-------
 (intensity, frequency and complexity) and  its temporal structure give rise to such a great
 diversity of stimulation, that it practically could never be tested in the laboratory.
     The  picture is further complicated  by the other independent variable-the routine
 activity in real-life situations or the laboratory task, to which we shall restrict ourselves here.
 The multiplicity of tasks resulting from the possible combinations of the different elements
 listed above  is in fact infinite and new associations are always possible. If we turn now to
 man himself, we encounter a great number of variables which have to be taken into account
 if noise influence is to be understood. First  there is a biological level which is of interest to
 us insofar as it constitutes  the basis of behavior and yields indications in case the overt
 behavior cannot be specifically interpreted. Secondly, there is the psychological level which
 accounts for a great part of the variance in noise investigations, in particular the role played
 above and beyond the actual stimulation, by the familiarity with both  the noise and the
 type of activity, motivation, attitude etc. Each of these variables can be studied separately
 and yields more or less specific results. However, no one has ever ventured  to look into all of
 them simultaneously and  draw a complete picture of their influence, even though each of
 them mediates in  a particular way the  influence of noise on the final output. The  strong
 interrelationships among these variables is a widely known fact although in empirical studies
 it is sometimes overlooked. Finally, we are faced with  the dependent variables, the final
 concrete output of the various above-mentioned factors-the performance, and the subjec-
 tive annoyance.
     Several  points need to be clarified here.  First of all, performance is a rather vague term;
 it refers to many measures, each having a different meaning as a function of the type of
 task, although they all may be divided into two main classes: those measuring accuracy and
 those measuring speed of performance. What I would like to point out is that the diversity
 of tasks imposes a diversity of  measures which are comparable only  in a very general
 manner.
     Maybe  the inclusion  of annoyance among the  dependent variables seems somewhat
 peculiar, but  my contention is, however—and I shall try to substantiate it—that regardless of
 the final effect of noise on performance (positive, neutral or negative) a certain annoyance is
 always present. Of course, annoyance has a feedback on the psychophysiological  state and
 thus on performance, just as performance level cannot fail to exert a certain influence on
 annoyance level.
     How  then does noise affect the dependent variables? I suggest that there are two main
 ways, depending on man's activity. In case the individual is engaged in an activity, the noise
 acts concurrently with the task  and its effect on performance depends on different task
 characteristics as well as on those of the intermediate variables. On the other hand, if man is
 not actively  engaged  in a professional  task, but is either resting or performing a simple
 relaxing home activity the noise  affects him more directly and its main consequence is a
 psychic disturbance, the subjective annoyance.
     In view of these complex  interrelationships,  the  divergent results in noise  effects
investigations could obviously be  anticipated. In fact, as shown in Figure 2, this is precisely
what happened; performance has been either  impaired, unaffected, or in some cases even
 improved  under noise conditions. The fact  that performance in noise showed no effects,
 detrimental effects or improvements  at the same sound pressure levels (between 50 &  110

                                        364

-------
                          NEGATIVE EFFECTS


                          Jurist, 1-J..9     Mmitoriitn
                          Turn-Pri- and
                          U'isncr, IS62     ViKilam-v


                          Ml., J'JC-I      liatii*
                          Mark and &)., IBtl PsycSomuli.r
                          llritiidhutt and   Visual mu-
                          Crufjory, 19(t.~>   nitarmp
                          Hclrmann and Osier-
T:-VS HMJ
.U Vs 75 uh
                          WilttTshctm a
                                , b.
                                    llwntH-v
Mrnnomann, jyos
WiU-cW, IbTO
NcMtal, 19TI
tu-nlr-sch, 1S71
Uyliuidvr and at..
19T2
POSITIVE EFFECTS
McCralh, 1963
Kirk and Hecot,
19631
Psychamotur
Simple KT
Simple AT
Psyehomotnr
Trackinfi
Attention

VfciUac*
Detection
^:,-io(i
30 -W
yo
7(t-9h

0.4788

72
C4
                          Waiktns. 1964    Detection
                          Samuel, 1964    AtLcnUon
                          Wvinstcin and    »IlK,AATE
                          U^cKvnziv 19M  H'nJpilnt.
                          O*Hallc> Mid Poplvw-
                          hky, 1971     Stroop
                          K'ttima, 1'jTl    Attention

                          SEl'THAL EFFECTS


                          .k-risiwi, l(J-'.7    Vifilant-c
                          Licni-it InJ
                          -l.uifttii, t'JM    AtU-nLii*i
    MU-THItl.


   industrial
     varied
     steady
     Inter rntt.
  «»t. Intermit.
 v» 110 «hitv
     «hitc
     airi-rad
                                                              Arouxal
                                                              Dislrat-tkm
CO
En

pcrfo.

TOT
CD
                   Arousal

                   ArouMl
CitllOn, !!*(.!.

Mi*Cujint IW.L

I.uL:i!« an«lal.l:P7ll
Itl:ii-I.uc11 .ind IV-lt,
1!*71
Sli'%i'Hf*. 1-I7J

Vinilani-t- Tlt-'Kl

Vi|>ilm^i- *
-------
 suggests that the major cause of the disruption is the muscular reflex associated with startle
 (May and Rice, 1971; Thackray and Touchstone, 1970). Tasks involving complex perceptual
 and/or  cognitive processes may be impaired for longer periods-up to 30 sec. (Woodhead,
 1959, 1964)—the detrimental effects of noise resulting from an interference with either
 information processing (Broadbent, 1971) or information reception (Woodhead, 1964).
     While all these  studies show some degree of impairment  associated  with  impulsive
 noise, investigations employing real or simulated sonic booms show divergent results ranging
 from performance decrement (Lukasetal., 1970; Rylander et al., 1972; Woodhead, 1969), to
 generally  non-significant effects (Lukas et aL, 1971; Harris, 1970) to performance improve-
 ment (Thackray et al., 1972).
     Concerning intermittent noise of lower intensity, Teichner et al. (1963) predicted that
 the same  noise  stimulus could either facilitate or disrupt performance. The  predominant
 effect at  any point in time would depend upon duty cycle, the ratio of noise on-time to
 periods of silence between noise presentations. Significant effects  were observed on speed of
 visual  target  detections  for  all  on-off ratios  except the 70% when compared  to  a
 control, no-noise  condition. Warner (1969), testing different noise intensities (80, 90 and
 100 dB) in an attention-demanding task at the 70% duty cycle, found no effect of intensity
 level  on detection time,  and fewer errors as a function  of noise intensity.  In  a  further
 experiment (Warner and Heimstra,  1971), it was found that the particular effect attributable
 to varying ambient noise ratios (0, 30, 70, 100%) on target-detection  time  is dependent
 upon the degree of difficulty of the inspection task.
     Here we  run  up  against a vital question, the fact now I think well-established that the
 deleterious effect  of  noise on performance increases as a function of increasing  task com-
 plexity. Task difficulty can be manipulated in different ways: (1)  One way is by multiplying
 the stimulation sources as was shown by Broadbent (1954) in the difficult 20-dials test and
 in the detection of the easily seen 20-light task in 100 dB noise, and by Jerison (1957,
 1963) in the  three-clock task vs. the one-clock task. (2) Another way is by  changing the
 intrinsic difficulty of  the task. Thus Hsia (1968) using six difficulty levels finds that a 65-dB
 noise exerts a detrimental effect on information processing only when the stimulus material
 is difficult,  while  Houston  (1968), manipulating  two difficulty levels, found that noise
 facilitates  performance in the high-difficulty condition for the more difficult  task but not
 for the easy one. Gulian (1972), comparing percentage of  errors  and reaction time (RT) in
 three  vigilance tasks  of different  difficulty level, under three noise conditions, found  a
 clear-cut interaction between noise and difficulty level, particularly with respect to the very
difficult one (Figure 3) (3) Another source  of difficulty arises from the temporal structure
of the task. It was found that noise acts adversely on performance as a function  of high
variability of  intersignal interval  (Dardano,  1962), or high signal  rate (Broadbent and
Gregory, 1965)  but that  lower signal rates  either improve or  do not change  detection
performance despite high noise level (95 dB) (Davies and Hockey,  1966).
     There is finally a fourth method of manipulating difficulty which seems to detect best
 the effect of noise  the method of simultaneous tasks.
    Indeed, results in an experiment reported by Boggs  and Simon  (1968)  with a two-
complepdty-level four-choice RT task and a secondary auditory monitoring task showed  that
noise produced a significantly greater increase in secondary task errors when the secondary

                                         366

-------
                    100
                                   TOdB  BOdB
                                   Cont  Intermit.
Quiet  7OdB  9OdB
       Cont  Intermit
Figure 3.  Effects of Noise on Performance level as a function of task difficulty
    A — an auditory vigilance task • the most difficult one. S discriminated a pure tone (400 Hz) among 4
different pure tones
    B — a word recognition task of medium difficulty. S reacted to one trigram (doc) among 4 different
Digrams
    C -  an easy visual and auditory reaction time task. S reacted to a visual and an auditory stimulus and
discarded another visual and auditory stimulus.
     The noise was  the same in all  3 experiments:  70 dB white continuous noise  and 90 dB varied
intermittent noise.
task was paired with the complex than with the simple primary  task. Finkclman and Glass
(J970) showed  that while performance on  the primary task is unaffected by predictable vs.
unpredictable noise, only the unpredictable  noise resulted in performance degradation in the
subsidiary  task. Hockey (1°-70 a. b) using a primary tracking and a secondary multisource
monitoring task, showed that the tracking  task improved in noise (100 dB vs. 70 dB) as did
the  location  of centrally located  signals in the monitoring task, but that detection of
peripheral  signals is impaired.
     These and other similar experimental results arc explained by Broadbent by his arousal-
filtering  hypothesis. Noise increases arousal and it  affects  perception only when there is a
competing stimulus from which the reaction stimulus has to be discriminated  that is, when
a filtering  process has  to take place. Arousal affects filtering in the sense  that the aroused
system devotes a higher portion of its time to  the intake of information from dominant
sources and less from  relatively  minor ones. Of course  the higher  the arousal level, the more
                                           367

-------
 adverse the effects of noise, whereas at moderate levels of arousal, performance maintains its
 efficiency.
      That effects of a noise cannot be accounted for only by its physical characteristics or
 by those of the task is emphasized in several experiments which evidence modifications in
 performance, arousal and annoyance through manipulation of the relevance of the stressor.
 In two consecutive experiments, Glass et al. (1969, 1971) showed that adverse postadaptive
 effects following loud unpredictable noise (110 dB) were substantially reduced if the subject
 believed he had control (vs. no control) over the termination of the noise.
      Approaching the problem from another  point of view, Munz et al. (1971) tested
 subjects who had either high or  low involvement in  a pursuit rotor tracking task while
 simultaneously exposed to a task-related or task-unrelated 80 dB noise. They found no
 effects of noise on performance, but highly-motivated  Ss reported experiencing greater dis-
 comfort under task-unrelated noise as compared to the other condition of the control one,
 and their statement was supported by their post-experimental ranking of working condition
 performances.
      Finally,  the meaning of noise is an important variable-and  it has been shown that
 speech impairs performance more than a neutral noise. That is why Wisner (1971) concludes
 that  laboratory experiments using meaningless  noises  cannot explain performance of sub-
 jects working in a place where there is considerable conversation.
      This evidence fully justifies Kryter's suggestion that in one way or another the task and
 its completion are dependent upon the presence of noise  and all the observed effects of
 noise are due to psychological factors related to stimulus and response contingencies associ-
 ated  with the noise by individuals. Individual differences in  reaction to noise arise because
 of inappropriately interpreted stimulus  and response contingencies, but these tend to be
 eliminated with learning and experience.
     To summarize: (1) the level of noise needed to show adverse effects is high-95 dB-and
 high  frequencies  seem to be more noxious  than low ones; (2) the harmful effect of noise
 seems to  be on accuracy rather than on speed; (3) monitoring that requires time sharing
 among several potential signal sources is affected by high levels of noise, as is (4) monitoring
 that requires the operator to translate delayed data.
     If noise and task characteristics only partially explain the various shifts in performance
 efficiency, it follows that their causes should be sought elsewhere, as well. It is suggested
 that level of arousal, peculiarities in auditory perception, and annoyance produced by noise
 are the main variables involved. Individual differences in all  these variables, correlated with
 personality measures, introduce an important cause of variation.
     What evidence is  there to support these assumptions? We refer first to the hearing
mechanism in  order to emphasize that if differences in auditory perception and processing
of sounds are established, they probably are at the basis of differences in noise suscepti-
bility, and ultimately would have consequences on performance.
     Several investigators have linked  personality with two  aspects of the auditory thresh-
 old: the  absolute sensitivity and  the variability of the measures  obtained (see Stephens,
 1972, for a review).
     Eysenck  (1970)  proposed that sensory thresholds, tolerance levels,  and preference
levels for  sensory stimulation will  differ in  introverts and extroverts. Indeed some studies

                                        368

-------
showed that introverts might have more sensitive auditory thresholds (Smith  1968), less
variability of the audiometric threshold (Reed, 1961, Reed and Franci 1962; Farley and
Kumar, 1969; Stephens,  1971), and that less extraneous stimulation is required to produce
uncomfortable loudness level (Stephens and Anderson, 1971) for them, but that extroverts,
whether children (Elliot,  1971) or adults (Hockey, 1972), prefer higher levels of sensory
input.
     Independent evidence of individual differences in auditory  perception correlated with
personality measures comes from Soviet researchers. With respect to differences in sensory
threshold, subjects with a strong nervous system are shown to have higher thresholds than
do those with weak nervous systems (Nebylitsyn, 1957, 1966, Borisova, 1967). Specific
differences were found  in individual loudness functions, which suggests the need to intro-
duce a concept of susceptibility to noise (Barbenza et al., 1970a) which might be correlated
to anxiety (Stephens,  1970) and excitability on the MMPI scale (Barbenza etal., 1970b).
     Obviously more research is necessary in order to assess the influence of these factors on
performance level.
     Arousal. It is widely  assumed that noise, by increasing the  amount of stimulation
reaching the CNS, has the effect of raising the level of arousal, so that the Ss feel more alert
and perform better; but in the extreme, when noise exceeds a certain SPL, Ss become more
tense and in this case arousal may result in inefficient behavior. These relationships are best
expressed by the inverted-U hypothesis but its validity is sometimes questioned, particularly
with respect to the concept of over-arousal which Broadbent considers as lacking in preci-
sion.
     Evidence about  the arousing effect of noise comes  from physiological and behavioral
studies, especially from  studies about interaction of noise with other agents such as loss of
sleep, knowledge of results, alcohol, etc. (Wilkinson 1969; Hamilton and Copeman 1970).
The way in  which noise induces physiological and behavioral arousal can be clearly followed
up in Kryter's (1970) diagram (Figure 4). Even though changes in arousal level  provide a
satisfactory  explanation of the changes in performance efficiency, the diversity of results in
studies  using more or less the  same  task  and noise parameters points to the  need for
additional clarification. This is even more apparent when considering those studies  where no
effects of noise  whatever could be  detected. For instance, in a study  by Gulian (1970),
performance in  a vigilance task of  Ss exposed to noise (70 and 90 Db,  continuous and
intermittent) showed  no overall effects of noise. However, when Ss were divided according
to their arousal  level, established  through several EEC parameters, significant differences
were found  between  hypo- and hyper-reactive Ss not only in performance efficiency but
also in evolution of performance (Figure 5).
     Different individuals manifest not  only a distinctive basal level of arousal, but also
differ in arousability toward noise. Extroverts are thought of as chronically less highly
aroused than introverts, and it is argued that when subjects enter the task situation at a low
level of arousal,  their performance in noise should improve to a greater extent than would
that of subjects who begin work at comparatively high arousal levels.
     Data supplied by different investigators support this viewpoint. Thus, it was found
(Davies and  Hockey,  1966) that  the facilitating effect of high-intensity noise was signifi-
cantly  greater for extroverts than for introverts, that extroverts make significantly fewer

                                         369

-------
                                                No 2 MMMTOKV-AUTONOMIC  7
                                                   NfMVOUS SYSTEM   /
                                                                /
Figure 4. Schematic diagram of primary auditory (hearing) and secondary auditory (nonauditory) systems
(After Kryter, 1970)
errors in a variety of auditory conditions (Davies et al., 1969; Blake, 1971; Di Scipio, 1971),
etc. Yet, when the demands of the task are slight, no significant differences in performance
appear between extroverts and introverts, although significant differences in arousal level
(skin conductance) are apparent (Gulian 1971, 1972). Thus it seems that the lack  of diffi-
culty blurs the differences between introverts and extroverts in performance, although they
remain present at the physiological level. On the other hand, recent studies (Hockey 1970 a,
b) have outlined that in complex simultaneous tasks, introverts tend to emphasize the high
priority demands more under normal environmental conditions. Thus differences in arousal
level certainly act differently on performance level. Perhaps a deeper insight could be gained
in the mechanism of noise-induced  arousal  and its effects on performance level if  the
relationships between selective and diffuse arousal were taken into consideration.
    Annoyance.  With anoyance, at last we enter a field of noise investigation where we no
longer meet with conflicting evidence. Everybody complains about noise.
    Annoyance is the final product of noise, whether or not it impairs performance, but of
course many factors, psychological, educational etc. influence its extent and its expression.
Hawel  (1967)  devised a sophisticated scheme for defining  the  complex  relationships
between activity, noise type  and intensity,  the individual's momentary disposition,  and
specific reactions and then" impact on annoyance level. As was stressed by Anderson (1971)

                                         370

-------
                 HYPOREACTIVE
                             HYPERREACTIVE
    */.
    1001-

  u
  cs»
  SJ  90
  o
  u
     85
     80

                      correct detections
V
                                                         arousal
                     arousal
                                                       correct detections
                                                        110
                                                                              100
          Q     WC     LC     WI     LI
              experimental conditions
                              WC     LC     WI     LI
                           experimental conditions
                                                           c
                                                           a
  o
  i/i
  •D
  O
800
                                                                               70
                                                                               60
                                                         50
Figure 5. Differences in arousal and performance level in hyporeactive and hyperreactive individuals. Q -
quiet; WC - weak continuous noise 70 dB; LC - loud continuous noise 90 dB; WI - weak intermittent noise
70 dB; LI - loud intermittent noise 90 dB.
there appear to be at least four aspects of noise annoyance, including social awareness of
noise, personal sensitivity, annoyance  toward specific noise  in a particular situation, and
annoyance toward a set of specified noises in unspecified situations.
     Borsky (1954) found differences in annoyance due to changes in attitude as large as 6
dB as did other authors (e.g, Sorensen, 1970). Atherley et al. (1970) stated that the subjec-
tive importance of certain noises  would influence the  attitude towards them and induce,
accordingly, changes in physiological measures, while  Hermann et al. (1970) demonstrated,
in a  pseudo-tracking task, that annoyance, muscle tension, and TTS are dependent on  the
Ss' emotional attitude toward noise.
     Anderson  and  Robinson  (1971)  advance a two-factor explanation of annoyance:
annoyance is partly produced by changes in arousal caused by purely physiological response
and partly by a so-called cognitive element.
     Perhaps another factor should  be added, namely, aversion  to noise (Sullivan et  al.,
1970) which depends on anxiety  (Broadbent,  1957; Sullivan, 1969), previous experience
with different  noise environments (Spieth, 1956), the individual subjective definition of
aversiveness (Wolff, 1964), task involvement (Kryter, 1966), etc.
     Several studies  have stated a  considerable intersubject variability in individual annoy-
ance, which led Moreira and Bryan (1972) to claim that there  exists  a noise-annoyance
                                         371

-------
susceptibility and individual noise functions, the greatest difference between noise-sensitive
and insensitive Ss occurring at quite moderate levels of noise (Figure 6). Becker et al. (1971)
emphasize that noise-sensitive persons rate all noises, irrespective of their intensity, as being
more intrusive in their daily activity, and rated everything in their environment much more
unacceptable than did  the  noise-insensitive.  The noise-sensitive  subjects were also more
likely to perceive themselves as being more sensitive than the average person, and believed
that it affected their health.
     The differences in sensitivity to noise annoyance are stable, and do not depend upon
age, sex, education, job responsibility or such personality traits as determined by the EPI
and the MMPI, but are correlated with anxiety, and with various measures of personality as
given by the Rorschach Projection Test (Moreira and Bryan, 1972).
     Clearly more studies  are needed to delimit the  annoyance  produced by noise, its
psychophysiological and personality correlates, its effects on performance—the more so as
people tend usually to consider "noise"  the sounds encountered at work, while the noises
experienced at home are considered as merely sounds.
     Noise is annoying, it is a nuisance. That is why man started to study noise and the ways
of reducing it. And this is  the  only hard  fact, the only undisputable one;  for all  the
                          NOISr ANNOYANCE SUSCEPTIBILITY
                          6O
70          80
 Noise level (dBA)
                                                                9O
 Figure 6. Individual noise functions for 6 subjects (3 of the most noise sensitive and 3 of the most
 insensitive to annoyance by noise). Each curve is the mean  rating of all 3 noises by the subject. Noise
 sensitive: m	m,S,   RBH;»	•.S,  AC>	*£,  EN.  Noise  insensitive:D	a£>
 SC; o	o, S,   MLDu	AS,  DW. (After Moreira and Bryan, 1972)

                                          372

-------
behavioral studies of noise effects disclosed simply that certain noises are harmful, certain
tasks and activities are sensitive to noise, certain persons are more affected  than others by
noise, etc. Therefore,  annoyance appears as the crux of the psychological consequences of
noise  and I  feel that  it deserves  much more attention from the psychologist than it has
received up to now.
    Although  the results  in noise  studies are  quite conflicting and although it is rather
improbable that new studies would produce facts contradictory to those already established
it is certainly worth while to continue the investigations, because many aspects have been
neglected:
    - no data exist on very long-term habituation to noises;
    - only few data exist on sensory interaction and its effect on performance;
    - almost no data, except a study by  Lukas et al. (1971) exist on effects of noise on
persons who  are accustomed to noise and are given a task which usually is sensitive to noise;
    - almost no data exist on effects of noise in performing a habitual activity, which is
usually accomplished in a quiet environment;
    - only  few data are available  about individual differences in response to noise and
about evolution of performance and annoyance level over long periods of time in individuals
with differing reactivities:
    - only few data are available about the individual's compensatory effort, both psycho-
logical and physiological,  in the  process of adaptation to  noise and  its consequences on
subsequent performance.
    Many  more problems  exist  and  qualified answers could have a great practical and
theoretical impact. It is our privilege to seek these answers.
                                     Bibliography

   1. Anderson, C. M. B., The measurement of attitude to noise and noises. NPL Acoustic
     report Ac 52, Oct. 1971.
   2. Anderson, C. M. B., Robinson, D. W., The effect of interruption rate on the annoy-
     ance of an intermittent noise. NPL Acoustic report Ac 53, Oct. 1971.
   3. Atherley, G. R.  C., Gibbons, S. L., Powell, J. A., Moderate acoustic stimuli; the
     interrelation of subjective importance and certain physiological changes, Ergonomics,
     1970,75,536-545.
   4. Barbenza, C. H., de, Bryan, M. E.,  McRobert,  H., Tempest, W., Individual loudness
     susceptibility, Sound, 1970(b),4, 75-79.
   5. Barbenza, CM.,  de, Bryan, M.  E.,  Tempest, W., Individual loudness functions, J.
     sound&vibr., 1970 (a), //, 399-410.
   6. Becker, R. W.,  Poza, F., Kryter, K., A study of the sensitivity to noise, AD - 728 332,
     1971.
   7. Bieri, J., Meyers, B., Differential effects of two types ofaversive arousal on discrimina-
     bility,  Psychon.Sci, 1968,13, 203-204.
   8. Blake, M. J. F., Temperament and time of day, in W. P. Colquhoun (ed.), Biological
     rhythms and human performance, London, Academic Press, 1971.  109-148.

                                         373

-------
  9. Blackwell, P. J., Belt, J. A., Effect of differential levels of ambient noise on vigilance
     performance. Percept motor skills, 1971,32, 734.
 10. Boggs, D.H., Simon, J. R., Differential effect of noise on tasks of varying complexity,
     J. appLpsychol., 1968,52, 148-153.
 11. Borisova, M. N., Individualinye razlichiia v. pbrogah razlicheniia gromkocti zvukov, in
     Tipologicheskie osobennosti  vysshei   nervnoi  deiatelinosti  cheloveka,  Moskow,
     Prosveshchenie, 1967.
 12. Borsky, P, Noise and the community. Chaba Report nr. 4 (p. 11).
 13. Broadbent,  D.  E.,  Some effects  of noise on  visual performance.  Quart,  j.exp.
     psychology, 1954, 6, 1-5.
 14. Broadbent, D.E., Effects of noise on behavior, in C. Harris (ed.), Handbook of noise
     control, New York, McGraw-Hill, 1957.
 15. Broadbent, D. E., Decision and stress, London, Academic Press, 1971.
 16. Broadbent,  D. E., Gregory, M., Effects of noise and of signal rate upon vigilance
     analysed by means of decision theory. Human Factors, 1965, 7, 155-162.
 17. Dardano, J. F., Relationship of intermittent noise,  intersignal interval and skin
     conductance to vigilance behavior, J. appl. psychoL,  1962,46, 106-110.
 18. Davies, D. R., Hockey, G. R. J., The effects of noise and doubling the signal frequency
     on  individual differences  in visual vigilance performance. Britj.psychol., 1966, 57,
     381-389.
 19. Davies, D. R., Hockey, G. R. J., Taylor, A., Varied auditory stimulation, temperament
     differences and vigilance performance. Br.j.psychol.,  1969, 60,453-457.
 20. Diespecker, D. D., Davenport, W. G., The initial effect of noise on a simple vibrotactile
     learning task. Percept. & Psychophis, 1967,2, 569-571.
 21.  Discipio, W. J., Psychomotor performance as a function of white noise and personality
     variables. Perceptmbtskills, 1971, 33, 82.
 22.  Elliott, Colin D., Noise tolerance and extroversion in children. Br.j.psychol., 1971, 62,
     375-380.
 23.  Eschenbrenner, A. J., Effects of intermittent noise on the performance of a complex
    psychomotor task.  Human Factors, 1971,13, 59-63.
 24.  Eysenck, H. J., Personality and tolerance for noise, in Proceed. Symp.Psychological
     effects of Noise.  W. Taylor (ed.), Dept Social &  OccupatMed., Univ. of Dundee,
     1970.
25. Farley, T. H.,  Kumar,  K. V., Personality and audiometric response  consistency.
    Lauditres., 1969, 9, 108-111.
26. Finkelman, J. M., Glass, D.  C, A  reappraisal of the relationship between noise and
    human performance by means of a subsidiary task measure.  J:'Appl.psychol., 1970, 54,
    214-220.
27. Glass, D. C., Singer,  J. E., Psychic cost of adaptation to  an environmental stressor.
    J.personal. & Soc. psycho!., 1969,12, 200-210.
28. Glass, D. C, Reim, B., Singer, J. E., Behavioral consequences of adaptation to control-
    lable and uncontrollable noise. JJexp. social psychol., 1971, 7, 244-257.
29. Gulian, E., Effects of noise on an auditory  vigilance task. Rev.roum.sci.sociales - sene
    de Psychologic, 1966,10, 175-186.

                                       374

-------
30. Gulian,  E.,  Effects  of noise  on reaction  time and induced  muscle tension, Rev.
    roum.sci.$ociales - serie de Psychologic, 1967,13, 371-385.
31. Gulian,  E., Effects of noise on arousal  level in auditory vigilance, in A. F. Sanders
    (ed.), Attention and Performance HI, Amsterdam, North-Holland, 1970, 381-393.
32. Gulian,  E., Psychophysiological correlates of auditory vigilance under noise conditions
    in introverts and extroverts. Rev.roum.  sci.sdciales - serie de Psychologic, 1971,  15,
    125-136.
33. Gulian,  E., Cu privire la ponderea factorului "dificultate  a probei" in determinarea
    niveluluiperformantei In conditii de zgomot.Rev.Psihol., 1972,75,323-333.
34. Gulian,  E., Focusing  of attention and arousal level under interaction of stressors in
    introverts and extroverts.  Rev. roum. Sci.sociales -  serie de Psychologic,  1972,  16,
    153-167.
35. Hack, J. M., Robinson, H. W., Lathrop, R. G., Auditory distraction and compensatory
    tracking. Perceptmbtskills, 1965, 20, 228-230.
36. Hamilton, P.,  Copeman, A., The effect of alcohol and noise on components of a
    tracking and monitoring task. Rr.j.Psychol., 1970, 61, 149-156.
37. Harris, S.  C., Human Performance effects or repeated exposure to impulsive acoustic
    stimulation,  AMRL, 1970, nr.70- 38.
38. Hawel, W., Untersuchungen eines Bezugssystems fur die psychologische Schallbewer-
    tung. Arbeitswissenschaft, 1967, 6.
39. Hockey, G.  R. J., Signal probability and spatial location as possible bases for increased
    selectivity in noise. Q.j.exp.psychol., 1970, 22, 37-42.
40. Hockey, G.  R. J., Effect of loud noise on attehtional selectivity, Q.j.exp.psychol.,
    1970,22,28-36.
41. Hockey, G.  R.  J.,  Effects of noise on  human efficiency and some individual  dif-
    ferences. J.sound & vibr., 1972, 20, 299-304.
42. Hermann, H., Mainka, GM Gummlich, H., Psychische  und physische Reaktionen auf
    Gerausch verschiedener subjektiver Wertigkeit. Psychol. Forsch., 1970,33, 289-309.
43. Houston,  B. K., Inhibition and  the facilitating effect  of noise on interference tasks.
    Percept-motskills, 1968, 27, 947-950.
44. Hsia, H. J., Effects of noise and difficulty level of input information in  auditory, visual
    and audiovisual information processing. Percept, mot. skills, 1968,26,99-105.
45. Jerison, H. J., Performance on a simple vigilance task in noise and quiet. J.acoustsoc.
    Am., 1957, 29, 351-353.
46. Jerison, H. J., Effects of noise on human  performance.  J.appl.  psychol., 1959,43,
    96-101.
47. Jerison, H.  J.,  On  the decrement function in  human vigilance, in Buckner,  D.  N.,
    McGrath, J. J. (eds), Vigilance, a.symposium, New York, McGraw-Hill,  1963.
48. Kirk, R. E., Hecht, E., Maintenance of vigilance by programmed noise. Percept.mot.
    skills, 1963,76, 553-560.
49. Kodama, Habuku, Psychological  effect of aircraft noise upon inhabitants of an airport
    neighbourhood. Proceed. 17-th Int. Cong. Applied Psychol., Liege, 1971.
50. Kryter, K., The effects of noise on man. J.speech dis. Monog., 1950, suppl.l.
51. Kryter, K. D., Psychological reactions to aircraft noise. Science, 1966, 757, 1346-1355.
    1346-1355.

                                        375

-------
 52. Kryter, K., The effects of noise on man. New York, Academic Press, 1970.
 53. Lienert, G.  A., Jansen, G., Larmwirkung und Testleistung. Int.Z.  angew.Physiol.
     einschl. Arbeitsphysiologie, 1964, 20, 207-212.
 54. Lukas, J. S., Peeler, D. J., Kryter,  K. D., Effects of sonic booms and subsonic jet
     flyover noise on skeletal muscle tension and a paced tracing task.  NASA CR - 1522,
     1970.
 55. Lukas, J. S., Dobbs, M.  E., Peeler, D. J., Effects on  muscle tension and tracking task
     performance of simulated sonic booms -with low and high intensity vibrational compo-
     nents. NASA CR-1781, 1971.
 56. Mariniako, A. Z.,  Lipovoi, V.  V.,  Ucet summarnovo vremeni otdelinyh shumovyh
     vosdeistvii prigigienuceskoi otsenke preryvistyh shumov. gig. truda, 1972,5, 15-18.
 57. May, D. N.,  Rice,  C.  G., Effects of startle due to pistol shots  on control precision
    performance. J: sound Avibr., 1971,15, 197-202.
 58,  McCann,  P.  H., The effects of ambient noise on  vigilance performance.  Human
     Factors, 1969,11, 251-256.
 59.  McGrath, J. J., Irrelevant stimulation and vigilance performance, in Buckner, D. N.,
    McGrath, J. J. (eds), Vigilance: a symposium, New York, McGraw-Hill.  1963.
 60. Moreira,  N., Bryan, M.  E., Noise annoyance susceptibility.  J: sound vibr.,  1971, 21,
    449-462.
 61. Munz, D. C., Ruffner,  J. W., Cross, J. F., Reduction of noise annoyance through
    manipulation of stressor relevance. Perceptmbt. skills, 1971,32, 55-58.
 62. Nebylitsyn, V. V., Individualinye razlichiia v zritelinom i sluhovom analizatorah po
    parametru sila - chiuvstvitelinostl Voprosy psihogii, 1957,4, 53-62.
 63. Nebylitsyn,   V.  :D.,  Osnovnye svoistva  nervnoi  sistemy cheloveka.  Moskow,
    Prosveshchenie, 1966.
 64. Nosal, Effects of noise on performance and activation level. Polish Psychol.Bull., 1971,
    2, 23-29.
 65. O'Malley, J.,  Poplawsky, A.,  Noise induced  arousal  and  breadth of  attention.
    Percept.motor skills, 1971,33, 887-890.
 66. Parrot, J., Wittersheim, G., Effets  transitoires  d'echelons  de bruit sur la motillite
    palpebrale et divers criteres de performance,  au cours de  I'execution d'une tdche
    psychomotrice de longue dure'e.  in Problemes  actuels de la recherche en ergonomie,
    Paris, Dunod, 1968.
67. Reed,  G.  F., Audiometric response  consistency, auditory fatigue, and personality.
    Perceptmbtor skills, 1961,12, 126.
68. Reed, G. F., Francis, T. R., Drive, personality and audiometric response consistency.
    Perceptmbtskills, 1962,15,681-682.
69. Rentzsch, M., Einfluss von Larm aufeine Kombination geistig-korperlicher Tatigkeiten
    unter Betriebsbedingungen. Wiss.Z.Techn. Univers-Dresden, 1972, 21, 567-572.
70. Rylander, R. (ed.), Sonic boom exposure effects. J.sound&vib., 1972,20, 477-544.
71. Rylander, R., Sorensen, S., Berglund, K., Brodin, C., Experiments on  the effect of
    sonic-boom exposure on humans, J.acoust Soc.Am., 1972, 51, 790-798.
 72. Samuel, W. M. S., Noise and the shifting of attention.  Quart.j.exp. psychol., 1964,16,
    123-130.

                                        376

-------
73. Sanders, A. F., Bunt, A. A., Some remarks on the effects of drugs, lack of sleep and
    loud noise on human performance. Nederlands Tijdschrift voor de Psychologie, 1971,
    26, 670-684.
74. Smith, S.  L.,  Extroversion  and sensory  threshold. Psychophysiology,  1968,  5,
    293-299.
75. Sorensen,  S., On the possibility  of changing the annoyance reaction to noise  by
    changing the attitudes to the source of annoyance. Nordisk Hyg. Tidsk., 1970, suppl.
    1.
76. Spieth,  W., Annoyance  threshold  judgments  of bands  of  white  noise. J.acoust.
    Soc.AmM 1956, 28, 872-877.
77. Stephens, S. D. G., Personality and the slope of loudness function. Q. J.exp.PsychoL,
    1970,22,9-13.
78. Stephens, S. D. G., The value of personality tests in relation to diagnostic problems of
    sensorineural hearing loss. Sound, 1971, J, 73-77.
79. Stephens, S. D. G., Hearing and personality: a review. J: sound & vibration, 1972,  20,
    287-298.
80. Stephens, S.  D. G., Anderson,  C. H. B., Experimental studies on  the uncomfortable
    loudness level J. speech hearing Res., 1971,14, 262-270.
81. Stevens, S. S., Stability of human performance  under intense noise.  Jlsound & vibr.,
    1972, 21, 35-56.
82. Sullivan, R., Subjective matching of anxiety to intensity of -white noise. J:abnormal
    psychol., 1969, 74, 646-650.
83. Sullivan, R., Warren, R., Dabice, M., Minimal aversion thresholds for white noise:
    adaptation. Am. j. psychol., 1970, 83, 613-620.
84. Tanasescu,  Gh., Stanciulescu, EL, Domilescu, M., lonescu, M., Predescu, M., Catoranu,
    V., Popescu, E., Influenza zgomotului produs de copii in pauza asupra unor procese
    simple de  invafare  ale acestora, m Combaterea  zgomotului si  vibratiilor  - a  2-a
    Consfatuire, Bucuresti, 1964.
85. Tarriere, C., Wisner, A., Effets des bruits significatifs et non significatifsau cours d'une
    epreuve de vigilance. Trav.humain, 1962,25, 1-28.
86. Teichner,  W.  H., Arees,  E.,  Reilly,  R., Noise  and human performance, a psycho-
    physiological approach. Ergonomics, 1963,  6,83-97.
87. Thackray, R. I., Touchstone, R. M., Recovery of motor performance following startle.
    Percept, mot. skills, 1970, 30, 279-292.
88. Thackray, R. I., Touchstone, R. M., Jones, K. N., Effects of simulated sonic booms on
    tracking performance and autonomic  response. Aerospace Med., 1972,45, 13-21.
89. Warner, H. D., Effects  of intermittent noise  on human target  detection,  Human
    Factors, 1969,11, 245-250.
90. Warner, H. D., Heimstra, N. W., Effects of intermittent noise on visual search  tasks of
    varying complexity, Perceptmbtskills, 1971, 32, 219-226.
91. Watkins, W., Effect of certain noises upon detection of visual signals. J.exp.psychol.,
    1964,67,72-75.
92. Weinstein,  A., Mackenzie, R. S., Manual performance and arousal. Percept.mot.skills,
    1966, 22, 498.

                                        377

-------
 93. Wilkinson, R., Some factors influencing the effect of environmental stressors upon
     performance. PsychoLBnlL, 1969, 27, 260-272.
 94. Wisner,  A., Dutheil, R., Foret, J.,  Paraht, G, Signaux  sonores en  milieu bruyant.
     Complexite de la situation reette et experiences de laboratoire. Proceed 17-th  Int.
     Congr.Appl.Psyehol., Liege, 1971.
 95. Witecki, 1C, Determination of the  time of simple reaction to visual and acoustic
     stimuli under  the  conditions of interfering noise. Polsk.tygod. lekarst,  1970,  25,
     1012-1015.
 96. Wittersheim, G., Hennemann, M. C., Effets de deux types de bruits industriels sur la
     precision et la rapidite de choix dans  une tache psychomotrice de longue duree, in
     Problemes actuals de la recherche en ergonomie, Paris, Dunod, 1968.
 97. Wolff, B. B., The relationship of experimental pain  tolerance to pain  threshold: a
     critique of Gelfand's paper. Canadian j. psychol., 1964,18, 249-253.
 98. Woodhead, M. M.,  Effect of brief loud noise on decision making.  J.acoustsoc.Am.,
     1959,31, 1329-1.
 99. Woodhead, M. M.,  Searching a visual display in intermittent noise.  J.sound & vib.,
     1964,7,157-161.
100. Woodhead, M» M., The effect of bursts of noise on an arithmetic task. Am.j.psychol.,
     1964, 77,627-633.
101. Woodhead, Muriel,  M., Performing a visual task in the vicinity of reproduced sonic
     bangs. J.sound. &vib.,  1969, 9, 121-125.
                                        378

-------
          SIMILAR AND OPPOSING EFFECTS OF NOISE ON PERFORMANCE

                                    L. R. Hartley

     Medical Research Council, Applied Psychology Unit, 15 Chaucer Road, Cambridge

     Since the  1950's there have been many reports of effects of continuous broad-band
noise on performance. The general observation has been that performance in levels of less
than about 95  dBC is often improved by noise, but  that performance in levels of greater
than about 95 dBC is often impaired by noise. In this present series of experiments we have
continued to use broad-band noise with a spectrum of equal energy  per octave. In all the
following  experiments it has been presented at 95 dBC (the noise conditions) or at 70 dBC
(the quiet conditions).
     In many studies x>f performance in continuous noise, the adverse  or beneficial effect of
the noise  has been observed most conspicuously towards the end of the test, following
about 15  min of performance in noise. One question arising from these studies concerns
whether this effect of the noise arose because of an interaction with time-on-task or whether
it arose because there was  a cumulative effect of the noise, independent of time-on-task.
One way  we have attempted to answer this question is by comparing the performance  of
four groups of subjects. Two groups were exposed to 10 min of noise  or quiet, respectively,
during which time they did two short  tests. The other two groups were exposed to noise or
quiet, respectively, for a further 20 min and allowed to read magazines, before performing
the same two tests. Hence one group received 10 mm of noise and the other group 30 min
of noise, but both groups were tested for only 10 min.
     The experimental test was a modified version of the Stroop color interference test.
This version of the Stroop test was devised by Ray Adams of the Applied Psychology Unit.
In this test, color names are printed in incongruous hues. The subject is asked to select the
hue of the name on the left of the sheet and then cross out  the name of that hue from the
set of names on the right of the sheet.  He is instructed to work as quickly and accurately as
possible, completing as many lines of the test as he can. To assess the degree of interference
caused by having hue and  name attached to  the same response, performance on the coloured
sheets was compared with performance on a similar  test  where all the colour  names are
printed in black ink. In this latter control test the subject  simply crosses out one of the
names on  the right of the sheet that matches the name  on the left of the sheet. Both tests
are  performed  for 5 min each, and scores are the number  of lines  completed. Order  of
presentation of control and experimental tests is counterbalanced as is the order of pre-
sentation of noise and quiet.
     If, as has been argued by Broadbent (1971), noise  affects the filtering of one stimulus
from another, when both are present at the same time,  selection on the basis  of hue when
the predominant color name is also present should be impaired by noise.
     Figure  1 shows  the number of lines of material correctly completed in the  two tests.
The upper part of the figure shows performance on the control sheets on which the names
were  written  in black   ink. The lower  part of the figure  shows the experimental
(incongruously-colored) sheets. The main point of interest is the comparison between 10

                                        379

-------
 MEAN  NUMBER  OF LINES  COMPLETED  IN NOISE AND QUIET


o
UJ
H. ,
y
j
Q.
5
8
V)
5
HI
u.
O

££
HI
CO
2

-

IOU

I7O

/ ~

I6O

ISO


I4O

I3O


I2O


no
IOO
177
,72 	 -- CONTROL
— — 	 SHEETS
170 (69


.

IO MIM FYPfVllinF

	 	 3OMIN. EXPOSURE

-
132
"""~— - — —
"~* "^ — 	 . |25
-— EXPERIMENTAL
	 T24 SHEETS
^^^- 	 '
115
"

QUIET
                                           NOISE
Figure 1 The number of lines completed  in the experimental interference  and  on the control
noninterference tests in noise and quiet.
minutes of quiet or noise and thirty minutes of quiet or noise. The group of subjects who
had only 10 minutes of noise during the tests showed reduced interference in noise com-
pared to quiet; noise speeded up performance on the colored experimental sheets. In com-
parison, the group who had 30 min of noise in all, 20 min of it prior to the  tests, showed
increased interference in noise as compared to quiet. The longer exposure to noise slowed
down performance on the colored sheets, as compared  to quiet. This interaction between
exposure duration and  interference implies that there  are two different effects of noise
depending on the duration of exposure: an initial beneficial and a later detrimental effect.
These effects appear prominently hi the  experimental test involving perceptual selection,
but not in performance on the control test which is purely a measure of speed. These results
are consistent with  the view  that noise  has  a cumulative adverse effect independent  of
time-on-task. It is this cumulative effect of noise which  appears to impair perceptual selec-
tion and probably makes the organism increasingly distractable.
                                       380

-------
     A further noteworthy feature of this test is that there are no overt auditory cues
associated with performance. Hence the  effect of noise could not have been mediated by
auditory  masking, particularly in view of the different effects of the long and short expo-
sures.
     A large proportion of the results demonstrating  the adverse effect of noise  and its
interaction with incentive and  sleep loss have employed the 5-choice serial reaction test,
shown here  in Figure 2. In this test there are 5 brass discs corresponding to 5 possible light
sources. Using a stylus  the  subject taps  the disc appropriate to the light illuminated.  The
light promptly extinguishes and another source is lit. The  subject works as quickly  and
accurately as possible tapping as many correct discs and making as few errors as he can. A
third score is the number of pauses or gaps greater than \1A sec. between responses. Noise
over 95 dBC has almost  invariably shown  an adverse effect on the number of errors and gaps
made. Figure 3 shows the mean number of errors made at various intervals following the last
response, in noise and in quiet. The figure  on the left shows the mean absolute number of
errors made in noise and in quiet by a group of n subjects. The later part of the distribution of
errors is  quite similar to that of correct responses,  but as the figure shows, a  number of
errors are made with a latency of less than  200 msec. The figure on the right shows  the
errors plotted as a percentage  of the totals in quiet and in noise.  This  shows that noise
increases by an equal amount the number of errors made at each latency following the last
response. Noise  does not selectively increase the anticipatory or the  slower misplacement
type of error.
     Figure  4 shows similar latency distributions of correct responses as a function of noise
on  the left of the figure  and  as a  function of time-on-task on the right of  the figure.
Considering the effect of noise, there is an increase  in the proportion of responses with a
latency of 1000 msec or more  in noise as compared to quiet. Comparing the distribution of
responses in the first and  second halves of a test on  the right of the figure, time-on-task
causes a  similar increase in the proportion of responses with a latency  of 1000 msec or
more. Both  noise and time-on-task are similar in this respect. The gap score in the 5-choice
test records  the number of responses in the extreme tail of these distributions.
     The fact that noise can have at least two different effects on performance depending
on  exposure duration may be related to  qualitatively different aspects of noise; namely the
loudness or annoyance  experienced  in noise on the  one hand and the monotony and per-
ceptual isolation experienced on the other.  The results of the following experiments, involv-
ing  the  5-choice  test,  go  some way to support this dichotomy between loudness and
monotony in noise.
     In the  following experiment, the interaction of noise with headphone and free-field
presentation was  considered.  This  interaction is  of interest  since monophonic noise
binaurally presented over headphones is less variable and more isolating perceptually than
free-field  noise  and the perceived loudness may be less. Noise and  quiet presented  over
headphones was compared with the same sound pressure levels presented in the free field.
Subjects  performed the 5-choice test for 40 min under each of these 4 conditions.  The
subjects wore Knowles  miniature microphones  at the entrance  to each external auditory
meatus in ah1 conditions. Sound pressure  level was adjusted to 95  or 70 dBC in the free-field
and the same sound spectrum was presented over headphones at the same sound pressure

                                         381

-------
 5  Neon  lamps
 5 cm. apart
       Stylus
   18 cm.  long
1-5  cm. diameter
             30'
5 Circular brass discs
3 cm.  diameter  and
9 cm.  between centres
                            Figure 2 The 5-choice test.
levels as recorded by the microphones in the external meatus. Sound pressure level was
monitored continuously and adjusted as necessary during  the test.  Figure 5 shows the
difference scores between quiet and noise presented over headphones or in the free-field.
Gaps are shown on the 'left of the figure  and errors on  the right. The two modes of
presentation have approximately equal adverse effects overall although the pattern of
impairment is rather different. There is  a greater adverse effect of headphone noise upon
gaps and a greater adverse effect of free-field noise upon errors in the test. This interaction
in the-effect upon errors and gaps may be related to the  greater perceived loudness of
                                   382

-------
         24    6    8    IO   12
         RESPONSE TIME x  IOO MSEC
14
                                                                = Q-Q
                                                        Q_
                                                                                  Q N
 24    6     8    IO    12    14
RESPONSE TIME  x IOO MSEC
        Figure 3 Latency distribution of mean errors and mean percent errors in quiet and noise.
free-field as compared  to  headphone noise on the one hand and the greater perceptual
isolation and monotony in headphone as compared to free-field noise on the other.
     In a further experiment, the perceptual isolation and monotony accompanying con-
tinuous noise was reduced and the arousal or annoyance quality of the noise maintained by
presenting  the  noise  intermittently. Subjects  performed the 5-choice test for 40 min  in
continuous quiet, continuous noise and in intermittent noise presented in free-field. Intermit-
tent noise consisted of bursts from 1-5 sec long with an average  duration of 3 sec. Average
length of the quiet intervals was  1.5 sec. One group of subjects performed the test with
immediate knowledge of results and a second group performed without this incentive. The
difference between continuous and intermittent noise lies in gaps as Figure 6 shows. Con-
tinuous noise produced approximately twice the increase in gaps that intermittent noise did,
but overall, intermittent noise produced as large an increase in errors as continuous noise,
although there were minor differences  between  incentive conditions. Reducing the
                                         383

-------
mm
2*T-3
gl-9
1*8
.z
< 1-6
g 1-5
O ''4
1- 1-3
Z
ujl.2
O , ,
it '•'
gio
£ .9
*) -8
0 -7
bl
a -6
n
z
< -3
w
5 2
fi .t
V3
O /-v
•
'
•
.
•
-
.

•

•

• •
1
•

•
•

-

•
_


FL
I










i

2 4
?2-0
_ Z .

Q








1







i



1
-1-8
i _j
Q-Q-Q H,.6
N=N-N H(.S
01-4
















i


-

0
















6 8
H-l-3
Z
ui|.2
0
ail
al-O













Q











«2 Q
1 < ->
,2 -8
0 .7
Ul
CC 6
ffl ^4
Z
p ^ -3
Q!W (^J
1 2 -2
1 —
P O

•
'
•

:
.


•

•
-
•
-
-

.
•

.

•



i

























IM
1





































.


= FIRST HALF IN Q-Q
2=SECOND HALF IN Q-Q

i



















a





H
























mm
a











'





m
| |
«m
i
'• •• %
n
i i
J ;
§• 1
IO 12 14 -» ~ 2 4 6 8 IO 12 14
           RESPONSE TIME x IOO MSEC
RESPONSE TIME x IOO MSEC
      Figure 4 Latency distribution of mean percent correct as a function of noise and time-on-task.
monotony and isolation accompanying noise by varying it appears to greatly reduce the
adverse effect of noise upon gaps but leave the adverse effect upon errors unchanged.
     Finally, subjects performed  the 5-choice test in continuous free-field noise and quiet
with and without ear-defenders on. In this experiment, the perceptual isolation of noise was
maintained but the sound pressure level at the ear reduced by wearing ear protection. The
main finding was that, as Figure 7  shows for the difference scores between quiet and noise,
ear-defenders interacted with noise, greatly reducing the adverse effect of noise in the first
half of the test but causing as large an adverse effect as continuous noise in the second half
of the test. These results appeared  reliably in gaps but in this experiment subjects  failed to
show an effect of the 95  dBC noise upon errors in the test. Hence, reducing  the sound
pressure level by wearing ear protection appears to be beneficial initially, but the accom-
panying monotony and perceptual isolation may nevertheless have a detrimental effect later
in the test.
     In summary,  the results of these experiments are consistent with the view  that noise
has a cumulative, adverse effect, increasing with exposure duration, independent of  time-on-
tiSk. The cumulative effect may, however, have two rather different components corre-

                                         384

-------
   DIFFERENCE IN  GAPS AND ERRORS  BETWEEN  NOISE AND QUIET
    15
    14
    13
    12
    II
    IO
     9
     8
§
cc
£
UJ
DC
O
tn
a.
z
<
UJ
7
6
5
4
3
2
 I
O
                                   13-92
                              HEADPHONES
    IO-5O
        6-92
                                           6-56
                                  6-26
JREE-FIELD
      5-74

HEADPHONES
                                         O-3C«
       FIRST HALF           SECOND HALF   FIRST HALF             SECOND HALF
              TEST  HALVES                       TEST  HALVES
 Figure 5 Difference between noise and quiet with headphone and free-field presented with gaps on the left
 and errors on the right.
 spending to the loudness on the one hand, and to the monotony and perceptual isolation
 accompanying noise, on the other. The effect of loudness on performance may predominate
 in the short exposure, whereas the adverse effect of perceptual isolation and monotony may
 predominate following many minutes of exposure to continuous noise.
                                      385

-------
      MEAN  GAPS  AND ERRORS
         IN QUIET AND NOISE
o.
<
O
UJ
2
24
22
2O
18
16
14
12
10
8
6
4
                       GAPS
GC
O
CL
CL
UJ
 UJ
18
17
16
15
14
13
12
II
IO
9
8
7
6
                       19-28
      IO-I7
          INCENTIVE
                           16-50.
                                            22-61
                                           16-23
                           7-94
                          NO INCENTIVE
                      ERRORS
                  CN
                      12-28
     IO-94
     963
              IN
     8-11
                       7-78
12-83
12-16
                          7-83
    FIRST
    HALF
                 SECOND FIRST
                 HALF   HALF
              HALVES OF TEST
                                          SECOND
                                          HALF
Figure 6 Gaps and errors in continuous (CN), intermittent noise (IN) and quiet (Q).
                       386

-------
    DIFFERENCE  BETWEEN  QUIET
       & NOISE WITH & WITHOUT
            EAR-DEFENDERS
S.
      12
      II
      to
S   8
*   7
3   6
2   5
i   4
<   3
iu   2
2   I
" 9-56
                 NO  DEFENDERS
                                 6-5O
                                 4-28
                 EAR-DEFENDERS
         O-79
      FIRST HALF
                            LAST HALF
               TEST HALVES
     Figure 7 Difference between noise and quiet in gaps, with and without ear defenders.
                  REFERENCES

Broadbent, D. E. Decision and Stress, Academic Press, 1971.

                    387

-------
       THE EFFECTS OF DIFFERENT TYPES OF ACOUSTIC STIMULATION
                               ON PERFORMANCE

                                 C. Stanley Harris
                       Aerospace Medical Research Laboratory
                     Wright-Patterson Air Force Base, Ohio, USA

                                   ABSTRACT

    Most studies conducted in our laboratory on the effects of acoustic stimulation on
human performance have produced results showing: (1) no adverse effects,  (2) transient
effects (the adverse effect did not continue throughout the testing period), or (3) effects so
small that the subjects would be expected to adapt with repeated exposure. However, a few
experiments showed adverse effects. High-intensity broadband noise in which subjects wore
ear protection in  levels to 140 dB SPL adversely affected performance on a rail balancing
task and on a Hand-Tool Dexterity task (HTD). The adverse effect on the HTD task resulted
in part from the noise directly vibrating the test apparatus.
    Although an adverse effect  of high intensity noise  on performance was not easy to
demonstrate when measuring  for short time  periods, a lower  intensity level (105 dB) of
broadband noise presented for a longer  time period was found to adversely affect perform-
ance on a continuous task.

INTRODUCTION

    In the Aerospace  Medical  Research Laboratory  (AMRL), the effects  on  human
performance of several types of acoustic stimuli have been studied. However, the emphasis in
the present  paper was limited primarily to  a discussion of the effects of two types of
broadband noise. The first type was a free-field, low-frequency broadband noise which was
presented  in a  reverberation chamber to subjects  wearing  ear protectors. This noise  was
presented  at intensity levels from ambient to 140  dB SPL,  and is characteristically experi-
enced  by personnel working in the vicinity of operating jet engines. The second type was
relatively high-frequency broadband noise presented through earphones. This noise  was
presented  at intensity levels up to 115  dB. Most studies in  the literature on the effects of
noise on human performance have used noise similar to the second type of broadband noise.
The effects of several other types of acoustic stimuli have been studied as a followup to the
results obtained in  the reverberation  chamber. Research has also been conducted using
impulsive noise and noise combined with vibration. A discussion of the results of the latter
experiments has either been omitted or included under the task used for measuring perform-
ance.  However, in order to give  a complete  picture  of the research effort, all studies are
summarized in Table 1.

HIGH-INTENSITY BROADBAND NOISE (TO 140 dB)

    Nausea, vertigo, incoordination, and fatigue have been reported by individuals exposed
to jet  engine noise  (1, 19).  These  reactions, as well as  mental confusion (4), have been

                                        389

-------
 attributed to vestibular stimulation and to reflexes elicited by vibration of the skin, muscles,
 and joints. Any of these symptoms could lead to a reduction in performance efficiency.
 Furthermore, since many proprioceptive reflexes occur  with little or no conscious aware-
 ness, performance efficiency may be  affected by noise  intensity levels lower  than those
 necessary to elicit subjective symptoms.
     Several years ago the authors of the Benox  Report  (1) summarized the problems and
 explored  the effects of high intensity noise on man. Since that report, few studies have been
 conducted using intense noise. The urgency in understanding the effects of this noise on
 man has been reduced since ear protection is currently worn in intense noise environments.
 Nevertheless, intense  noise may adversely affect the performance of personnel even when
 they wear ear protection. In AMRL performance studies, the noise levels in the ear canals
 (up to  115  dB) of the  subjects were  no higher than have been used by several previous
 investigators. However, extra-auditory  effects of noise at  ambient levels to 140 dB added to
 the auditory stimulation. These studies were conducted in a reverberant noise chamber.
 Subjects either wore the same ear protection in noise intensity levels of 120 dB, 130 dB, and
 140 dB or were exposed to an ambient level of 140 dB with different types of ear protec-
 tion. In every study, each group was presented with four noise conditions in four counter-
 balanced orders. Testing within each condition was for a  brief period of time, usually 5 min
 to  20 min. Eight to 20 subjects were used in each group. A practice session was given to all
 subjects on the day prior to experimental testing. Subjects were also given audiograms and
 those without normal hearing  were rejected. Those accepted as subjects were briefly ex-
 posed to the acoustic stimuli used subsequently in the experiment. The noise presented in
 the reverberation chamber peaked at the low frequency end of the spectrum. Both  the
 intensity levels and the spectra are given in Figure 1 for  the ambient conditions and when
 the levels were attenuated by earplugs or earplugs plus muffs.


Cognitive Performance in Intense Noise

     Dickson and Chadwick asked jet mechanics to  describe their subjective experiences
when standing close  to an accelerating jet engine.  Most  reports were  vague, but were
described  best by "one  of the engineers who said  he experienced a momentary sense  of
imbalance accompanied  by  a lack of power to think (4)." In addition, the authors of the
Benox Report (I) have stated that people in high-intensity noise tend to forget or neglect to
follow instructions and to work hurriedly but require more time to complete a task. Both
studies suggested impairment of cognitive ability during exposure to intense noise.


Short-Term Memory

     The procedure used by Korn and Lindley (17) was  adopted for measuring short-term
memory (STM). This  task required the subject to  remember the order of presentation of
nine consonants. The  consonants were projected  on a screen for six seconds, then removed,
and the subject was allowed 10 seconds to write his answer. Testing was conducted in blocks
of 15 trials, and each  slide presentation of the nine consonants was a trial. This task seemed

                                        390

-------
                     s
150;
130-
110 -
90 •
70 -

7
140-
120-
K)0 ;
80-
60 -
40 r

130 •
110 -
90 -
70 -
50 •
                                   EAHMUFFS 8 PLUGS
                        SO-
                                                    AMBIENT
                                   EARMUFFS 8 PLU6S
                            J	I
EARMUFFS A PLUGS'

 I     I     I
                           OVER 75-
                            ALL ISO
             150-
             300
     300-
     600
600- 1200-
1200  2400
2400- 4800-
4800  9600
                                   OCTAVE PASS BANDS (Hz)
Figure 1 Ambient noise spectra and calculated noise levels in ear canals after reduction of noise by ear
protection.
an appropriate one to use because little or no learning was involved, the instructions were
easy to understand, yet the subjects seldom remembered the position of all nine consonants.
     Four experiments were conducted in noise. The results are summarized in study 9 in
Table  1. The STM  task showed no sensitivity to  broadband noise in the reverberation
chamber either with consonants chosen on the basis of high usage (Group 1) or low usage in
the English Language (Group  2).  Similarly, no sensitivity was  shown,  using high usage
consonants, to broadband  noise presented through  earphones (Group 3) or to broadband
noise that varied in the low cut-off frequency (Group 4).

Discrimination Task

     In this task six symbols were presented in one-inch-square boxes. The subject com-
pared each of four boxes with one centered above them as to whether the same or different
symbols occupied the same relative spatial  position in the respective boxes.  He then noted
the number of differences on a line directly under each block (see  Figure 2). After the
completion of a set of comparisons, the subjects advanced to other sets of five boxes until
the time limit expired. Performance was measured in two 4-min periods, with a 1-min rest
given between periods. The score for each period was the number of boxes completed minus
the number of errors made. A symmetrical exposure group and an asymmetrical exposure
                                         391

-------
                                      0

                                      H
V

C
/ V
S D
I



0 X
S C
N
0
H
V
0
z
                                                                  X

                                                                  C
                         Figure 2 Example of Discrimination Task.
group performed  the task on two practice days and during exposure to noise on four test
days. Noise at 130 dB and 140 dB had a small detrimental effect on the performance of the
asymmetrical group during  the first 4-min period but not  during the second 4-min period
(Figure 3). No adverse effect was obtained for the symmetrical exposure group. Therefore,
only a transient  decrement occurred and only for the asymmetrical group. Prior to the
experiment, the subjects had never been exposed to this type of noise and it seems unlikely
that more experienced  subjects  would  have  shown  even  a transient  decrement  in
performance.

Summary

     At  the  time these studies were conducted, an adverse effect of noise on  cognitive
performance seemed assured. Most of the literature on  the effects of noise on  human
performance did  not seem relevant, and  anecdotal  reports of people  working in noise
strongly suggested that mental processes were adversely affected. However, little evidence
was found to indicate an adverse effect of this type of noise on cognitive performance (see
studies 3, 9, & 10 in Table 1). Testing has, of course, been limited. More sophisticated  tasks
presented for longer periods of time may show more sensitivity  to noise in this range.
Nevertheless,  the  effect of high intensity jet  noise on cognitive performance is  less  than
many have previously believed.

Psychomotor Performance in Intense Noise

     Noise was  expected to have a greater effect on  psychomotor  tasks  than on  cognitive
tasks, because most reports of individuals working around jet engines refer to equilibratory
and  postural  disturbances rather than to mental confusion or disturbances. Parrack  (19)
points out that personnel working in noise levels up to  160 dB report heating of the  skin,
strong sensation of vibration in various parts of the body, sensations of muscular weakness,
                                        392

-------
                       50
                     S
                       45
                     (0
                     s
                     z
                       40
                        CONTFK
                                             O PLUGS ONLY
                                             • PLUGS 8 ONE MUFF
ITROL 120  I3O I4O CONTROL 120 130  140
 NOISE CONDITIONS (dB re. O.OO02 Dyne /cm*)
Figure 3 Mean corrected score of noise conditions for asymmetrical and symmetrical exposures for both
4-minute periods.
and excessive fatigue. Several cases of staggering, falling, and feelings of forced movement
have also been reported (1).  Therefore, high-intensity noise can adversely affect psycho-
motor performance through  effects on human physiology or through mechanical inter-
ference with motor movements.

Rail Task

     The rail task initially used in our studies was based on an ataxia test battery developed
by Graybiel and Fregley (5). Various parts of the task were dropped  because they showed
no sensitivity to noise conditions (6, 12, 18). Since most of the results obtained with the
task have been published, detailed experimental procedures are not presented. To perform
the task, the subject was asked to stand on a narrow rail in a heel-to-toe manner with his
arms folded across his chest and with his eyes open  (see Figure 4). Two rails were used, one
1% inch  wide and the other % inch wide. The balance time on both rails was scored to the
nearest second, beginning when the subject assumed the correct position on the rail, and
ending when he lifted a foot, unfolded his arms, or fell off the rail. The maximum score for
each trial was 60 seconds; if the subject was still balanced on the rail at the end of this time,
the trial  was discontinued. Five trials were presented on each rail, in most experiments.
                                          393

-------

                       Figure 4  Subject performing on the Rail Task.
     The rail task has been shown to be quite sensitive to acoustic stimulation (6). In the
reverberation chamber, decrements were found to be related to intensity of noise stimula-
tion and to stimulus asymmetry (6).  Detrimental effects on performance were found at 140
dB  regardless of the type of ear protection worn. As shown in Figure 5,  the asymmetrical
exposure produced decrements in performance at the 120 dB and  130 dB levels whereas
improvement occurred at those levels with balanced exposures. After-effects of exposure to
noise were found for asymmetrical exposures but not for symmetrical exposures (6). Using
pure tones presented  through earphones, evidence was obtained for a frequency (Hz) sensi-
tivity of the task. Frequencies around 1000 Hz to  1500 Hz seem to have  more detrimental
effects than either the higher or lower frequencies in the range of 100 Hz to 2500 Hz (8).
Furthermore, an asymmetrical stimulus presented intermittently (2.86 pulses per second) at
a frequency  of 1000 Hz has produced  large decrements on the rail task. The decrement
produced by intermittency added to the effect produced by asymmetry' (15) (see Figure 6).
Finally, an experiment has demonstrated that stimulus asymmetry has an adverse effect on
rail  task performance because one ear is  always stimulated more strongly than the other anil
not because the same ear is always stimulated rnore intensely (see study 6 in Table  1).
                                       394

-------
                      3O
                      20
                       IO
                       10
                     -20
                     -30
 AO BALANCED EXPOSURES (plugs only)
 AC BALANCED EXPOSURE (muffs 8 plugs)
- A« UNBALANCED EXPOSURE (plugs 8 one muff *
                               120         130        140
                          NOISE CONDITIONS (dB re. 0.0002 Dyne /cm*)

       Figure 5 Percent change from control means for each type of ear protection (Rail Task).
Hand-Tool Dexterity Task (HTD)

     The HTD was designed to measure motor performance largely independent of mental
factors. The equipment for the task  consisted of three horizontal rows of nuts and bolts
mounted on a wooden stanchion (see Figure 7). The nuts and bolts were of three different
sizes and  four  of the same size were mounted in a row. The subject's task, using two
adjustable wrenches and a screwdriver, was to remove all of the bolts from the left upright
and transfer them to the corresponding rows on the right upright. The score was the time
taken to make the transfer. Two groups were tested on the task during exposure to noise
(see  study 7 in Table  1). Significant increases in the time  taken to  complete the task
occurred at 130 dB  and 140 dB for a symmetrical exposure  group and at  140 dB for an
asymmetrical group  (see  Figure  8). The large difference in the initial  ability  and in the
different rate of learning of the groups made it impossible to determine the relative effects
of symmetrical versus asymmetrical exposures (7). Nevertheless; noise at the 140 dB level
adversely affected the performance of both groups. Furthermore,  the adverse effects may
have resulted from either auditory stimulation or extra-auditory or a combination of both.
The extra-auditory stimulation may have produced adverse effects either  by vibrating the
body  and limbs or  by vibrating the  task.  The latter explanation was supported by the
                                         395

-------
                        10
                     rt
                     <0
                    O
                    C
                    U.

                    g
                        10
                       20-
                        30-
                       4C -
 I            I
 O ASYMMETRICAL
 A SYMMETRICAL
	CONTINUOUS
	INTERMITTENT
                                 1
 1
            105
                                85          95
                                    INTENSITY MB
    Figure 6 Percent change relative to 65 dB for subjects tested using the 1000 Hz tone (Rail Task).
statements of three subjects who mentioned that they were bothered by the shaking of the
smallest bolts in their stations at 140 dB. Direct observations suggested to the experimenter
that the shaking of the nuts and bolts did add a few seconds to the time taken to complete
the task at 140 dB but probably had no effect on the completion time at 130 dB.
     In a subsequent experiment (21) subjects were tested on the HTD task during auditory
stimulation alone. The noise spectra and the levels approximated those delivered to the ear
canals  of subjects  wearing ear protectors  in  a free-field broadband  noise  of 140  dB.
Specifically, the subjects were presented: control, 100 dB, 115 dB,  and 115 dB in the left
and  100 dB  in the right earphone (see study 8 in Table 1).  The noise presented in  this
manner did not adversely affect  performance on the task.

Summary

     The rail task was quite sensitive to acoustic stimulation; however, these results  are
more of theoretical than practical interest since this sensitivity would not have occurred
using wider rails. On Jhe other hand, the Hand-Tool Dexterity  test was more representative
of tasks performed by individuals working in high  intensity noise.  Therefore, decrements
obtained on the HTD test are of greater concern than decrements obtained on the rail task.
                                        396

-------

                                               .,-

                   Figure 7 Subject performing the Hand-Tool Dexterity Test.
We are still uncertain of the manner in which the HTD test is affected by noise. It" perform-
ance was adversely affected by  a direct  mechanical effect of noise  on the task, then the
results are not as important as an adverse effect produced  by interference with the motor
coordination  of man.  The reason is that changes in the  task equipment, dimensions, or
procedures often can  reduce  or eliminate problems caused by  direct mechanic;!, inter-
ference.

BROADBAND NOISE  (EARPHONES)

     In these studies, broadband noise was presented through earphones at intensity levels
to 115  dB. The experiments described earlier in which noise levels and spectra, presented
through earphones,  approximated  the acoustic stimuli in  the ear canals of the subjects
exposed to an ambient noise level of 140 dB wearing different types of ear protection, fit
the category of this  type of noise. Although these were control studies with the purp<
comparing the results  with those obtained in ambient noise in the reverberation chamber,
performance was measured during short  exposures to broadband  noise. No adverse elfects
were obtained on the short-term  memory test, or on the Hand-Tool Dexterity  test; however,
there was a small but statistically significant reduction in the amount of time subjects would

                                         397

-------
                         310
                         300
                       E.290
                         280
                        270
                        260
                                   	ASYMMETRICAL EXPOSURE
                                   -- SYMMETRICAL EXPOSURE
                               CONTROL   120    130     140
                            NCISE CONDITIONS (dB re 0.0002 Dyne/cm2)

Figure 8 Mean time to complete Hand-Tool  Dexterity Test at each  noise level for asymmetrical and
symmetrical exposures.
balance on the rails. This noise was of a lower frequency than the noise used with other
tasks measured under this heading, and there was probably less reason to expect an effect on
performance.

Serial Search Task (SST)

    All two digit numbers from 10 to  99 were used twice in constructing a test sheet for
the serial search task (see Figure 9). The numbers were presented in pairs in six different
columns and 15 pairs were in each column. Each two-digit number occurred once as ,the first
member of a pair and once as the second member of a pair.  The subject started each  sheet
by looking for the  number  10  as the first member of a pair. When he found it, he wrote
down the  number which was the corresponding second number of the pair on a piece of
paper and  proceeded to look for this new number as the first member of a pair, and so on.
    In two experiments noise had an adverse  effect on the serial  search task. In one
experiment the effect was transient (13), and in the other the effect was constant across the
testing period and increased  with days of presentation of the task (14). In the first experi-
ment, the performance of one group of subjects was measured during exposure to a broad-
band noise of 105  dB (see Figure 10), while the other group served as a control. All the
                                        398

-------
12
41
56
51
84
69
66
47
95
20
40
16
71
45
13
61
47
41
83
30
73
35
69
42
74
96
49
76
31
25
79
)0
49
83
50
65
53
21
82
94
80
91
22
44
63
60
88
97
75
92
91
72
50
39
32
22
48
33
64
38
19
23
62
81
26
25
85
37
48
64
96
78
67
98
53
39
93
44
54
40
53
15
94
95
14
16
55
51
56
59
52
14
6)
30
57
77
70
86
74
87
13
29
55
11
72
70
37
36
98
52
57
43
82
23
71
34
90
45
29
46
36
93
27
68
46
60
43
17
97
54
42
73
76
75
59
23
58
11
89
27
62
26
68
87
31
21
20
65
86
84
92
33
99
15
24
33
31
32
88
39
34
28
90
89
35
63
85
G6
19
77
80
24
17
18
67
79
10
78
13
12
                        Figure 9 Sample sheet of the Serial Search Task.
subjects were tested on the task during 5 days, a preliminary training day and 4 test days.
The pattern of testing on the preliminary day was: test 6 min, 1st sheet—rest 3 min—test 12
min, 2nd sheet-rest 3 min—test 12 min,  3rd sheet-rest 3 min-test 12 min, 4th sheet. This
same procedure was followed throughout the remaining four days of testing with the excep-
tion that the noise group was informed  on the  second day  that they would be asked to
perform the task in noise each day after the six minute warm-up. The noise was turned on 1
min before testing began in the  first  12-min testing period. The noise was presented con-
tinuously throughout the remaining testing periods and the rest intervals.
     An analysis of variance calculated on the  data revealed no significant effect for noise;
however, there were significant effects for the  noise x trials interaction, and for the noise x
days x  trials interaction. The noise x trials interaction was due to the large difference
between the noise and control group on the first trial, which was statistically significant (see
Figure 12). The three way interaction was obtained mainly because the difference between
groups on the  first trial was larger on the 4th  day than it was on the preceding three test
days.
     In  the second experiment, three groups of subjects were tested in the same manner as
the previous groups, except that the rest periods were omitted. One group was tested during
exposure to intermittent noise. This group was presented the same spectrum and intensity
level (105 dB) as the continuous noise group; however, the noise was interrupted at a rate of
2.86 times per  second and a duty cycle of 50% was used. The results showed an effect for
noise that approached significance (p<.10) and a statistically significant effect for the noise
x days interaction. This effect is illustrated in Figure 11. The difference between the control

                                         399

-------
                    i  no
                    9
                    §  100
                    d
                    s
                    §  90

                    ij  so
                    Id
                    oc
                    w
                       70
                       60
                       50
                       40
                                I   I    I    I   I
                                       I
                           3L5 63  125  250  500  IK  2K  4K  8K  I6K
                                 OCTAVE BAND FREQUENCY (Htrtz)
                       Figure 10 Octave band levels under the earphones
group and  the  noise groups  increased as a function of days with the largest  difference
occurring on  the last day of testing. No significant differences were obtained on day 1 or
day 2, but both noise groups differed significantly from the control on day 3 (beyond the
.05 level), and  on  day  4 (beyond the .01 level). No significant differences were obtained
between the continuous and the intermittent noise groups. In these experiments, the rest
periods produced more learning on the SST and allowed the subjects in the noise group to
perform more like the control group on the last two trials on each day.


SUMMARY AND DISCUSSION

     High-intensity broadband noise (up to 140 dB with ear protection) had little effect on
cognitive performance.  No adverse effects were obtained with a short-term memory task, a
paper & pencil  maze task (see study 10 in Table 1), and only a transient effect was obtained
using a  discrimination  task. The transient effect was obtained only for an asymmetrical
exposure group and not for a symmetrical exposure group. It may be  that for this type of
noise to adversely affect cognitive performance, the same conditions apply as Broadbent (2,
3) has suggested are necessary for obtaining a detrimental effect on performance at lower
intensity levels  of noise. These conditions are: (a) testing should be continuous  for a mini-
mum of  30 to  60  minutes, (b) noise of 100 dB or above should be used, and (c) the task
should be experimenter paced or one that requires the continual attention of the subject.
     Psychomotor  performance showed  more sensitivity to noise. The rail task was more
sensitive to noise than any other task used. However, the part of the rail task battery used in
                                        400

-------
                       70
                      65
                    O
                    tit
i
8
                      60
                    i
                      55
f
                                   O CONTROL
                                   A CONTINUOUS NOISE
                                   Q INTERMITTENT NOISE
                                         ,
                            ,
                                 I       2       3
                                          DAYS
          Figure 11 Mean number completed per trial on each day for rest and no-rest groups.
most experiments was deliberately selected for its sensitivity and was a difficult task. There-
fore, the practical implications of these findings are unknown. On the other hand, there is
little doubt that Hand-Tool Dexterity performance is adversely affected by noise, and it
seems reasonable to assume that such performance will be adversely affected by intense
levels of jet engine noise. Of course, the conclusion would not be that personnel cannot
perform the HTD task in the noise, but that it would take them a little longer to complete
the task; approximately  10%  longer at 140 dB as indicated  by our studies. Whether the
increase in the time to complete the task was caused by extra-auditory effects of the noise
or by a combination of auditory and extra-auditory effects  cannot be determined at the
present  time.  There was  some indication that the effect  may have been due to direct
interference with the task itself, primarily  because of the shaking of the nuts and bolts in
their stations.  It seems unlikely that this could account for all the decrements observed but
it cannot totally be disputed at the present time.
     In  the studies using the  serial search task we deliberately applied the conditions sug-
gested as necessary for producing a detrimental effect of noise on human performance.
There is no question that noise either adversely affected performance on the SST or slowed
down the rate  of learning/ The effect was one that increased with repeated days of exposure.
If testing had  been stopped  after two days of exposure to the noise then no statistically sig-
nificant effect of noise would have been demonstrated since only on days 3 and 4 did the
                                         401

-------
                       70
                       65
                     o
                     UJ
                     III
                     8 60
                       55
                       50
                     	1	

                      NO REST
                     ./TEST
O CONTROL
A CONTINUOUS NOISE
D INTERMITTENT NOISE
                                 I
                                 I           2           3
                                      TWELVE MINUTE TRIALS

              Figure 12 Mean number completed for trials for rest and no-rest groups.
difference between the noise and control groups reach statistical significance. These results
suggested that if the conditions recommended by Broadbent for producing an effect of noise
on performance are applied, and if a fairly large number of subjects are tested during re-
peated exposure to noise, then a significant effect of noise on performance can be demon-
strated.
                                         402

-------
                       Table 1

NOISE STUDIES CONDUCTED IN AEROSPACE MEDICAL RESEARCH
               LABORATORY, 1965-1973.
/
Study



1




2





3









U





5








6



Author



Ehoenberger
& Harris
(1965)


Nixon, Harris,
& von Gierke
(1966)



Harris & von
Gierke (1967)








Harris St Som-
mer (1966)




Harris (1972)








larris
(unpublished)



Group



1




1





1

2
3

ft



5
1


a


i

2

3


4

1

2


N



16




21





S

12
16

a



8
24


24


10

10

10


10

10

10


Tajis



Tsai- Partington num-
bers, draw a line be-
tween successive num-
bers (1 - 25) randomly
scattered on a paper.
Rail task (CT)
Eyes open (EO)
Eyes closed (EC)
Rail walking (BW)


Rail task (EO & EC)

Rail task (EO & EC)
Rail task (EO 8r EC)

Discrimination task
(DT) (during exposure)
Rail task (after exp-
osure) .
Same as above
RT (EO & EC)


BT (EO & EC)


RT (EO)

RT (EO)

RT (EO)


RT (EO)

Rf CEO.) ' •

RT (EO)

Whole Body
(MB) or
Earphone
Exposure
(&>
EE




WB





WE

WB
WE

WB



WB
EE


EE


EE

EE

EE


EE

EE

EE

A pproxijna t e
Exposure
Time Per
Condition
(Minutes )
30 - 15




25





13

S
8

10



10 '•
10


10


10

10

10


10

10

10

Exp. Conds.
on Same or
on Differ-
ent Day a

Different




Different





Different .

Different
Different

Different



Different
Different


Different


Same

Same

Same


Same

Same

Same


Moise Conditions



(l) Quiet - Quiet
(2) Quiet - 110 dB
(3) 85 dB - 110 dB
(4) 95 dB - 110 dB

(1) Control (70 dB)
(2) Earplugs & muffs in 120 dB
noiae (symmetrical).
(3) Earplugs & 1 muff over right
ear (RE), in 120 dB noise
(asymmetrical) .
Earplugs «t muffs in 70, 120, 130,
and 140 dB.
Earplugs in above noise levels.
Earplugs & 1 muff (RE) in above
noise levels.
Earplugs & 1 muff (RE) in above
noise levels.


Earplugs in above noise levels.
Control & 95 dB to both ears -at
frequencies of 100, 260, 590, 1500,
and 2500 Hz.
Control & asymmetrical exposure of
75 dB (RE) - 95 dB (LE) of above
frequencies.
Symmetrical presentations of 65,
85, 95, 105 dB of 1000 Hz tone.
Asymmetrical presentation (above
tones, presented only to LE).
Symmetrical presentation of above
stimuli with intennittent, 2.86
pulses/second (pps), presentation.
Asymmetrical presentation of
stimuli used for group 3.
1500 Hi tone, 2pps, 65, 85, 95, & E
105 dB. f
Above stimuli with pulses alter- c
nated between ears.

Results



Small but significant effect
when noise level switched from
Quiet to 110 dB (condition).


Small but significant effect on
RT (EO) during asymmetrical
condition.



Effect on RT (EO) at 140 dB.

Effect on RT (EO) at 140 dB.
Effect on RT (EO) at 120 & 140 dB.

Transient effect on DT & margin-
al effect on fiT (EO) at 130 &
140 dB (p< .10).

No effects on either task.
No effect.


Asymmetrical exposure on RT (EO)
1500 Ha significantly different
from control, 100 Hz & 590 Hz.
No effects.

Significant effect, 95 dB & 105
dB different from 65 dB.
Significant effect, 105 dB dif-
ferent from 65 dB.

Significant effect, 85, 95, 105
dB different from 65 dB.
jignificant effects for both
;roups with no significant
Lifference between groups.


-------
TABLE 1 (CONT)
7



8

9











10




11





11




12






13







Harris (1968)



Sooner t Har-
rlB (1970)
Harris
(unpublished)










Harris
(unpublished}



Harris
(unpublished)




Harris
(unpublished)



Harris & Fil-
eon (1971)





Harris (1972)







1

Z

1

1



2

3

-. 4



1




1





2




1



2


1


2

3


16

16-

16

12



12

12

12



12




20





10




15



15


20


20

20


Hand tool dexterity
task (HID).
HID

HID

Short tern memory (STK,
for relative position
of high frequency con-
sonants.
SIM, low frequency
consonants.
SIM, high frequency
consonants.
Sane as above.



Finding path through a
series of mazes, pre-
sented via paper it
pencil. Also finding
most direct route test.
Minnesota Rate of Man-
ipulation (turning).
Medium Taping.
Two Plate Taping.
Purdue Pegboard (both
hands).
Purdue Pegboard (both)
Purdue Pegboard (right)
Purdue Pegboard (left)
O'Connor Finger Dex-
terity test.
Visual search for num-
bers, Serial Search
task (SST).

SST


SST


SST

SST


WB

WB

EE

WB



WB

EE

EE



WB




EE





EX








EE





EE

EE


6

6

6

5



5

5

5



10




6





6




44



44


38


38

38


Different

Different

Same

Same



Same

Same

Same



Different




Same





Same




Different



Different


Different


Different

Different


Earplugs in 70, 120, 130, &140 dB
noise.
Earplugs & 1 muff (RE) in above
noise levels.
Same as group 3, below *

Earplugs in 70 dB It 140 dB noise.
Earplugs & muffs in 1W> dB, ear-
plugs St 1 muff (HE) in 1AO dB.

Same as above.

"Simulations of above noise levels
in ear canals.
105 dB, broadband noise varying
in lower cut-off frequency, 600
Hz, U200 Hz 4 2,00 Hzj control,
300 Hz cut-off presented at 70 dB.
Earplugs in 70, 120, 130, and
140 dB.



1500 Hz, 2 pps, at 65, 85, 95,
and 105 dB.




1500 Hz, 2 pps, at 65, 85, 95,
and 105 dB.



Control group, perform SST for
16 min. with 2, 3 rain, rest per-
.ods between trials (U days of
testing).
ilxp. same as above but 105 dB
noiae presented throughout test-
ing.
Control group, SST presented con-
.inuously on each of i daya for
36 consecutive min. (3 trials).
Same as above, but with 105 dB
iroadband noiae.
Same as above with intermittent
2.R6 pps) 105 dB broadband noise.

Significant effect, 130 
-------
TABLE 1 (CONT)

IJt
15
16
17
18

Sonuner
(unpublished)
Harris & Sch-
oenberger
(1970)
Harris & Som-
mer (1973)
Harris (1970)
Harris (1970)

1
1
1
1
2
1

16
8
12
10
10
6

SST presented for 30
min. at 1.5, 3-5, 5-5,
and 7-5 hours.
Compensatory tracking
(vertical & horizontal
scores) St 2 response
time tasks.
Compensatory tracking
(vertical & horizontal
scores) & 2 response
time tasks.
Pursuit Rotor task
(PR)
PR
PR

WE
EE
EE
WB
WB

480
20
80
30
30
30

Different
Different
Different
Different
Different
Different

Control, 65 dB(A), 90 dB(A), 95
dB(A), aircraft fly over at 1 min.
between peaks.
(1) 85 dB noise (2) 85 dB noise
with 5 Hz vertical vibration at
.25 gz (3) 110 dB noise (4) 110
dB noise with 5 Hz vertical vib-
ration at .25 gz.
60 dB and 110 dB presented alone
and combined with vibration (6 Hz
at .10 gz).
Control, 60, 20 sec. trials on
PR per day.
9 impulsive noise stimuli (112 dB
peak, 400 milliseconds duration)
presented per day over a 4 day
period .
Practice day, 1 test day with
above stimulus.

No effects.
Significant effect on vertical
dimension of tracking task.
Significant effect on horizontal
tracking & approached signif-
icance on vertical tracking.
(p<.10).
Very small but significant effect
which was considerably reduced
by the 4th day of testing.
No significant recovery in sen-
sitivity to impulsive noise
after 5-3 month period.

-------
                                  REFERENCES

  1.  Benox Report: An exploratory study of the biological effects of noise.  ONR Project
     NR 144079, Univer. of Chicago, December, 1953.
  2.  Broadbent, D. E., Perception and Communication. New York: Pergamon Press, 1958a.
  3.  Broadbent, D. E., Effects of noise on an intellectual task. Journal of A coustical Society
     of America, 30, 824-827, 1958b.
 4.  Dickson, E. E. D., and Chadwick,  D. L. Observations on disturbances of equilibrium
     and other symptoms induced by jet-engine noise,/. Laryngol. andOtol, 65: 154-165,
     1951.
 5.  Graybiel,  A., and Fregley, A.  R.}  A New Quantitative Ataxia  Test Battery, BuMed
     Project MR 005.13-6001,  Subtask 1, Report No. 107 and NASA Order No. R-37,
     Naval School of Aviation Medicine,  Pensacola, Florida, 1963.
 6.  Harris, C. S., and von Gierke, H. E., The Effects of High Intensity Noise on Human
     Equilibrium,  AMRL-TR-67-41,  Aerospace Medical  Research  Laboratory,  Wright-
     Patterson AFB, Ohio,  1967.
 7.  Harris, C.  S., The Effects of High Intensity Noise on Human Performance, AMRL-TR-
     67-119,  Aerospace  Medical  Research  Laboratory,  Wright-Patterson  AFB, Ohio,
     1968.
 8.  Harris, C.  S., and Sommer, H. C., Human Equilibrium during Acoustic Stimulation by
    Discrete  Frequencies,  AMRL-TR-68-7,   Aerospace  Medical Research Laboratory,
     Wright-Patterson AFB, Ohio, 1968.
 9.  Harris, C.  S., Human Performance Effects of Repeated Exposure to Impulsive Acoustic
    Stimulation,  AMRL-TR-70-38,  Aerospace Medical  Research  Laboratory,  Wright-
     Patterson AFB, Ohio, 1970.
10.  Harris, C. S., Long Term Adaptation of Pursuit Rotor Performance to Impulsive Acous-
     tic  Stimulation,  AMRL-TR-70-92, Aerospace  Medical Research  Laboratory, Wright-
    Patterson AFB, Ohio, 1970.
11. Harris, C.  S., and Shoenberger, R. W., Combined Effects of Noise and Vibration on
    Psychomotor Performance,  AMRL-TR-70-14, Aerospace Medical  Research Laboratory,
    Wright-Patterson AFB, Ohio, 1970.
12. Harris, C.  S., Effects of Acoustic Stimuli on the Vestibular System, AMRL-TR-71-58,
    Aerospace Medical Research Laboratory, Wright-Patterson AFB, Ohio, 1971.
13. Harris, C.  S., and Filson, G. W., Effects of Noise on Serial Search  Performance, AMRL-
    TR-71-56, Aerospace Medical Research Laboratory, Wright-Patterson AFB, Ohio, 1971.
14. Harris, C.  S., Effects  of intermittent and continuous noise on serial search perform-
    ance, Perceptual & Mo tor SkiUs.  35,627-634, 1972.
15. Harris, C. S., Effects of increasing intensity levels of intermittent and continuous 1000
    Hz tones on.human equilibrium, Perceptual & Motor Skills,  35, 395-405, 1972.
16. Harris, C.  S., and Sommer, H.  C.,  Interactive effects of intense noise and low level
    vibration on tracking performance and response  time, Aerospace Medicine, 1973 (sub-
    mitted for publication).
17.  Korn, J. H., and Lindley, R. H., Immediate memory for consonants as a function of
     frequency of occurence and frequency of appearance. Journal Experimental Psychol-
    ogy. 66, 149-154, 1963.

                                      406

-------
18.  Nixon, C. W., Harris, C. S., and von Gierke, H. E., Rail Test to Evaluate Equilibrium in
    Low-level Wideband Noise, AMRL-TR-66-85, Aerospace Medical Research Laboratory,
    Wright-Patterson AFB, Ohio, 1966.
19.  Parrack, H. O., Eldredge, D. H., and Koster, H. F.,  Physiological Effects of Intense
    Sound, Memo.  Report No. MCREXD-695-71B, U. S. Air Force, Air Material Com-
    mand, Wright-Patterson AFB, Ohio, 1948.
20.  Shoenberger, R. W., and Harris, C. S., Human performance as a function of changes in
    acoustic noise levels, Journal of Engineer ing Psychology, 4 108-119, 1965.
21.  Sommer, H. C., and Harris, C. S., Comparative Effects of Auditory and Extra-Auditory
    Acoustic  Stimulation  on Human  Equilibrium and Motor Performance,  AMRL-
    TR-70-26,  Aerospace Medical  Research Laboratory, Wright-Patterson AFB,  Ohio,
    1970.
                                       407

-------
             BEHAVIORAL EFFECTS AND AFTEREFFECTS OF NOISE1

                                     David C. Glass
                               Department of Psychology
                            The University of Texas at Austin

                                    Jerome E. Singer
                               Department of Psychology
                      State University of New York At Stony Brook
                                     Introduction

     Many aspects of urban life can be viewed as work under stress. People have roles, duties
and tasks to perform  while all around them there is noise, crowding, litter, and traffic. A
number of social critics have commented upon global aspects of these factors, but there is
little  research  of an analytic  nature directed toward ascertaining the specific effects of
urban-like stressors. This paper reports results of approximately two dozen laboratory and
field  experiments,  conducted  over a five-year period, which systematically explored the
effects of stress in man.
     Broad-band noise was the  principal stressor used in our research, and we will, therefore,
limit our discussion to the behavioral consequences of noise exposure. An audio tape con-
sisting of a melange of indistinguishable sounds was  prepared, and, when played back at
intensities up to  108 dbA, served as the stimulus.2 The high-intensity noise thus generated
proved stressful;  the ability of subjects to work  under this stress,  as well as adverse after-
effects of noise exposure were noted.

                                      Adaptation

     The most reliable result  was that people adapted to the noise. When noise was pre-
sented in intermittent bursts over a 24-minute session, few disruptive effects were shown
after  the  first  few noise  trials. This result is, of course, consistent with a good deal of
previous research in the area (cf.  Broadbent, 1957; Kryter, 1970).  Indeed, it is easier to list
the special circumstances under  which noise does produce  an  immediate effect than to
 'Preparation of this paper was made possible by grants from the National Science Foundation (GS-34329
 and GS-33216), Russell Sage Foundation, and the  Hogg Foundation  for Mental Health.
 2The noise consisted of a tape recording of the following sounds superimposed upon one another (1) two people
 speaking Spanish", (2) one person speaking Armenian; (3) a mimeograph machine; (4) a desk calculator; (5)
 a typewriter. We selected this particular concatenation  of sounds as an  analogue of the spectrum of
 complex noise  often present in the urban environment. A sound-spectrographic analysis of the noise
 recording showed that energy did indeed range broadly from 500 Hz to 7,000 Hz, with the mode at about
 700 Hz. Free field stimulation was used throughout  most  of the research.

                                          409

-------
 catalogue the general cases where it does not have an effect. We found that people do not
 adapt to noise under at least two particular arrangements:
      1.   If a person is in a situation of cognitive overload, working on more than one task
          and straining his ability to cope with nonstressful stimuli, the addition of noise
          produces performance decrements.This can be seen, for example, as an increase in
          errors on the subsidiary task in a two-task  situation (e.g., Finkelman and Glass,
          1970).
     2.   If a person is working on a vigilance task requiring constant monitoring or atten-
          tion,  the presence of high-intensity  noise is  apt to be disruptive. Someone who is
          tracking a  pursuit rotor or monitoring a series of dials will do his job less effi-
          ciently under noisy than under quiet conditions, (e.g.,  Glass and Singer, 1972;
          Broadbent, 1957).
     In our research, adaptation was noted by comparing the performance  of people sub-
jected to noise  with  that of people not so subjected on a variety of tasks ranging from the
boringly simple to the  oftentimes challenging and  interesting. Over the course of several
experiments, matching stimulus  configurations with motor movements, clerical aptitudes,
spatial relations, and  driving an automobile simulator all showed adaptation—no decrement in
performance under conditions of loud intermittent noise. A typical set of data are shown in
the first table. These data illustrate adaptation and accompanying lack of adverse behavioral
effects. Under most circumstances, task performance under noise does not differ from task
performance without noise, past the first few bursts.
     Adaptation can be defined in other ways, however (cf.  Lazarus, 1968). In findings
parallel to those obtained with performance measures, people showed adaptation or habitua-
tion (we use the terms interchangeably)3 on several psychophysiological indices, including
phasic skin conductance (GSR), muscle tension in the neck,  and finger vasoconstriction.
These autonomic measures  failed to show continued high reactivity to spasmodic bursts of
noise over a 24-minute noise session.  Figure 1 shows GSR adaptation data as average log
conductance change  scores within each of 4 blocks of noise  trials. There is a significant
decline in GSR on successive blocks in each noise condition. Since initial reactions to loud
noise (108 dbA) were greater than to soft noise (56 dbA), the magnitude of GSR decline is
understandably greater in the former condition. However, the magnitude and  rate of adapta-
tion is virtually identical in the predictable and unpredictable conditions within each noise-
intensity treatment;  that is, subjects  were  equally  reactive at the beginning of the noise
session and equally unreactive at the end.

                        Noise Aftereffects and Unpredictability

     The finding that noise had no routine effects upon task performance in  the laboratory
can be considered an  example of art imitating life, for in the city, despite noise and a host of
other stressors,  work goes on. This finding does not imply that noise has no adverse effects;
to the contrary, our research suggests that it is deleterious to routine functioning subsequent
3See Thompson and Spencer (1966) and Lazarus (1968) for discussions of adaptation, habituation, and re-
lated concepts.

                                        410

-------
                                       Table 1
        AVERAGE NUMBER OF ERRORS ON PART 1 AND AVERAGE DECREMENTS IN
           ERRORS FROM PART 1 TO PART 2 OF THE NUMBER COMPARISON TEST
                                          Experimental  condition

                                  Loud noise      Soft  noise      No  noise
                                   (108  dbA)       (56  dbA)        control
                                   (n =  18)        (n = 20)       (n  = 10)

  Part 1 errors                     3.28            3.30            2.80

  Decrement  in errors             -1.85           -1.48           -0.20
     from Part 1 to
     Part 2
to its  occurrence.  In other words, despite lack  of direct effects, noise had disruptive
aftereffects. These aftereffects occurred whether or not adaptation took place and were
demonstrated on a variety of performance measures. The ability of people to find errors
when proofreading, to continue working on difficult graphic puzzles, and to work effi-
ciently on a competitive-response task were all adversely affected by having been previously
exposed to noisy conditions.
     These aftereffects, surprisingly, were not only a function of the physical intensity of
noise, but  also  depended  upon the social and cognitive context in which noise occurred.
Indeed, the reduction of noise level from 108 to 56 decibels did not have  as large  an
ameliorative effect as any of several cognitive factors. Two of these factors—predictability
and controllability—had a particularly powerful impact on noise aftereffects.
     Exposure  to unpredictable noise, in contrast  to predictable noise, was followed  by
greater impairment of task performance and lowered tolerance for post-noise frustrations.
Despite equivalent adaptation in the two noise conditions, the magnitude of adverse after-
effects was greater following unpredictable noise.  Typical results are shown in Figure 2.
These data are the average number trials taken by subjects on insoluble (though seemingly
soluble) graphic puzzles (Glass and Singer,  1972). Since the task is inherently  frustrating,
fewer trials represent lower tolerance for frustration.
     As can be  seen, subjects showed less persistence on the puzzles following exposure to
loud unpredictable noise than to loud and soft predictable noise. And this effect was true
for both puzzles. It would appear,  then, that lowered tolerance for frustration is a con-
sequence of exposure  to the presumably more aversive unpredictable noise. There is also an

                                         411

-------

     •  -

j-

c
* CJO
u
•
Jj wo
o
a
*l C)0
0
U
5 OJO
• o>0
I
.







































N t
•0^








1



                                                            N 10


             Loud           Soft
         Unpredictable  Molt*
Loud            Soft
 Predictable  Nolee
  Control
No  Nol»e
       Figure 1. Mean log conductance change scores for four successive blocks of noise bursts.
unexpected  tendency for this effect  to  appear even when the unpredictable noise i-
particularly  loud. Soft unpredictable noise was associated with lower frustration-tolerance
than loud predictable noise.
     We have obtained, as I mentioned earlier, similar results with aftereffect measures other
than frustration tolerance. It would appear that unpredictability is indeed a potent  factor in
the production of  noise  aftereffects. The  case for the  existence of this phenomenon is
further strengthened by the range of conditions over which it has been obtained in various
replications  (Glass  and Singer, 1972i These include:  (a) different ways of manipulatin;-
predictability, such  as periodic noise schedules as well as signalling noise onset; (b)different
levels of noise intensity,  such as  108 dbA  and 56  dbA; (c) different subject populations,
such as male and female college students and middle-aged white  collar workers; and  (d)
different laboratory settings.
     Our emphasis on the unpredictability variable is not meant to minimize the importance
of intensity in producing noise aftereffects. We have recently completed a  field  study ot
traffic noise in New York City which clearly demonstrates the importance of the  inteiiMi\
                                         412

-------
     28
     24
     20
      16
      12
E
bl
00
       8
UJ
C5
<
QC
hi
             INSOLUBLE PUZZLES

             • 1st
                  nd
                LOUD      SOFT      LOUD     SOFT    CONTROL

          Unpredictable  Noise    Predictable Noise    No  Noise
               Figure 2. Average number of trials on the insoluble puzzles.


                                  413

-------
 parameter (Cohen, Glass and Singer, 1973). The results produced correlations of the order
 of .45 between impaired auditory discrimination in children (i.e., the ability to differentiate
 speech sounds) and the ambient noise level in their high-rise apartments. These levels ranged
 from a low of about 55 dbA on the 32nd floor to a high of about 75 dbA on the 8th floor.
 Moreover, the magnitude of the  relationship between noise and discrimination was greater
 the longer the children had  resided in  the apartments. There was also evidence linking
 impaired auditory discrimination to deficits in reading achievement.
                   Noise Aftereffects and Perceived Uncontrollability


     Another cognitive  factor mediating noise aftereffect phenomena is the individual's
belief that he can escape or avoid aversive sound, i.e., perceived controllability. In a series of
laboratory experiments (Glass and Singer,  1972), subjects who were given  a switch with
which to terminate noise (Perceived Control Condition) showed minimal aftereffects com-
pared to other subjects exposed to the same noise without the switch (No Perceived Control
Condition). This reduction in aftereffects occurred even though the switch was not in fact
used. Merely perceiving control over noise was sufficient to ameliorate its aversive impact.
     Figure 3 shows the relevant results. It is immediately obvious that tolerance for frustra-
tion was appreciably increased by the perception of control over noise termination. These
effects have been obtained with a number of experimental variations of perceived control,
including the induction  of a perceived contingency between instrumental responding and
avoidance of the stressor.
     But, what specific  stress-reducing mechanisms _are aroused by the manipulation of
perceived control? In answering this question, we reasoned that uncontrollable and unpre-
dictable noise confronts the individual with a situation in which he is powerless to affect the
occurrence  of the stressor and he  cannot even anticipate its occurrence. The individual is
likely to give up his  efforts  at controlling the stimulus under these circumstances, and we
may thus describe his psychological state as one of "helplessness" (cf. Seligman, Maier, and
Solomon, 1971). Perceived Control subjects label their psychological state as one in which
they have control over their  environment, and, thereforei are not helpless. By contrast, No
Perceived Control subjects label themselves  as having minimal environmental control. Task
performance after noise stimulation is affected in a way that is consistent with prior experi-
ence, when control was or was not perceived as available.
     We tentatively conclude,- ,therefore, that unpredictable and uncontrollable noise pro-
duces adverse aftereffects because unpredictability and undontrollability lead to a sense of
helplessness which manifests itself as lowered motivation in subsequent task performance.
David Krantz and I have just completed two experiments designed to test aspects of this
helplessness interpretation. Preliminary analysis of the results indicates that manipulated
helplessness does indeed produce lowered motivation which transfers from one experimental
task to  another.  It should  also be emphasized that the same effect occurred following
exposure to both high and  moderate noise intensities (i.e., 105 dbA and 75 dbA).

                                        414

-------
              45
              40
              35
              30
              25
              20

               15
         M
         IB
               10
                        Insoluble  Puzzles
                        EH  2
                               nd
                               Perceived
                                Control
No  Perceived
   Control
Figure 3. Average number of trials on the insoluble puzzles for perceived control and no perceived control
        conditions.
                                         415

-------
                              Summary and Conclusions

      In summary, noise appears to have few direct effects. People adapt to aversive sound.
 But noise does have disruptive  aftereffects, and these are in large measure a function of the
 cognitive circumstances  in which acoustic stimulation  occurs. These conclusions do not
 mean that  aftereffect phenomena are  the  "psychic price" paid by the individual  for his
 adaptation  to noxious noise (cf. Dubos,  1965; Selye,  1956; Wohlwill, 1970). It is entirely
 possible that noise  aftereffects are  as much post-stressor phenomena  as postadaptation
 phenomena. Further analysis and experimentation enabled us to reach a partial adjudication
 of this theoretical issue. Our  current position  is that after-effects represent  behavioral
 consequences of cumulative exposure to aversive stimulation.  It is not the adaptive process
 itself that causes deleterious aftereffects,  but  the fact  of mere exposure in spite of adapta-
 tion.
                                   REFERENCES

Broadbent, D. E. Effects of noise on behavior. In Ch. 10 of C. M. Harris (Ed.), Handbook of
    Noise Control. New York: McGraw-Hill (1957).
Cohen, S.,  Glass, D. C.,  and Singer, J. E. Apartment noise, auditory discrimination,  and
    reading ability in children. /. exp. soc. Psychol. (1973, in press).
Dubos, R. Man Adapting. New Haven, Conn.: Yale Univ. Press (1965).
Finkelman, J: M., and Glass, D. C. Reappraisal of the relationship between noise and human
    performance by means of a subsidiary task measure. /.  appl. Psychol,  54, 211-213
    (1970).
Glass, D.  C., and Singer, J. E. Urban Stress: Experiments in  Noise and Social Stressors. New
    York: Academic Press (1972).
Kryter, K. D. The Effects of Noise on Man. New York: Academic Press (1970).
Lazarus, R. S. Emotions and adaptation: Conceptual and empirical relations. In W. J. Arnold
    (Ed.),  Nebraska Symposium on  Motivation.  Lincoln, Nebr.:  Univ.  of  Nebr.  Press,
     175-266(1968).
Seligman, M. E. P., Maier, S. F., and Solomon, R. L. Unpredictable and uncontrollable aversive
    events. In F. R. Brush (Ed.), Aversive Conditioning and Learning. New York: Academic
    Press, 347^00(1971).;
Selye,H. The Stress of Life. New York: McGraw-Hill (1956).
Thompson, R. F., and  Spencer, W.  A. Habituation: A model phenomenon f6r the study of
    neural substrates of behavior. Psychol. Rev., 73, 16-43 (1966).
Wohlwill, J. F. The emerging discipline of environmental psychology. Amer. Psychologist, 25,
    303-312(1970).

                                        416

-------
                         EFFECTS OF NOISE ON A SERIAL
                         SHORT-TERM MEMORY PROCESS.

                            G. Wittersheim and P. Salame

                       Centre d'Etudes Bioclimatiques du CNRS
                                 Strasbourg, France
1. Introduction
     The advancement achieved these last 10 years in our knowledge of memory processes
has numerous implications in the design  of communication systems. Working processes
involving short-term memory are  very frequent in industrial work situations as well as in
private life.  Classical examples of this are the telephone-girl's job or the dialing of a tele-
phone number which, after having been picked out from a telephone directory, needs to be
kept at least for a short time in memory until its dialing. Recent ergonomical investigations
have led, on one hand, to  the design of the format, structures and codes for material to be
memorized and, on the other hand, to the design of new keyboards. However, our knowl-
edge concerning the effects  of environmental  factors  such as noise on the information
receiving, storing and transmitting process is still incomplete.
     The cycle of such a process can be split into three main phases:
     1. An acquisition phase for material to be memorized and which, generally, involves a
relatively high perceptual  load, either visual, auditory,  or coming from some other sense
organ;
     2. A retention phase, which can be very short, but during which rehearsal may be
performed;
     3. A reproduction or response phase requiring motor activity, either verbal or manual,
but  also,  in most cases, perceptual  control. In a few situations where the cycles to be
processed are regularly repeated or paced as was the case in our investigation, a fourth phase
has to be added:
     4. An expectation phase before feedback of information as  to the correctness of the
response, during which no active mental operations are required.

2. Experimental conditions

     The aim of the present investigation was to study  the effects of noise on short-term
memory  depending on whether the noise was produced during the first, second, third or
fourth phase
     A 95 dB(lin) pink noise in an open field condition was used in this experiment.  The
spectral composition  of the noise is shown in Figure  1. The noise had been previously
recorded on a magnetic tape and  its emission synchronized with  the onset  and the end of
each phase by means of a device controlled by the signal programming machine.
     The task was a sequential machine-paced memory task. Each session lasted 30 minutes
during which 140 ± 2 cycles were displayed one after the other to the subject (s). The time
structure of one cycle was the following (Figure 2).

                                        417

-------
                          Figure 1  Pink noise Spectral composition.
     - Acquisition phase: Six digits taken from the vocabulary of the 5 first digits were
displayed in a random sequence to  s.  Each digit was displayed for 500 msec and separated
from the next by  140 msec. So the whole acquisition phase lasted for 3700 msec.
     - Retention  phase: This phase  also lasted for 3700 msec. Overt and or covert rehearsal
were allowed. At the end of  that phase, the letter  R (on a cold cathode tube, as the digits)
was automatically turned on,  indicating to s that he should respond.
     - Response phase: This phase lasted for 4480 msec.
     In  case of an error or an omission, a small white light signal flashed at the end of the
cycle, but no error correction had to be performed. After a one-second delay, the next cycle
began.
     The subject  sat in a small soundproof cubicle (figure  3). He was instructed  to repro-
duce  the  6 digits as rapidly and as  accurately as  possible by pressing on the keys  of a
keyboard. Visual control of  keyboard operations was recommended in order  to minimize
motor errors.

                                         418

-------
         ACQUISITION
RETENTION
                          3700
          REPRODUCTION
1
1
-innnnr-

"
!4
T
u
I
I
I
I
1
5
j EXPECTAT.
! O*
I II
V 1 II
h j *
Ull U U 11 ll DEL.
-RESPONSE--*!
1
7400
                                         11880
Figure 2 Time structure of a complete acquisition reproduction cycle. The reproduction phase <1>5 can be
split into two subphases $3 and $4. In phase $2 noise would be produced only until the 6th response-key
press. This particular condition has not been studied in the present investigation.
     Twenty-one Ss participated in the experiment. A replicated latin square design was
used. Thus each S had to perform four 30-min. sessions, each corresponding to noise in one
of the four phases. Practice sessions took place in the morning and experimental sessions in
the afternoon. These were  separated by  pauses lasting 35 minutes during which Ss were
tested on a audiometer and  were required to fill in a questionnaire concerning the subjective
ratings of the task difficulty. A more detailed questionnaire concerning the subjective  noise
effects was carried out at the end of the experiment.
     All data were directly processed by an on-line PDP-8 computer.

Results:

     The main results are shown in figure 4, where $ 1 stands for noise during the acquisi-
tion phase, O2 during the retention phase, $5 during the reproduction phase and $4 during
the expectation phase.Because during the expectation phase session did not interfere with
any noise active mental processing the performances during that session were used as com-
parisons for the three other  phases.
     Performance  on  the accuracy scores, with  errors and  omissions grouped together,
deteriorated  significantly (p< .05) when noise was produced during the acquisition and
retention phases, but there was no difference when noise was  produced in the  response
phase.
     Speed scores were computed by a special procedure: the time between the beginning
and the end of each response list was divided by the total number of keys swept over by the
hand while responding. Thus these elementary time scores obtained from different digit lists
could easily be compared. There were only very small differences  between the four  obtained
means of speed scores. However, by combining accuracy and speed scores through a T-score

                                         419

-------



Figure 3 Inside of the Subject's cubicle. The stimulus sources are located in front of the S on a semi-
circular screen. The keys are arranged on the keyboard (all Ss were righthanded) in such a manner that they
can be easily reached through mere forearm movements.
procedure, a hierarchical effect can be shown: deterioration appeared to be the most impor-
tant when noise  was produced  in  the acquisition  phase and the least  important  in  the
expectancy phase.
     If now we examine the results concerning the subjective ratings (Figure 5) we observe
that the most  unpleasant and most difficult session was the session during which noise was
produced  during  the acquisition phase. In  that phase too, memorization  was judged  to be
the most  difficult,  whereas it was judged  as facilitated  when noise was produced  in  the
response phase.

Discussion:

     There seems to exist now enough  experimental evidence  for  the hypothesis that
memorizing implies the  translation by  the central  mechanisms of the visual message into an
auditory  message which is processed and stored by the brain. We  should  then expect noise

                                         420

-------
               ERRORS
                p
               10
               ms
               205
               200
                50
                45
 ACCURACY

       t   p(bn)
01 vsOt: 2,143 <05
01 vs05 2,235 4: 2,130 <.05
02 vs $5: 2,046 
-------
                 <)>1     $2     1     (J>2     5    $4

MOST  DIFFICULT TRIAL
            10
                                                                r
             MEMORIZATION MADE EASIER
  -MADE  MORE  DIFFICULT
                               SUBJECTIVE  RATINGS
Figure 5 Subjective Ratings, n = number of subjects judging at the end of the experiment which session
had been the most unpleasant, the most difficult, etc.
below 95  dB. Schonpflug and Schafer (1962) found that a  1000-Hz sound at 95 dB im-
proved memory relative to performance with the same sound but at only  55 dB and they
observed  that  the differences in  retention reflected differences in the  activation level.
Hermann  and Osterkamp (1966) confirmed the hypothesis that intermittent white noise at
95 dB has a harmful effect on both the level of retention and organization of material to be
memorized by interrupting logical and associative connections.
     Rabbitt (1968) found that when  Ss  tried  to remember lists of digits played to them
through pulsed white noise, the number of errors they made was greater than in normal
conditions. According to that author, the digit-recognizing process in a noisy condition may
preempt channel  capacity which is necessary  for efficient retention  in immediate memory
storage. The present experimental conditions were not quite the same as those of Rabbitt, as
the digit  lists in  that  author's experiment were spoken  through noise.  Thus no  visual-
                                         422

-------
auditory  translation was necessary. Nevertheless in the retention phase, noise interference
with rehearsal was clearly present.
    An interesting point to be mentioned is the sound fitting of subjective ratings to the
performance scores. Thus both efficiency and  comfort in memory-task performance are
liable  to  be  seriously impaired by noise produced while information is being taken in and
edited in  storage.

                                 BIBLIOGRAPHY

HORMANN, H., OSTERKAMP, U. (1966, a): Uber den Einfluss von kontinuierlichem Larm
    auf die Organisation von Gedachtnissinhalten. Z. exp. angew. Psychol., 13, 31-38.
HORMANN, H., OSTERKAMP, U. (1966, b):  Uber den Einfluss von diskontinuierlichem
    Larm auf  die Organisation von Gedachtnissinhalten.  Z. exp.  angew.  Psychol., 13,
    265-273.
MILLER, H. (1957): Effects  of high intensity noise on retention. J.  appl. Psychol., 41.
    370-372.
SCHONPFLUG,  W.,  SCHAFER,  M.  (1962): Retention und Aktivation bei  akustischer
    Zusatzreizung. Z. exp. angew. Psychol., 9, 452-464.
SLOBODA,  W., SMITH, E. E. (1968):  Disruption effects  in  human short-term memory:
    some negative findings. Percept, mot. Skills, 27, 575-582.
RABBITT, P. M. A (1968): Channel-capacity, intelligibility and immediate memory. Quart.
    J. exp. Psychol., 20, 241-248.
                                       423

-------
   THE EFFECT OF ANNOYING NOISE ON SOME PSYCHOLOGICAL FUNCTIONS
                                 DURING WORK

                                 Irena Franszczuk
                         Central Institute of Work Protection
                                  Warsaw, Poland


    Psychological investigations concerning the influence of noise on human performance
were carried out with the cooperation of our Acoustical Department. Engineers dealing with
technical acoustical problems often met, in practice, some psychological and physiological
questions, as follows:
     1)   How does noise act on human performance  and the feeling of comfort, and what
         is the psychological mechanism of this effect?
    2)   Which kinds of work are most disturbed by noise?
    3)   On which physical characteristics of noise does its disturbing influence on human
         comfort and performance depend?
    The solutions of these problems, as we know from the literature, are rather ambiguous.
    The causes of this ambiguity have been very well described in the report of Dr. Gulian.
In our investigations we tried to avoid  ambiguity or at least some of the factors influencing
it. All conditions were constant in each of our experiments, except for the noise. The noise
was generated by loudspeakers  driven by a  tape recorder during all experiments except the
control  experiments  in silence. We used various natural noises recorded  in factories and
different bands  of  white  noise  from  generator. The acoustical  variables  were level,
frequency, and bandwidth. The level and spectrum of noise used were measured in all our
experiments.
    We  used no noises greater than  90 dB  SPL. Experiments were  carried out under
laboratory conditions with  students performing for some hours a task requiring attention
and finger dexterity. The level of psychic  performance  during the work was measured by
means of different psychological tests. The first problem in these investigations was to find a
test sensitive enough to measure subtle small changes. All our results were statistically
tested. The annoyance caused by the noise was evaluated by each investigated person on a
six-level scale.
    In  this report  I would like  to summarize the  most important results of our four
experiments, all  of which are  published in the Quarterly Journal of our  Institute, "Prace
CIOP", (1, 2,3,4). They are shown in Figures 1 and 2.
     1.  A noise  band  of median frequency 4000 Hz prolongs simple reaction time. An
annoying noise in which sounds of frequencies near 4000 Hz predominate, even though the
level is not greater than 85 dB, produces an increase of simple reaction time to both light
and sound stimuli. Independent of differences in the individual sensitivity to noise, the
phenomenon of  prolongation of reaction time has been observed in each of our 24 test
persons  in all four of the experiments cited (Figures 1 and  2). This prolongation is very
small but statistically significant (p < 0,01).
     2.  Broad-band white noise has more effect on simple reaction time than narrow bands
of white noise with the median frequency 250 Hz (3). This dependence can be seen in

                                        425

-------
 21
Figure 1 Simple reaction time under different acoustical conditions (4). Median values of the reaction time
in hundredths of second obtained from 8640 measures  of 12 persons:  (a) to light stimuli (b) to sound
stimuli. Silence S; Noise A: octave band of white noise with median frequency 4000 Hz; Noise B: octave
band of white noise with median frequency 250 Hz.
Figure 2, in which we have the average results of 720 measures of simple reaction time* of 6
persons, in each of the experimental noises and in silence.
     3.  Under certain conditions, noise (the level of which does not exceed 85 dB, without
dominant components of high frequencies) may be an activating factor, and may shorten the
reaction time both to light and to auditory stimuli. We have obtained the shortest reaction
time with narrow bands of white noise with the median frequency 1000 Hz (ref. 3) (Figure
2). Perhaps  these results may be explained  by  some arousal hypothesis. The  similarity of
changes in simple reaction time to light and sound stimuli supports such a hypothesis.
     4.  The noise reduces perception efficiency.
     Using in our investigations the octavebands of filtered  white noise with the median
frequency 4000 Hz and level 80-85 dB, we found a statistically significant prolongation of
                                         426

-------
24
23
22
21
20
 19
 18
                            CM


i
       b)
21
20
19
18
17
16
-
v>

-

.0

!NJ

^
C\j
i
i

1
1


|

^

~,
^.

i
i

i

figure  2 Comparison  of simple reaction  time  under different acoustical  conditions and in silence
{Franaszczuk, 1968,1971). (In hundredths of seconds).
    Median value obtained from 720 measures, 6 persons:
    (a) to light stimuli;
    (b) to sound stimuli.
     Silence          —S
     White noise      -WN
     Tone            — T
     1/3 octave band  -1/3
     Octave-band      — 0
     250,1000,4000  - median frequency in c/s.
                                             427

-------
 average simple  reaction time  both to  light  and to  sound stimuli in comparison with the
 average reaction time in silence. The increment of the average reaction time to sound stimuli
 is greater than the increment  of reaction time to light stimuli (4). This difference was very
 little, 0.01 sec, but statistically significant.
     We have obtained similar differences in some other experiments too (2).
     The simple reaction process to the stimuli is composed of the phase of perception and
 of the  motor phase (pressing the key).  The motor phase is  identical in both kinds of
 reaction, but the  phase of perception is different,  because the sound stimuli are partially
 masked by the noise. This explains the greater increment of the reaction time by auditory
 stimuli in noise and suggests that the perception process is the most disturbed by the noise.
     5.  All experimental noise conditions causing a prolongation of simple reaction time
 were  evaluated  as more annoying than noises  not causing it. The subjective feeling of
 annoyance  is a  signal of decreasing psychic performance even when  these changes  are not
 measurable and not observed.
     Experiments  involving other psychological tests and tests of choice reaction time did
 not give us any differences between noise and silence conditions.
     The simple reaction time is the simplest measure, giving much data in a very short time.
 With this test it is not possible to compensate for a decrement  of performance by exerting
 additional effort,  as one can in  spite of tiredness, during other work. The simple reaction
 time measure is closely connected with the subjective feeling  of annoyance and may be
 considered an indicator  of general influence of noise.

                                     References

     1.  Franaszczuk, I., Examination of Changes in the Effectiveness of Mental Functions
 During Work Under Noisy Conditions. Prace CIOP Nr 49. 1966.
     2.  Franaszczuk, I., Investigation  on Fluctuations  of Reaction Time to Visual and
Auditory Stimulus During Work Under  Noise Conditions. Prace CIOP Nr 54,  1967.
     3.  Franaszczuk, I., Effect of Width of  Frequency Band of Noise Upon Reaction Time
of Man To Light And Sound Signals. Prace CIOP Nr 58, 1968.
     4.  Franaszczuk, I., Effect of Annoying Noise of Determined Frequencies on Simple
and Composed Reaction Times. Prace CIOP Nr 64, 1970.
     5.  Franaszczuk, I., The influence of disturbing acoustic stimuli on  some psychic
 functions. Prace CIOP 71,1971, 335-343.
                                        428

-------
                        SESSION 5
NON-AUDITORY PHYSIOLOGICAL AND PATHOLOGICAL REACTIONS

               Chairmen: E. Grand jean, Switzerland
                       S. Kubik, Czechoslovakia
                            429

-------
                       NON-AUDITORY EFFECTS OF NOISE
           PHYSIOLOGICAL AND PSYCHOLOGICAL REACTIONS IN MAN

                                    Gerd Jansen
                                Inst. Hyg. Arbeitsmed.
                           University of Bochum, Germany


     During the Conference on "Noise as a Public Health Hazard" in 1968, an earlier report
on "Effects of Noise on Physiological  State" was presented. It was shown  that a formula
exists, established by systematic research, that allows prediction of the vegetative reactions
by means of sound levels and bandwidths (Jansen, 1967),  Moreover,  it was possible to
establish the limits of normal vegetative reactions (Fig.  1).  These limits have been applied to
concrete noise situations,  especially for assessment of noise-induced disturbance of health
around airports. These limits were  confirmed by two investigations:
     1.   noise-exposed steelworkers showed more vegetative disturbances due  to noise
         than those workers from "quiet" factories (Jansen, 1959);
     2.   by means of vasodilative medications which produce a contrary reaction, it was
         shown that pathological reactions  during noise applications  began  beyond the
         limits mentioned above. The medication  does not work in healthy men; therefore,
         we saw no influence on noise-induced reaction during application of subcritical
         noise bursts, whereas during application of supracritical noise a significant  com-
         pensation occurred (Jansen, 1969).
     In Dec. 1969, the AAAS organized a symposium "Physiological  Effects of Noise"; in
the  course of this Symposium physiological  and  patho physiological  noise reactions  were
reported. The main problems described were: cardiovascular noise reactions, the influence of
noise on adaptation processes, resistance against disease, endocrine and metabolic functions,
biochemical and pharmacological  influences, effects on reproductivity and some neurologi-
cal and sleep disturbances (Welch and Welch, 1970).
 Investigation Concerning Activation

     The summary of all published results, especially those of Washington and Boston, is
 that noise acts as a stimulator for activation of a group of physiological functions designated
 as "arousal reactions".
     Arousal, therefore, means an elevation of excitation level of certain systems of the
 body. The activation-theory, as experimentally affirmed by Hebb (1955), Malmo (1959)
 and many others, differentiates between cortical, autonomic, motor, endocrine and affective
 arousal.  The last concept is closely linked to emotional stress. Acoustic stimuli are con-
 ducted to the cortex via  the ascending  reticular activation system (ARAS). The latter
 responds to both qualitative and quantitative changes in an ongoing stimulus; it is influenced
 by and at the same time influences the connections to the structures of cortex, subcortex,
 cerebellum, sensory neurons,  motor innervation and vegetative centers. It plays a specific
 role in regulating vegetative and affective behavior.

                                         431

-------
       us

       dB
       105


       95


       65


       75


       65


       55
Schadigung '
deutliche
vegetative Reaktion
mogliche
vegetative Reaktion
                                             •t-
                125    200   320    500    800    1250   2000   3200    5000   0000 Hz
             WO    160    250    400    640   1000    1600   2500   4000   6400
                        Grenzwerte  zur  Beurteilung
                   larmbedingter vegetativer Reaction
                                       Figure 1
Classification of Activation Reactions

     The activation reactions due to high and moderate sound levels include: inhibition of
gastro-intestinal peristaltic activity,  inhibition of secretion of gastric juices and saliva, dila-
tion  of the pupil,  temporary increase  of blood pressure, diminution of GSR, moderate
decrease of stroke volume of the heart, influence on the pulse rate (especially during sleep),
increased production of catecholamines  and steriods. a decrease in skin temperature, and an
increase of cortical blood volume.
     The activation reactions mentioned above comprise in large part the so-called "orient-
ing reflex" according to Sokoloff (1963) and Lynn (1966).  Orienting reactions are necessary
to increase  the readiness for action  to a high  level and to  guarantee the possibility of
immediate  reaction. Repetitive stimuli, however, produce more or less rapid habituation, so
that  in  the course of time  the reaction pattern diminishes or disappears. It  should  be
                                         432

-------
pointed  out  that the orienting reaction and its ability to habituate are necessary for the
human life.
     From experimental studies with noise-induced peripheral vasoconstriction in the skin,
the relations between intensity and frequency of stimuli can  be presented in  a severely
simplified, schematic figure (Fig. 2). This figure is only a basis for discussion and not a result
of research, but it may lead to further research using the concepts of orienting and defensive
reaction. By "defensive  reactions"  we mean  those noise-induced  reactions that do not
habituate with increasing intensity and/or with frequent stimulus presentation.
     In my own experiments (Jansen, 1973) it was possible to  demonstrate orienting reac-
tions that rapidly become defensive  reactions with stimulus repetition. If the sound level is
high enough, defensive reactions will appear immediately. To prove this we simultaneously
recorded peripheral blood volume at the extremities and at the  head. We found that orient-
ing reactions were characterized by constant or even augmented pulse amplitudes (vasodila-
tion) at the head, while pulse amplitudes at the fingers showed a  decrease (vasoconstriction).
          :0
                   keine Reaktion
                                                      TTT
                                                          w
                                                           m
HI: Defensivreaktion
  IffMtiOT!
                             Orientierungsreaktion
                                                        Intensitat
         Beziehung zwischen  Reizhdufigkeit und -intensitdt
                            (Gewohnungseffekte)
                      - schematise?)  nach  Sokoloff-
                                      Figure 2

                                       433

-------
 When the orienting reaction changes to a defensive reaction, then vasoconstriction occurs at
 the head as well as in the fingers.
     As you know,  peripheral blood volume begins to decrease at sound levels of 60 to 70
 db(A); another classical activation or orienting reaction is a change in the GSR (or skin
 resistance) and this begins at a lower level. The skin resistance indicates that physiologically,
 a certain level of activation exists; it is released by affective or emotional stress. Due to its
 high capability of habituation, GSR seems to be characteristic of orienting reactions. It is a
 matter of fact that females show stronger reactions in GSR recordings during noise applica-
 tion than men.                     .
     The indicator of activation so far discussed, the orienting response, indicates, of course,
 something about the severity of the noise exposure, and that there is an influence on certain
 functions, but says nothing about the harmfulness of the noise; this is more appropriately
 indicated by the defensive reactions, which, however, are to be expected first at very high
 sound levels.


 Problem-oriented Experiments

     In order to apply the results and knowledge of noise investigations to the assessment of
 noise in  different practical situations, research in certain  areas should be developed. One
question  in urgent need of an answer is  the question of noise influence during phases of
decreased activation due  to  biological  rhythms, especially circadian  rhythms. Another
important problem is how noise may affect those people who biologically need quiet (for
instance,  shift workers after a night shift). Finally, how noise affects ill persons or those in a
state of convalescence is also an important question.
     The enumeration of these problems leads to the central questions of noise research; the
enumerated results are only small parts  of the complexity of noise effects. Therefore,  it
seems necessary to introduce multivariate research methods to psychosomatic noise research
as is done in other scientific disciplines.  Concerning the question  of noise and biological
rhythms  Griefahn (1973) will report in this  Congress on the  different noise reactions
during  certain periods of the  ovarian cycle. In another project in our  laboratory,  Dams
(1972) studied some moderator-variables in noise situations. He tested the influence of age,
sex, ambient noise and medication with a vasodilative  substance in the modification of aural
and extraaural noise reactions.  He investigated  peripheral pulsations, the pulse rate and the
breathing rate. He thereby demonstrated that in young  female persons, stronger vegetative
reactions were caused than with young male subjects; the reactions in old male and female
persons were  the same but they were much weaker than those of the young. Concerning the
influence of ambient noise, Dams demonstrated that  the "depth of modulation" is impor-
tant for  the amount of the vegetative reaction; older females show stronger reactions in a
high ambient noise (Fig. 3). The latter fact may be caused by the more accentuated connec-
tions between neuro-vegetative functions and emotions in female subjects.
     In my own investigations (Jansen, 1970),  e.g. the results presented at Boston, I found
that there is on the average a negative relation between the vegetative reaction and the TTS.
This means that a subject showing a  high TTS has only  small vegetative reactions when
exposed to higher sound levels than the limits for normal vegetative reactions, and vice versa

                                        434

-------
      Q.

      O
3  50
    A3CD
                                                                   I
                                                                Inilialreaklion

                                                                durchscriniltlichf
                                                                Gfsamtreaktion
                                                                    5 mat 5 mm Ruhf und
                                                                    5 min BreitbondgfrOusch
                                                                    105 dBIA) im Wtchsel
                                                                   t • n : Jhtophyltin - nicol/nof
                                                                         Grundpegel

                                                                   600 mg I - n

                                                                   Gesehlecht dtr Vpn

                                                                   Alter dtr Vpn
                            Vegetative  Schallreaktionen
                Alter  Geschlecht  Theophyllin- nicotmat und Grundpegel
                                        Figure 3
(fig. 4). These experiments were repeated by Dams (1972), Bergmann (1973), Meier (1971)
and Hezel (1972). They found the same relation. Bergmann, in addition, demonstrated that
subjects with good hearing and those with slightly poorer hearing showed different behavior
concerning the relation  between pulse amplitude and  TTS: he concludes that the  quality
(elasticity) of the regulation of blood vessels is different in  the two groups. In particular, the
results showed that the better-hearing subjects showed a negative and the worse-hearing
subjects a positive  correlation between pulse amplitude and TTS (fig. 5).
     Summarizing, it may be concluded that noise stimuli beyond the critical curve limit for
normal vegetative  reaction is 99 dB(A) at maximum, and  that between 90 dB(A) and 100
dB(A), a general hazard to human health must be considered. These  possible disturbances
might be found or manifested in various manners, even in psychic behavior, as there  is no
function in the human  body exclusively affected by noise.

Psycho-physiological Investigations

     Similar to the correlation  between vegetative reactions and TTS,  there might be corre-
lations between vegetative and  psychic reactions  caused by noise.  In an  interdisciplinary
                                         435

-------
              •« 5
.

B

•



20
              E
              i
              i
              -
                                         — Ohr mil der grofleren  TTS
                                         --Mr with grater TTS
                                         y-. 32.i -0.15 FPA
                                         r--0.62
                         —Ohr mil der genngeren TTS
                         —ear with smtller TTS
                         y*27.7-0.19 FPA
                         T: -0.72
                          •10  0  -10 -20  -30 -tO -50  -60 -70 -80 -90  -700%
                           d Finger-Pull-Amplitude IFPA) 20 see ntch Ltrmbeginn
                       A Finger-Pulse- Amplitude (FPA) 20 tec past beginning of noise
            Beziehung  zwischen Finger-  Puts - Amplitude  und  TTS
              Relation  between  Finger -  Pulse - Amplitude and  TTS
                                        Figure 4
study. *•_• investigated this  possibility with  psychomotor test procedures, psychological
classifications and physiological reactions (Jansen, 1962). We tried to define psyc'iic dimen-
sions of the subjects and their reactions to annoying noise by using moderator variables such
as ncuroticism, test anxiety, social desirability, etc. Simultaneously with the noise stimula-
tion we  recorded the  pulse  amplitudes. We confirmed its dependence  or intensity  and
bandwidth. A factor analysis of all psychological and physiological parameters resulted in a
factor "pulse amplitude" (beside other factors) which was not correlated very highly with
other test  factors. We interpret this result as  an indication  that even without psychic con-
comitants-i.e., without a positive or negative  attitude to the noise source  the possibility of
harmful noise influence on the human body exists.
     When meaningful noise  stimuli were applied, it was found that physiologica' reactions
depended on the above-mentioned dimensions of personality (Fig.  6). First of all. extremely
labile subjects reacted  physiologically more strongly  if the noise wa; meaningful. Stable
persons reacted less. If the noise lost its meaningfulness. this difference vanished: there was
                                         436

-------
                       ohnt  t.-n
                                            mil t.-n.
                                                       '—— / --O.I09'
                                                     /	/ . 0.261 J
                                                                            30
                                                                            dB
                                                                            2L


                                                                            :.


                                                                            •i


                                                                            rfl
                                                                      •-
                                                                      \
           '-tO  -20   0   20   tO   60 V. 80  -to  -20   0   20   40  60 '/. 80
                  mittlert Reaction der Fingerputsamplitudc in emer 8 mm Larmphase
iorm 6 rrun Srf/rfiontfperouicft 705 3B(A}
t -n i 750 fig Ttwophytlin • nicofi^of 160 ffim  tor
                                                   atr Vrrloubuny
             6 Vpn. mitttere Ruhehorschwette  8,0 15.7 dB  »Hlr (t096 Hi / 5732 Mi/bnar  Otr»r\//t vp < Vrrtuchrl
                        Vegetative   Reaction  und  TTS2
                      bei  Versuchen   ohne  und  mit  t.-n.
                                         Figure 5
seen only the dependence on intensity and  bandwidth. In a practical sense, this means that
all new and unexpected noises may cause different psychological and physiological reactions
in men depending on the dimensions of personality, but that the reactions become homoge-
neous  if the noise has lost its meaningfulness by habituation, frequent repetition or positive
attitude.
     In additional experiments, Hoffman (Jansen. 1972) found out  that the overall sensa-
tion of annoyance due to a noise is aggravated at peak levels of 90 dB(A).
     We therefore propose establishing "limiting  ranges" instead of a single limit for total
noise exposure. Within this range it might be possible to establish a criterion as "representa-
tive" of expected noise effects.
     As human health is endangered by single noise events as well, it seems justifiable to
demand an assessment of noise not only by the calculated equivalent continuous noise level
l^q, but also by limits for single noise events which must  not  be exceeded even if  the Loq is
below the criteria fixed in standards or laws.
                                          437

-------
F/ngerpulsamp/iludp CM
(18 sec noch Larmbeginn)
^-1 <» <0 5
o o o o
informationslos
75 dB 95 dB
U




Signifikanzen
75ns-


zwischen vegetaliven Schollreaktionen
75 dB 95 dB Ind ^B 1 	 1



ns




ns
5
5
ns
5

S
ns
5



Neuroliker Nichtneuroliker
hohe geringe
Prufungsongsl Prufungsongsl
hemmende fordernde
Prufungsongst Prufungsongst
geringe Soz/ale hohe soziale
Se/bsleinschofzg Selbstemscholzg
Introverliertheil Extrover tiertheil
(weiblich) (monnlich)
in/ormationslos
weifles Rauschen
informationshaltig
Industriegerousch und Ton 3 200 Hz
s = signifikanl
ns = nichl significant
Personlichkeitsmerkmate und vegetative Reaction auf
informationslose und informationshaltige Gerausche
                                       Figure 6
Conclusions

     Reviewing the results described above, it is clear that the relations between high sound
levels and their psychophysiological influences  are quite unequivocal. There are reactions
that  may be judged to endanger human well-being and health. It is obvious, too, that single
noise events  whose  intensities exceed established limits are as important as  equivalent
continuous sound  levels. The  risk  of  hearing damage seems on the other hand to be de-
scribed fairly well  by  Leq, whereas this assessment does not correspond to the extraaural
reactions.
     Concerning the  middle range  of sound intensities, the  assessment should  be based on
psychomotor disturbances (and in  addition with physiological methods) and -  most impor-
tant  - by determining the degree of annoyance.  In the middle range, Leq and especially the
rate of presentation of stimuli are just as important as maximum level assessment.
     In regard to low noise levels, only psychological classifications (and perhaps GSR) may
be used for evaluation of the exposure. Leq here seems to be the preferable method of noise
assessment.

                                        438

-------
     It should be mentioned that a decreasing sound level does not lose its meaningfulness
at corresponding rate, so we have to consider, besides the acoustical characteristics, the
information coupled to the noise.
     Though  the pathophysiology and pathopsychology of noise exposure is not so well
developed as  its  psychophysiology, we know on the other hand some facts of noise-induced
disturbances of human well-being. One of these facts will be reported later when Rohrmann
presents an abbreviated report  on our German Research Society Airport Study. It will be
demonstrated that by means of epidemiological methods and concepts, the complex noise
reaction—especially in the middle range of intensities—can be described much better by
psycho-sociological than by physiological research alone.
     The existing standards and criteria in different countries are in most cases compromise
agreements. These agreements have sometimes been partly  verified;  this refers especially to
the 15 dB(A) difference between day and night  standards. It should be the aim of noise
research to verify the existing standards or to show how we should modify them. But this is
only possible with a coordinated cooperative attack  in many lands by scientists from many
disciplines. Only by cooperation and integration of results  can the complex effects of noise
be assessed adequately.
                                        439

-------
          INDUSTRIAL NOISE AND MEDICAL, ABSENCE, AND ACCIDENT
                     RECORD DATA ON EXPOSED WORKERS*

                                  Alexander Cohen
                 National Institute for Occupational Safety and Health
                                 Public Health Service
                   U.S. Department of Health, Education and Welfare
                                Cincinnati, Ohio 45202

                                    Introduction

     This  paper describes the  first findings  in a  project seeking to determine evidence
coupling severity of occupational noise exposures with occurrences of extra-auditory prob-
lems of consequence to worker health and safety. Specifically, comparisons are reported of
the frequency of medical disorders, absences, and job accidents entered in company records
of workers subjected to high and low noise levels at their workplaces. This evaluation was a
retrospective one, using data contained in worker files of two  manufacturing firms located
in the southeastern  United States. In each company, the entries of interest were extracted
and tallied for the 5-year period, 1966-1970, which was just prior to the establishment of a
hearing conservation program for those exposed  to the higher-level noise.

                                 Sources for the Data

     Approximately 90% of the worker records evaluated  in this study were drawn from a
plant complex which manufactured large pressure boilers.  This facility, referred to as Com-
plex A, consisted of four manufacturing buildings, each divided into numerous shop areas,
bays, and  offices. Key sources of noise within work areas were generated by machinery used
in vertical turning, boring and  facing  of large-diameter boiler sections.  Other high noise
emitting equipment included arc-air flame-cutting tools, air compressors, heavy presses, and
many automatic panel welding machines. Also typical  of the high noise producing opera-
tions in Complex A were chipping and grinding on large nuclear reactor vessels and compo-
nents weighing up to 1,000 tons. Many of these noise operations took place simultaneously,
radiating into many work areas within each building in this complex.
     A secondary source of  record data was a plant engaged in the production of electronic
missile and weapon parts. High noise levels in this plant, called Complex B, were generated
in the  operation of boring  machines, grinders, pneumatic  presses, air  compressors, and
riveting machines. Complex  B was less  than one-fifth the size of Complex A in manufactur-
ing area, and utilized production equipment and machinery far smaller in scale and massive-
ness. Noisy operations in Complex B were  also more localized to those areas where indi-
vidual tools were in  use.
*This paper is a condensation of a report prepared by the Raytheon Service Company (1972) which under-
took this records study via contract (HSM 099-71-6) with the National Institute for Occupational Safety
and Health. Mr. Robert Felbinger served as project director for the Raytheon Service Company.

                                         441

-------
     The  choice of these establishments took account of some considerations which are
mentioned in connection with the procedural aspects described below.

                                 Procedural Aspects

     Repeated noise measurements had been made in various work areas within each com-
plex spanning the same time period as the record data to be evaluated. These noise measure-
ments permitted the division of such work sites into a high noise classification, defined by
the presence of sustained or interrupted noise levels of 95 dBA or more, and  a low noise
classification,  defined  by levels  not exceeding 80 dBA, regardless of temporal pattern.
Division of work areas into more  than two noise classifications was precluded by incomplete
information relating actual exposure time to .specified noise levels, and the probable fluctua-
tion of such exposure conditions from year to year due to changing production schedules or
other factors. Noise surveys  were performed at the outset of this study to confirm previous
readings and to make final decisions regarding work sites to be assigned to the high and low
noise classes. Figure 1  describes  the average 'sound  levels in dBA (re 0.0002 microbar) for
continuous or intermittent noises found in the areas finally chosen. Each data point repre-
sents a work area of one or more exposed persons. The range  and distribution of sound
levels for  the high noise workplaces in Complex A reflected more intense noise conditions
than those shown for Complex B.
     Both plant complexes  included a sizable complement of employees who had worked
for many consecutive  years  in the areas defined as having highland low noise  levels. This
assured realization of a study goal which was to evaluate record data for a minimum of 500
workers with prolonged experience in noisy jobs and 500 workers with comparable experi-
ence in quieter ones. Medical,  accident and  attendance files on all personnel working in the
classified high and low noise level areas in each complex were made available after measures
were  taken  to insure their confidentiality. The record data of interest had  been logged by
medical and administration staffs in each establishment which had remained intact over the
years, lending consistency to the record-keeping process. Initial screening excluded those
persons who were not employed  in the high and low noise jobs for the entire 5-year period
of this  records study,  and/or whose pre-employment health examinations  indicated ear
trouble (hearing problems) or active health disorders. The remaining workers in each com-
plex were  then sampled to form groups of comparable size in high and low noise areas  that
were matched as much as possible in.age, experience at  the workplaces designated, work-
shift, etc.  Table I shows the, age-job experience makeup of the final groups constituting the
sources  of record  data for  the  noisy  and  quiet work locations in Complexes A and B.
Equivalence  in  these variables for the  high  and low noise groups is only  approximate.
Complex A workers were all males but a few females were included in both the high and low
noise  groups of Complex B. No blacks were entered in the  worker groups selected for
Complex B.  Less  than  10%  were blacks in  the Complex A samples. This small number of
females and blacks precluded any efforts to evaluate separately the effects on extra-auditory
 - •   '•-":"    •        "  —""   "     ,*-•'-,".         • - ,.  i -     1
problems of sex or race. Hourly and salaried persons covering a wide range of salary levels
were represente'd in the high and low noise groupings of each complex. Match-ups between
actual job  functions by workers in the high  and low noise sites were few. More will be  said

                                       442

-------
                                      -140-
                                      -130-


                                      -120-
                                      _J	
                   HIGH     «:..           I                    HIGH
                  NOISE    ji.       -10-              NO|SE
                            • • • ••          i              •
                 (95dBAOR ;  !'	
                  HIGHER)   H:       -100-

                            	
                                      -90  -
                                 ••    -  80  -
                                           1
                   LOW     •              I              .     LOW
                  NOISE    -        -  70  -              NOISE
                 (80dBA OR              	!	

                                      -60-
                                      -50-

                                      — i —
                                      -  40  -
                       PLANT COMPLEX         I         PLANT COMPLEX
                           A
 Figure 1 Sound levels in dBA for continuous or intermittent noise conditions observed in the work areas
'constituting the high and low noise groups in complexes A and B. Each data point represents a work area of
 one or more exposed persons.
about this point in the course of the paper. At least 45% of the workers in the high and low
noise groups of each complex worked the first or early shift of the workday.
     The medical, attendance and accident-files for all workers selected for this survey were
searched and  relevant  entries  collated  by research  assistants subject to additional close
review by other members of the investigating team. To protect against recorder bias, none of

                                          443

-------
                                       Table I
         COMPOSITION OF WORKERS BY AGE AND EXPERIENCE IN HIGH AND LOW
                        NOISE GROUPS IN COMPLEXES A AND B.
IGE IN
reARs
BELOW 26
26-35
36-45
46-55
56-65
TOTAL
SAMPLE
DUMBER OF WORKERS
COMPLEX A COMPLEX B
HIGH LOU HIGH LOW
NOISE NOISE NOISE NOISE
60 9 7 0
155 102 21 8
89 138 19 27
89 141 17 28
61 59 2 2
454 449 66 65
YEARS AT
PRESENT
JOB
5-10
11-15
16-20
21-25
26-30

COMPLEX A COMPLEX B
HIGH LOW HIGH LOW
NOISE NOISE NOISE NOISE
345 269 50 10
44 102 16 54
43 41 0 1
6 14 0 0
15 21 0 0

these persons was permitted knowledge as to whether a given set of files  belonged to a
worker in a high or low noise group. All record entries were accepted at face value. In regard
to occurrences  of- accidental injuries, there were notations of only the  type of injury and
body part  affected. Circumstances surrounding such reported accidents were typically not
detailed. Both minor accidents, necessitating dispensary treatment only, as well as  major
ones, involving lost time, were tallied.
     Record entries reflecting health problems included references to disorders diagnosed  by
the attending physician and symptomatic complaints reported by the worker. These entries
were classified into nine different categories of medical problems based on the nature  of the
diagnostic information or symptoms reported. These categories are shown in Tables  II and
III, to be discussed later. Absence data collected over the time period of this evaluation was
coded in two ways, namely, in total days and as the number of discrete events lasting one or
more days. Absences for reasons other than reported illness or injury were  not included—
that is,  only sickness-absenteeism  was considered. One cannot discount the  fact, however,
that a certain number of these reported absences may have been for  reasons other than
bonafide sickness.
                                        444

-------
                                  Results and Discussion
General Findings
     The  accident,  medical and  absenteeism data extracted from worker records for  the
5-year period, 1966-1970, were  evaluated  separately  for Complexes A and B. Figures  2-5
present such  data in  the  form of cumulative percent frequency distributions.1 There  are
plotted the percentages of workers in high and low noise job sites with specifiable numbers
of accidents (Figure 2), diagnosed disorders (Figure  3), and absences, both discrete (Figure
4) and total days (Figure  5), as logged for the 5-year period in each complex. For Complex
A, the distribution curves  for the high and low noise exposed workers show clear differences
for each of these problem indicators. Specifically, greater proportions of the worker group
exposed to high level noise reveal more accidents,  more health disturbances, and greater
amounts of absence than  in the comparison group  not so  exposed. As an illustration,  the
accident data in Figure 2 show  that fewer than  5% of the workers in the quieter areas of
Complex  A had  15  or more  accidents for the 5-year period of this records study. In
'In actuality, these distribution curves have been  plotted in an inverse way to display more clearly dif-
ferences in the  frequency of extra-auditory problems among workers in the high noise group relative to
those in the low noise group. Each point on a given curve represents the percentage of workers in the
specified group  having as many or more of the number of problem occurrences shown on the abscissa for
the 5-year period of record collation.
              PLANT  COMPLEX    A
                      HIGH NOISE GROUP (N-454)
                0	...    NOISE GROUP (N*449)
                -P-—O---6--.6-- -A..-A--.A.
                                       45  50
PLANT  COMPLEX   B
       H.  HIGH NOISE GROUP (N-66)
       -»  LOW NOISE GROUP (N*65)
    15  20  25  30  35 40 '45  50
                               NUMBER OF ACCIDENTS-5  YEAR TOTAL

 Figure 2  Cumulative percent frequency distribution of workers in high and low noise groups in complexes
 A and B with a specifiable number of accidents over the 5 year period, 1966-1970.
                                           445

-------
CO
             PLANT  COMPLEX   A
                      HIGH NOISE GROUP (N-454)
                          NOISE GROUP (N'449)
        PLANT  COMPLEX   B
           o	° HIGH NOISE GROUP (N-66)
           o	o LOW NOISE GROUP ( N-65)
          5  10  15 20  25 30  35  40  45 50       >l    5  10  15  20  25 30  35  40  45  50

                 NUMBER  OF DIAGNOSED  MEDICAL  PROBLEMS  -   5   YEAR  TOTALS

 Figure 3 Cumulative percent frequency distribution of.workers in high and low noise groups in complexes.
 A and B with a specifiable number of diagnosed disorders over a 5 year period, 1966-1970.
CO
              PLANT COMPLEX   A
                      HIGH NO|SE GROUP (N-454)
                      LOW NOISE GROUP (N-449)
          5  10  15 20  25  30 35  40  45 50
         PLANT  COMPLEX   B
             —» HIGH NOISE GROUP (N-
             ---o LOW  NOISE GROUPIN-
65)
2t  .5  10  IS, 20  25,30   35 40  45 50
                         NUMBER  OF DISCRETE ABSENCES  -  5  YEAR  TOTALS

Figure 4 Cumulative percent frequency distributions of workers in high and tow noise group in complexes
A and B with a specifiable number of total days absent over a 5 year period, 1966-1970.
                                            446

-------
CO
             PLANT  COMPLEX    A
             u       HIGH HOISE GROUP (N-454)
                 ----«>LOW NOISE GROUP (N-449)
          10  20 30  40  50 60  70 80  90
             HIGH NOISE GROUP (N*66)
             iow  NOISE GROUP (N« 65)
10  20  30  40 50  60 70 80  90
                         NUMBER OF ABSENCES  (DAYS) -5 YEAR  TOTAL

Figure 5 Cumulative percent frequency distributions of workers in high and low noise groups in complexes
A and B with a specifiable number of total days absent over a 5 year period, 1966-1970.
contrast, 35% of the workers at the noisier worksites had 15 or more accidents in this same
time span and 10% had as many as 40.
     The  medians of the distribution  curves shown in Figures 2-5 indicate that the 5-year
record entries for a typical worker in a noisy area of Complex A include 8-9 more accidents,
3-4 more diagnosed  medical  problems, 40 more days  of absence, and  25 more discrete
occasions of absence than that found for a counterpart worker in a less noisy area of the
same  facility. Statistical evaluation of these differences in medians  found them all to be
significant.
     Differences in  the cumulative frequency distributions of  the  accident-j health and
absence data recorded for workers in the high vs. low noise areas of Complex B, however,
were  riot1 as great as those seen in Complex A. Based on .-median values,  the 5-year record
data for a typical worker in the high noise area of Complex B, relative to a typical worker in
the quiet, show one  more accident, equal occurrences  of diagnosed medical problems, 2
more days  of  absence  and 4  more  discrete instances  of absence.  Though  slight, the
differences  in accidents and  discrete numbers of absence between the high and low noise
exposed workers attained statistical significance.
     The  median numbers of accidents, health disturbances and absences when computed
by individual years within the 5-year collection period in Complexes A and B yielded results
consistent with the overall totals as described above.
     Why Complex A showed greater  differences between high and low noise groups than
Complex  B in the frequency of extra-auditory problems  is conjectural. One possibility is
that the high noise  classification  for Complex A included areas  with  much higher sound
levels than those classified in the high noise group of Complex B (see Figure 1). Apart from
                                         447

-------
this noise factor, the differential risk of injury or illness specific to jobs and work areas
classified in the high versus low noise groups of Complex A might have been much greater
than that of Complex B.
Specific Evaluations

     Different comparisons of the record data  were made to define the influence of age,
length of job experience, work shift and other variables. Select results can be summarized as
follows:
     (1)   The number of accidents per worker was greatest for the younger persons in noisy
          jobs and/or those who had the least experience at such jobs in both complexes.
          This accident rate diminished with increasing age and job experience for workers
          in noisy workplaces, with similar though less obvious changes  noted for those
          located in  quieter ones. For the 5-year period, the youngest workers (25 years or
          below) with least experience (10 years or less) in noisy jobs showed typically 9-10
          more accident occurrences than their peers in quieter jobs, and 8 more than those
          found for the oldest (over  55  years), most experienced (greater than 25 years)
          workers in noise.  These  results agree with other findings in the literature which
          generally report more  accidents among younger, less experienced workers (Mann,
          1944; Hale and Hale,  1972, Freeman et al (undated)). That high  levels of noise
          may  act as an  additional potentiating factor in this context  seems plausible.
          Drawing such a conclusion, however,  presupposes the  same jobs or equally risky
          ones being performed by the subject workers in both the noisy  and non-noisy
          areas. Assurances of these conditions  were lacking for this study, as they seem to
          be for other research'concerned with more general effects of age and experience
          on accidents (Hale and Hale, 1972, Freeman et al (undated))."
     (2)   Younger workers in  both the high  and low  noise level groupings showed  the
          greatest number of diagnosed disorders entered in  their medical files for the
          5-year period. Differences revealing more frequent medical problems for workers
          in the high vs.  low noise jobs were only apparent in Complex A, and  became
          smaller with increasing age.  Variations  in these differences with job experience,
          apart from age, were uncertain.
     (3)   Sick-absences, either in terms of total days or discrete occurrences, were found to
          be greatest for the younger workers, especially those in the high noise level group.
          This amount of absenteeism tended to decrease in the middle age  groups only to
          increase again for the oldest workers. A similar U-type relationship was seen in the
          absence rates of workers in noisy areas  as a function of years of job experience.
          Absenteeism measures for the workers in the low noise group showed no change
          (Complex  A) or increased (Complex B) with advancing age or longer years of job
          experience. The higher rates of absenteeism among young workers in the noisier
         jobs can  be a natural consequence of the  increased  numbers of accidents  and
          health disturbances also noted  in the records of this group. Taken  together, these
          findings may depict the initial strain of coping with a work situation subject to
          intense noise and possibly other stressors as well. Older workers, though showing
          fewer  injuries  and  health  problems in their  files, may  be liable  to more

                                         448

-------
         absenteeism due to greater susceptibility to illness, not necessarily job connected,
         and some loss in recuperative capacity.
     (4)  No consistent differences emerged in comparing frequency differences in recorded
         accidents and medical  problems or amounts of  sick absence as a function of
         workshift for workers in the high and low noise groups in either complex.
     Additional evaluations were  performed to clarify certain aspects of the health data and
elaborate further on the overall results. For example, when sorted into diagnostic categories,
the medical entries filed for workers in both complexes revealed respiratory disturbances to
be most common, irrespective of workplace noise levels (see Tables II, III).
     For workers in the higher noise, however, more respiratory cases involved hoarseness,
laryngitis and sore throats. An undetermined  number of these  ailments could be attributed
to the shouting of workers in communicating in the noisy work sites. Other more frequently
noted disorders  for  workers in the high vs. low noise groupings  fell  into the allergenic,
                                       Table II
        NUMBER OF DIAGNOSED DISORDER BY MEDICAL CATEGORY FOR WORKERS
                            IN HIGH AND LOW NOISE GROUPS

                               COMPLEX A - 5 YEARS
CATEGORY
OF
DIAGNOSED DISORDERS
RESPIRATORY
ALLERGENIC
MLJSCULO/SKELETAL
CARDIOVASCULAR
DIGESTIVE
GLANDULAR
NEUROLOGICAL
UROLOGICAL
NUMBER
HIGH
NOISE
331
196
75
64
50
39
34
29
AFFLICTED
LOW
NOISE
146
86
31
37
21
10
11
14
NUMBER OF
HIGH
NOISE
2152
358
104
114
66
48
49
40
OCCURRENCES
LOW
NOISE
590
118
47
70
30
14
29
15
                                        449

-------
                                      Table 111
        NUMBER OF DIAGNOSED DISORDERS BY MEDICAL CATEGORY FOR WORKERS
                           IN HIGH AND LOW NOISE GROUPS
COMPLEX B - 5 YEARS
CATEGORY
OP
DIAGNOSED DISORDERS
RESPIRATORY
CARDIOVASCULAR
ALLERGENIC

MLJSCULO/SRELETAL
GLANDULAR
DIGESTIVE
UROLOGICAL
NEUROLOGICAL
NUMBER
HIGH
NOISE
56
27
13
2
2
1
1
0
AFFLICTED
LOW
NOISE
59
13
21
3
o
0
0
0
NUMBER OF
HIGH
NOISE
384
28
20
2
2
1
1
0
OCCURRENCES
LOW
NOISE
360
16
33
3
0
0
0
0
musculo-skeletaJ, cardiovascular and digestive categories, especially in Complex A. Symp-
toms and  diagnostic signs here were less specific in nature  or origin as related to noise. In
this regard,  health  examination surveys of workers in noisy  industries have also noted
increased incidence  of circulatory, allergenic and neurological problems of assorted descrip-
tions which have been ascribed to excessive occupational noise exposure (Jansen, 1961,
1969; Shatalov et al, 1962, Anticaglia and Cohen, 1970). At the same time, however, this
research has been  criticized for the inability to control other adverse workplace or job
factors,  apart  from noise, which  may  have influenced the results (Kryter, 1970, Miller,
1971).
    The present study can be similarly criticized since, as already noted, job situations for
workers in the high and low noise groups could not be matched on a one-to-one basis.
Partial equating was tried, using jobs with the same functional titles which were found in the
high and low groups of both Complexes A and B. Comparisons of the record data by select
                                        450

-------
job titles and noise levels yielded differences which were most often in directions showing
either more numerous accidents, medical problems or absences under the higher level noise
(see Table IV).  The magnitudes, and in a few instances, the directions of these differences
for the specified jobs were quite variable when compared to one another, and to the overall
differences based on the total group comparisons. This variation stresses the importance of
"the. job.factor and the need to better account for it in this type of research. On this latter
point, a  most effective approach would be to contrast the incidence of extra-auditory
problems in the same workers before and after noise reduction takes place, especially if such
controls did not materially change the  nature of the job operations or alter other non-noise
hazards attendent to the total work situation. There exists an opportunity to implement this
approach as  part of a follow-up effort  to this record study. Specifically, hearing conserva-
tion measures stressing the use of personal ear protectors have been in effect in Complex A
and B for the past two years, and it is  planned  to evaluate again the medical, accident, and
sick-absence entries of the subject workers subsequent to the establishment of this program.
Reduction in  individual worker noise exposures through ear protectors should diminish the
occurrences of medical, safety, and related sick-absenteeism problems if, in fact, noise was a
causal factor. Positive findings here would also indicate the extent to which efforts designed
to reduce noise hazards to hearing can also offset extra-auditory problems as well.
     This additional work is slated to be undertaken only at Complex A. This is to capitalize
on the large  group of workers available for study, and the  fact that their initial record
entries, as reviewed above, showed the  clearest indications of increased health and accident
problems among workers  in the high noise workplaces. Any conclusions regarding noise as a
major or contributing cause of these extra-auditory problems would be incautious at this
time, and should be deferred pending the outcome of the follow-up  study. Indeed, the data
available at present offer only circumstantial evidence.

                                       Summary

     Entries  in  medical,  attendance, and accident files for over 500 workers situated in
noisy plant areas (95  dBA or higher) were compared with 500 others in quieter workplaces
(80 dBA or less) gathered over a 5-year period in two plant complexes. Most of the record
data were taken from the larger  of the two  establishments which manufactured  boiler
equipment, and  which was  also  found to have generally more intense noise conditions.
Workers subjected to the high workplace noise  here showed greater numbers of diagnosed
medical  problems,  absences for illness,  and job related  accidents  than were  noted for
workers in the quieter areas of the same plant. Medical diagnostic categories showing signifi-
cant differences between high and low noise level jobs were respiratory (hoarseness owing to
shouting in noise) and  non-specific allergenic,  musculoskeletal, cardiovascular and gastro-
intestinal disturbances.  Differences between high and low noise  level groups showed wide
variation when sorted by job type, suggesting that the increased frequency of extra-auditory
problems can be greatly affected by this variable, regardless  of noise level.  Evidence for
increased medical, absence, and  accident problems in comparing the high and low noise
exposed groups in the second plant complex, which produced electronic missile and weapon
parts, was not as prominent as that noted in the first one. A follow-up study is planned to

                                         451

-------
                                     Table IV
      TYPICAL OCCURRENCES OF MEDICAL PROBLEMS,SICK-ABSENCE, AND ACCIDENTS
        CLASSIFIED BY JOB TITLES FOR WORKERS IN HIGH AND LOW NOISE GROUPS.
                          COMPLEX A - 5 YEAR TOTALS
JOB
TITLE
FOREMEN
PESTS &
[NSPECTS
ADMINIS-
TRATIVE
TOTAL
SAMPLE
NOISE
LEVEL
HIGH
LOW
"HIGH"
LOW
HIGH
LOW
HIGH
LOW
N
34
138
48
38
10
45
459
449
AVERAGE OCCURRENCE PER WORKER
MEDICAL
2.0
3.7
3.9
3.6
4.5
0.8
3.9
0.4
DISCRETE
ABSENCE
9.1
3.1
.6
5.7
74.7
4.7
30.3
4.2
TOTAL
ABSENCE
21.6
8.3
.5
18.1
107.1
9.4
49.8
8.8
JOB
ACCIDENTS
4.5
4.8

3.4
7.6
0.3
9.0
0.4
                          COMPLEX B - 5 YEAR TOTALS
JOB
TITLE
&CHINE
)PERATORS
ASSEMBLY
WORKERS
TOTAL
SAMPLE
NOISE
LEVEL
HIGH
LOW
HIGH
LOW
HIGH
LOW
N
16
16
46
38
66
65
AVERAGE OCCURRENCE PER WORKER
. MEDICAL
PROBLEMS
4.7
4.7
7.3
7.3
4.8
5.5
DISCRETE
ABSENCE
14.0
4.4
16.6
7.5
10.8
7.0
TOTAL
ABSENCE
17.5
9.9
26.3
38.2
18.2
16.5
JOB
ACCIDENTS
1.8
0.3
2.0
1.4
1.7
0.7
   NOTE:   VALUES  BASED ON CELL SIZES OF 10 OR MORE
evaluate entries in the records of the same workers over a period subsequent to the establish-
ment of an ear protection program in the first plant complex studied. Reduction in indi-
vidual worker noise exposure through  ear protectors should diminish the  occurrence of
medical, sick-absence, and accident problems if, in fact, excess noise was a causal factor.
                                   References

Anticaglia, J.  R., and Cohen, A. Extra-auditory effects of noise as a health hazard. Amer.
    Indust. Hyg.Assoc. J. 31, 277-281 (1970).
Freeman, F., Goshen, C. E. and King, B. The role of human factors in accident prevention.
    (Operations Research Inc. Contract Report SAph-73670) Public Health Service, Dept.
    Health, Educ., Welfare, Wash. D.C. (undated).
                                      452

-------
Hale,  A.  R. and Hale, M. A review of industrial accident literature. National Institute for
    Industrial Psychology, London, England (July 1972).
Jansen, G. Adverse effects of noise in iron and steel workers (in German). Stahl u. Eisen, 81,
    217-220, 1961.
Jansen, G.  Effects of noise on the physiological state. Proceedings of National Conference
    on Noise as a Public Health Hazard,  Amer. Speech and Hearing Assoc. Kept. No. 4,
    89-98, (Feb. 1969).
Kryter, K. D. Effects of Noise on Man (Academic Press, N.Y. 1970), p. 508.
Mann, J. Analysis of 1009 consecutive accidents at one ordnance depot. Indust. Med. 13,
    368-374, 1944.
Miller, J. D. Effects of noise on people.  Kept. No. NTID 300.7, Environmental Protection
    Agency, Wash. D.C. 20460 (Dec. 1971).
Raytheon Service Company. Industrial  noise and worker medical, absence,  and accident
    records. Contract No. HSM 099-71-6, Burlington, Massachusetts (1972).
Shatalov, N. N., Saitanov, A., and Glotova, K. V. On the state of the cardiovascular system
    under conditions of noise exposure. Labor Hyg.  Occup. Diseases, 6 (7), 10-14 (1962).
                                         453

-------
        FACTORS INCREASING AND DECREASING THE EFFECTS OF NOISE

                                   D. E. Broadbent
                  Medical Research Council, Applied Psychology Unit,
                                 Cambridge, England

     Human beings have a limit to the number of features of their surroundings which they
can perceive in any limited period  of time, and therefore anything which happens in the
environment has  to  compete  with  other events for their attention. Until about 1960, a
number of the effects of noise  could be explained simply by considering that intense sounds
have a tendency to win in such a competition. On this view, a man in  a noise would show
failures of perception  because important signals would fail to be analyzed while  he  was
being 'distracted' by the noise. Physiological changes  could then be explained as  due to
compensating mechanisms which attempt to combat this distracting effect.
     Since 1960,  however, it has become increasingly  clear that this analysis may confuse
cause and effect; it may be that exposure to noise produces a change in the state of the man
and that this changed state is reflected in failures of selective perception. The evidence for
this changed interpretation comes from a number of experiments, but  for those who have
specialized in other areas it will be sufficient to quote a result from Wilkinson (1963). In
this experiment, men were asked to perform a task with and without the presence of  100-dB
noise, each condition being met when the  men were in a normal  state  and when they  had
been deprived of sleep for 24 hours. Three main points appeared in the results. First, the
usual harmful effects of sleeplessness were reduced by the presence of the noise. Second, the
effects of noise itself on the task  were harmful if the men had slept normally, but  if they
were sleepless, noise  actually improved their performance.  These  two findings suggest that
noise  creates some general  state of arousal which reduces  the effects of sleeplessness,  and
which only impairs efficiency if the man is already as highly aroused as is desirable. Such a
conclusion is supported by a great deal of related evidence (Broadbent, 1971).
    Wilkinson's third finding  is similar  to that of many other experiments on noise: the
effects are greater when the task has been continued  for a prolonged  period in the noisy
conditions. There are two  possible explanations for this. One is that  the work  produces
some kind of change in the man, which we may call 'fatigue' if we can avoid defining that
word  too precisely, and that the noise affects the man more when he has been 'fatigued'.
The other possibility is that the noise gives rise to some cumulative effect, so that the longer
one stays in the noisy environment, the more incapable one becomes of performing even a
novel  task. Since  Wilkinson's experiment,  like most others, started the noise at the same
time that the man started  work,  it is impossible to  distinguish these possibilities from his
results. In this  paper, I am going to outline three recent  as yet  unpublished experiments
from  our laboratory which show that noise changes  the  general state of the perceptual
system rather than merely distracting it;  and the first is one which indicates that noise gives
a cumulative effect on the man which may persist even when the noise itself has ceased.
     Hartley has used the same task as Wilkinson, in noise and in quiet,  and in each case tor
a work period of 20 minutes.  The main interest of the experiment lay  in the condition to
which the man was exposed during the 20 minutes before the measured session; he might be

                                        455

-------
 reading, or he might be performing the task, and  in either case he might be in quiet
 conditions or else in 100-dB noise. There were therefore eight experimental conditions by
 which one can separate the different theoretical possibilities already mentioned.
     At first sight the results are complex, but this is only a superficial difficulty. Perform-
 ance is worse if the man has worked for a previous 20-min period, and also if he spent the
 previous 20 min in a noise environment. (This latter  fact in itself shows that the noise has
 changed the state of the man rather than simply acting as  a 'distractor', since it makes him
 inefficient even when he is subsequently working in quiet and there is no noise to distract
 him). Thirdly, performance is worse when noise is present  than when it is absent. The  key
 point however in distinguishing the different theoretical explanations  is whether the effect
 of noise is bigger when the man has previously worked than when he rested; and it is not.
 On the other hand, the effect of noise is greater when the man has previously been exposed
 to noise. The combination of all these findings looks complicated, but in fact the conclusion
 is simple; noise affects people at the end of a work-period because it produces a cumulative
 change in  their state, and not because  it affects them more  when they are 'fatigued' by
 work.
     The next experiment I wish to discuss considers whether the general state which noise
 produces is one which might change the function of the senses and perceptual mechanisms.
 McLeod has  devised a method  of measuring the integration time of the  eye, following
 techniques introduced by Allport (1968). The  basic method is to present a series of lines on
 a cathode-ray tube, one after another, each separated by 1 cm from the previous one. The
 man controls the number of lines presented before the equipment returns to the original  line
 and repaints  it. His task  is to set the number  of lines present  at such a  value that  the
 addition of one more would cause the whole display to appear to flicker. At this point the
 man is seeing simultaneously a number of lines which have all beea presented to the  eye
 successively, and this is therefore  a method of assessing temporal resolution in the visual
 system.
     As is well known, in low levels of illumination the integration time of the eye increases,
 which is obviously adaptive in extracting as much visual information as possible from a weak
 signal. McLeod's results show however that a similar change occurs in loud  noise, the two
 effects  interacting so that the  effect  of noise is  only  statistically significant  at 0.25
 foot-lambert and not at 40 foot-lamberts.
     We thus have evidence that noise produces a general change in men exposed to it,  and
 that this change affects the intake  of sensory information.  The last point I wish to make is
 that the perceptual changes are of such  a kind that they would resemble 'distraction'. In a
 series of studies by myself and  my wife, we  presented  visually mixtures of relevant  and
 irrelevant information, and found that noise impaired the ability to select the one from the
 other. In the most definitive trials, we used words interleaved so that the  odd-numbered
letters   came  from  one  word  and  the  even-numbered ones  from the  other,   e.g.
 LjEuAnDgEIRe; a  difference of colour between the two words was also introduced, so that
some men could be asked to identify  the black and word and  some the red one. If now the
exposure duration was increased until correct identification  of the word took place, noise of
about 100 dB had no harmful effect on threshold for the easy word (capital letters, black
print, common word), the threshold for the  difficult word (small letters,  red print, rare

                                        456

-------
word)  was increased by  about thirty per cent. Control experiments using the words sepa-
rately showed no such effect; thus the effect of the noise was to impair the functions which
would normally suppress  the large and conspicuous, but irrelevant, letters.
     To summarize, these experiments show changes which cannot be due to distraction by
noise, but which would have the effect of producing failures of perception;  noise may not
always distract,  but rather make men more distractible. This view fits well  with results
discussed elsewhere in  this meeting, and particularly with the complexity of the effects on
performance  described by Dr.  Gulian;  tasks in which the maintenance of attention is no
problem will  show no impairment by noise,  nor will tasks performed when other conditions
are unarousing. The disruption of performance in tasks which do require selective percep-
tion such as the Stroop test is however to be expected and so is the increased deterioration
which  will follow arousing conditions such as deprivation  of control  over the situation
(Glass and Singer, 1972). It is particularly worrying that Hartley's results, like those of Glass
and Singer, show a persistence of the effects  of noise after the stimulation has ceased.  One is
reminded  of  the finding of Jansen (1959)  that family disturbances are significantly more
common amongst those who work in noise, and of the higher rate of admissions to Spring-
field Hospital from streets with a high exposure  to aircraft noise (Abey-Wickrama  et al,
1969). In each  case, there may be factors other than  noise which might  be alternative
explanations  of the effects; but equally there is the  speculative possibility  that there is a
chronic effect of noise in distorting perceptual input, which disturbs personal relationships
as well as laboratory tasks. There is a need for further work  on chronic effects of noise, if
only to eliminate this possibility.
                                     References

Abey-Wickrama, I., a'Brook, M. F., Gattoni, F. E. G., and Herridge, C. F. Mental hospital
     admissions and aircraft noise. Lancet, 1275-77 (1969).
Allport, D.  A.  Phenomenal simultaneity and the perceptual moment hypothesis.  Brit. J.
     Psychol,. 59, 395-406(1968).
Broadbent, D. E., Decision and Stress, Academic Press: London (1971).
Glass, D. C., and Singer, J. E. Urban Stress: Experiments on Noise and Social Stressors.
     Academic  Press: New York and London. (1972).
Jansen,  G.,  Vegetative functional disturbance caused by noise. Archiv. fur Gewerbepatho-
     logie und Gewerbehygiene. J 7, 238-261 (1959).
Wilkinson,  R. T. Interaction of noise with  knowledge of results and sleep deprivation. J.
     Exp. Psychol., 66, 332-337. (1963).
                                         457

-------
           EXAMPLES OF NOISE-INDUCED REACTIONS OF AUTONOMIC
             NERVOUS SYSTEM DURING NORMAL OVARIAN CYCLE

                                 Barbara Griefahn
                               University of Bochum
                                    Essen, BRD

    The extent of reactions of the autonomic nervous system caused by ergotropic stimuli
are dependent on the vegetative status of the test person. People with trophotropic circula-
tion function, which means those with small pulse rate, small  cardiac output, and great
peripheral resistance, show greater vasoconstriction during noise exposure than people with
ergotropic circulation function (Jansen 1969, Jansen and Schulze  1964, Matthias and Jansen
1962, Oppligerand Grandjean 1959).
    Women with normal ovarian cycle show in the premenstruum a characteristic increase
of pulse rate, minute output  respiration frequency and basal  temperature; in the post-
menstrual period, they show a decrease of these values. Thus, the vegetative functions of a
fully-developed woman are characterized by a cyclic succession of the trophotropic follicle
phase  and the ergotropic corpus-luteum-phase (Brehm  1959, Doring 1948, 1953, Goodland
and Pommerenke 1953, Artner 1960).
    The question  concerned here is  whether these changes of vegetative status are great
enough to cause different responses to the same stimulation.

Method

    12 females were tested in  57 experiments, 2-4 days before and  after the beginning of
menstruation and 2-4 days before and after ovulation  for two complete cycles. The tested
persons were 17-39 years old, their cycle lasted between 27 and 32  days. They were
healthy; none of them had a hearing loss greater than 20 dB. None of the test persons took
hormone preparations or circulatory preparations.
    None of the test persons had measured basal temperature, so it was  necessary to find
other  parameters which point to  the existence of normal or anovulatory  cycles. In accord-
ance with the results of other authors (Doring and  Feustel 1953, Brehm  1959) we found the
values of pulse rate and respiration  frequency  in the  corpus-luteum phase significantly
greater than in the follicle phase. Therefore and because of the careful selection of the test
persons it is very probable that we tested only within normal ovarian cycle.
    During the experiments, the test persons sat in a comfortable chair in a sound-proofed
room.  After  a quiet period of 15 min a  white  noise of 95 dB(A)  with a duration of 2
minutes was presented 5 times, each time followed by a quiet period of 3 minutes.
    During the experiments we recorded finger-pulse amplitudes, pulse rate and respiration
frequency.

Results

    On each day of examination (Figure  1) we found a great initial  decrease of the finger-
pulse  amplitudes. In the first half of the cycle (the follicle phase; 2 - 13 days or, in Figure 1,
                                       459

-------
                 0123

                      Breitbandgerausch  95 dB(A)
                                                                     :
                                                                 2-£  Tag
                                                                des Cyclus

                                                                16 Versuche
                                                                 8 Vpn
                                                                    b
                                                                 11-13  Tag
                                                                des Cyclus

                                                                15 Versuche
                                                                11 Vpn
                                                                17-19 Tag
                                                                des Cyclus

                                                                IS Versuche
                                                                10 Vpn
                                                                    d
                                                                25-27 Tag
                                                                des Cyclus

                                                                11  Versuche
                                                                6  Vpn
                  Fingerpulsamplitude  bei  Larmbelastung
                      zu  unterschiedlichen Cycluszeiten
Figure 1  Relative amplitude of the expansion of the finger (vasoconstriction effect) associated with heart
beat in response to a 95-dB(A) white noise burst of 2 rnin duration (cross-hatched). The value from 0 to 0.5
min on the abscissa is taken as lOO^. The parameter is the number of days since menstruation.
                                         460

-------
curves a and b) this initial decrease is followed by a gradual increase, in the second half (c
and d: the corpus luteum phase) the finger-pulse amplitude remains near 85%. (The dif-
ferences between the  follicle and corpus-luteum phases of the cycle are significant until the
24th second of noise exposure).
     The results will  be much clearer after calculation of only one average value for each
day of examination  (Figure 2). Contrary to the values within one phase (a-b  or c-d) we
found that the values in the follicle phase are very significantly smaller than those of the
corpus-luteum-phase.
     These results are concordant to the results of some other authors, who found a greater
vasoconstriction in trophotropic  than in ergotropic  people (Jansen 1970, Jansen  and
Schulze 1964, Heinecker,  Zipf and  Losch 1960); and they are in agreement with the rule
found by Wilder (1931),  which  postulates that the reactions of the autonomic nervous
system will be smaller, the greater its excitation.
     After dividing the total reaction into the initial and the residual reaction (the latter
being the value at the end of the  noise burst) we found that the curve of the residual
reaction  follows  closely the total reaction; the values of the initial reaction show great
deviations, so we could suppose a psychic reason. But after calculation of the initial reaction
within the second cycle after beginning of the experiments (Figure 3) we found the same
curve. Psychic conditioned reactions are dependent on habituation, so that the extent of the
reaction  becomes  less after a lot of stimuli.  Therefore it is very  probable  that psychic
displeasure has only an insignificant influence on the results.
     We thought that the cause of the different reactions are the ovarial hormones. There-
fore,  with the use of the crosscorrelation function we calculated the delay  between the
curve of the reaction values and that of the hormonal level.
     Usually, crosscorrelation functions will be calculated  in  order to discover a periodic
event within an apparent  stochastic curve. In  this case,  with well-known periodicity, the
hormonal curve is displaced against the residual reaction until the best agreement is found,
as shown by a maximum  of the crosscorrelation function. The place of the maximum, in
this  case  the number of  days until temporal agreement, is therefore the degree for the
probability of a causal relation.

     The data used for crosscorrelation (Figure 4) are those of the remaining reaction, the
daily  urinary excretion of estrogens, after Brown, Klopper and Loraine (1958) and the daily
gestagen level in plasma, after Neill et al (1967).
     The maximum of the crosscorrelation function between estrogens and reaction is r = 4,
between  gestagens and  reaction r =  1  (Figure 5). That means that the alteration of the
reaction follows the increase of the estrogens after 4 days, the increase of the gestagen level,
however,  after only  1  day. Therefore  it seems that the lower reaction is  caused by the
appearance of the gestagens.
     If the extent of vasoconstriction caused by noise exposure is really an indicator for the
degree of excitation of the autonomic nervous system, the ergotropic situation in the second
half of cycle will be caused by the increase of the gestagens, the trophotropic situation in
the first half by the decrease of the gestagens and not by the increase  of the estrogens.

                                        461

-------
*/•} bei BB 95 dB(A
Ql
tn

o wi
                         cr
                                                           •A
es


«
                                                               sr.
                                                               f X
                        a/b    a/c
a/d    b/c
b/d
c/d
                   or 2 - 4  Tag  des Cyclus   c = 17-19 Tag des Cyclus
                   b:11.-13.  Tag  des Cyclus   d-25-27 Tag des Cyclus

                   • Haltewert    O Gesamtreaktion    A  Imtialreaktion
                Larmbedingte  periphere  Vasokonstriktion

                  an  verschiedenen   Tagen   des   Cyclus
Figure 2  Noise-dependent peripheral vasocon strict inn on different days of the menstrual cycle. The dotted

curve shows the change in finger volume immediately after onset of the 95-dB(A) noise, the solid line the

effect at the end of the 2-min noise burst. The dashed curve is the average change over the entire 2 min.
                                        462

-------
75
s
5 *
o» 70
CO
CD
S 65
•o
1 »
O
3
h.
&
c
so








/
/•
9 '
*
6


P"
/ ,
/ /
.//
'/

.U.
• *m
/N
f






^
\
a



Imtiolrtoktion
















t t 8 12 16 20 2t 28
Tag des Cyclus
Figure 3 Comparison of initial vasoconstriction effect during the first (solid curve) and second (dotted
curve) month of observation.
Respiration Frequency

     Ovarial  hormones have a stimulating  effect on respiration frequency (DOring  1953,
Doringet al,  1950, Wilbnuul ot al.. 1<)51>).
     According to Harmon (1933) and Stevens (1941). noise also effects a small increase of
metabolism and by  that a higher level of  carbonic acid concentration in blood, which is
accompanied by an  increase of respiration  frequency. Though effects of hormones and of
noise on metabolism and therefore on respiration rate are very small, it seems possible that
both together will effect a significant reaction.
     Two  to four days after the beginning of menstruation, when there is only a small
production of estrogens and  gestations, respiration rate shows, according to this theory, an
insignificant change (figure 6): 2 - 4 days before ovulation, when the first peak of estrogens
appears,  we found a significant increase. Two to four days after ovulation, when the estro-
gen level decreases, and at  the same time gestauen level  increases, the reaction is significant
too. Shortly  before  the beginning of the menstruation,  after the decrease of the estrogens
and gcstagens. the reaction is insignificant.
     The maximum of the crosscorrelation  function (Fig. 7) between estrogens and respira-
tion rate appears at T  = 27  or   1. This precocious agreement is explainable, because the
values of the estrogens are excretion values.
                                         463

-------
        -
        I









— Cestagene
— Oestrogene


...
r^m
^__
• •••


«





^_ . _
...









....
— • -

- - -




. _ -


_.-





_ . _

_ _ _




-•-





....

...




.._
^^^




• . —

— —


• . V


^_



- . *






- -
- - _>





*»••
1
— • —



....







1 •
_



—
...
_._


• V •*

— —


•
	

IV.


-^— —
_ _ «

r:.:


...
—
...




_ . -,



- --
•^^M
_ • —
	 1
Fingerpuls-Hattewert
Atmungsreaktion











^r



irrr

                             S
J2
20      2i       28
       Tag des Cyctus
          Werte  fur  die  Kreuzkorrelationsfunktion
Figure 4 Data used to calculate cross-correlation functions between hormone levels (gestagens, solid lines,
estrogens, dot-dash) and autonomic response to noise (residual vasoconstriction, dashed lines; change in
respiration rate, dotted).
                                       464

-------
• -
)
Relativwerte
1 o +



'



'


N,
--


\



\



\
N


^.
\


X
\
\
N
\
N

X
\
\
\

\

\
->.


S



\




/



7


.

4 8 12 16

.

**

/
'-
•^IMM
^•M «
/
—
	
.

'
-

t
9
•
•


•
•


'
•


^
'
•


'
•

\

•

N

H

\

i

\
\
•

\
.• •
A


•


'
'



•

'•
•
28 32 36 r
(Toge)
Kreuzkorrelationsfunktionen
                           Figure 5 Cross-correlation function between hormone levels and vasoconstriction cause by noise.

-------
        t>
        HI
        01
107,5

 *
105.0
        !T   102.5
         e
         c
            KW.O
         c
         •)
         o»
            97.5
            S5.0
                            .XI-
                            t *
                                                 16
                                             20
                                                                 2t       28
                                                                Tag des C/c/us
        0
        A
        o
—  Durchschnittlicrie Grsamtrtaktion alter Versuchspersonen
• -  Gesomtreaktion der Versuchspersonen Nr 2. Nr 3. Nr 4 und Nr 10
—  Gesamtreoktion der Versuchspersonen im  I  Cyclus der Versuchsreihe
    Gtsamtreaktion der Versuchspersonen im  II Cyclus der Versuchsreihe
                       Durchschnitttiche  Gesamtreaktion
                 der Atmungsfrequenz  und  Vergleichskurven
            bei  Larmbelastung zu unterschiedlichen Cyctuszeiten
Figure 6  Change  in respiration rate caused by 95-dB
-------
                                            Oestrogene/Atmungsrcoktion
                                                             •.
                                                                  36    T (Togel
                     Kreuzkorrelationsfunktion
Figure 7 Cross-correlation function between estrogen level and change in respiration rate induced by noise.
                                 REFERENCES

1.  ARTNHR. J.:  Vegetative  Ausg.mgskige uiul Cyclus.  Arch, (.ynaek.  /'/.?  (>0)
    379-392
2.  BRE11M.  U.:  Der  Kreislauf der Fran, scin Alternsuang. seine Anderung  with rend
    Puhertat,  Klinuikterium.  Zyklus. Sehw;inj:ersch;irt uiul  Wochenbett und seine  Heein-
    flussung durch  Zufuhr  von Ovarial-  und   tioiuuloiropen  HDnnoneii. Habil-Schrift.
    l-Yanklurt Main. 24. Sept. ll>5'>
3.  BROWN.  H. B.. A.  KLOPPI R und .1. A LORAIM  The Urinary l-xcrction olOestro-
    genes. Pregnaediol and (ionadotrophins During the Menstrual Cycle. J. Fmiocrin. /7
    (1958)401-410.
4.  DORINC. (i. K., H.  II. I  ()1 S( IIKI  und B. ()( H\V\I)I  Ueitere  rntcrMichiinpon iiber
    die Wirkung der Sexualhormone aui" die Atmung.  Plliigcrs Arch.  252. < I(»5(M 216-230
                                       467

-------
 5. DORING, G.  K. und  E.  FEUSTEL:  Uber Veranderungen der Pulsfrequenz im
    Rhythmus des Menstruationszyklus. Klin. Wschr. 31II, (1953) 1 000-1  002
 6. DORING, G.  K.:  Uber  rhythmische  Schwankungen von  Atmung und  Korper-
    temperatur im Menstruationscyclus. Pflugers Arch. 250, (1948) 694-703.
 7. DORING, G. K.: Uber Veranderungen der Atmung wahrend des Cyclus. Arch. Gyneak.
    ; 82 (1953) 746-751.
 8. GOODLAND, R. L. und W. T. POMMERENKE: Cyclic Fluctuations of the  Alveolar
    Carbon Dioxide  Tension During the Normal Menstrual Cycle.  Fertil. and Steril. 3,
    (1952)393-401
 9. HARMON, F. L.: The effects of noise upon certain psychological and physiological
    processes. Arch. Psychol. (N.Y.), 747(1933) 140-164.
10. HEINEKER, R., K.-E. ZIPF und H.-W. LOSCH: Uber den Einfluss kSrperlichen Train-
    ings auf Kreislauf und Atmung. II. Mitteilung: Trairiingseinflub auf die Reaktionsweise
    des Kreislaufs bei Anwendung von Kalte-, Flickerlicht und Larmbelastung. Z. Kreislauf-
    forsch. 49, (1960)924-935.
11. JANSEN, G.: Relation between Temporary Threshold Shift and Peripheral Circulatory
    Effects of Sound. Physiological Effects of Noise, 67-74 Plenum Publishing Corporation
    (Vortrag Boston) AAAS, 28.12.1969 Edited  by: Bruce L. Welch and Annemarie S.
    Welch.
12. JANSEN, G. und J. SCHULZE: Beispiele von Schlafstoningen durch Gerausche. Klin.
    Wschr. 42, (1964)132-134.
13. MATTHIAS, S.  und G. JANSEN:  Periphere Durchblutungsstorungen durch La'rm bei
    Kinderri. Int. Z. angew. Physiol. 19, (1962) 201-208.
14. NEILL, J. D., E. D. B. JOHANSSON, J. K.  DATTA und E. KNOBIL:  Relationship
    Between the Plasma Levels of Luteinizing Hormone and Progesterone During the Nor-
    mal Menstrual Cycle. J. Clin. Endocr. 27/II, (1967) 1167-1173.
15. OPPLIGER,  G. und E. GRANDJEAN: Vasomotorische  Reaktionen der Hand  auf
    La'rmreize. Helv. Physiol. Acta 17, (1959) 275-287
16. STEVENS, S. S.: The Effects of Noise and  Vibration on Psycho-Motor Efficiency.
    Harvard University, Psychological Laboratory,  OSRD Report 32 and 274 (1941)
17. WILBRAND, U., CH. PORATH, P.  MATHAES und R. JASTER: Der Einfluss der
    Ovarialsteroide auf die Funktion  des Atemzentrums. Arch. Gynaek. 191,  507-531
    (1959)
18. WILDER, J.: Das "Ausgangswert-Gesetz" - Ein unbeachtetes biologisches Gesetz; seine
    Bedeutung fur Forschung und Praxis. Klin. Wschr. 10, (1931) 1889-1893.
                                     468

-------
        THE INFLUENCE OF NOISE ON AUDITORY EVOKED POTENTIALS

                           J. Gruberova, S. Kubik, J. 2alcik
                          Institute of Occupational Hygiene
                             Bratislava, Czechoslovakia
     The auditory  evoked potential, which is defined as an electrical response of brain to
acoustical stimuli,  has recently attracted the attention  of audiologists for the purpose of
objective audiometry.
     The auditory evoked potential detected from the scalp is a nonspecific response widely
distributed over the scalp, with  a maximum in the vertex region.  A typical example of
auditory evoked potential is shown in Figure 1. The sequence of negative and positive waves
in characteristic.
     Seventeen healthy experimental persons were investigated before and after noise expo-
sure. The recording electrodes were placed, according to the ten - twenty  system (Jasper,
1958), in positions O^, P£, C^ andT3- The reference electrode was placed on the chin.
     Recorded  potentials were amplified by a Schwarzer EEC apparatus and added by a
multichannel analyzer NTA  512  of KFKI Budapest. One hundred  responses were always
summed up.
     A 500-msec acoustical stimulus of level about 90 dB with irregular pauses (from 0.5 to
5.0 sec) was used.  Auditory evoked potentials were investigated at three frequencies: 500,
1000 and 2000 Hz. Stimuli were delivered by earphone directly to the ears.
                       100
200
300
400
                     Figure 1 Typical average auditory evoked potential.
                                        469

-------
     Each experimental subject was investigated in four sessions. At the beginning of every
session the auditory evoked  potentials to the three frequencies were recorded and then a
white  noise of level about 90 dB was applied.  The period of white noise application was
changed at every session. Four periods of 0.5, 1,  1.5 and 2 hours were used. Immediately
after termination of the noise, the auditory evoked potentials at three frequencies were
again investigated.
     Ten of the subjects were also tested under the same conditions by classical audiometry.
In comparing the auditory evoked potentials recorded before and after noise, we  concen-
trated  our attention on  amplitude  differences of waves N2 —  ?2- The amplitude was
measured  peak to peak. The mean differences of the whole group were calculated. The
results are shown in Fig. 2.
     A statistically significant difference was found in parietal, central and temporal records
after 0.5, 1, and 1.5 hours of exposure. No statistically significant difference was found
after 2 hours of exposure.
     In Figure 3 can be seen the mean shifts of acoustical threshold of ten persons, obtained
by classical audiometry.
     In  the evaluation of results  obtained after  the noise,  it is  necessary to  take into
consideration the fact  that the white noise has,  besides an influence on the hearing appara-
tus, also an influence on the state of vigilance of the subject, and the amplitude of auditory
evoked potential does depend on vigilance.
                                                                        T,
Figure 2 Decrease in the peak-to-peak difference between N2 and P2 of the auditory evoked potential after
noise exposures of various durations (abscissa) for four different electrode placements (parameter). Stimu-
lus frequencies 500 Hz (thin line), 1000 Hz (dashed line), & 2000 Hz (thick-line).
                                         470

-------
                  dB

                  10 —
                              -I—
                               30
Figure 3 Changes in behaviorally-determined auditory sensitivity (TTS) after exposure of various durations
{abscissa) to 90 dB(A) noise, at 500 Hz (thin line), 1000 Hz (dashed line), and 2000 Hz (thick line).
     The changes of amplitude of auditory evoked potentials with changes of vigilance are
explained by Fruhstorfer and Bergstrom (1969) in terms of a decline in activity of certain
brain functions which are essential for the maintenance of vigilance.
     The Roumanian author Edith Gulian showed also that noise application produces a
decrease of vigilance of experimental persons. She found in her experiments that continuous
noise of 90 dB after 1.5 hours produces a clear decrease of vigilance. In EEC she found a
clear decrease of alpha index at the end of a 1.5-hr session.
     Jerison  (1959) observed the  mental performance of experimental persons exposed to
noise levels of 85 and 115 dB in four 0.5-hr intervals. He found the greatest decrease of
performance after  1.5 hour.  In the last half hour an  improvement in performance was
observed.
     Those  time  relations can  also be  seen  in  our experiments. While after 1.5 hr, the
decrease of amplitude is statistically significant, after 2 hr of noise exposure the decrease
was not significant.
     Amplitude of auditory evoked potentials is influenced, then, by the state of vigilance
on the one hand and changes of hearing produced by white noise exposure on the other.
     We can still compare the changes of auditory evoked  potentials and the changes of
auditory threshold measured by classical audiometry. From classical audiometry results we
can see the maximal decrease of hearing at the end of the first hour, while in  auditory
evoked responses there is a statistically  significant decrease of amplitude even at the end of
1.5 hr. This difference  can be explained in no other way than by changes of vigilance.
     The results obtained by the method of classical audiometry are responses mediated by
the specific auditory pathways, while the auditory evoked potentials are mediated by non-
specific pathways-that is, by the ascending reticular formation of the  brain stem and the
nonspecific thalamic system — and these structures are essential for the state of vigilance.
                                         471

-------
                                    Literature

Fruhstorfer, H., Bergstrom, R. M., Human vigilance and auditory evoked responses. Electro-
     enceph. din. Neurophysiol., 27, 346-355 (1969)
Gulian, E., Effects of noise on an auditory vigilance task, Rev. Roum. Sci. Social. Psychol,
     10,175-186(1966)
Gulian, E., Effects of noise on reaction time and induced muscular tension, Rev. Roum, Sci.
     Social Psychol, 13,33^5 (1967)..
Gulian, E., Effects of noise on arousal level in auditory vigilance, Acta Psychol (Amst.), 33,
     388-393, (1970)
Jasper, H.  H., The ten twenty electrode system of the  international federation, Electro-
     enceph. din. Neurophysiol, 10,371-375 (1958).
Jerison, H. J., Effects of noise on human performance, J.appl psychol, 43, 96-101 (1959).
                                        472

-------
 SOME DATA ON THE INFLUENCE OF NOISE ON NEUROHUMORAL SUBSTANCES
                          IN TISSUES AND BODY FLUIDS

                                  Lech Markiewicz
                        Department of Physiology and Hygiene
                           of the Central Research Institute
                               for Labour Protection
                                  Warsaw, Poland


     Noise, especially at high intensities, is a strong bionegative stimulus acting as a stressor
on the organism.  Changes in sympathetic nervous system reactivity and endocrine activity
constitute non-auditory effects of the noise. Yet  long-lasting  exposure to noise of high
intensity  has a considerable effect  on endocrine system reaction. This was pointed  out in
previous research  work (see references).
     The level of catecholamines can  be the indicator of stress magnitude as well as the
modified sympathetic  nervous  system reactivity.  Many authors have found an  increased
excretion of catecholamines in urine due to noise of high intensity, especially when it comes
unexpectedly and is short-lasting.
     The results of experiments given here prove that noise of high intensity and various
frequencies have some bearing on the catecholamine level in the organism.
     Experiments were carried out  on white rats by exposing them to noise for three hours
daily. The noise frequencies 50, 4,000 16,000 and  20,000 Hz. and intensities from  100 to
130 dB were used. The  catecholamines were estimated in tissues (brain, heart, suprarenal
gland), blood and  urine.  The urine for catecholamine determination was collected in
metabolic cages during 24 hours, beginning immediately  after the exposure to noise. Blood
and  tissues were collected on finishing the experiment. Determinations were performed after
1, 3, 6, 8 and 24 weeks of exposure. In this report some of the more interesting results are
presented.

1. Catecholamines in urine

     Stimuli of frequency 50 Hz cause an increased level of excreted noradrenaline (NA) in
urine only within the  first week of exposure. In the later  phases of the experiment the
excretion of NA remains at the level of  the control value (Fig.  1). On the other hand,
concentration of  adrenaline (A) is higher not only within the first week but also in the third
week of experiment.
     Similar results were  also obtained in the second series of experiments, when  a noise of
frequency 4,000  Hz and  of intensity 100 dB was applied. An increase of NA excretion was
evident within the  first  week only, becoming slightly lower than the control value in the
third week of experiment, but the  more intensified A excretion still  remained in the third
week of experiment. However, the influence of an acoustic stimulus of frequency 16,000 Hz
leads to a decrease of the excretion of both catecholamines; although there is  a marked
tendency for the return of noradrenaline to the control value within the sixth week of
experiment, adrenaline is still being excreted at a low level. Prolonged experiments with

                                         473

-------
  10
M9/24h
1.25-1
            50
   1.00-
   0.75-
   0.50-
   0.2 5i
-^
£
  0.3-
  0.2-
  0.1-
     0  1  3
frequency c/s 50
intensity dB 108
                                    4.000 16.000 20.000
                     NoradrenaLine
   111!     |ll  Lili
   0136     013   0136   01
                         0136
                    Adrenaline
                                         I
                                    01    6  24 (weeks)
                013   0136
                 4.000      16.000
                  100      128

                  Figure 1 Catecholamines in urine
                                    0 1
                                       20.000
                                        130
1
24 (weeks)

-------
20,000 Hz show a low level of catecholamines in the sixth week of experimentation, and
their recovery only after 24 weeks of applying  the acoustic stimulus. Such a long lasting
deviation from normal indicates a lesser degree of adaptability of the adrenergic system to
an acoustic stimulus of high frequencies than to a  stimulus of, shall we say, 4,000 Hz.

2. Catecholamines in blood

     Since the previous experiments have shown different catecholamines excretion in urine
when lower (50 and 4,000 Hz) and higher (from 16,000 to up 20,000 Hz) frequencies were
applied, the level of NA in blood was examined just after termination of a long-lasting series
of acoustic stimulus activity. Figure 2 shows the results, indicating an increased level of NA
in the third as well as  in the first week of the experiment, when noise of 4,000 Hz and  100
dB intensity was applied.
     On  the other hand, long-lasting  stimulus activity of  20,000 Hz did not lead  to an
increase, but on the contrary, to a slight decrease  of NA level in the blood. Nevertheless,  it is
interesting that the A level in blood was, in the initial phase of experiment, increased in
spite of the fact that the excreted amount of A in urine was insignificant.

3. Noradrenaline in brain

     Application of low or high frequencies leads to increased amount of NA in brain tissue,
which is illustrated in fig. 3.  This increase is observed with 50  Hz frequency in both the
third  as well as the first week and it returns to normal not sooner than after 8 weeks time. It
is worthwhile to emphasize  that  with the 4,000-Hz stimulus  the changes are the least
marked. However, in this case an increase of NA in, the brain is found.
     Stimuli of 20,000 Hz applied cause a two-phase reaction to appear; first there is an
increase of NA level in the brain, lasting till  the 6th week of experimentation. Only the
determinations done after 24  weeks, when the experiment has been stopped, have shown
outstanding decrease (nearly by half) of NA level  as compared with the control value.
     Experiments  concerning behavior  of  serotonin carried  out  together  with  Dr.
 Markowska have shown that an increase of this compound occurred in the brain as well as in
 blood with simultaneous excretion of its metabolite 5-HIAA (5-hydroxy-indolacetic acid) in
 urine.
 Comment

     The experimental results, as described above, have proved that metabolic disturbances
 of catecholamines and serotonin are produced under the influence of long-lasting acoustic
 stimuli. They point out the lack of adaptability to this harmful factor of human environ-
 ment.  Stress reaction at the initial phase is characterized by the increase of catecholamines
 in urine, with high level these compounds found in blood and brain at the same time.
     Longer-lasting stimuli, especially of high frequencies, lead to changes in synthesis and
 degradation  of biogenic amines.  Indeed, the influence of noise causes  reduction of

                                        475

-------
       ug/1
      20.0 -I


      15.0-
S     10.0H
NA
        5.0-
                                           03        013 (weeks)

  frequency c/s 4.000                                20.000

  intensity dB  100                          122              122
                                Figure 2 Catecholamines in blood

-------
                                                                 I  T
        10     20      50    100            500   1.000        4.000          20000

        /g
     1.00-


     0.75-
-^

     0.50 H
     0.251 M  •  •     M                M M  M      MM      mm



           0138                  013        06        4 (weeks)

     frequency c/s   5O                           4.000              20.000

     intensity dB  108                            100                 130
                                     Figure 3 Noradrenaline in brain

-------
monoamineoxidase activity with simultaneous intensified excretion of cortisol and thyroid
gland hormones.
     The negative stimuli of this kind, as applied in our experiments, become factors that are
able to cause serious changes in  functioning of the whole organism which, if repeated
habitually, may lead to disease.
     A particular part is played by acoustic stimuli which are on the border of audio- and
ultrasonic devices and are getting more' common. As Konarska and I have pointed out, even
destruction of Corti's organ does not prevent metabolic disturbances of catecholamines and
corticosteroids.

                                  REFERENCES

1.   Arguelles A. E. - Introduction to Clinical Endocrinology editor: E. Bajusz, Basel 1967
     p. 123.
2.   Hrubes V,Benes v. - Activities Nervosa Superios, 7 p. 165, 1965.
3.   Markiewicz L. Konarska M. - Prace Centralnego Instytutu Ochrony Pracy, 19 p. 233,
     1969.
4.   Markiewicz L., Markowska L., - Prace Centralnego Instytutu Ochrony Pracy 18, p.365,
     1968.
5.   Sackler A. M., Weltman A. S., Jurtshuk P. -Aerospace Med. 31, p. 749, 1960.
                                        478

-------
     STRESS AND DISEASE. IN RESPONSE TO EXPOSURE TO NOISE - A REVIEW

             Gosta Carlestam, *Claes-Goran Karlsson** and Lennart Levi**

     Noise has been defined as any unwanted sound, the most prevalent "waste products"
of pur age. Numerous authors claim to have shown that noise provokes physiological stress
reactions, not only as concomitants to the distress reactions implicited in the very definition
of noise, but also through reflex stimulation of the auditory nerves and on to the hypothala-
michypophyseal system. It is  occasionally claimed that exposure to  noise  can cause  a
number of diseases belonging to the  field  of psychiatry and internal medicine, either by
these or by some other mechanisms.
     The  purpose of this  paper is to examine critically the evidence in favor of these
hypotheses and to report,  in summary, a study  conducted  at the laboratory for Clinical
Stress Research.** At the  National Institute of Building Research*, David Wyon and his
associates are studying noise as a component in  the indoor environment.

Noise and physiological stress

     The  term "stress"  is used here in the sense  that Selye described it, namely, the non-
specific response of the  body to any demand made upon it; a stereotyped, phylogenetically
old adaptation  pattern primarily preparing the  organism for physical activity, e.g. fight or
flight.
     It is conceivable that in the dawn of the history of mankind, noise very often was a
signal of danger or else  of a situation requiring muscular activity. In order to survive, the
human organism  had to prepare itself for activity, inter alia by  the non-specific adaptive
reaction  pattern  defined as stress. More often than not, noise  in  today's  industrialized
societies  has a meaning very different from what  it had  during stone age. Yet, according to
one  hypothesis, our genetically determined psychobiological programming still makes us
react as if muscular activity would be an adequate  reaction  to any sudden, unexpected or
annoying noise stimulus. True, it can be argued that some authors have demonstrated not an
increase but rather no reaction or even a decrease in hormonal activity in response to noise
(Bugard,  1955; Sakamoto, 1959). One explanation for  this  controversy might be that the
measurements have been made  at varying intervals after noise exposure. Various endocrine
systems can react after various intervals or even in different directions, some of the reactions
being diphasic.  Accordingly, some reactions,  present immediately after the exposure, may
have disappeared or changed direction in some instances but not in others.
     As one may expect, the reaction pattern  to noise  is not entirely non-specific but is
partially  conditioned by the specific  characteristics of  the reacting  organism. One  man's
meat may be another man's poison. Comparing adrenal hormone reactions in response to
noise in  healthy controls  with those of patients  with cardiovascular diseases  or schizo-
phrenia, Arguelles et al. (1970) found increases in hormone excretion in all three groups, the
reactions  in the two patient groups, however, being significantly more  pronounced.

 •National Institute of Building Research Box 27 163, S-102 52 Stockholm
**Laboratory for Clinical Stress Research, Karolinska sjukhuset, Pack, S-104 01 Stockholm
                                        479

-------
     Horio et al. (1972) exposed rats to various noise levels, measuring corticosteroid levels
in the adrenal glands. They found rapid increases in concentration reaching a maximum
after  15 minutes of noise exposure. At moderate noise levels, the corticosteroids soon re-
turned  to  initial levels. At  higher noise levels, however, corticosteroid concentration re-
mained elevated over longer periods, interfering  with the circadian rhythm.
     Measuring 17-ketosteroid excretion in urine in response to meaningful and meaningless
noise of moderate intensity, Atherley et al..(1970) found that the meaningful but not the
meaningless variety did induce physiological stress reactions.
     In an experiment conducted at our laboratory, 22 young female IBM operators were
studied in  their usual work. In half of the group, the  noise level produced by their IBM
machines increased 6 dB from one day to the next during four consecutive days, the noise
levels being 76, 82,  88 and 94 dB-C, respectively. The other half were subjected  to the
same noise levels but in the opposite order (i.e., 94, 88, 82, and 76 dB-C, respectively). The
noise level normally  prevailing in the office was 76 dB—C. Every  working day started with
two hours  of rest without noise exposure, followed by three 2-hour work periods with noise
exposure as indicated.
     Contrary to what might be expected, the subjects reported only minor increases in
self-rated fatigue  (figure  1) and  "distress" (figure  2).  Although these ratings  increased
slightly with increasing noise,  the  rating differences between the highest and lowest noise
levels were conspicuously small. The corresponding epinephrine and norepinephrine excre-
tion levels (figures 3 and 4) were low or moderate and the changes from control to noise
periods and from low to high noise levels were usually non-significant. Thus,  not even the
objectively rather considerable noise levels used were particularly potent as stressors. This
may be due  to the familiarity of the noise and to the generally positive attitudes of these
subjects to the job per se and to the experiment. It is conceivable that such factors may have
counteracted the stressor effects of the noise. Briefly, then, noise may be a potent stressor
under some circumstances and in some individuals, but need not generally be so.

Noise and disease

     Sakamoto (1959) found  that more than 50%—i.e. a rather high proportion-of the
inhabitants living close to an airport complained of  various  types  of somatic distress,
possibly induced by the aircraft noise.
     In  epidemiological  studies,  several  authors (Mjasnikow,  1970; Andriukin,  1961;
Shatalov et al., 1962; Ratner et al., 1963) report an increased incidence of hypertension in
workers exposed to  high noise levels. According to Mjasnikov, this increase  in morbidity
manifests itself after 8 years of exposure, reaching a maximum after 13 years of exposure.
     Similarly, other authors (Jerkova and Kremarova, 1965; Andrukovich, 1965; Strakhov,
1966;  and Dumkina,  1970) found  an increased incidence of "nervous complaints" in
workers habitually exposed to higher noise levels. Living in areas close to a noisy airport was
accompanied  by increased number of admissions to psychiatric hospitals (Abey-Wickrama et
al., 1969 and 1970). However, the causal implications of this statistical relationship can be
seriously questioned (Chowns, 1970).
     Jensen and Rasmussen Jr.  (1970) inoculated mice with various infectuous agents,
before or after exposing them to noise.  It was found that those inoculated with stomatite
                                        480

-------
oc
                    Ext'eme 5 i
                  Notot all 1
                                       94 dB (C)
                                    A
 08-10  10-12 12-14 14-16
[: jFRIDAY
   ~~l TUESDAY
                                                                    FATIGUE
                                                               88dB(C)
                                                         82 dB (C)
                                                        J]
                                    i
08-10  10-12 12-14 14-16
[jUJTHURSDAY
[     [WEDNESDAY
                                 4
08-10 10-12 12-14 14-16
  [J WEDNESDAY
  ~~I THURSDAY
                                                                                                                  76dB(C)
                         i
                                                                                                                   t
                                                                                                           08-10 10-12  12-14  14-16
                                                                                                           (•   J TUESDAY
                                                                                                           [     1 FRIDAY
                       Figure 1. Self-rated "fatigue" under different noise conditions and during different times of the day.

-------
 EXTREME
                                       DISTRESS  SCORE
12

11

10
9
oo 8
K)
7






- T
A





'





011.1
•








































NdB(C) I24BIC] naBK.


J
fl1

NOT AT All 6' — •— •- *— • — f~* — ? • • I'
ot-w r-u u-u u-w ot-io w-i2

1



12 -U
I | rfiinAv [ 1 THURSDAY

1


J

U-K 01-10 10

;
.
u

\



»



u-u u-w


1



M-K)
| jWfliNESDAY |

i



.



i'




W-t2 12 -V U-W
| TUESDAY
^]TUEf.rAY [ | WEDNESDAY f^^] THURSDAY [ [FRIDAY
Figure 2. Self-rated "distress" under different noise conditions and during different times of the day.

-------
         Adrenaline
1
to

I
7
t
5

*
3
2


"3/nin





. 76dB-C 82







1


X


X
x
X
x
X
X
X
X
X
X
x








X
x
X
X
x
X
X



X
X

X
X
X
X
X
X
x
x
OI-M »-
X
X
X
X
X


L i


X
X
x
X
X
X
X
X

x
X
X
X
X
X
X
x
x
1






xT
>
X
X
X
x
x
X
X
x
X
X





dB-C 88c

^
X
X
x
X

x
x
x
x
x
x
x
x




X
X
X
X





X
X
X
X
x
X"
x
X
X
X
x


I ,
x
'
x
X
X
X
X
n n-n «-i*  io-u n-w w-







*

X
X
X
X
X
X

X
X
X






IB-C



p

X
X
X
X
X
x
X
X
X
X
It M-M /«•






x




'
X
X
X
X
X
X
X
It 11-






X
X



'
X
X
X
X
X
X
n w







X
X








?fdB-C



X
X
X
x
X
X
X
x
X
•H M-
TUESDAY WEDNESDAY THURSDAY







X
x




LX
X
X
X
X
X
X

ro ;o-
X



x
X
X
x
x
x
X
X
X
X

1 tt-
X


X
3
X
X
x
X
X
X
X
x
X
A* *-








» Tim*
FRIDAY
     II
         Noradrenaline
      t


X

x

X

x
x1
X

X

X

X

X

X
x
X
X
x
X
X
x
X
x

x^





















X

X"

x

X

X

X
X

x

x

X

x

X
X
X
X
X
X
x
x

^










X
X
X
XI








X5
X
X
X
X
X
X
X
X
X
X











X
X
X
x^







X
X
X

xl
X
X
X
X
X
X

                                     X
                                         x
X ~
X
X
X
x
X
x
x*
X
X
x
-X
X
.x
x
x
X
x












X
x

x

x

x




x1
X
X

X

X

X
x

x

X

X

••

x
x

x

X
x
X
X









x
X
X
X
x
X
x
x





X
X
X
X
X
x
xl
X
X
X
X
X
X
X
X

X

X

>

X

X
ff
X
X
X
x
X
x
x
X
X
X
X







X
X
X
x-






T
X
X
X

x

X
X
X
X
x
X
x
X
X
XI
X
X
X
X
X
X
X
X
X
x
           O*-M io-i»  a-*  H-I*    «-»  W'li a-*  «-<*    «-io  »-«  /»-*  *-/«   »-»  »-«  /a-»  «-« TfMl
             TUESDAY      WEDNESDAY      THURSDAY        FRIDAY

Figure 3.  Urinary excretion of adrenaline and noradrenaline during four consecutive days with increasing
noise level.
                                           483

-------
      Adrenalin*
s
7
6
5
3

4
1



I
• "ftdB-C 88dB-C


it
/
/
/
/
1 r4-1
J
X

X
x
X
X
X
X
X
T
>
X

X
X





s>

x
X
X

X
X
x






1
'lx
x*
x
X
X
X
X
x
X
X

X
X
X
X
X

X
X







X

x
X
X
X
X
x







/
x
X
x
x
X
x
x
x

82dB-C 76dB-C


X
X
X
X
X
X
X
X

pL.
x
X

X
x
X
X
X
X
X
>
^
X
x
X
x
_L






t
X
X
X
x
£
x
x
X
X
5
X
X
X
x
xK
X
X
X
^x





OHO w-12 ft-w NH» M-W n-B a-n n-it ot-w w-a n-n *H« «-/• w-a a-w «-« r/«f
TUESDAY WEDNESDAY THURSDAV FRIDAY
      Noradrcnallne
4"9/"
"1
It
n
i«
ii
15



n

•

.





J-

X
X
run



"
x
X
X
x
x
x
x
x








7
X
X


;
X
X
XI
X
x





1
xj
<
X

X
X
X
.Xl
X
X
X
X
X
X
v>\


/

s

'•

x1

X

X
X
s
X
X
X
X
X
X
X
x<
f^


• *



X
X
X
X
x
x
x
X
X
r^>1








*
r.
                                          x
                                            X
                                           -X
                                            X

                                          X

                                          X
        ono w-n n-H H-W
           TUESDAY
oi-io to-a  Q-H «-/*
 WEDNESDAY



X





X

X
X
X
X




xH LX
^ K
X
X
<^1 ixj
7
;
X
x
X
x
X
&d








x
X
x
X
1
X

-------
virus just before noise exposure were more susceptible, whereas those inoculated after the
exposure were less susceptible, than non-exposed controls.
     Reviewing studies  on noise and mental disease, Lader  (1971) concludes  that noise
exposure does not generally increase psychiatric morbidity but might be of some etiologic
significance in neurotic and anxious subjects.
     Briefly, then, some of the physiological reactions found  in response to noise exposure
seem  to be closely related to the non-specific  physiological reaction pattern  defined as
"stress". Stress has been hypothesized to be one of several pathogenetic mechanisms acting
by increasing  the "rate  of wear and tear" in the organism. Although some circumstantial
evidence has been presented, there is still no proof.
     Some epidemiological studies seem to indicate a higher occurrence of "psychosomatic"
and mental disorders in  subjects  exposed to prolonged and rather intense noise. However, it
should be kept in mind that such an exposure is often accompanied by exposure to a variety
of other potentially noxious stimuli. In addition, various segregational forces  may "sort
out"  particularly susceptible individuals to noisy, unpleasant and/or pathogenic environ-
ments.
     Accordingly,  we have  to conclude that the  evidence in favour of noise as a major
pathogenetic  environmental agent is rather shaky. To solve this controversy, future research
should focus  on controlled intervention studies with  an interdisciplinary and multifactorial
design.
                                     References

Abey-Wickrama, I., Brook, M. F., Gattoni,  F. E., and Herridge, C. F.: Mental - hospital
     admissions and aircraft noise. Lancet 1275-1277 (1969).
Abey-Wickrama, I., Brook, M. F., Gattoni,  F. E., and Herridge, C. F.: Mental - hospital
     admissions and aircraft noise. Lancet 467 (1970).
Andriukin, A. A.: Influence of sound  stimulation on the development of hypertension.
     Clinical and experimental results. Cor Vassa 3:285-293 (1961).
Andrukovich,  A.  L: Effect of industrial noise  in winding and weaving factories on the
     arterial pressure in operators of the machines. Gig.Tr.Zabol. 9:39-42 (1965).
Arguelles, A.  E., Martinez, M. A., Pucciarelli, E., and Disisto, M. V.: Endocrine and meta-
     bolic effects of noise in normal, hypertensive and psychotic subjects. In: Welch and
     Welch (Eds.): Physiological Effects of Noise, Plenum Press, New York-London, pp.
     43-56(1970).
Atherley, G.  R. C., Gibbons, S. L., and  Powell, J. A.: Moderate acoustic stimuli: the
     intervelation of subjective importance and certain physiological changes. Ergonomics,
     13:5:536-545(1970).
Bugard, P.: Presse med., Paris 63:24 (1955).
Carlestam, G., and Levi, L.: Urban Conglomerates as Psychosocial Human Stressors. Stock-
     holm, Royal Ministry for Foreign Affairs, (1971).
Carlestam, G., Noise - the Scourge of Modern Society Ambio Vol. No. 3 1972 Stockholm.
Chowns, R. H.: Mental-hospital admissions and aircraft noise. Lancet 467 (1970).

                                        485

-------
Dumkina, G. Z.: Some clinico-physiological investigations made in workers exposed to the
     effects of stable  noise.  In: Welch and Welch (Eds.): Physiological Effects of Noise,
     Plenum Press, New York-London, p. 346 (1970).
Horio, K., Sakamoto,  H., and Matsui, K.: Adrenocortical response to noise exposure. Re-
     print from: Joint Meeting of International Societies for Hygiene Preventive and Social
     Medicine, 29 Oct. -1 Nov (1972).
Jensen, M. M., and Rasmussen Jr., A. F.: Audiogenic Stress and susceptibility to infection.
     In: Welch and Welch (Eds.): Physiological Effects of Noise, Plenum Press, New York-
     London, pp. 7-20 (1970).
Jerkova,  H., and Kremarova, B.r Observation of the effect of noise on the general health of
     workers in large engineering  factories;  attempt at evaluation. Pracovni Lekarstui
     17:147-148(1965).
Lader, M. H.T Responses to repetetive stimulation. In: Levi, L. (Ed.): Society, Stress and
     Disease: The Psychological Environment and Psychosomatic Diseases, London,  New
     York, Toronto: Oxford Univ. Press, pp. 425429 (1971).
Levi, L.  (ed.): Society, Stress and Disease - The Psychosocial Environment and Psychoso-
     matic Diseases. Oxford University Press, London (1971).
Levi, L.  (ed.):  Stress  and Distress in Response to Psychosocial Stimuli. Pergamon Press,
     Oxford (1972).
Mjasnikow, A. L.: In:  The pathogenesis of essential hypertension. Proceedings of the Prague
     Symposium 153-162 (1970).
Ratner, M. V., Medved, R. A., Film, A. P., Skok, W. I., Rodenkow, W. F., und Makarenkow,
     N. A.: Thesen des Berichtes der allunionswissenschaftlichen Tagung iiber methodische
     Probleme  der Larmwirkung auf den Organismus. Institut fur Arbeitshygiene und
     Berufskrankheiten, AMW, UdSSR (1963).
Sakamoto, H.: Endocrine dysfunction in noisy environment. Report L Mie Medical Journal.
     9: 1:39-58(1959).
Sakamoto, H.:  Endocrine dysfunction in noisy environment.  Report  II. Mie  Medical
     Journal. 9: 1:59-74(1959).
Shatalov,  N. N., Saitanov. A. O., and Glatova, K. V.: On the State of the cardiovascular
     System  under conditions  of exposure-to continuous  noise.  Report T-411-R,
     N65-15577 Defense Research Board, Toronto, Canada (1962).
Strakhov, A. B.: Some questions of the mechanism of the action of noise on an organism.
     Report N 67-11646, Joint Publication Research Service, Washington, D.'C. (1966).
                                       486

-------
                   SOME LABORATORY TESTS OF HEART RATE
                          AND BLOOD VOLUME IN NOISE

                                   Karl D. Kryter
               Stanford Research Institute, Menlo Park, California 94025
Introduction
     Two soundproof rooms, decorated as residential living rooms, were arranged so that
two subjects placed in each room could be monitored continuously for EKG and peripheral
blood flow. The subjects in each room were separated by a drape so that they could not see
each other.  The subjects, seated in easy chairs, were asked to read  novels or magazines
during the test sessions. By means of hidden loudspeakers, noise could be introduced into
the test rooms.
     Each daily test session  lasted for two continuous hours (except for one 5-minute
break). The subjects were instructed to try to behave as  though they were resting and
reading in their own homes. They were also told tha,t they, would not be exposed to noises
at a  level any. greater than they might hear in a home and that there was to be no regular
pattern of noise, or no-noise, during any  session. Six adult housewives, age  25-45 years,
served as subjects in the two pilot studies.                                         .
     The data for  the blood  volume from a photoelectric plethysmograph, and the heart
rate, determined by means of an electrocardiograph, were each averaged for each successive
10-sec epoch during the test sessions except for a forced 10-sec period that occurred every 5
min to clear the storage register of the computer.

Results

     Study /-Repeated Sessions of Exposure to a Similar.Pattern of Quiet-Noise-Quiet.

     Figure  1  gives the  average heart rate in beats per second for  4 subjects who were
exposed to a "quiet" (35 dBA) ambient background with 1-4-min bursts of noise at a level
of 90 dBA interspersed at more or less random intervals. (See top section of Table 1 for
details of procedure.)
     It is seen that during the periods of quiet  for any one session, the average heart rate
was less  than during the noise interval. This would, of course, suggest that the bursts of
noise caused an increased stress-arousal reaction in the subjects as compared to their some-
what more relaxed physiological state during the periods of quiet. Of equal interest is that
the average heart rate for both the quiet and noise conditions showed a progressive decrease
from the first session to the third daily session.
     As seen in  Figure  2, the noise also caused an increase (although  of smaller relative
magnitude)  in  physiological  tension measured  by peripheral  blood flow. However, the
average blood volume during each of the different sessions was very similar. It is to be noted
that  the units for measuring blood volume were subtracted from the number "50" in order
that  an increase in the  score, as for heart rate,  would indicate an increase in "stress", i.e.
greater peripheral vasoconstriction with reduced peripheral blood volume.
                                        487

-------
           Q = 35 dBA Ambient (Quiet)
          N* - Bursts of Pink Noise at 90 dBA
               Bursts of Noise of 1 — 4 minute Duration
               Period Between Bursts, 7 — 10 minutes
81.0 	


80.5 	

/y.b 	
78.6
78.5 	
Q
o 78-°
o
UJ
uj -IT r
cc
HI
a.
77 n
<
"- 76.5 —
a:
*£ ir n
Uj /U.U
X
75.5 	
75.0 —
•JA n 	
73.5 —

Q

N*




AVERAGE

Q



76.7





—

Q


N*








AVERAGE


75.6


Q


	
a

N*
•
	 	 76.9

AVERAGE
1





Q
—
Q

N*




AVERAGE



Q
	


	
                                SESSIONS
   Average
of All Sessions
         TA-8755-5
          Figure 1 Average heart rate during pilot study.
Each of four  subjects in  three daily sessions of quiet  interrupted
by  bursts  of noise.
                              488

-------
                                   Table 1 Sequence of test sessions and acoustical conditions
                                                   for pilot studies
                                Study I:   Four Subjects (A, B,  C,  and D)
Session
1
2
3
Acoustic
35 dBA Pink 90
7-10 minute ' 1-4
duration
Same
Same
Condition
dBA Pink
Q
minute '
duration
as 1
as 1
                                                                               35 dBA Pink
                                                                               7-10 minute
                                                                                    duration
oo
Study  II:   Four Subjects  (A,  B,  E, and  F)
      1  Session
                 Acoustic Condition
1

2


3
l.-M -.._
1
: 4

1
:5

A & B
E & F
A & B
E & F
A & B
E & F
A & B
E & F
A & B
E & F
Q (35 dBA,
N (85 dBA,
NQN
QNQ
QQQ
«5 
-------
     Study II—Daily Sessions of Exposure to Different Patterns of Quiet and Noise.

     The basic data of this study for heart rate are summarized in Figure 3 and the results
for peripheral blood volume  in  Figure 4. (See lower section of Table  1 for details of
procedure.) We see in Figures 3 and 4 that:
     (1)  Unlike the findings  in  Study I, there is  no consistent  relationship  between the
         presence of noise and increased stress as revealed by increased heart rate or
         decreased peripheral blood volume; for example, bursts of quiet during ambient
         noise, or burst of noise during ambient quiet both caused some apparent decrease
         in stress.
     (2)  for the average overall sessions there was a somewhat greater heart rate when the
         ambient  was noise, whereas the blood volume indicated that the least amount of
         stress was present when the ambient was the noise.                    :
     We  feel, however, that the results cannot be interpreted  in a meaningful way with
respect to  the  effect of the  presence  or absence of noise  on these physiological stress
reactions, but rather that there was an interaction effect with the alternation from session to
session of  the  patterns of noise and  quiet that occurred within any  one session. This
possibility is revealed in Figure 5 where it is seen that during the second or (repeat) session
for a given acoustic sequence  there is a decided increase in stress as measured by heart rate.
This increase, as is seen oh  the right-hand portion of Figure 5, occurred for all of the
individual conditions  scored  separately; that  is, the heart rate as found during the two
ambient conditions and also between the two "burst" conditions.
     This result is contrary to the adaptation that occurred with repeated sessions of the
quiet-noise-quiet sequence in  Study I  (Figure 1) and must be related to the interposition
between  the repeat sessions, for a given acoustic sequence,  of sessions utilizing different
acoustic sequences or patterns of stimulation. Qearly, changing test conditions from session
to session had a decided effect upon the heart rate of the subjects, regardless of the ambient
and non-ambient acoustic conditions that were utilized within  any one session.
     On  the other hand, peripheral blood volume, as shown in Figure 6, showed a small
decrease  in  "stress" between  the .first  and second sessions of a given sequence of acoustic
conditions. However,  it is suggested that these peripheral blood volume data are probably
not as reliable an indicator of what is usually considered in the present context as "stress,"
as is the measure of rate of heart beat, for the, following reasons: (a) the changes in blood
volume, as measured, are insignificantly small; (b) there appeared to be no consistent trend
of the average  blood volume from session to  session during Study  I; and (c) the  rather
extreme degree  of vasodilation (as indicated by  a lower number on our blood volume
measure) during all sessions of Study  II (about a  25 percent lower score than found in
Study I)  was perhaps due  to  the fact that during the  period Study II was conducted the
outdoor ambient temperature was generally high, on some days exceeded  100°F.  Even
though the test  chamber was  air-conditioned, the temperature in the room was somewhat
higher during extremely hbt days and we suspect that the general vascular condition of the
subjects was influenced to some extent by these conditions.      .
     Accordingly, it is hypothesized that although  changes in peripheral blood volume may
be indicative of relative conditions of stress within rather short spans of time,  participation

                                        490

-------
   N*  = Bursts of Pink Noise at 90 dBA
    Q  = 35 dBA Ambient (Quiet)
         Bursts of Noise of  1 — 4 minute Duration
         Period  Between Bursts 7 — 10  minutes
   NOTE:  The larger  number of the  vertical ordinate the less the blood volume (the
           greater constriction  of  the peripheral blood vessels).
44.0

42.5 	
X 41 5 41-4
I
LU
O
>
Q
O
O
m 39.5 —
39.0 —
38.5 —
38.0 —
37.5 —



39.6
Q


•

42.9
N*

AVERAGE





\ AVERAGE




41.6
Q

1



V



—
—
—
—
—


41.5
Q


41.8

42.0
N*




41.9
Q





—
—
—
—
~*






41.0
Q

2
SESSIONS

;
41.4


41.7
N'

AVERAGE


41.4
Q


—
-
—
—
—
~~*



40.7
Q


41.5


42.2
N*






41.6
Q



	
—
—
—
—
~
3 : Average
of All Sessions

                                                                             T A-8 7 55-6
         Figure 2 A measure of peripheral blood volume during pilot study.
           Each of four subjects in three  daily sessions  of quiet  interrupted by
           bursts of  noise.
                                        491

-------
                      Four Subjects Each Session = 2 hours
                      N = Pink Noise at 85 dBA
                      Q = 35 dBA Ambient (Quiet)
                     N* - Bursts of Pink Noise at 85 dBA Ambient
                          Bursts of Noise of 1 - 4 minute Duration
                          Period Between Bursts,  7-10 minutes
   77.0
8
tr
LU
Q.
HI
I
76.5

76.0

75.5

75.0

74.5

74.0

73.5

73.0

72.5

72.0

71.5

71.0

70.5

70.0

69.5

69.0

68.5

68.0
                                                                          AVERAGE
                                        Averages for  Conditions,
                                                  Pilot Study I
                                                                  79.0-
                                                                  78.0
                                                                       77.0
                                                                  76.0-
                                                                  75.0-
                                                                  74.0-
                                                                              lf>
	 ; — — 72.9- . — . -

AVERAGE
7.1.6
	





^^
ffl
LU
cc
9-
Q

L 9 DURING 71.7



n
t-
O
a
Q







o
to
p.
LU
CC
a.
N

r><
CN
r~
0
z
£
o

«
<-
to
O
N
AVERAGE

Q N-Q.





	 /1 .4


n
695!
I*

jjE
r
r>
O'
O
2
CC
Q
N-
AVERAGE
l-
V)
O
a.
Q
CQ



CO
*
a
6

en
oi
Z
i
b
i



AVcnAljc 	 —





CO
8
o
•
2
O
              AMBIENT
               QUtET
              35  dBA
                           AMBIENT
                            85 dBA
                          PINK NOISE
                         WITH  BURSTS
                          OF QUIET
   AMBIENT
    QUIET
     WITH
   BURSTS
OF PINK NOISE
    AVERAGE
      OVER
PRE-DURING-POST
                                                                          TA-8755-7
                   Figure 3  Average heart rate during pilot study.
                                       492

-------
     N - Pink Noise at 85 dBA Ambient
     Q = 35 dBA Ambient  (Quiet)
     N* =  Bursts of  Pink Noise at 85 dBA Ambient
           1—4 minute Duration
     Q* =  Intervals of Quiet 1—4 minute Duration
     Period Between N* and Q* Bursts 7-10 minutes
     Each Session =  2 hours

     NOTE:  The larger the number on the vertical ordinate  the less the blood volume,
             the greater the constriction of the peripheral blood vessels.
*M ft
T3 5
*>o n . . ,i ,
*V7 ^ 	

X
i
s 31-&
UJ
31.0
_l
O r^^
> JU.t>
Q 30.1
o 	
O 30.0 	








•^

















AVERAGE

— i -wi n










Average for Conditions,
Pilot Study I









30.5
AVERAGE













42.0 -
41.5 	
41.0-
--IS
UJ
-40.0- £
—
DURING 42.2 J
41.5 AVERAGE

CD
CO
o
Q.
Q N* Q





	 l AVERAGE

30.2





AVERAGE


m
   29.0

   28.5

   28.0



   27.0

30.1

	 —



]m^m
8
UJ
DC
Q.
Q
AVERAGE
| O DURING 30.1
R
1-
O
a.
Q
30.0






in
8
UJ
CC
a.
N
AVERAGE

| 9 DURING 29.5
a
H
w
O
a.
N






UJ
CC
a.
Q
| 2 DURING 30.4 |
— ^
cp
1-
in
O
a.
Q

30.2








o"
n
a
9
a

p
z
9
z

in
8
C3
*
Z
a
AVERAGE







             AMBIENT        AMBIENT      AMBIENT
               QUIET          85 dBA          QUIET
              35  dBA       PINK NOISE   WITH BURSTS
                           WITH  BURSTS OF PINK NOISE
                             OF QUIET
                                                                          TA-8755-8
          Figure 4 A measure of peripheral blood volume during pilot study.
                                        493

-------
  80
   78
   76
ui
D
u
UJ
I
   72
   70
   68
   66
O  Q"N"Q
X  QQQ
•  N( )N

O  QUO

X  "Q"

O  "N"
                                                          _L
         FIRST SESSION
            FOR A
          CONDITION
                SECOND     MIDDLE
                SESSION     SESSION
                 FOR A    (QQQ ONLY)
               CONDITION
FIRST SESSION
    FOR A
  CONDITION
SECOND SESSION
    FOR  A
  CONDITION

       TA 8755 2
                    Figure 5 Average heart rate in study for various test sessions
 of the peripheral vascular system in homestatic functioning of the body may make interpre-
 tation of the relation of blood volume to stress rather difficult.

 Summary and Conclusions

      (1)  It would appear that heart rate and peripheral blood flow are possibly correlated
          measures of physiological stress reactions, but that peripheral blood flow, as
          detected and measured in  these studies, is probably a somewhat less sensitive and
          reliable measure of physiological reaction to "stress" than heart rate.
      (2)  When tested in successive sessions with a quiet ambient, subjects showed increased
          heart rate and a small decrease in peripheral blood flow when occasional bursts of
          noise (90 dBA) were presented.
      (3)  When the ambient conditions were varied between test sessions, either bursts of
          "quiet," in  a noise-ambient, or  bursts of "noise," in a quiet ambient, resulted in
          an apparent slight decrease hi stress as measured by decreased heart beat rate or
          increased peripheral  blood volume.  However, the change  was  larger  with the
          bursts of "quiet" than with the bursts of "noise."
                                          494

-------
   34
   33
   32
X
 i
8
-  31
tu
5
D
O

Q
O
g
03
30
   29
   28
   27
                                     T
                          •  N"Q"N
                          O  Q"N"Q
                          X  QOQ
             I
                        I
                                           I

                                        • N( )N
                                        O Q{ )Q
                                        X "Q"
                                        D "N"
                                           I
       FIRST SESSION
          FOR A
        CONDITION
                     SECOND
                     SESSION
                      FOR A
                    CONDITION
QOQ ONLY
                    FIRST SESSION
                       FOR A
                     CONDITION
SECOND SESSION
     FOR A
  CONDITION
  Parameter  is acoustic  condition  (see Table  1 for  specifics).
           Figure 6 Average of a measure of blood volume in study II for various test sessions.
      (4)  The average heart rate and freedom of peripheral blood flow of the subjects were
           as much, if not more, related to the experimental design of the test sessions as to
           the presence of quiet  and noise per se. For example, average heart rate of the
           subjects increased  with continued testing day after day when (Study II) quiet-
           noise-quiet sessions were alternated on successive days with so-called noise-quiet-
           noise  sessions, or  a completely quiet session; whereas, when the subjects were
           exposed on consecutive days only to  the quiet-noise-quiet pattern (Study  I), a
           significant progressive decrease was evident in this stress response (i.e. adaptation
           or  habituation occurred). In that  regard, it could be  suggested that in Study I
           (with  only the quiet-noise-quiet sequence being used) the subjects in addition to
         ,  being  apprehensive about the general experimental situation viewed the noise as
           the obvious change in the environment that was potentially stressful and to which
           they might be expected to respond. Whereas in Study II, it is conceivable that the
           subjects, from session to session, became somewhat confused as to what were the
                                          495

-------
         true experimental variables that were being manipulated. This presumably could
         cause a general increase in tension as the sessions progressed without the adapta-
         tion or habituation that occurred in Study I.
    (5)  These findings must be considered as tentative in view of the small number of
         subjects involved, the exploratory nature of the studies, and the lack of consist-
         ency between decreased peripheral blood volume and increased  heart  rate as
         measures of "stress."

Acknowledgment

    This paper was prepared under Grant NIMH 18161-02  from the National Institute of
Mental Health, Department of Health, Education and Welfare.
                                 REFERENCES

 1.  BERGMANN, F.  J., Untersuchungen zur Frage der Kompensation vegetativer Schall-
     reaktionen durch eine vasoaktive Substanz (unpublished Dissertation, Essen).
 2.  DAMS, E. H., Kompensationserfolge larmbedingter Funktionsstorungen durch Theo-
     phyllin - nicotinat  in  Abhangigkeit  vom  Umgebrunspegel so'wie  von  Alter und
     Geschlecht. Dissertation Essen (1972).
 3.  GRIEFAHN, B.,  Examples of Noise induced  Reactions of the Autonomic Nervous
     System during Normal Ovarian Cycle. (Int. Congress on  Noise as  a Public Health
     Problem. Dubrovnik, 1973).
 4.  HEBB, D. O., Drives and the C.N.S. (Conceptual Nervous System) Psychol.  Rev.  62
     (1955)
 5.  HEZEL, J. S. Zur vasokonstriktorischen Reaktion bei Impulsschallexposition (110  dB
     (A))  und unterschiedlicher zeitlicher  Energieverteilung  (unpublished Dissertation,
     Essen).
 6.  JANSEN, G. Zur Entstehung vegetativer Funktionsstorungen durch Larmeinwirkung.
     Arch. Gewerbepath. u. Gewebehyg. 17, 238-261 (1959).
 7.  JANSEN, G. Zur  nervosen Belastung durch La'rm. Beihefte z. Zentralbl. f. Arbeitsmed.
     und Arbeitsschutz, 9 (1967) Dr. Dietrich Steinkopff Verlag
 8.  JANSEN, G. Relation between Temporary Threshold Shift and Peripherical Circula-
     tory Effects of Sound.  Physiological Effects  of Noise, 67-74 (1969) Plenum  Press
     Publishing Corporation,  New York.  Edited by: Bruce L. Welch and  Annemarie S.
     Welch.
 9.  JANSEN, G., Grenz- und Richtwerte in der Larmbekampfung und ihr psychophysio-
     logjscher Aussagewert. Tagung: "Der La'rm als Gefahr fur den Menschen in der zu
     planenden und gebauten Umwelt" DAL, Stuttgart 10. -11.10. 1972.
10.  JANSEN, G., Untersuchungen Uberdie psychophysiologische Wirkungvon Gerauschen
     mit unterschiedlichem Bedeutungsgehalt. Forschungsbericht an das Bundesministerium
     deslnnern(1973)

                                       496

-------
11.  JANSEN,  G., Fluglarm.  Eine interdisziplinare Untersuchung iiber die Auswirkungen
    des Fluglarms auf den Menschen.-Arbeitsphysiologische Sektion.- (unpublished).
12.  JANSEN,  G., W.  KLOSTERKOTTER und  R.  REINEKE:  Experimentelle Unter-
    suchungen   zur   Kompensation   larmbedingter   Gefassreaktionen.  Schriftenreihe:
    Arbeitsmed. Sozialmed. Arbeitshyg. 29, 303-330 (1960).
13.  LYNN, R., Attention, Arousal, and the Orientation Reaction. Pergamon Press (1966).
14.  MALMO,  R. B., Activation:  a Neuropsychological  Dimension.  Psychol. Rev.  66
    (1959).
15.  MEIER,  F. J.,  Untersuchungen  viber  die Wirkung von Schallreizen mit gleicher
    Energiesumme bei unterschiedlichem Zeitgang auf Gehor und peripheren Kreislauf-
    widerstand. Dissertation, Essen (1971).
16.  SOKOLOFF, N., Perception and  the Conditioned Reflex. Pergamon Press, Oxford
    (1963).
17.  WELCH, B. L.,  and A.  S. WELCH, "Physiological Effects of Noise".  Plenum  Press
    Publishing Corporation, New York (1970)
                                       497

-------
             SESSION 6

SLEEP AND ITS DISTURBANCE BY NOISE

       Chairmen: B. Metz, France
                M. Levi, Yugoslavia
                499

-------
                     EFFECTS OF NOISE ON SLEEP: A REVIEW

                                 Harold L. Williams
                               University of Minnesota

     The effects  of acoustic stimulation on behavioral and physiological phenomena of
human sleep depend on several factors, including: (a) the stimuli: their physical parameters,
qualitative aspects and scheduling; (b) stage of sleep and accumulated sleep time; (c) instruc-
tions to the subject and his psychophysiological and motivational state; and (d) individual
differences on such variables as age, sex, physical condition and psychopathology. Stimulus-
response relations vary, not only for different physiological systems and  behavioral events,
but also with subtle differences in measurement criteria for specific physiological systems or
behavior. Furthermore,  habituation and adaptation to acoustic stimulation during sleep
depend on both  the programming of stimuli and  the response system under study.  This
paper reviews a number of recent investigations which indicate that the measurement of
responsiveness during sleep is a multivariate problem of considerable  complexity.

                                  Sensory Processes

     In general, with neutral and brief acoustic stimuli such as clicks or tones, the intensity
required to awaken a sleeping subject is somewhat higher than his sensation threshold during
wakefulness.  This well-known fact has been taken as evidence that in  sleep, either the
thresholds of sensory transducers are  raised or conduction in afferent pathways is dimin-
ished.  However, several investigators, recording various behavioral and physiological re-
sponses to  sensory  input, have  concluded that neither sensation thresholds  nor the
information-handling capacity of sensory systems is impaired during sleep, at least in man.
For example, Davis et al. (1939) reported that during stages 1 and  2  the K-complex of the
sleep electroencephalogram (EEC) was regularly evoked by rather faint tones, about 20 dB
above the noise level of the bedroom,  Williams et  al. (1964) found that in sleep stages  1, 2,
and 3 (see Rechtschaffen and Kales, 1968), simple acoustic stimuli no more than 5 dB above
waking sensation  threshold  could evoke statistically reliable EEC, autonomic and behavioral
responses, while  slightly higher  intensities  were usually  required for stages 4 and REM.
Keefe  and  his  co-workers  (1971), delivering increasingly intense  tones to young males,
reported that the  average threshold for nighttime awakening was about 25 dB (range 5 to 35
dB)  above the sensation threshold established during waking,  or about  60 dB referred to
.0002 dynes/cm2. Awakening thresholds for daytime sleepers were about  15 dB higher  than
those for night sleepers,  and thresholds (at night) were slightly lower in stage 2 than in REM
or high-voltage slow-wave (delta) sleep. In that study, as in the investigation by Williams et
al. (1964), statistically reliable EEC and cardiac responses occurred to stimuli only 5 to 10
dB above waking sensation threshold.  Thiessen's  studies of recorded truck noise found
significant shifts  toward  EEC awakening in 10%  of truck sounds  at 40 dB(A) and some
subjects awakened more than half the time at peak noise levels no  greater than 50 dB(A).
(See also Thiessen's report in this Symposium.) We can conclude that the frequent failure to
awaken the sleeping human with sensory stimuli whose intensity is above waking sensation

                                        501

-------
levels is  not generally due to raised thresholds at the periphery or to impaired sensory
transmission. (See also, Rechtschaffen et a!.,  1966; and Watson and Rechtschaffen, 1969).
                              Properties of the Stimulus

Brief Sounds.

     The likelihood and magnitude of EEC, autonomic and behavioral responses to brief
(msec  to sec) neutral acoustic stimuli delivered during sleep is a monotonic function  of
stimulus intensity (e.g., Williams et al> 1962; Watson and Rechtschaffen et al. 1969; Keefe et
al} 1971; Anch, 1972). However, different response systems show differential sensitivity to
stimulation. For example, in the Keefe et al, 1971, study, EEC responses were most sensi-
tive to 1,000 Hz tones, followed by cardiovascular variables. Electrodermal responses were
found  only in the presence of full arousal, and respiratory periodocity was not altered by
stimuli sufficiently intense to cause EEG or behavioral awakening. Despite variations among
response measures, each of these studies showed that humans are capable  of perceiving
graded auditory stimuli and of responding proportionately to their intensity in all stages of
sleep, throughout the night. (See also Derbyshire and McDermott, 1958, for a simjlar con-
clusion.)
     Responsiveness during sleep also varies with other physical parameters of simple audi-
tory stimuli. Vetter and Horvath (1962)  found that with brief acoustic stimuli, relatively
low frequencies (100 Hz) and fast rise times (1 msec) were most effective for eliciting the
EEG K-complex in  stage 2, and unpublished pilot studies in our laboratory confirmed these
same effects for frequency of responding on a microswitch taped to the hand. Similarly, Hutt
et al. (1968) reported that square-wave tones were more potent than  sinewave tones for
eliciting electromyographic responses in human neonates, especially square-wave tones with
low-frequency fundamentals (i.e., 100 Hz).
     Several investigators have observed that the arousal value of a stimulus depends on its
quality as well as its strength (See Koella, 1969). Weak stimuli that are novel or unexpected
(e.g., offset of a continuous sound), relevant to biological drives (e.g., the smell of meat to a
hungry animal) or that possess acquired significance (e.g.,  one's own name, or the whimper
of one's child) can cause immediate awakening. In some stages of sleep, humans can analyze
and respond differentially to such complex sounds as spoken names (Oswald et al., 1960),
sentences  (Lehmann and Koukkou,  1971), or complex instructions (Evans et al, 1966).
Furthermore, differentiated responses acquired during waking to specific acoustic stimuli,
persist  during sleep in both  animals and humans (e.g., Buendia  et al., 1964; Granda and
Hammack,  1961; Zung and Wilson, 1961; Williams et al., 1966; Schicht et al., 1968). These
results indicate that during some stages of sleep, sensory analyzer mechanisms in  the brain
remain operative  so that incoming signals  can  be  encoded and  categorized. Moreover,
instructions or information  acquired prior to sleep  must remain in long-term  memory,
available to the analyzer mechanism.

                                        502

-------
Fluctuating and Continuous Sounds.

     Crescendos of white noise rising over a period of seconds, sounds of airplane flyovers,
and  fluctuating sounds of automobile traffic can cause gross alterations in sleep, including
inhibition  of  delta sleep  (stages 3 and 4), increased body movements,  wakefulness,  and
delayed onset of sleep (Lukas and Kryter,  1970; Schieber et al, 1968; Pearsons et al [with
Globus and co-workers] 1973). The  results of Schieber and his colleagues (1968) in Metz's
laboratory with recorded traffic noise are particularly interesting.  They found that low-
density traffic sounds averaging 61  dB were more disruptive of sleep than high-density
traffic averaging about 70 dB. These data suggest that relatively infrequent sounds (one or
two per minute) which exceed background noise levels may cause more general disturbance
of sleep then relatively frequent sounds with higher average intensity. Perhaps in the latter
case where the surprisal value of the stimuli  is less, adaptation is more likely to occur.
     In Thiessen's experiments, where  sleeping subjects were exposed to recorded noise
from a passing truck at selected sound levels, there was a 5% probability of behavioral
awakening at 40 dB(A) and a  30% probability at 70  dB(A),  but with wide individual
differences. Some subjects awakened more than half the time at 50 dB(A) whereas others
almost never awakened, even at 75 dB(A).
     In a study by Scott (1972), continuous high-intensity noise (95  dB) that was turned on
at bedtime caused a loss of stage REM,  but had no substantial effect on non-REM states or
on  other  measures of sleep  disruption. By the  second  night of stimulation, stage REM
percent was returning toward baseline control level. Scott estimated that adaptation would
have been complete after a few nights of noise exposure. Other evidence concerning adapta-
tion and habituation will be discussed in another section of this review.

                              Properties of the Response

     As was  mentioned, different response systems are differentially sensitive to neutral
acoustic stimuli of moderate intensity. Whereas, during waking, EEC, autonomic and motor
responses occur simultaneously to tones adjusted to sensation threshold, during sleep these
responses  show a consistent hierarchy. The EEC is most sensitive, followed by the cardio-
vascular system (i.e., heart rate and peripheral vasoconstriction), followed by electrodermal
activity, respiration and motor behavior. Further, this hierarchy of sensitivity is consistent
over the several stages of sleep. Factors such as  accumulated sleep time, time of day or
night, or the presence of phasic  activity such as rapid eye movements during stage REM
apparently do not alter the ranking or responsiveness (Keefe et al., 1971).
     As would be expected, thresholds for the awakening response depend on its definition.
Full  EEC arousal often  occurs  without a  specified motor response, particularly during
high-voltage sleep stages 3 and 4 (Keefe  et al., 1971). Investigators disagree, however, as to
whether the converse can occur during  sleep—that is,  a designated or conditioned motor
response without associated EEC arousal. Keefe et al. (1971) found no such events, whereas
Williams et al. (1966) reported  that many  motor responses to  specific auditory signals
occurred without prior signs of EEC awakening. It is agreed that the  obtained threshold for
behavioral awakening increases with  the  complexity of the required motor response.  For

                                         503

-------
 neutral stimuli, pressing a microswtich taped in the hand occurs systematically to noise or
 tones at about 25-35 dB above waking sensation threshold (Williams et al., 1964; Keefe et
 al.,  1971). The threshold for obtaining complex verbal responses signifying recognition of
 specific properties of the stimulus appears to be about 65 dB above background noise levels
 (Rechtschaffen  et al., 1966), as does  that for more complex motor responses  such as
 reaching for and pressing a button on the headboard of  the bed (Lukas and Kryter, 1970).
 (See Miller, 1971, for a review of these and other aspects of the problem.)

                      Stage of Sleep and Accumulated Sleep Time

 Sleep Stage.

     The EEC stages of sleep 1 through 4 in the Dement  and Kleitman (1957) classification
 were labelled  for their order of occurrence  after sleep onset, and for their apparent ordinal
 relation to threshold for arousal. Stage 1, characterized by loss of the alpha rhythm of quiet
 waking,  and stages 3 and 4 with their slow high voltage delta waves were classified as the
 "lightest" and deepest" states of sleep respectively. The  discovery  by Aserinsky and Kleit-
 man (1953) of the periodic state of REM sleep, associated with a low-voltage stage 1 EEC,
 complicated the situation because arousal  thresholds in  this stage were higher than in the
 stage 1 episodes found at the onset of sleep. In general,  however,  the likelihood of behav-
 ioral responding to neutral stimuli is a decreasing function of the amplitude and period of
 the  background EEC rhythms (Zund and Wilson, 1961; Williams et  al., 1964; Rechtschaffen
 et al.,  1966; Keefe et al., 1971). Although simple acoustic stimuli of moderate intensity can
 elicit specific  physiological responses in any stage of sleep, during  high-voltage delta sleep,
 the  likelihood of a specified, easily executed motor response is low. For example, Williams
 et al. (1966) found instrumental responding on a microswitch taped in the hand to tones at
 35 dB above sensation threshold only during low-voltage EEC stages 1, 2, and REM. More-
 over, Evans et al. (1966) were able to elicit relatively complex motor responses to verbal
 instructions only in stage REM. Although the reasons for this are not entirely understood,
 we can no longer conclude that  the stages of sleep,  1 through 4, define a universal con-
 tinuum either of depth of sleep or thresholds of arousal. When awakening is defined as full
 EEC arousal, the awakening thresholds in stages 2 through 4 and REM are apparently nearly
 identical (Keefe et al, 1971). The relative  loss of designated motor responses in the high-
 voltage stages  of sleep is not due to raised sensory thresholds, and probably not to failure of
 the signal analyzer system. Schicht et al (1968) found that discriminated classically condi-
 tioned cardiovascular responses  acquired during wakefulness  could be elicited regularly
 during extinction trials in stage 4 sleep. This finding, if confirmed, is evidence that signal
 analysis is still possible during stage 4 and that previous failures to observe discriminated
responding during that stage occurred only because stage 4  is relatively incompatible  with
the organization and execution of motor responses. Thus, Keefe and his colleagues (1971)
reached the tentative conclusion that impaired motor responding during high-voltage slow-
wave sleep was due to the disorientation and confusion which accompany sudden awakening
from that state rather than  to raised response thresholds. Broadly speaking, the human is
neither deafferented nor de-efferented during sleep.

                                        504

-------
Sleep Time.

     Several investigators have found that thresholds for awakening decreased as time asleep
increased  (e.g., Williams,  1966; Rechtschaffen et al., 1966; Watson and Rechtschaffen,
1969; Morgan and Rice, 1970; Keefe et al.,  1971). However, in all of these studies amount
of accumulated  sleep was confounded with chronological time. Thus, it is  not known
whether  the amount of  accumulated sleep or a circadian biological rhythm, relatively
independent of time asleep, is the principal factor in this effect. Williams (1966) proposed
the latter explanation on grounds that the curve of behavioral responsiveness over the night
is  reminiscent of the circadian  curve of body temperature  described  by Kleitman and
Ramsaroop  (1948).  However, Keefe et al.  (1971) reported  the same temporal trend in
awakening thresholds for both day and night sleepers. But the day sleepers had been on the
reversed sleep-waking schedule for at least seven days so that circadian biological cycles were
probably also reversed.
                            Motivation and Pre-Sleep State

     Studies summarized earlier in this review provide clear evidence that motivational and
incentive factors can influence the  probability of either physiological  or behavioral re-
sponses to noise. Sounds which are relevant to survival, or which, through conditioning, or
instructions, acquire signal properties  are more likely to arouse  the sleeper  than neutral
sounds. As Miller (1971) suggested, for weak stimuli, the effects of motivation depend on
the stage of sleep.  For example, Williams  et al. (1966)  found that as the motivation to
respond to  designated 35-dB tone stimuli  was enhanced by instructions and contingent
punishment, instrumental responding on a microswitch increased about  five-fold in low-
voltage stages  2 and REM, but very little  in high-voltage delta sleep. On the other hand,
Zung and Wilson (1961) showed that for moderately intense stimuli, instructions and finan-
cial incentives induced a marked increase in frequency of EEC arousal and  waking responses
in all stages of sleep.
     As in wakefulness, small differences in instructions given prior to sleep can have sub-
stantial effects on behavioral  and  physiological responding. For example,  the FAA (CAMI)
group in Oklahoma  City*, employing simulated sonic booms of 1.0 psf, did not label sounds
as sonic booms.  The  subjects  were told that  the investigators  were  interested  in sleep
behavior, moods, and performance; that noises might occur during the night (including the
presence of experimenters in the test room);  but  that the subject's task was to ignore
disturbances of any kind and get  the best night's sleep possible. The frequency of full EEC
arousal to hourly booms in this investigation was considerably less than that found by Lukas
and Kryter  (1970), even in  elderly subjects. The  latter investigators were more explicit
about the nature of the stimuli, and the response requirement of pressing a button attached
to the headboard of the bed.
*Personal communication from Dr. Wm. Collins. See also Collins' report in this Symposium.

                                        505

-------
     Other aspects  of the  pre-sleep  state alter the sleep EEC profile, and possibly the
subject's responsiveness to disturbing stimuli. For example, Lester et al (1967) reported that
a moderate increase of daytime stress, such as that occasioned by a college examination, was
associated with increased spontaneous arousal and inhibition of delta sleep. Jansen (1969)
cites evidence that emotional factors, stress and  neuroticism influence responsiveness to
noise in waking subjects. It is reasonable to predict similar positive relationships between
disturbed emotional states  and responsiveness to  noise during sleep. Indirect evidence for
such a relationship comes from studies showing that 64 hr of sleep deprivation caused a
systematic reduction in behavioral and physiological responsiveness to noise stimuli in all
stages of sleep (Williams et al, 1964). Keefe and his colleagues (1971) suggest that the higher
awakening thresholds found in their daytime sleepers may also have resulted from chronic
loss of sleep.

                                Individual Differences

     As  mentioned .earlier in this review,  responsiveness to noise during sleep varies in
relation to the age of the subject, sex, psychopathology and physical condition. The series
of studies by -Lukas and his co-workers used simulated sonic booms ranging in "outdoor"
intensities from  .06 to 5.0 psf, and recordings of subsonic jet flyovers, ranging in "outdoor"
intensity from  101-119 PNdB. They found  that children 5-8  years old  were  relatively
undisturbed by  either type oftooise,  whereas elderly men were much more disturbed than
younger subjects (Lukas and Kryter,  1970a  and 1970b). In general, this age effect was
confirmed by Collins' group, using simulated sonic booms with "outdoor" intensities of 1.0
psf. However, the average magnitude of boom effects was considerably less in Collins et al's
investigation than in  the studies by  Lukas et al.  (See Collins' report in this symposium.)
Possible reasons for this difference include differences in instructions, scheduling of subjects
and variation of the boom intensity parameter. Steinicke (1957)  reported that both the
elderly and people under thirty were more readily  awakened by noise than the middle-aged,
and that manual workers were  more susceptible  to noise awakening  than intellectual
workers. He concluded, incidentally, that the  noise in bedrooms should not exceed 35
dB(A).
     Although the sleep of small  children and normal infants (e.g., Gadeke et al, 1969) is
less disturbable  by acoustic  stimuli  that that of adults, babies subjected to gestational
difficulty or birth trauma may be hyperresponsive. Murphy (1969) on the basis of clinical
observation suggested that the short gestation, anoxic or brain-injured infant, in particular,
displays  exceptional responsiveness to sounds. Bench and Parker (1971), however,  in an
interesting application of signal detection theory, failed  to confirm this assertion. In fact,
their short-gestation babies tended to have higher awakening thresholds than full-term in-
fants.
    For neutral auditory  stimuli delivered during sleep, the threshold  for EEC  arousal
responses is lower in women than men (Steinicke, 1957; Wilson and Zung, 1966). Lukas and
Dobbs  (1972) found  similar greater sensitivity in middle-aged women to the sounds of
subsonic jet aircraft flyovers and simulated sonic booms. The women were particularly
responsive to  the sound of aircraft  flyovers. Wilson, and Zung (1966)  suggest that this

                                        506

-------
tendency toward hyperacuity in women may have adaptive significance for the mothering
role.
     There is evidence that EEC arousal thresholds differ for different types of psycho-
pathology.  For example, Kodman and Sparks (1963) reported that schizophrenic patients
showed a marked elevation of auditory sleep thresholds, whereas Zung et al. (1964) found
markedly reduced EEC arousal  thresholds in the depressive disorders. In fact the auditory
sensitivity during sleep in depressed males was greater on average than that found by this
same group in normal middle-aged females (Wilson and Zung,  1966). As had been men-
tioned, it is probable also that the sleep-disturbing effects of acoustic stimuli increase with
neuroticism. Monroe (1967) found that the sleep of neurotic subjects was grossly disturbed,
even in a quiet environment, and Jansen and Hoffmann (1965) reported that subjects high
in neuroticism were generally more sensitive to and disturbed by noise than normals.

                  Short-Term Habituation and  Long-Term Adaptation

     Whether short-term  habituation or long-term adaptation can occur during sleep is a
subject of  debate. For the EEC and autonomic responses which comprise the orienting
reflex,  Johnson's group in San Diego has found  no evidence of habituation over a few trials
or adaptation over many  nights  (e.g., Johnson and Lubin, 1967; Townsend et al, in press).
Similar findings are reported by Lukas and  Kryter  (1970) and Collins' group (this Sym-
posium) for simulated sonic booms,  and by Hutt et al. (1968) for EMG responses in the
human neonate.  Firth (1973) did find some habituation trends for autonomic and EEC
responses to short  runs of closely spaced 1000-Hz (70 dB) tones. However, this result, if
replicated, is more significant for theories of brain functioning during sleep than for applica-
tion.
     Anecdotal evidence suggests that the frequency and duration of behavioral awakening
or gross disturbances  of  sleep should show long-term adaptation. We are all familiar with
accounts of soldiers sleeping undisturbed in the presence  of artillery fire, or city  people
sleeping in the presence  of high levels of urban noise. Yet, so far, neither laboratory nor
field studies have produced unequivocal evidence of long-term adaptation. Lukas and his
associates  did  report some adaptation in college students, but only  to sonic booms of low
intensity (about 0.7 psf, "outside") and only in stage 2 sleep (Lukas,  1969). Townsend et al.
(in press) suggest that the relatively small effects on sleep found in their study of young men
exposed day and night to "pings" may be due to pre-sleep adaptation. However, as will be
reported by Dr.  Friedmann in this Symposium, the sleep of middle-aged couples who had
resided in  the vicinity of Los  Angeles International Airport for more than 5 years was
considerably disrupted by jet flyovers.

                              Summary and Conclusions

     During wakefulness, the presence of  raised  thresholds  for psychophysiological or
behavioral responding often permits an inference about the status of the sensorium.  During
sleep, however, interpretation of, the same finding requires more complex analysis.  Raised
thresholds  could be due  to alterations in sensory analyzer systems, in systems which link

                                        507

-------
sensory and motor processes, or in mechanisms which mediate the selection and execution of
psychophysiological or motor behaviors. Taken together, recent studies of responsiveness to
acoustic stimulation indicate that the states of sleep are not accompanied either by in-
creased thresholds for sensory transduction, diminished conduction in afferent pathways,
gross impairment of sensory analyzer systems or loss of sensory-motor links. Differentiated
and systematic EEC and cardiovascular responding are found during sleep either to condi-
tioned  or  biologically relevant stimuli whose intensities are  very  near waking sensation
thresholds. Thus, the sleeping human can encode, categorize and respond differentially to
near-threshold simple and complex acoustic stimuli. The increased thresholds for behavioral
responding often reported for sleeping subjects are probably due to the fact that some states
of sleep are generally not compatible with the selection and  execution  of certain motor
responses.
      Although the likelihood of behavioral responses to neutral acoustic stimuli is lower in
high-voltage than in low-voltage states of sleep, the notion that the stages of sleep, 2 through
4, represent a universal continuum of depth of sleep is no longer tenable. When awakening is
defined as full EEC arousal, stages 2 through 4 and REM are nearly identical. Moreover, the
relative impairment of motor responding found during high-voltage stage 4 sleep may be due
to the disorientation and confusion which accompany awakening from that state rather than
to raised response thresholds per se.
      The interpretation of stage-of-sleep data is further complicated by the fact that respon-
siveness varies with chronological time. Whether this is a function of amount of accumulated
sleep or the phase of  a circadian biological rhythm  is not known.  Finally, as in  other
psychophysiological studies, responsiveness during sleep is altered by subject variables such
as age, sex, instructions, motivation,  medical illness and psychopathology. Thus, it is not
surprising that investigators in the field have been unable to recommend uniform guidelines
for the regulation of noise in the environment.
                                    Bibliography

Anch. M., The auditory evoked brain response during adult human sleep. (Abstract) Paper
     delivered at the  11th annual meeting of The Association for the Psychophysiological
     Study of Sleep (1972).
Aserinsky, E., Kleitman, N., Regularly occurring periods of eye motility, and concomitant
     phenomena during sleep. Science, 118: 272-274(1953)
Bench, J. and Parker, A., Hyper-responsivity to sounds in the short-gestation baby. Develop.
     Med. Clin. NeuroL,  13: 15-19 (1971)
Buendia, N., Sierra, G., Goode, M. and Segundo, J. P., Conditioned and discriminatory
     responses in wakeful and sleeping cats. Electroencephal. Clin. NeurophysioL, Suppl.
     24,199(1964)
Davis, H., Davis, P. A., Loomis, A. L., Harvey, E. N. and Hobart, G. Electrical reactions of
     the human  brain to auditory stimulation during sleep. /. NeurophysioL,  2: 500-514
     (1939)

                                        508

-------
Dement, W. C, Kleitman, N., Cyclic variations in EEC during sleep and their relation to eye
    movements, body motility and  dreaming. Electroencephal. Clin. Neurophysiol,  9:
    673-690(1957)
Derbyshire, A. J. and McDermott, M., Further contributions to the EEC method of evalua-
    ting auditory  function. Laryngoscope International Conference on Audiology, 68:
    558-570(1958)
Evans, F. J., Gustafson, L. A., O'Connell, D. N., Orne, M. T., Shor, R. E., Response during
    sleep with intervening waking amnesia. Science, 152: 666-667 (1966)
Firth, H., Habituation during sleep. Psychophysiology, 10: 43-51 (1973)
Gadeke, R.,  Doling, F. K. and Vogel, A., The noise level in a children's hospital and the
    wake-up threshold in infants. A eta Paediat. Scand. 58:  164-170 (1969)
Granda, A. M. and Hammack, J. T., Operant behavior during sleep. Science, 133: 1485-1486
    (1961)
Hurt, C., von Bernuth, H.,  Lenard, H. G., Hutt,  S. J.,  Prechtl, H. F. R.,  Habituation  in
    relation to state in the human neonate. Nature, 220: 618-620 (1968)
Jansen, G., Effects of noise on physiological state. ASH A Reports No. 4 (1969)
Jansen, G. and Hoffmann, H., Larmbedingte Anderungen der Feinmotorik und Lastigkeit-
    sempfindungen in Abhangigkeit von bestimmten Personlichkeitsdimensionen. Z. Exper.
    Angew. Psychol,  12: 594-613 (1965)
Johnson, L.  C. and Lubin,  A., The orienting reflex  during waking and sleeping. Electro-
    encephal. Clin. Neurophysiol, 22: 11-21 (1967)
Keefe, F. B.,  Johnson, L. C. and Hunter, E. J., EEC and autonomic response pattern during
    waking and sleep stages. Psychophysiology,  8: 198-212(1971)
Kleitman, N.  and Ramsaroop,  A.,  Periodicity  in  body temperature and heart  rate.
    Endocrinology, 43:  1-20(1948)
Kodman, F.  and Sparks, C., The auditory sleep threshold of the catatonic schizophrenic./.
    Clin. Psychol, 19: 405 (1963)
Koella, W. P., The central nervous control of sleep.  In  W. Haymaker, E. Anderson, W. J.
    Nanta (Eds.) The Hypothalamus,  Springfield: Charles C. Thomas. Pp. 622-644
Lehmann, D. and Koukkou, M. Das EEC des Menschon beim lernen von neuem und bekann-
    tem material. Arch. Psychiat. Nervenkr,, 215:  22-32  (1971)
Lester, B. K., Burch, N. R. and Dossett, R. C., Nocturnal EEG-GSR profiles and influence  of
    presleep states. Psychophysiology, 3: 238-248 (1967)
Loomis, A. L., Harvey,  E. N., Hobart, G. A.,  Cerebral states  during sleep as studied by
    human brain potentials. /. Exp. Psychol., 21: 127-144 (1937)
Lukas,  J. S.,  The effects of simulated sonic booms and jet flyover noise on human sleep.
    Proceedings of the  Sixth Congress on Environmental Health, Chicago, A.M.A., April
    (1969)
Lukas, J. S., and Kryter, K. D., Awakening effects of simulated sonic booms and subsonic
    aircraft noise. In  Physiological Effects of Noise, B. L. Welch and A. S. Welch (Eds.)
    Plenum Press, New York, pp 283-293 (1970a)
Lukas, J. S. and Kryter, K. D., Awakening effects of simulated sonic aircraft noise on six
    subjects  7 to 72 years of age. NASA  Report No. CR-1599, Washington, D.C. (1970b)
                                       509

-------
Lukas, J.  S. and Dobbs, M. E., Effects of aircraft noises on the sleep of women. NASA
     Contractor Report CR-2041 (1972)
Miller,  J.  D.,  Effects  of Noise on  People. U. S.  Environmental Protection  Agency,
     NT1D300.7(1971)
Monroe, L. J. Psychological and physiological differences between good and poor sleepers.
     /. Abnorm. PsychoL 72: 255 (1967)
Morgan, P. A.  and Rice, C. G., Behavioral awakening in response to indoor sonic booms.
     Institute of Sound and Vibration Research,  University of Southampton, Technical
     Report No. 41(1970)
Murphy, K. P., Differential diagnosis of impaired hearing in Children. Develop. Med.  Child.
     NeuroL, 11:561(1969)
Oswald, I., Taylor, A. M. and Triesman, M., Discriminative responses to stimulation during
     human sleep. Brain, 83: 440-553 (1960)
Pearsons,  K. S., Fidell, S., Globus, G., Friediiiann, J. and Cohen, H. The effects of aircraft
     noise on sleep electrophysiology as recorded hi the home. Bolt, Beranek and Newman
     Report 2422 (1973)
Rechteschaffen, A., Hauri, P. and Zeitlin, M., Auditory awakening thresholds in REM and
     NREM sleep stages. Perceptual and Motor Skills, 22: 927-942 (1966)
Rechtschaffen,  A. and Kale, A. (Eds.): A Manual of Standardized Terminology, Techniques
     and Scoring System for Sleep Stages of Human Subjects.  Public Health Series, U.S.
     Government Printing Office, Washington, D.C. (1968)
Schicht, W. W., McDonald, D. C.  and Shallenberger,  H. D., Progression of conditioned
     discrimination from waking to sleeping. (Abstract) Psychophysiology, 4: 508 (1968)
Schieber,  J. P., Mery, J. and Muzet, A., "Etude Analytique en  Laboratoire de Plnfluence
     due Bruit  sur Le Sommiel," (Rept. of Centre d'Etudes Bioclimatique du CNRS, Stras-
     bourg, France), (1968)
Scott, T. D., The effects of continuous, high intensity white noise on the human sleep cycle.
     Psychophysiology, 9: 227-232(1972)
Steinicke, G., Die Wirkungen von Larm auf den Schlaf des Menschen. Forschungsberichte
     des Wirtschaftes- u. Verkehrsministerium Nordrhein-Westfalen No. 416 (1957)
Steinicke, G., Cited in Urban Traffic Noise:  Status of Research and Legislation in Different
     Countries.   Draft report on  the Consultative  Group on  Transportation Research,
     DAS/C51/68.47 Revised; Organization  for Economic Cooperation  and Development,
     Paris, France, March (1969)
Thiessen, G. J., Effects of noise during sleep. In Physiological Effects of Noise. B. L. Welch
     and A. S. Welch (eds.) New York: Plenum Press, pp. 271-275 (1970)
Thiessen, G. J.  and Olson, N. Community noise-surface transportation. Sound and  Vibra-
     tion,  2:10-16 (1968)
Townsend, R. E., Johnson, L. C. and Muzet, A. Effects of long term exposure to tone pulse
     noise on human sleep. Psychophysiology. (In press)
Vetter, K.  and  Horvath, S. M. Effect of audiometric parameters on K-complex of electro-
     encephalogram. Psychiat. NeuroL, Basel, 144: 103-109(1962)
Watson, R. and Rechtschaffen, A. Auditory awakening thresholds and dream  recall in
     NREM sleep. Perceptual and Motor Skills, 29:635-644 (1969)

                                       510

-------
Williams, H. L., Hammack, J. R., Daly, R. L., Dement, W. C. and Lubin, A. Responses to
     auditory stimulation, sleep loss and the EEC stages of sleep. Electroencephal.  din.
     Neurophysiol,  16: 269-279(1964)
Williams, H. L. The problem of defining depth of sleep. In S. S. Kety, E. V. Evarts, H. L.
     Williams (Eds.), Sleep and Altered States of Consciousness. Baltimore, Md.: Williams
     and Wilkins, pp. 277-287 (1966)
Williams, H. L., Morlock, H.  C. and Morlock, J.  J. Instrumental  behavior during sleep.
     Psychophysiology, 2:208-215 (1966)
Wilson, W. P. and Zung, W. W. K. Attention, discrimination and arousal during sleep. Arch.
     gen. Psychiat., 15:523-528(1966)
Zung, W. W. K. and Wilson, W. P. Response to auditory stimulation during sleep. Arch. gen.
     Psychiat., 4:548-552 (1961)
Zung, W. W. K., Wilson, W. P. and Dodson, W. E. Effect of depressive  disorders on sleep
     EEC  responses. Arch. gen. Psychiat., 10:439-445  (1964)
                                        511

-------
              PREDICTING THE RESPONSE TO NOISE DURING SLEEP
                                  Jerome S. Lukas
                             Stanford Research Institute
                            Menlo Park, California  94025

                                  INTRODUCTION

     Recently, auditory stimuli  with very different spectra  have  been used to study  the
effects of noise on human sleep. In most of these studies the physical characteristics of the
stimuli were  described only  partially,  making questionable  direct comparisons of results
obtained in the  several laboratories. To provide a technique for comparing stimuli and to
estimate  relative sensitivities  of  subjects, a burst of pink noise was recommended (Rice,
1972) for use in laboratories, and tape recordings of that noise were distributed (Lukas) to
some.
     This paper  describes  a study  that used the recommended pink noise burst as one of
three different stimuli, and that correlated several physical  descriptors of the noise, with
different measures of response to those noises.

I     METHOD
     A. Procedure
     The SRI sleep laboratory consists of two identical, acoustically isolated rooms in which
four subjects, divided into two pairs, are tested simultaneously. A test period for one pair of
subjects was considered a control period for the other pair. Typically, test periods alternate
with control periods in each room. In any given room, stimuli are presented randomly with
respect to sequence, intensity, and interval between stimuli, but any two stimuli are  not
presented at  intervals of less than twenty minutes. On the average, stimuli occurred once
every forty minutes.
     The first stimulus on any test night was presented only after both subjects in the room
are in sleep stage 2, at least, or about one hour after the subjects went to bed.
     For any subject the procedure included, first, three accommodation nights in  the
laboratory, next two nights at home, then fourteen consecutive  nights in the laboratory.
The  first two nights of the fourteen, as well as nights 9, 10, and 14, were considered control
nights, during which the subjects were permitted undisturbed sleep. Stimuli were presented
during the remaining nine test nights.
     The subjects were instructed  to sleep as normally  as possible, but to push a button
attached to the headboard of each bed (the "awake switch") if they should awaken for  any
reason. They were never told when stimuli would  be presented  or how many had been
presented.
     Both electroencephalographic (EEC)  and  behavioral responses  to  the noises were
scored. The behavioral response was reserved exclusively for  the use of an "awake" switch
attached to the  headboard of the bed, while the EEC (central, C3, with reference to  the
contralateral  mastoid, A2) responses were scored on the basis of the criteria presented in
Table 1.

                                         513

-------
                                      Table 1

          CRITERIA FOR SCORING THE ELECTROENCEPHALOGRAMS VISUALLY
  Score
                          Response Required
    0
No change in EEC. This category also includes "K complexes," brief
bursts of Alpha (about 10 Hz activity), spindles, and eye movements, as
appropriate for the subject's sleep stage.*

 Sleep stage change of one or two steps, but without arousal. The change
 must occur within 30 s of stimulation and continue for at least an addi-
 tional 40 s.

 Arousal of at least 10 s duration, but without use of the "awake" switch.
 Typically such a record shows brief bursts of Alpha, 10 or more s  of
 low-amplitude Beta (20-40 Hz) activity, and gross body movements.

 Awake response, in which the subject, after arousal, will move about and
 use the "awake" switch. Usually the response occurs within one minute of
 stimulus termination.
*"K complexes," Alpha, spindles, and eye movements occur normally in the EEC in some
sleep stages. If such activity were scored as a response, the subjects in those stages would
appear to be overly sensitive to stimulation as compared to stages in which the activity does
not normally occur.
     B. Stimuli
     The three stimuli were (1) landing noise from a DC-8 without acoustically treated
engine nacelles, (2) landing noise from a DC-8 with acoustically treated nacelles (Langdon et
al., 1970), and (3) a burst of pink noise. The aircraft noises were originally recorded out of
doors, but for the purposes of the study were shaped to simulate noises as they would  be
heard indoors. The time-courses of the stimuli heard by the subjects are illustrated in Fig. 1,
and various physical descriptions of the noises are presented in Table 2.
     It is important to note,  in Table 2, that although the stimuli had nearly identical
nominal intensities (79 or 61 dBA maximum), as progressively more information about their
physical characteristics was added into the descriptors, the stimuli became relatively more or
less severe ("noisy"). For example, adding the tone correction to EPNdB (compare columns
EPNdB and EPNdBT) makes the noise from the jet without treated nacelles about 2 dB
more "noisy" than the jet with treated nacelles, and both these noises are at least 4 dB more
noisy than the pink noise, although most (about 3.5 dB) of this 4 dB difference is due to the
                                       ;514

-------
      <•)  DC-8 WITH UNTREATED ENGINE NACELLES LANDING AT  159.1 m I522 ft)  ALTITUDE
       Ibl  DC-8 WITH TREATED ENGINE NACELLES LANDING AT 154.4 m (507 ft) ALTITUDE
                                    (cl  PINK NOISE  BURST
Figure 1:   Time histories of the three test stimuli measured in a test room near the subject's ears.
                                             5I5

-------
                                                           Table 2

                                          PHYSICAL DESCRIPTIONS OF THE STIMULI



Stimulus
DC-8 with
untreated
nacelles
DC-8 with
treated
nacelles

Pink noise
Nominal
Overall
Duration
(s)

|30
30

(30
1 30
1 4
\
1 4

Rise Time
to Peak
(s)

16.5
16.0

19.0
18.0
1.0
1.0
Duration
to 10 dB
Down Points
from Max dBA

7.5
7.5

9.0
10.5
3.5
3.25

Nominal
Level
dBA

79
>- 61

79
61
79
61


Max
dBA

78.9
61.1

78.4
60.4
78.0
59.9


Peak
dBA

79.8
62.0

79.5
61.5
78.3
60.3


Max
dBD2

83.8
66.2

82.6
64.6
81.9
63.7


Max
PNdB

91.5
73.8

90.0
71.5
89.7
71,9


EPNdB
(1)

85.0
66.7

84.8
66.3
81.3
63.5


EPNdBT
(2)

87.5
69.3

85.6
67.1
81.7
63.8


EPNdBTM
(3)

88.6
70.2

87.2
68.7
83.1
65.4


EPNdBTM-ic
(4)

88.6
70.2

87.2
68.7
95.1
70.4
ON
    Definitions  and techniques for calculating the various physical units may be found in Kryter (19708, 1970b).

    (1)  Calculated using 15 s as a reference duration.
    (2)  Tone correction recommended by Kryter (1970a, 1970b).
    (3)  Modified to account for the critical bandwidth of the ear at frequencies below 355 Hz (Kryter, 1970a, 1970b)
    (4)  An impulse correction applied to Col. 5 to account for a rise of some 32 dB above background noise level
        (about 35 dBA) in the first 0.5 s of the "high" intensity pink noise burst, and a rise erf some 14 dB above
        background level in 0.5 s by the "low" intensity pink noise burst.

-------
differences in duration of the stimuli (see the columns labeled EPNdB and Duration to 10
dB ... dBA). The pink noise typically is less noisy than are the two jet noises; however, if
account (Lukas, Peeler, and Dobbs, 1973) is taken of its impulse characteristics (see Column
EPNdBTM-ic), the pink noise becomes 6 to 8 dB more noisy than the nominally high level
jet noises, and approximately equivalent to the jet noises at nominally low levels (61 dBA).

     C. Subjects
     Four middle-aged (46 to 58 years) males were studied. Three of the four had been
subjects in a previous study; thus they had slept in the laboratory and were familiar with the
procedures  used. All thought themselves  to  be reasonably normal sleepers, without any
particular bias for or against aircraft noise, and none of the subjects lived near or in the
flight paths to the local airports. Audiometry showed their hearing was within normal limits.
H     RESULTS
     Typically, during the control trials the subjects did not show any response (Response
0), although a few spontaneous changes in sleep stage (Response 1) occurred. As shown in
Table 3, in only one case was an arousal (Response 2) observed, and the awake switch was
not used during the  control trials.  It may be concluded that,  in the main, responses  to
stimuli rather than spontaneous changes during sleep are discussed below.

                                       Table 3

                  RESPONSE FREQUENCIES DURING CONTROL TRIALS
                          (Numbers in parentheses are percentages)
Test
Room
Number
1

2

Number of
Control
Trials
158

162



0
152
(96.2)
158
(97.5)


1
5
(3.2)
4
(2.5)


2
1
(.6)
0



3
0

0

Number
of Test
Trials
166*

159

     *During some test trials the control subjects may have been still awake from their previous test
     trial or were moving before and during stimulus occurrence. Such instances were not counted as
     control trials. Hence, the numbers of test and control trials are not equal.

     An earlier study (Lukas, Dobbs, and Kryter, 1971) suggested that response frequencies
 to  a noise during sleep are normally  distributed, particularly when  about eight or more
 subjects are studied. Since three of the four subjects were studied previously, it is reasonable
 to  assume that their responses also are within those normal limits. Consequently, the re-
 sponse frequencies of the four subjects are combined in the data presented below.
                                         517

-------
     A. Effects of Stimulus Intensity
     For each of the three stimuli a change of 18 dBA in stimulus intensity decreased the
frequency of 0 responses and increased the frequencies of 2 and 3 responses; but, as shown
in Table 4, the  magnitude  of these  changes was not  similar for the three stimuli. For
example, with the pink noise an increase of 18 dBA resulted in an increase of approximately
44 percentage points in the frequency of 3 responses, while in response to the treated jet
noises an increase  of only  7 percentage points  was observed.  Somewhat  less disparate
changes were observed with  respect  to the other responses for nominally equivalent
variations in stimulus intensity.
     In Table 5, the data are reorganized  to facilitate comparison  of the response fre-
quencies to the  three stimuli when they are of nominally equivalent intensity.  It should be
noted that, at 61 dBA, pink noise resulted in the lowest frequency  (6.3  percent) of be-
havioral awakenings (Response 3), while the untreated jet noise had the highest (about  24
percent), and no great differences were observed  in the frequency of 0 responses. At the
higher stimulus  intensity, however, the frequencies of behavioral awakening were about
                                        Table 4

                       RESPONSE FREQUENCIES TO THREE STIMULI
                               EACH AT TWO INTENSITIES
                            (Numbers in parentheses are percentages)
Stimulus
Untreated Jet
Treated Jet
Pink Noise
Nominal
Intensity
(dBA)
79
61
79
61
79
61
Responses
0
16
(29.1)
30
(65.2)
15
(44.0)
51
(70.8)
11
(15.7)
30
(62.5)
1
3
(5.5)
4
(8.7)
6
(17-6)
6
(8.3)
7
(10.0)
6
(12.5)
2
9
(16.4)
1
(2.2)
4
01.7)
3
(4.2)
17
(24.3)
9
(18.7)
3
27
(49.0)
11
(23.9)
9
(26.5)
12
(16.7)
35
(50.0)
3
(63)
X2
16.37*
7.5 6t
f
35.36*
       *3 df (degrees of freedom); p < 0.001.
       t3 df, 0.10 > p > 0.05, not significant.
       *3df,p<  .001.
                                         518

-------
                                         Tables

                   RESPONSE FREQUENCIES TO STIMULI OF NOMINALLY
                                 EQUIVALENT INTENSITY
                           (Numbers in parentheses are percentages)
Nominal
Intensity
(dBA)
79
61
Stimulus
Untreated Jet
Treated Jet
Pink Noise
Untreated Jet
Treated Jet
Pink Noise
Response
0
16
(29.1)
15
(44.0)
11
(15.7)
30
(65.2)
51
(70.8)
30
(62.5)
1
3
(5.5)
6
(17.6)
7
(10.0)
4
(8.7)
6
(8.3)
,6
(12.5)
2
9
(16.4)
4
(11.4)
17
(24.3)
1
(2.2)
3
(4.2)
9
(18.7)
3
27
(49.0)
9
(26.5)
35
(50.0)
11
(23.9)
12
(16.7)
3
(6.3)
X^
15.71*
16.08t
      *6 df, 0.02 >p> 0.01.
      t6 df, 0.02 >p> 0.01.
equal (about 50 percent) for the pink and untreated jet noises, and the pink noise resulted
in fewer (about 14 percentage points) 0 responses than did the untreated jet noise.
     These data indicate quite clearly that the jet noises  emanating from aircraft with
nacelle treatment are less disturbing than those from jets without the nacelle treatment. At
high noise levels, for example,  treated jet noise was associated  with an higher incidence of
no changes (0 response) in the EEC, and a lower incidence of behavioral awakenings and
arousals (Responses 3 and 2, respectively) than was the case with untreated jet noise. An
essentially similar result was observed when the two jet noises  occurred at the lower inten-
sity. The  practical importance of these results is illustrated  in Figure  2,  in  which the
frequency of 0 responses (no disruption of sleep) is plotted against the nominal intensities
of the treated and untreated jet aircraft noise. It will be seen that the treated jet noise at 75
dBA disrupted sleep to about the same degree as did the untreated jet of about 68.5 dBA.

     B.  Adaptation
     The  trend of the data, shown in Table 6, suggests that some  adaptation to the  noise
from the  treated jet occurred during the first six nights  of tests,  but adaptation to the
untreated  jets did not occur. If anything, the subjects were awakened more frequently by
the untreated jet noises as the experiment progressed into its final phases. In Table 6, it may
                                         519

-------
  100
CO
uj
CO

O
Q.
W
UJ
EC
UJ
O
<
z
UJ
O
cc
UJ
a.
   50
40 —
   30  —
   20  —
   10  —
    0
     60
                                              75
80
                                         68.5  70
                                       INTENSITY IN dBA
           NOTE:  Intensities for equivalent sleep disruption by the two stimuli
                  are indicated on the abscissa.

 Fig. 2 Frequency of no  sleep disruption responses to DC-8 jet aircraft with and without acoustically
 treated engine nacelles.

 be seen that in the case of the untreated jet an increase of about 4 percentage points in the
 frequency of behavioral awakening occurred between test nights 1, 2, and 3 and nights 4, 5,
 and 6, but an increase of about 8 percentage points in the frequency of no EEC change
 occurred over the same period. However, after two nights in the quiet, test nights 7, 8, and
 9, the frequency  of behavioral awakenings  (and arousals) increased about  12 percentage
points over the frequency observed during test nights 1, 2, and 3, and the frequency of 0
responses decreased  below the percentage observed during the first three test  nights. In
contrast, noise from jets with treated nacelles resulted in a reduction of about 17 percentage
points in the frequency of behavioral awakening and an increase of about 20 points in the
                               520

-------
                          Table 6

RESPONSE FREQUENCIES TO THREE STIMULI DURING COMBINATIONS
            OF TEST NIGHTS INDICATING ADAPTION
Stimulus
Untreated Jet
Test
Nights
1, 2, 3
4, 5, 6
7, 8, 9
Responses
0
14
(42.4)
19
(54.3)
12
(36.4)
1
5
(15.2)
0
2
(6.1)
2
3
(9.D
3
(8.6)
4
(12.1)
3
11
(33.3)
13
(37.1)
15
(45.5)
  Nights 1,  2,  3 versus 4, 5, 6 - X2 = 5.870, 3 df, not
    significant.
  Nights 1,  2,  3 versus 7, 8, 9 - X2 = 2.199, 3 df, not
    significant.
Treated Jet
1, 2, 3
4, 5, 6
7, 8, 9
14
(60.9)
25
(80.6)
18
(48.6)
2
(8.7)
3
(9.7)
7
(18.9)
1
(4.3)
0
6
(16.2)
6
(26.1)
3
(9.7)
6
(16.2)
  Nights  1,  2,  3 versus 4,  5,  6 - TT = 3.799,  2 df
    (Responses  2 and 3 combined), not significant.
  Nights  1,  2,  3 versus 7,  8,  9 - X2 = 3.789,  3 df, not
    significant.
Pink Noise
1, 2, 3
4, 5, 6
7, 8, 9
12
(28.6)
15
(38.5)
13
(36.1)
2
(4.8)
6
(15.4)
6
(16.7)
12
(28.6)
5
(12.8)
7
(19.4)
16
(38.1)
13
(33.3)
10
(27.8)
  Nights 1,  2,  3 versus 4, 5, 6
    significant.
  Nights 1,  2,  3 versus 7, 8, 9
    significant.
- X2 = 5.422, 3 df, not

- X2 = 4.304, 3 df, not
                           521

-------
frequency of no EEC changes when nights 4, 5, and 6 are compared with nights 1, 2, and 3.
However, in this case, after sleeping two nights fii the quiet, the subjects were still awakened
less frequently than they were during nights 1, 3, and 3, but more frequently (about 10
percentage points) than during test nights 4^ 5, and 6; and itha frequency of no changes in
the EEG was reduced about  12 percentage points below the 60.9 percent frequency ob-
served during nights 1, 2, and 3.

     C.  Response Prediction
     It is of interest to estimate how the different physical measures of noise, which have
been found useful and accurate in predicting annoyance in the awake subject, may predict
sleep disturbance or awakening. To this end, various physical measures of the stimuli used in
the present study  were correlated (Pearson's product-moment coefficient) with  the per-
centage  of responses in each of several response categories. On the basis of those preliminary
correlations, physical measures that appeared  ty add to the magnitude of the coefficient
were correlated with appropriate response data from our earlier studies (Lukas, Dobbs, and
Kryter,  1971; Lukas and Dobbs, 1972) and some other studies (Thiessen, 1970; Berry and
Thiessen, 1970; Morgan and Rice, 1970; Ludlow and Morgan, 1972; Collins and lampietro,
 1972) in which the stimuli were reasonably well described and in which the responses were
similar or identical to those used herein.
     The.correlation coefficients obtained with the three groupings of data are presented in
Table 7. Before proceeding it must be noted that the coefficients, because of sample size,
are considered to be only preliminary estimates and indicative of a possibly  fruitful ap-
proach to predicting the effects of noise on sleep.
     The coefficients shown in the upper third  of Table 7, which were obtained using data
from the most recent study,  suggest, that Max  dBA and PNdB were somewhat better pre-
dictors of the effects on sleep than were EPNdB> EPNdBT, and EPNdBTM, but  that the
addition of the impulse correction (Kryter, '1970a, 1970b) resulted in the highest correla-
tions. However, when data from several studies were 'combined (middle and lower thirds of
Table 7), EPNdB (and presumably also EdBA)  was found to be a slightly better predictor
than Max PNdB,  or Max dBA, but the apparent value of the impulse correction diminished.
     With respect to  the type of  response that correlated most highly with the different
physical descriptors, the results suggest that the percentage of 0 responses (no change in the
EEG) or, conversely, the percentage  of an EEG response of at least one sleep stage, is more
highly correlated with the physical descriptors than is behavioral awakening (Response 3) or
awakening and arousal (Responses 3 and 2).
  'Also worthy of note is  tiie  lack of consjlsteritly higher or lower correlations with
Response 3 versus  the combination  of responses 2 and 3. This result) in addition to the
generally small differences in  magnitude between  the  coefficients, is consistent with sub-
jective reports that on  occasion the subjects did not use the awake switch (Response 3)
because they"were "too5 tired," or "didn't have the Energy to turn over."
     Finally, it is_ clear that as  data' from other laboratories (at least as the physical descript-
ors were estimated for purposes; of this analysis) were included in the calculations, the
correlations between inost of the response measures and tne physical descriptors decreased.
However, the increase in the coefficient  between EPNdlB and the 0 response when the 0

                                        522

-------
                                         Table?
                COEFFICIENTS OF CORRELATION BETWEEN RESPONSES TO
          NOISE DURING SLEEP AND VARIOUS PHYSICAL MEASURES OF THE NOISE


Response
%0
%3
%2&3


%0
%3
%2&3
%0
%0,1&2§
%3
Physical Measure
i
Max
dB
-.90
.83
.85
Max
dBA
-.79
.64
.61
-.64
-.62
.53
Max
PNdB
-.90
.84
.85









EPNdB
-.84
.81
.79

EPNdB
-.81
.77
.75
-.67
-.78
.67

EPNdB
-.82
.82
,78









EPNdBTM
-.82
.82
.78









EPNdBTM-ic
-.97
.86
.93

EPNdB-ic
-.69
.50
.50
-.60
-.64
.49
Number
of
Data
Points
6*
6
6


20t
20
20
37*
37
39
   'Includes data only from study reported herein.
   tData from this study, plus that from middle-aged men and women (Lukas et al., 1971, Lukas et al., 1972).
    Stimuli were other aircraft noises and simulated sonic booms at several intensities each.
   *Data from Thiessen (1970), Berry and Thiessen (1970), Morgan and Rice (1970), Ludlow and Morgan
    (1972), Collins and lampietro (1972) added to the above, EPNdB and EPNdB-ic are estimates.
   §Responses 0 and 1 combined from Thiessen (1970) and Berry and Thiessen (1970) and Responses 0,1,
    and 2  combined from Collins and lampietro (1972) since by their definitions responses 0, 1, and 2 are
    identical to our Response 0.
response included  complexes and  bursts of alpha that were  observed in the studies of
Thiessen (1970), Berry and Thiessen (1970), and Collins and lampietro (1972), suggests that
EPNdB and presumably EdBA may continue to be a reasonably good  predictor of general
sleep disturbance, and perhaps better than Max PNdB or dBA.

ni  DISCUSSION
     That  sleep was less disrupted by noise from jet aircraft with acoustically treated na-
celles than by aircraft without acoustical treatment was demonstrated in this study. This
result is consistent generally with the magnitude of  the physical  descriptor (EPNdBT),
which takes into account the "pure tone" characteristics of the untreated jet noise. Since
these tones are suppressed by nacelle treatment, the magnitude of EPNdBT for the treated
jet is less  than  that for the untreated jet. The result is also consistent, generally, with the
                                          523

-------
finding that the treated jet  noise was less annoying to the awake subject than  was the
untreated jet noise (Langdon et al., 1970).
     The magnitude of the correlation between the responses and the physical descriptors
was less than might be hoped for. Although the various physical measures used in this study
have been found to be correlated to a greater or lesser degree with subjective annoyance in
awake subjects, the general diminution of the coefficients as  sample size increased suggests
that  the commonly used descriptors of noise may not be appropriate for prediction of its
effects on sleep. On the other hand, the stimuli used in the various studies were generally
inadequately described for our present purposes, and hence the estimates we used may have
included some  error. In addition, response data from other than middle-aged subjects were
used, although it has been demonstrated (Lukas and Dobbs, 1972; Lukas, 1972) that re-
sponse frequencies vary as a function of subject age. Despite these possible errors, certain of
the correlation coefficients remained sufficiently high to suggest that  the correlational ap-
proach may have value. However, in future studies the stimuli must be described in greater
detail than has been the case heretofore.

IV   CONCLUDING REMARKS
     Because of the small number of subjects and types of noises studied, the conclusions,
presented below, should be considered  tentative.
     1.   For equivalent sleep disruption (ie., a change of at least one sleep stage)the level
         of noise from the untreated jet engine must be about 6 PNdB or 6 dBA less than
         the noise from the jet with treated engine nacelles.  Since  the jet with treated
         nacelles is about 10 dBA (Langdon  et al.,  1970) less intense than the jet without
         nacelle treatment performing the same maneuver, the treated jet can be expected
         to result in much less (perhaps less than &) sleep disturbance than the untreated
         jet
     2.   Predictions of the effects of noise on sleep appear to attain their highest accuracy
         when the physical  descriptor of that noise includes information about its more
         long-term spectral content

V    ACKNOWLEDGMENT
     The work reported herein was accomplished under Contract NAS1-11243 between the
National Aeronautics and Space Administration and Stanford Research  Institute.
                                  REFERENCES

Berry, B., and G. J. Thiessen, The Effects of Impulsive Noise on Sleep, National Research
    Council of Canada, Publication No. N.R.C. No. 11597 (1970).
Collins, W. E., and P. F. lampietro, Simulated Sonic Booms and Sleep: Effects of Repeated
    Booms  of  1.0  psf. Office of Aviation Medicine Report No.  OAM-72-35, Federal
    Aviation Administration, U.S. Department of Transportation (1972).
Kryter, K. D., Possible Modifications  to the  Calculation of Perceived Noisiness, NASA
    Report No. CR-1636 (August 1970a).

                                       524

-------
Kryter, K. D., The Effects of Noise on Man (Academic Press, New York, 1970b).
Langdon, L. E., R. F. Gabriel, and A. H. Marsh, Investigation of DC-8 Nacelle Modifications
    to Reduce Fan-Compressor Noise in Airport Communities,  Part IV-Psychoacoustic
    Evaluation, NASA Report No. CR-1710 (1970).
Ludlow, J. E., and P. A. Morgan, Behavioral Awakening and Subjective Reactions to Indoor
    Sonic Booms, J. Sound and Vibration, 25, pp. 479-495 (1972).
Lukas, J. S., M. E. Dobbs,  and K. D. Kryter, Disturbance of Human Sleep by Subsonic Jet
    Aircraft Noise and Simulated Sonic Booms, NASA Report No. CR-1780 (July 1971).
Lukas, J.  S., and  M. E. Dobbs, Effects of Aircraft Noises on the Sleep of Women, NASA
    Report No. CR-2041 (June 1972).
Lukas, J. S., Awakening Effects of Simulated Sonic Booms and Aircraft Noise on Men and
    Women, J. Sound and  Vibration, 20, pp. 457-466 (1972).
[for a fuller account of the rationale underlying use of the impulse correction  in this case
    see]  J. S. Lukas, D. J. Peeler, and M. E.  Dobbs, Arousal from Sleep by Noises from
    Aircraft  With  and Without Acoustically  Treated  Nacelles, Final Report, NASA
    Contract No.  NAS1-11243, prepared by Stanford Research Institute (January 1973).
Lukas, J. S., A recording of the standard noise and a complete list of its physical descriptors
    is available  from J. S. Lukas,  Stanford  Research  Institute, Menlo Park, California
    94025, USA. It may be noted that the standard noise burst provided is different from
    that recommended at the sonic boom symposium, because upon test the recommended
    noise was found to be too short to be reliably reproduced or to elicit reliable responses
    from test subjects. The recommended noise was therefore changed from a  one-half
    second burst to a four second burst.
Morgan, P. A., and C.  G. Rice, Behavioral Awakening in Response to Indoor Sonic Booms,
    Tech. Report  No.  41, Institute of Sound and Vibration Research,  University of
    Southampton, England (1970).
Rice,  C. G., Sonic Boom Exposure Effects II. 2: Sleep Effects, J. Sound and Vibration, 20,
    pp. 511-515(1972).
Thiessen, G. J., Effects of Noise During Sleep,  in B. L. and A. S. Welsh (eds.) Physiological
    Effects of Noise, Plenum Press, New York, pp. 271-275 (1970).
                                       525

-------
  THE EFFECTS OF NOISE-DISTURBED SLEEP ON SUBSEQUENT PERFORMANCE

                           M. Herbert1 and R. T. Wilkinson
                              Medical Research Council,
                              Applied Psychology Unit,
                                Cambridge, England.

Summary
     To investigate the performance effects of sleeping in a noisy environment, 10 Ss slept
for 5 nights each, during one of which sleep was disturbed by playing pairs of randomised
clicks at one of four intensities of 65, 75, 80 and 90 dBA through a speaker in the bedroom.
Performance tests, lasting all day,  consisted of alternations of the  1-hour Wilkinson Vigi-
lance and  Adding tests and  a 30-min short-term  memory  test. The sleep profile was
examined in terms of (a) sleep stages and (b) REM cycle rhythmicity.

Effects of noise:
     1.   Increase in Stage 1 and time spent awake.
     2.   An insignificant tendency for REM and SWS to be reduced.
     3.   No uniform effect on rhythmicity.
     4.   Fewer responses of any kind in the first vigilance test and relatively fewer sums
         done in the middle of the first adding test.

Correlational Analysis:  The effect of nighttime  noise on performance during the day corre-
lated with  REM cycle  rhythmicity  during sleep, but not with total sleep time or minutes
spent in REM or SWS.

Conclusion:
     The effects upon performance of noise-disturbed sleep were relatively small and con-
fined to the early part of the day. Significant correlations of REM cycle rhythmicity with
various performance indices suggests the importance of the regularity of the sleep profile for
the diurnal  cycle in performance and recovery from the effects of disturbed sleep.

Introduction
     Much  recent research into the effects of sleeping in a noisy environment has concen-
trated upon changes  in  the electroencephalogram (EEC) manifested by either a change in
the sleep state towards  a 'lighter' sleep, or by awakenings. (Lukasand Kryter, 1970;Thies-
sen, 1970; Lukas, 1972; Lukas and Dobbs, 1972). For various reasons (Morgan, 1970) other
authors  prefer the criterion  of behavioural arousal in  which the S has to press a switch
placed close to the bed when he wakes for any reason (Ludlow and Morgan, 1972; Rylander
etal., 1972).
    1This work was carried out while the author was in receipt of an MRC Scholarship, which is gratefully
acknowledged.

                                        527

-------
     The common assumption upon which most  of this work appears to be predicated is
that such periods of temporary EEC desynchronisation or awakenings are detrimental to the
sleeper either because they induce some degree of sleep loss or because they could interfere
with the proportions of the two 'major' types of sleep for which there appears to be a need,
i.e. Slow Wave Sleep (SWS) and REM sleep (Webb 1969).
     The evidence, however, from studies of partial sleep deprivation suggests that, at least
for young, male adults, performance is not affected until sleep is reduced  to around the
2-hour mark on a single night, or 5 hours on two successive nights (Wilkinson, Edwards and
Haines 1966; Wilkinson 1969; 1970). So unless the noise results in quite severe sleep loss it
seems  that noise-disturbed sleep, on these grounds, would have little effect.
     Selective deprivation studies in which Ss have been deprived of SWS or REM sleep are
equally inconclusive. In a recent report, Chernik (1972) could  find no effects on recall or
learning of two nights of REM loss, and Johnson and his group (1972), after a study in
which  the recuperative value of REM or SWS after total deprivation of sleep was examined,
concluded that perhaps any type of sleep was sufficient for recovery in these circumstances.
One might therefore expect that performance  after sleep in which these two 'dominant'
stages were present, albeit in a reduced amount, would not be drastically affected.
     Yet common experience tells us that we do not function as effectively after a disturbed
night or one of poor sleep quality.  LeVere et al. (1972) have shown that sleep disturbed by
jet aircraft noises was followed by a  decrement in a reaction time test with  an added
memory  component  and by a concomitant increase in the amount of EEC delta activity
recorded in the morning after final awakening. Roth et al. (1971) also suggested that there
was a decrement  in  time  estimation, arithmetic  and memory tasks performed  after
noise-disturbed sleep.
     This somewhat paradoxical situation has brought about  a reorientation to the sleep
EEC.  In addition to  the  conventionally defined  sleep stages, the EEC shows a cyclical
tendency. REM periods occur, on average, once  every 1% hours. This tendency appears to
be quite strong in that it is not  destroyed by 180° inversion of the sleep cycle—a condition
which  affects many  other temporal aspects of the EEC (Weitzmann et al., 1970) and that it
seems  to take  a profound condition such as chronic alcoholism  to severely fragment  it
(Johnson et al., 1970).
     The present research was therefore designed  to extend testing into the day to see firstly
whether effects of noise-disturbed sleep on standard performance tests were demonstrable
and, if so, how long into the day they persisted, and secondly to examine the EEC aspects
of sleep in conjunction with subsequent daytime performance.


Method
     Ten healthy, male Ss with a mean age of 20.9 yrs (range 18-35), tested in pairs, slept in
the laboratory for 2 adaptation nights followed by  3 experimental nights. Ss were requested
to refrain from alcohol and napping during the experiment. Each S spent one night during
which  a  tape  on which  were recorded two clicks  spaced  one sec apart at one of four
intensities of 65, 75, 80 and 90 dBA was played through a speaker at the foot of the bed.
The pair of clicks was randomised with respect to intensity and inter-pair-interval, the latter
having  a mean of 20 sees. Ss were given an example of the noise before going to bed.

                                        528

-------
AWAKE

  REM
            YOUNG   ADULTS
                            2345

                                  HOURS  OF  SLEEP

                         Fig. 1 Typical Sleep Profile of Young Adults
                                (taken from Berger, 1969)
8
     To control for practice and order effects, Ss were tested in a simple crossover design, so
that one member of each pair had his disturbance on Night 3, the other on Night 5. As we
were also interested in subjective reactions,  self-report scales  were completed before lights
out and in the morning on awakening.
     The lights were turned out  at  approximately 2300 and  EOG,  EMG and monopolar
EEC were  monitored at a paper speed of 1 0 mm/sec for the following  8 hours. Records
were subsequently scored 'blind' according to the  Rechtschaffen and Kales (1968) manual
by  two scorers. The noise was turned on shortly after the  S was judged to have fallen
asleep-usually in descending Stage 2 sleep-and  was finally  turned  off just prior to final
awakening in the morning at 0645. Performance  testing began at 0745.
     The tests we used were the Wilkinson Vigilance and Addition Tests (Wilkinson, 1969)
which lasted  for one hour each. The vigilance test yields a measure of correct detections and
false reports  every  15  min; the  adding test, by means  of pen colour  changes, yields the
number of  sums done every 10 min. The final test was a short-term memory test (STM) in
which the S listened to prerecorded strings of 8 digits lasting 4 sec per string which he wrote
down on a provided sheet of paper  in the 6 sec before the start of the next string. Apart
from the STM test which was supervised by the experimenter, testing took place in separate
rooms. Knowledge  of results, which was given  for alternate  tests, was balanced across Ss
and, to control for time-of-day  effects, the  test  schedule was identical on each  day and
followed the programme shown in table 1.
Results
     All significance levels are assessed by using nonparametric tests (Siegel, 1956).

                                         529

-------
                                      Tablet
                                 TEST PROGRAM
TIME
0745
0850
0950
1020
1045
1145
1300
1405
1440
1540
1600
TEST
Vigilance 1 '
Adding 1
STM 1
- Break -
Vigilance 2
LUNCH
Adding 2
STM 2
Vigilance 3
- Break -
Adding 3
Sleep Data
     Our choice of time constant was shorter than is usually recommended, but was con-
stant throughout all Ss. The scorers averaged 89% agreement.
     (a) Stages: There was a significant increase due to noise in the percentage of time
spent in Stage  1 after sleep onset (p = .01) and in the amount of time spent awake (p < .01)
both in total (p < .01) and after sleep onset  (p < .01), but no S had less than 6 hours total
sleep (range 6h,7m to 7h,30m). The tendencies towards a reduction in SWS and REM were
not significant.
     (b) Rhythmicity: In view of Globus et al.  (1972) we adopted as our measure of
rhythmicity the coefficient of variation (CV) of the intervals between  REM period onsets,
taking episodes of REM occurring within  15 min of each other to be part of the same REM
period. (Weitzmann et al., 1970). The tendency for sleep to be less rhythmic during noise
was not significant (T=10:N=9). We were  unable to calculate the CV for one S as he had an
insufficient number of REM periods.
                                       530

-------
40
 5Q


 4Q


 30
i

 2Q


 IO
                               Stage I
             noise       control                none        control

  Fig. 2 Minutes spent in stage 1 and awake during noise and control nights with 95% confidence limits
IOQ
m
80.
(
6Q
1 4Q
20
O


i

4
•
i
i
.^
^^
m
»
sws





T 	 , REM
<~ 	 •-





             noise        control                noise       control

   Fig. 3 Minutes spent in SWS and REM during noise and control nights, with 95% confidence limits

                                      531

-------
Performance Data
     Table 2 shows that there is little difference due to noise in composite measures on the
various tests but we expected to have to analyse the performance data in more detail than
looking solely at signals detected and the number of sums done. The analysis followed two
basic strategies: —
     1. Decremental Analysis.  Here  we  compared  levels  of performance  after noise-
disturbed and control sleep.
     l(a). Adding. The practice effect over days was sufficiently strong to mask changes in
the number of sums  done. However, on the basis that initial and end effects might be
operating, we examined the relative decrements in the middle of the test compared with the
two ends. This revealed that on the first test of the day, there was a relatively greater decline
in output in the middle of the test after disturbed sleep (p < .02). This effect was present
only in the first adding test.
     l(b).  Vigilance. Analysis of correct detections and false positives combined showed
that on the first test of the day there were significantly fewer responses made after noise (p
< .05; Randomisation test). None of the other vigilance tests showed a consistent effect.
Other  measures appropriate for vigilance analysis, notably the Signal Detection Theory
(SDT) parameters of d' and j3 were not uniformly affected.
     2.  Correlational Analysis
     Stage Correlations. Ss were ranked both overall and within groups on total sleep time,
minutes spent in SWS and  REM and their ranks were correlated with the various within-test
measures on the two tests that showed a decrement after noise. Coefficients were generally
low and none were significant.

Rhythmicity Correlations
     Ss were ranked on the degree of change in the CV from noise to control night. Taking
the CV to be a measure of disturbance, a low rank obtained in this way indicates that
                                       Table 2
              MEANS AND STANDARD DEVIATIONS OF PERFORMANCE TESTS
Test No.
Noise
Control
Vigilance
No e Signals Detected
1 2 3
24.1 27.4 28.5
9.7 7.0 8.4
28.4 28.7 28.6
5.1 6.9 6.6
Adding
NoB Sums Done
1 2 3
281.8 294.5 292.7
108.9 160.7 120.0
308.8 301.7 300.3
92.8 111.7 112.0
STM
No. Errors
1 2
95.1 89.6
87.2 91.4
79.9 66.8
78.5 66.3
                                        532

-------
                                  Tables
         COEFFICIENTS 
-------
                                          Table 4



         COEFFICIENTS (Rs) RESULTING FROM CORRELATION OF CV WITH PERFORMANCE MEASURES.





                    Vig 1     Add 1     STM 1      Vig 2     Add 2     STM 2     Vig 3     Add 3
Common
Decline
Adding
Middle: Outer
Vigilance
Decrement in
Output
STM:
Decrement
over time
Variability:
Standard
Deviation Output
DIUHNAL
Test 1 - Test 3
.050


-.783 *






-.280
.750*
-.023

.200







.070
.150








-.650



.905*


-.200






.883*

.800*

-.047







.230









.483



-.500


-.516






-.866*

.300

-.583







.230

* :  p < .02




* :  p =» .02




* :  P < .03

-------
compared with his control sleep, the S has less disturbed sleep and so on, with high ranks
showing more disturbed sleep than control.
     (a) Adding. On the grounds that overall performance levels were affected at a time of
day when they are normally 'poorer' (Blake, 1967) and affected by sleep deprivation (Wil-
kinson, 1972), we elected to examine the 'post-lunch dip,' the period when a 'drop' in
performance against a  background  of relative improvement over the  day has been con-
sistently found  (Colquhoun, 1971).  It  was  hypothesised  that such  naturally-occurring
troughs in performance might partly reflect, in their degree,  the relative disturbance of the
previous night's sleep. To test this, the total adding scores per 10 min over the day were
normalised and the differences between the lowest scores following noise and control nights
during the period  1:00 pm to 1:40 pm (to avoid the 'end' effect) were ranked. The resulting
rs was .800  (p < .02), indicating that the degree  of post-lunch dip was significantly associ-
ated with the rhythmicity of the previous night's  sleep. The more unrhythmic the sleep, the
greater was  the decline. It was further postulated that where other tests showed a decline in
the normative score at  a  common temporal  point on the two days (termed  'common
decline'), this method might be applied.
     Such periods selected from the other tests yielded coefficients of-.02 (Add. 1) and .30
(Add. 3) which are clearly nonsignificant.
     (b) Vigilance. The similar analysis on  the normalised %-test raw output measure pro-
duced r? = .05 on test 1; rs = .905  (N=8) (p < .02)pn test 2; rs = -.500 on test 3. Although
correlational analysis on the  decline in output in test 1 shows a significant negative coeffi-
cient  (-.783), this is possibly a result of the positive correlation of rhythmicity and the
output in  the first half of the test (raw output rs .650; proportion noise:  control output rs
.684, p = .05) and a lack of relationship in the second half (rs -.075).
     (c) STM. Division of this 30-min test into 3 periods of 10 min rarely allowed an
adequate point of common decline to be isolated. Rhythmicity correlations were therefore
not calculable. However the decrement  over time  was  measureable. The corrected pro-
portion scores, a measure which takes into  acount the differential results obtained by using
the proportions of errors or corrects, of the number of items wrong in the  two halves of the
tests,  were compared and the difference correlated with the C.V. Thers of -.650 is reason-
ably strong  but  reaches significance only at the one-tailed level.  It is noteworthy that the
relationship  is negative,  i.e.  the increase in errors over the task is less with increasingly
unrhythmic  sleep.
     (d) Variability in performance. This is a generally underexplored area of performance.
Correlations of the CV with  the respective standard deviations are shown  in Table 4 which
shows a significantly positive  relationship on Vig 2 and a negative on test 3.

Discussion
     In looking at these results, the first finding to stress is that, in overall terms, the effects
on standard  performance tests of noise-disturbed  sleep appear reliably only in the first two
hours of the day, and that the pattern or responses suggests a common factor of decreased
motivation.
    Although the vigilance test is experimenter-paced, the measure found to be the most
sensitive i.e. the total output of responses of any kind, is the one that most closely approx-

                                         535

-------
 imates to a subject-paced measure. The SDT parameter 0, which is commonly held to reflect
 willingness to respond as distinct from capacity to discriminate the signal (Wilkinson, 1969),
 does show a rise after disturbed sleep i.e. increased unwillingness to report. This rise does
 not reach  significance probably because  SDT parameters are dependent upon false positives
 which, in the present setting, are small in number and highly variable from subject to subject
 due to the relatively low signal frequency.
      In the adding test where the S is free to generate his own pace, the effect appears more
 strongly, and the presence of strong initial and/or end effects supports an analysis in terms
 of reduced motivation.
      The location of these effects in the early part of the day also suggests an interaction
 with the diurnal cycle. Wilkinson (1963) looked at performance in Ss who had been allowed
 recovery sleep after  one  night's sleep loss, and  found that performance  effects were still
-present, predominantly in the  morning,  an effect which was  ascribed  to disruption of
 normal physiological rhythms. Furthermore, the nature  of the  performance changes was
 unlike those due to sleep deprivation in  that this aftereffect was apparent at the start of the
 test.
      The present experiment, taken in conjunction with this earlier finding, lends support to
 the hypothesis of the regulatory function of the ultradian cycle during sleep for subsequent
 performance. It is likely  that Wilkinson's Ss had a large increase in SWS during their re-
 covery sleep (Berger and Oswald,  1962) which  might be sufficient to  disrupt partly the
 cyclical nature of the sleep, with a consequent disturbance  of the diurnal rhythm in  later
 performance. In the present noise experiment, the degree of rhythmicity during the night is
 directly correlated with the output in  the first half of the vigilance test. As the relationship
 in the second half is  low and unsystematic, it therefore seems that the significant negative
 correlation of the CV with decrement over time  on task is due primarily to the first half of
 the test.
      Wilkinson also found that his performance effects were not as great in the afternoon.
 In overall  terms  this is the conclusion of the present experiment. Nevertheless secondary
 effects, further implicating the role of the rhythmical aspects of sleep, begin to appear after
 the initial effect has disappeared.
      The first suggestion  comes in the STM test in which the degree of increase in errors
 during the task is negatively correlated with the CV. The more disturbed the sleep profile,
 the less the increase in errors, although this is significant only  at the one-tailed level.
      Examining the point of common decline in the immediate pre- and  post-lunch tests,
 the correlations become stronger. In the last vigilance test  of the morning, the more dis-
 rupted the sleep profile, the greater is the decline at this point. Although  the 'yardstick' is
 the standard deviation of the total output over the day—a measure bearing no relation to the
 rhythmicity of the previous night's sleep-it may be that this result is a complex function of
 the degree of output variability within  this test, as the latter does  show a strong positive
 correlation with  the  CV.  However, the  post-lunch dip is  also similarly related and, in this
 test, the variability in output is only weakly related to the sleep measure.
     By the middle of the  afternoon, most of the effects have disappeared and begin to show
 a negative  relationship, in one case (variability of output in Vigilance 3) significantly so.
 This latter finding may again be reflecting disturbance of the diurnal rhythm  as the analysis

                                          536

-------
of the change over the day in vigilance variability shows a strong relationship to the sleep
measure, (rs = .750).
     In overview, the general pattern of performance that emerges after sleeping in a noisy
environment is one where the early tests of the day are affected in a manner that suggests
both decreased motivation and disruption of the normal diurnal cycle. Recovery has taken
place by mid-afternoon, but the speed with which this occurs depends to a large degree
upon the fragmentation of the cyclical nature of the preceding sleep.
     It has been argued by Hauri and Hawkins (1971) that total or selective sleep depriva-
tion may  not  be comparable with naturally-occurring disturbed sleep and indeed it seems
that the present  effects of disturbed sleep are not identical with those of sleep deprivation
for two reasons. The initial effect appears in the first half of the test and  in the morning,
and secondly because the pattern of STM responses is different from that  occurring under
total deprivation conditions (Wilkinson and  Spence, in press).  The usual  strategy to deal
with the recuperative aspects of sleep has  been to adopt a molecular approach and to
consider that the main restorative value lies in one or other  of the stages, particularly SWS
(Hauri,  1970). This strategy cannot yet be  ruled out, but we were unable to find  any
meaningful correlations of either  total sleep time or minutes spent in SWS or REM sleep
with early performance parameters, which suggests that these aspects of sleep have a func-
tion which is not immediately apparent. In dealing with disturbed sleep, a more molar
orientation to the EEC  appears  to  have value  in terms of understanding  its  effects on
performance.
     For the purposes of this  congress it  is important  to add a note  of caution.  Our
disturbing agent can be considered to be relatively constant, whereas much of the present
day  concern  about the consequences of nocturnal noise centers mainly  upon  much less
frequently occurring noises, particularly sonic  booms. We would expect that  where the
architecture of sleep, as measured by the CV technique, is broken by irregularly occurring
noises, that similar performance effects would be demonstrable. At the moment, however,
this remains a hypothesis.

                                   REFERENCES

BERGER,  R.  J.  The  Sleep and Dream Cycle. In A.  Kales (ed). Sleep:  Physiology  and
     Pathology. Philadelphia. Lippincott (1969).
BERGER,  R. J., and OSWALD, I. Effects of sleep  deprivation on behaviour, subsequent
     sleep and dreaming, Journal of Mental Science, 108, pp 457-465. (1962).
BLAKE,  M. J. F. Time  of day effects on performance in a range of tasks. Psychonomic
     Science 9, 349-350 (1967).
CHERNIK, DORIS, A. Effects of REM sleep deprivation on learning and recall by humans.
     Perceptual and Motor Skills, 34,  83-294 (1972).
COLQUHOUN, W. P.  Orcadian Viriations in Mental Efficiency in W. P.  Colquhoun (ed).
     Biological Rhythms and Human  Performance.  London.  English Universities Press.
     (1972).
DOBBS, MARY,  E. Behavioral responses to  auditory stimulation during sleep. Journal of
     Sound and Vibration, 20, 467-476. (1972).

                                        537

-------
GLOBUS, G. G., PHOEBUS, E. C, and BOYD, R.-Temporal organisation of night workers'
    sleep. Aerospace Medicine, 43, 266-268 (1972).
HAURI, P. What is good sleep? International Psychiatry Clinics, 7, 0-77, (1970).
HAURI, P.  and HAWKINS, D. R. Phasic REM,  depression  and the relationship between
    sleeping and waking; Archives of General Psychiatry, 25,  56-63, (1971).
JOHNSON,  L. C.,  BURDICK, J. A., and  SMITH,  J. Sleep  during alcohol intake and
    withdrawal in  the chronic alcoholic. Archives of General Psychiatry,  22, 406^18.
    (1970).
JOHNSON,  L., NAITOH, P., LUBIN, A. and MOSES, J. Sleep stages and performance. In W.
    P. Colquhoun (ed). Aspects of Human Efficiency. Diurnal Rhythm and Loss of Sleep.
    London. English Universities Press. (1972).
LEVERE,  T. E.,  BARTUS,  R.  T.,  and  HART, F. D.  Electroencephalographic and
    behavioural effects  of nocturnally occurring jet  aircraft sounds. Aerospace Medicine,
    43, 384-389. (1972).
LUDLOW, J. E., and  MORGAN,  P. A. Behavioural awakening and subjective reactions to
    indoor sonic booms. Journal of Sound and Vibration, 25, 479-495. (1972).
LUKAS, J.  S. Awakening effects of simulated sonic booms and aircraft noise on men and
    women. Journal of Sound and Vibration, 25, 479-495. (1972).
LUKAS, J.  S. and KRYTER, K.  D.  Awakening effects of simulated sonic booms and
    subsonic aircraft noise in WELCH, B. L. and WELCH, A. S. (ed). Physiological Effects
    of Noise, New York, Plenum Press (1970).
LUKAS, J.  S. and DOBBS, M. E. Effects of aircraft noises  on  the sleep of women. Final
    Report NASA CR-2041 (1972).
MORGAN,  P. A. Effects of noise upon sleep.  University  of Southampton. Institute of
    Sound  and Vibration Research. Technical Report, 40. (1970).
RECHTSCHAFFEN, A.  and KALES,  A. (eds).  A manual  of  standardised terminology,
    techniques and scoring system for sleep stages of human subjects. National Institutes of
    Health  Publication No. 204.  Public Health  Service. U.S. Government Printing Office
    (1968).
ROTH, J., KRAMER, M. and TRINDER, J. Noise, sleep and post sleep behaviour. Paper
    presented to  the 124th  Annual  Meeting of the American Psychiatric Association.
    Washington D.C (1971).
RYLANDER, R., SORENSEN, S., and BERGLUND, K. (1972). Sonic boom effects on
    sleep. A field experiment on military and civilian populations. Journal of Sound and
    Vibration. 24, 41-50 (1972).
SIEGEL, S.  Nonparametric statistics for the  behavioural sciences. New York. McGraw-Hill.
    (1956).
THIESSEN, G. J. Effects of noise during sleep. In WELCH, B.  L. and WELCH, A. S. (eds).
    Physiological Effects of Noise. New York. Plenum Press (1970).
WEBB, W.  B. Partial and differential sleep deprivation.  In  KALES, A. (ed).  Sleep:
    Physiology and Pathology. Lippincott. Philadelphia. (1969).
WEITZMAN,  E.  D.,  KRIPKE,  D. F.,  GOLDMACHERM D.,  MCFREGOR, P., and
    NOGEIRE, C. Acute reversal of the sleep waking cycle  in man: Effects on sleep stage
    patterns. Archives of Neurology, 22, 483-489. (1970).

                                     538

-------
WILKINSON, R. T. Aftereffect of sleep deprivation. Journal of Experimental Psychology,
    6(5,439-442(1963).
WILKINSON, R. T.  Sleep deprivation: performance tests for partial and selective sleep
    deprivation. In Progress in Clinical Psychology, Grune and Stratton. (1969).
WILKINSON,  R.  T.  Methods  for, research on  sleep deprivation  and sleep  function.
    International Psychiatry Clinics 7, 369-81 (1970).
WILKINSON, R. T. Sleep deprivation: Eight questions. In W. P. Colquhoun (ed). Aspects of
    Human Efficiency: Diurnal Rhythm and Loss of Sleep.  London. English Universities
    Press.
WILKINSON, R. T. EDWARDS, R. S. and HAINES, E. Performance following a night of
    reduced sleep. Psychonomic Science 5,  71-472. (1966).
                                        539

-------
                EFFECTS ON SLEEP OF HOURLY PRESENTATIONS
                     OF SIMULATED SONIC BOOMS (50 N/M2)

                        William E. Collins and P. F. lampietro
                Aviation Psychology and Stress Physiology Laboratories
                           FAA Civil Aeromedical Institute
                          Oklahoma City, Oklahoma 73125

     Relatively little research has been conducted  concerning the effects of sonic booms on
sleep behavior. However, there is a good deal known about the general influence of noise on
sleep patterns (e.g., Dobbs,  1970; Kryter, 1970; Williams,  1970). A  number of laboratory
studies, using different auditory stimuli, have shown that waking responses are a function of
the following variables: individual differences, intensity of  the acoustic stimulus, age of the
sleeper, sex of the sleeper, time of night, stage of sleep, amount of accumulated sleep, and
personal significance of the auditory stimulus (e.g., the sleeper's own name). Moreover, most
physiological responses (e.g., brain wave activity, heart rate) appear to  show little or no
"adaptation" to acoustic stimulation during sleep.
     Lukas and his co-workers (Lukas, 1970; Lukas and Dobbs, 1972; Lukas, Dobbs, and
Kryter, 1971; Lukas and Kryter, 1968, 1970a, 1970b) have conducted most of the studies
of sonic booms and sleep behavior. Those studies exposed a total of 22 male and female
subjects to simulated sonic booms ranging in "outdoor" intensities from 0.6 to 5.0 psf and
to recordings of subsonic jet flyover noise (ranging in "outdoor" intensity from 101-119
PNdB). With the exception of four middle-aged females, subjects were tested frequently but
on  non-consecutive nights;  the female  exceptions were tested for  14  consecutive  nights
(Lukas and Dobbs, 1972). Results of those studies, using criteria of arousal and awakening,
may be summarized as  follows:  children (5-8 years of age) appear  to  be undisturbed by
noise during  sleep; in general, younger subjects are less sensitive  to noise than are  older
subjects; irrespective of age,  individuals may show considerable variability in relative sensi-
tivity to noise during sleep; men appear to be less sensitive to noise than do women; and the
occurrence of behavioral awakenings is a function of the intensity of  the noise (Lukas,
1970).
    The present  study investigated the effects on sleep, mood, and  performance of simu-
lated sonic booms occurring regularly during the night over a consecutive 12-night period.
An overpressure level  of 50 N/m2 (1.0 psf), as measured "outdoors," was selected as an
"acceptable" boom stimulus based partly on determinations from other studies which indi-
cated that sonic boom  overpressure levels of 0 to 1.0 psf would produce no significant
public  reaction day or  night, while levels of 1.0 to  1.5 would  produce probable  public
reaction (von Gierke,  1966). The 50 N/m2 level has been likened to the sound  of moderate
thunder (Richards  and  Rylander,  1972)  or of moderate  to distant thunder (Ferri and
Schwartz, 1972).
     2The sleep data in this study were scored and analyzed by Milton Kramer, M.D., and Thomas Roth,
Ph.D., University of Cincinnati, under FAA Contract DOT FA70AC-1125-3.

                                         541

-------
METHOD

                                     Facility

    The sonic boom simulation facility (Thackray, Touchstone, and Jones, 1970) at the
Civil Aeromedical Institute has two main components: an electromechanical boom genera-
tor and a test room. The generator was set to provide simulated booms (directly to the test
chamber) of 50 N/m2; the rise time was 6.8 msec and the boom duration was 299.5 msec.
Within the  test room, the comparable data  were 6.5 N/m2, 12.1 msec, and 283.6 msec,
respectively (cf. Thackray et al.,  1971). Sound levels of the booms measured by a B&K
Impulse Precision Sound  Level Meter (Type 2204), were approximately 80 dBA on the
Impulse setting and 68 dBA on  the Slow  setting. The test room,  of standard dry  wall
construction, simulates a middle  bedroom in  a frame house. One  wall of the pressure
chamber forms the "outside" wall of the test room.

                            Physiological Measurements

     All physiological measurements (and the occurrence of each boom) were recorded on
8-channel polygraphs and on  magnetic  tape.  Recordings were made of the electro-
encephalogram  (EEC),   electro-oculogram  (EOG),  electromyogram  (EMG),  electro-
cardiogram (ECG), and the basal skin resistance (BSR). Electrode placement sites and the
recording techniques employed fon EEG, EOG, and EMG were those suggested by Rechts-
chaffen and Kales (1968).. ECG was derived from two standard EEG electrodes taped to the
subject's thorax; BSR tracings'were  obtained  from two electrodes taped to the palmar
surface of the distal segment of the right forefinger and right ring finger.

                                 Mood Assessment
    The Composite Mood  Adjective Check  List assessed affective states prior  to and
subsequent to each sleep  period. Each List was scored for 15 mood factors and an overall
index of mood (cf., Smith and Hutto, 1972).

                             Performance Measurement
    The CAMI Multiple Task Performance Battery  (MTPB), used before and after  each
sleep period, was programmed to present two active (mental arithmetic and pattern discrimi-
nation) and two passive tasks (monitoring lights and monitoring meters). Ten performance
scores were derived for each subject (cf., Chiles and West,  1972).

                                     Subjects
    A total of 24 male subjects was used; eight subjects each  were in groups aged 21-26,
4045, and 60-72.  Prior  to the experiment, subject-candidates were interviewed, given a
hearing test, and administered a health questionnaire. Subjects were not told that simulated
sonic  booms would be presented;  they were instructed to ignore disturbances of any kind
and get the best night's sleep possible.

                                       542

-------
                                     Procedure
     Two subjects from the same age group reported to the sleep laboratory at 2000 hours
each night for a total of 21 consecutive nights. The first five nights allowed subjects to adapt
to sleeping in the laboratory environment (nights 1 and 2), and provided "Baseline" data
(nights 3-5).  During the next  12 nights ("Boom"), the  subjects were presented  with a
simulated boom at  hourly intervals starting at 2300 hours and ending at 0600 hours. The
final four nights were termed "Recovery" nights (no booms presented).
     At 2000 hours and again at 0700 hours the  subjects were tested on the performance
battery for 30 minutes. At 2040 and at 0730 hours, the mood check list was administered.
Between  2100  and  2200 hours all electrodes were attached and  other preparations com-
pleted so that the subjects would be in bed at 2200 hours. Continuous recordings were made
for 500 minutes  until  the subjects  were awakened at 0620 hours. More details regarding
procedures and  results are presented elsewhere (Collins and lampietro, 1972).

RESULTS AND DISCUSSION

                                   Sleep Profiles
     Patterns of Sleep.  Percentages reflecting the mean amounts of time subjects in each
age group spent in four stages of sleep, in movement during sleep, and in being awake during
Baseline,  Boom, and Recovery phases appear in Table 1. Analyses of variance conducted on
each of these scores indicated no significant differences  at the .05 level among the three
phases; thus,  the  booms had no significant effect on the percentage of time spent in any
sleep stage. However, significant differences (p < .01  to  p < .001) among the age groups
were obtained for five of the six sleep stages (Stage 3-4 was excepted) indicating that age
influenced the proportion of time spent in one sleep stage or another (the sleep pattern of
the oldest group accounted for most of these differences).
     The distribution of time awake during the night, the latencies for onset of Stage 2 and
Stage REM, and the number of changes in sleep stages during the night are presented by age
group in  Table 2 for Baseline,  Boom,  and Recovery nights. Analyses of variance of these
scores yielded no significant difference  at the .05  level across the three phases. Thus, these
sleep profiles showed  no  effects which could be attributed to the boom presentations.
Significant  differences  which were obtained among the age groups for spontaneous time
awake (p < .001), latency to Stage REM (p < .01), and shifts in .sleep stages using 5-minute
and  10-minute time bases (p <  .001 in both cases) reflect differences in sleep patterns with
age and are independent of the presence or absence of the booms. Similarly, statistically
significant interactions (age groups by the three phases) obtained for the latency scores for
both Stage 2 (p < .001) and Stage REM  (p < .01) reflected  no  effects of the boom
presentations but, rather, increased latencies  for the youngest age  group from  the Baseline
through the Boom through the Recovery phases.
     Awakenings. /The tracings were used to calculate the nightly frequency of awakenings
for Baseline,  Booms and Recovery phases (Table 3).  A fractionally higher incidence of
awakenings occurred during boom nights for all  groups.  However, an analysis of variance
yielded only  a significant age effect (p < .01); thus, the frequency of awakenings increased
with age, but  no effect on awakenings can be attributed to  the booms.

                                        543

-------
                       Table 1

MEAN PERCENT OF NIGHT (500 MINUTES) IN THE VARIOUS STAGES
   OF SLEEP AND WAKE FULNESS FOR BASELINE, BOOM, AND
                   RECOVERY NIGHT

                         Age Group  (Years)
Stages

21-26
Baseline Boom
Total Time Awake
** Movement Time
During Sleep
Sleep Stage 1
Sleep Stage 2
Sleep Stage 3-4
Sleep Stage REN
Total
6.6
2.2
5.3
44.9
19.0
22.0
100.0
8.0
2.2
5.2
45.4
18.0
21.2
100.0

Recovery
11.1
2.4
5.2
44.9
16.5
19.9
100.0

Baseline
10.9
1.6
9.4
40.3
16.3
21.5
100.0 '
40-45
Boom
9.2
1.3
8.7
40.4
18.1
22.3
100.0

Recovery
7.9
1.3
9.3
42.0
17.3
22.2
100.0

Baseline
18.2
1.5
7.5
34.6
21.8
16.4
100.0
60-72
Boom
16.9
1.4
8.1
35.3
20.2
18.1
100.0

Recovery
16.2
1.4
8.4
36.8
20,1
17.1
100.0

-------
Ui
•&•
  Measures

Minutes Awake
Before Sleep

Minutes Awake
During the Night

Minutes Awake
After Sleep

Minutes Latency
to Stage  2

Minutes Latency
to Stage  REM

No. of 30-Second
Stage Changes

No. of 5 -Minute
Stage Changes

No of 10-Minute
Stage Changes
                                                   Table 2

                       DISTRIBUTION OF TIME AWAKE (IN MINUTES), MEAN NUMBER OF SLEEP
                            STAGE ALTERATIONS, AND MEAN LATENCIES (IN MINUTES)
                       TO STAGES 2 AND REM FOR BASELINE, BOOM, AND RECOVERY NIGHTS

                                                       Age Group  (Years)

Baseline
21.9
8.0
3.1
14.2
60.0
63.6
10.5
8.1
21-26
Boom
27.0
12.1
0.4
16.4
66.1
64.8
10.6
7.9

Recovery
27.9
15.4
1.8
20.4
78.3
63.9
10.3
7.3

Baseline
24.0
24.1
6.7
15.0
62.1
62.3
13.5
9.5
40-45
Boom
21.2
19.9
4.6
11.8
54.7
65.2
13.1
9.8

Recovery
19.3
15.2
3.4
11.7
49.9
64.5
12.7
9.7

Baseline
25.2
52.3
13.3
14.0
52.9
72.1
15.5
10.0
60-72
Boom
24.1
51.6
9.7
13.0
54.2
69.3
15.5
10.6

Recovery
25.3
47.8
7.7
14.1
53.4
69.7
15.3
10.5

-------
                                      Table 3

              MEAN FREQUENCY OF AWAKENINGS PER NIGHT PER SUBJECT
                    FOR BASELINE. BOOM, AND RECOVERY NIGHTS
Age Group (Years)
21-26
1.7
1.8
1.4
m
40-45
3.3
3.7
3.1
374
60-72
5.9
5.9
5.1
TTs
                                                                            Mean

 Baseline           1.7               3.3               5.9             3.5

 Boom                1.8               3.7               S.9             3.8

 Recovery           1.4               3.1               5.1             3.2

 Mean
     These awakenings were scored on the basis of evidence in the physiological tracings and
were not the "behavioral awakenings" (whereby subjects signal their waking state) reported
by Lukas and his co-workers (e.g., Lukas, 1970; Lukas and Dobbs, 1972; Lukas and Kryter,
1970a). That we instructed our subjects to ignore disturbances and to attempt to get the
best night's  sleep possible  might well account for the smaller number of responses to the
booms reported here, compared with other data (e.g.,  Lukas and Kryter,  1970a) obtained
under conditions in which  the subjects apparently were made more aware of the purpose of
the study and were asked  to  signal  whenever awakened. Moreover, the occurrence of age
differences in awakenings is well known (e.g., Kramer et al., 1971; Williams, 1970).
     EEC Changes in Response  to Booms. The seven-point scoring criteria established by
Williams (Table 4) were used to make direct comparisons of EEG responses to the booms
with responses to periods of pseudo-stimulus controls (i.e., to periods of sleep 30 minutes
prior to presentation  of a boom). From  these comparisons, presented in Table 5, it can be
determined that, while 74.2 per cent of the booms  produced an EEG response (i.e., a
non-zero Williams Score), only 36.2 per cent of the control periods showed an EEG change.
     More "zero" and "1" scores were obtained in control periods and more "2" through
"7" scores were obtained  in response to the booms; these differences  were significant by
chi-square analysis at the  .01 level. Although statistically significant, these effects were
functionally  mild since boom presentations produced awakenings only 5.5 per cent of the
time (compared with 0.7 per cent for non-boom controls) and resulted in shifts in stages of
sleep only 14.3 per cent of the time (compared with 4.2 per cent for non-boom controls).
    Age. Chi-square  analysis of the data in Table 5 yielded  a significant difference (p <
.01) in EEG responses as a function  of the age groups. This difference  is  due primarily to
more frequent responses at the higher scores ("5" and "7") in the oldest group. The two

                                       546

-------
                                     Table 4

         THE WILLIAMS CRITERIA FOR SCORING EEC TRACINGS (ADAPTED FROM
     LUKAS AND KRYTER, 1968). THESE SCORES ARE NOT INDEPENDENT SINCE A HIGH
  SCORE USUALLY INCLUDES ALL THE LOWER SCORES, E.G., A WILLIAMS SCORE OF THREE
                   INDICATES THAT K COMPLEXES ALSO OCCURRED

Williams
 Score              	Change Required  on  EEC  Record	

   0                No change.

   1                A K complex of low amplitude (less than 150 microvolts)
                    which occurs within one second after boom presentation,
                    but is usually coincidental with the boom.

   2                A K complex of high amplitude (above 150 microvolts) or
                    several K responses which occur within two seconds of
                    termination of the boom stimulus.

   3                The presence of an Alpha pattern or synchronization within
                    two seconds of termination of the boom stimulus
                                                />
   4                Body movement or movement of facial or eye muscles within
                    six seconds of termination of the boom stimulus.

   5                A one-step shift in sleep stage  (e.g. , from a Stage 3 to
                    a Stage  2) within  one minute of  termination of the boom
                    stimulus.

   6                A two-step shift in sleep stage  (e.g., from a Stage 4 to
                    a Stage  2) within  one minute of  termination of the boom
                    stimulus.  (This score was not assigned since we used a
                    combined Stage 3-4 and a shift of two stages resulted in
                    awakening.)

   7                Prolonged Alpha movement and an Awake response within one
                    and one-half minutes  of termination of the boom stimulus.
                    (The delay was recommended for studies which require the
                    subject  to signal  his  awareness  of  being awake; it allows
                    time  for the subject  to find the signalling device.)
youngest groups showed more "0" and "2" scores, and fewer scores of "5" (shift in stage of
sleep) and "7" (awakening) than did the oldest group.
    Adaptation.  To examine  possible "adaptation" effects  across the 12 boom nights,
analyses of variance were performed on Williams Scores which occurred at least 10 per cent
of the time following boom presentations; those scores were "0," "2," "4," and "5" (see
Table 5). The percentage of EEC responses under each of the four scores is  presented in
Table 6 for each boom night. No differences among the 12 nights were significant at the .05

                                      547

-------
                                      Table 5

            MEAN FREQUENCY OF OCCURRENCES (IN PERCENTAGES) OF EACH
             WILLIAMS SCORE IN EEG TRACINGS FOLLOWING PRESENTATION
                   OF BOOMS AND OF PSEUDO-STIMULUS CONTROLS
                                     Age Group  (Years)
Williams
Scores

   0   .

   1

   2

   3

   4

   5

   6
21-26
Boon
28.0
4.6
27.6
0.9
26.6
10.6
0.0
1.7
Control
69.0
8.2
13.1
0.0
6.0
3.4
0.0
0.0
40-45
Boon
27.8
5.1
14.9
6.6
25.8
13.7
0.0
6.1
Control
61.7
5.8
12.1
1.5
13.9
3.9
0.0
1.1
60-72
Boon
21.5
5.0
14.0
6.4
25.8
18.6
0.0
8.7
Control
60.7
12.6
7.7
0.5
12.1
5.4
0.0
1.0
level for any score or age group by analyses of variance. Thus, there was no evidence in the
EEG tracings of "adaptation" to the occurrence of the booms.
     Sleep Stage. Subjects might be more responsive to booms during certain stages of sleep
(cf. Williams et al., 1964). Thus, per cent responses for the four stages of sleep (REM and
Stages  1, 2, 3-4) of "0," "2," "4," and "5" scores (scores which occurred at least 10 per
cent of the time following booms) by the Williams Criteria were plotted in Figure 1 for all
subjects combined. Analyses of variance were performed (without collapsing the age groups)
for each of the four Williams Scores. Results indicated significant (p < .05 to p < .001)
differential responsivity to the booms during certain stages of sleep. Specifically, Sleep Stage
1 and Sleep Stage 2 yielded significantly fewer "0" scores and significantly more "4" scores
following boom presentations than were obtained in control periods. This result is con-
sistent with other studies (Lukas, 1972; Lukas and Kryter, 1970b) and may be attributed to
the fact that Stage 1 in particular is a transition phase between being awake and reaching the
deeper sleep of Stage 2. Williams Scores of "2" and "5" showed different results; for these
scores, Stages 2 and 3-4  were most sensitive  to the boom presentations. The latter finding
might be expected because the K complex, which defines a score of "2," and stage shifts,

                                       548

-------
vo
    40-45

    Score

     0
     2
     4
     5
    60-72
    Score

     0
     2
     4
     5
                                                   Table 6

                              MEAN FREQUENCY OF OCCURRENCE (IN PERCENTAGES) OF
                                WILLIAMS SCORES OF 0, 2. 4, AND 5 FOLLOWING BOOM
                               PRESENTATIONS DURING EACH OF THE 12 BOOM NIGHTS
                                             Boom Nights
   Age Group
     (Years)      1       2       3       4       5       6        7       8       9     _10_    _U_    _12_

    21-26

    Score
0
2
4
5
25.0
32.8
28.1
12.5
26.6
26.6
21.9
6.3
25.0
28.1
26.6
9.4
37.5
23.4
20.3
9.4
34.4
18.8
28.1
12.5
21.8
23.4
34.4
12.5
31.6
28.1
26.6
9.4
32.4
31.0
20.5 .
11.4
23.4
26.6
31.3
10.9
33.0
27.5
35.2
12.4
21.9
31.3
23.4
14.1
28.1
30.0
31.3
7.8
17.2
21.9
20.3
18.8
23.4
12.5
35.9
12.5
28.1
14.1
29.7
17.2
17.6
27.2
29.0
7.8
20.3
7.8
26.3
17.2
23.4
12.5
23.4
12.5
29.7
14.1
26.6
15.6
34.4
17.2
25.0
10.9
33.9
14.3
19.0
13.6
34.4
10.9
31.3
12.5
28.1
12.5
21.9
17.2
34.4
14.1
23.4
7.8
21.9
10.9
14.1
18.8
12.5
8.0
25.2
24.1
28.1
12.5
20.5
14.7
20.3
4.7
35.9
15.6
23.4
23.4
18.8
17.2
20.3
10.9
28.1
17.2
17.2
17.2
29.7
18.8
35.9
14.1
21.9
17.2
21.0
20.8
21.0
27.5
16.3
17.8
33.7
17.4
19.4
12.9
27.0
20.7
20.96
14.28
33.46
13.81

-------
                              80
                               60
                              40
                               20
   TRIALS
     BOOM
o—o CONTROL


SCORE  2
                        3-4
REM  I    2 3-4
                               40 r
                               20
                                   SCORE 5
            REM  I    2 3-4
REM  I    2  3-4
                    STAGES  OF  SLEEP
Figure 1 Frequencies of occurrence in percentages of Williams scores of 0, 2, 4, and 5 following Boom
presentations in different stages of sleep
                          550

-------
which define the score of "5," are much less likely to occur in REM or Stage 1 sleep as a
function of the boom presentations (a noise-induced stage shift from REM or Stage 1 would
probably result in awakening and would thereby be scored as a "7").
     Time-of-Night. Analyses similar to those noted above for sleep stages were conducted
for time-of-night. For these  purposes, each boom night (500 minutes) was divided into four
quarters (results for all subjects combined are depicted in Figure 2). No significant time-of-
night differences between booms and control periods  were obtained for Williams Scores of
"2" and "5"; however, significant increases in sensitivity  to the booms were obtained  for
"0" and "4" scores (p < .05 and p < .001, respectively). There were proportionately fewer
"0" scores  for boom periods during the first and third quarters  of the night than were
obtained for control periods, and proportionately more "4" scores following booms during
the second  and third  quarters of  the night. Although these data extend the periods of
maximal sensitivity into the third quarter of  the night, they are in general agreement with
other studies (Kramer et al.,  1971; Morgan  and Rice,  1970; Rechtschaffen et  al., 1966;
Williams et al.,  1964) which show that sensitivity to noise is greater during early rather than
later periods of the night.
                                      Heart Rate
     The oldest group had significantly higher heart rates (p < .05) than did the younger
subjects, but introduction  of the booms produced no overall change in this measure (Table
7). Heart rate variability scores (standard deviations) also yielded no significant differences.
Although there was a significant increase (p < .05) in heart beats immediately after boom
presentations for all subjects across all nights (by 0.8, 0.6, and 0.8 beats per minute for the
21-26, 40-45, and 60-72 year olds, respectively), there was no effect attributable to age and
there was no "adaptation"  evident.
     Mean  EMG  levels (difference scores based on the level measured during the first five
minutes of the night, cf. Collins and lampietro, 1972) appear in Table 8 by age group for
the nine periods of the night. An analysis of variance of the  components of these data for
Baseline, Boom, and Recovery phases yielded only  a significant period-of-night effect (p <
.001) due to a general decrease in level of muscle tone during most of the night. The sharp
drop after  the first period (first 20 minutes) is accounted for by the  decrease of muscle tone
which accompanies the onset  of sleep; subsequent declines probably reflect increasing REM
sleep (lower levels of EMG activity would be expected). During the last two hours, the EMG
level increased as the end of the sleep period neared. There were no significant differences
among age groups or across  the Baseline, Boom, and Recovery nights. Moreover, EMG
variability scores (standard deviations) showed no boom-related effects.
     Significantly more changes (p < .001) in EMG levels occurred in response to the booms
than in control periods (41.5 vs.  10.1, 43.0 vs. 14.2, and 58.1 vs.  18.7 per cent, respectively,
for the 21-26, 40-45, and 60-72 year olds), and  the age-related differences in the frequency
of such changes  were significant (p < .05). However,  all EMG levels returned to  their
Baseline values within ten minutes of any responses to the booms.  Furthermore, statistical
tests indicated no evidence  of "adaptation" across the 12 Boom nights.

                                         551

-------
       80
  ZIO
  ~ a:
  oo
       60
    CM
(O  .
LU O  40
O 01  20
O O
CD O
  LU CL
             SCORE 0
  O (0
       40
3H20
  LU
  0-
              H	1	1	1
              1234
           SCORE  4
              1234
                              80
                              60
                               40
                               20
                                     TRIALS
                                       BOOM
                                   o—o CONTROL
                                   SCORE 2
                              40
                               20
                                    1234
                                  SCORE 5
                                    1234
                QUARTER  OF THE  NIGHT
Figure 2 Frequencies of occurrences in percentages of Williams scores of 0. 2. 4, and 5 following Boom

presentations in different quarters of the night
                         552

-------
Age Group
  (Years)

21-26

40-45

60-72

Mean
                                       Table 7

             MEAN NUMBER OF HEARTBEATS PER MINUTE PER SUBJECT FOR
                       BASELINE, BOOM, AND RECOVERY NIGHTS


                                                 Nights
Baseline
63.6
62.4
69.7
65.2
Boom
62.4
62.0
68.9
64.4
Recovery
64.9
63.1
68.4
65.5
                                Basal Skin Resistance

     Mean BSR levels (difference scores, cf. Collins and lampietro, 1972) by age group for
the nine periods of the night appear in Table 9. A statistically significant age effect (p < .05)
is due to the consistently lower level of skin resistance during all periods and across  all
conditions  for  the  youngest subjects  (at  least  partially  attributable  to  differences  in
amplifier gain settings). A significant change (p  < .001) in  BSR across the  nine nightly
periods reflects  an expected general decrease in skin resistance during the night. Significant
differences (p < .0!) in BSR scores among the Baseline, Boom, and Recovery nights may be
accounted for by a general decrease in skin resistance (signifying increased  arousal) across
the three experimental conditions. Since skin resistance continued to decline during the
Recovery (non-boom) nights, the  effect  cannot  be attributed to boom presentations.  In
addition, no  boom-related  effects were  obtained  upon analysis of variability (standard
deviations) in BSR scores.
     Disregarding age, a mean change in BSR level of -5.0 kilohms occurred within five
seconds for  19.4 per cent of the booms; the mean latency for those occurrences was 3.2
seconds. Recovery to pre-boom BSR levels occurred within ten minutes for 51.0 per cent of
the boom-induced changes and, of these 51.0 per cent,  the mean latency for recovery was
47.9 'seconds. Analyses of variance conducted for each of these five BSR measures yielded
only one significant effect (p < .05); the youngest subjects showed less change to the booms
in BSR level than did either of the two older groups of subjects. There was no evidence of
"adaptation" to the booms across the 12 Boom nights.

                                        553

-------
                                   Table 8
         MEAN EMG LEVELS P^R SUBJECT DURING NINE PERIODS OF THE NIGHT
             FOR BASELINE, BOOM, AND RECOVERY NIGHTS. EACH VALUE
          REPRESENTS A DIFFERENCE SCORE FROM MEASUREMENTS MADE IN
            THE FIRST 5-MINUTE EPOCH. MINUS SIGNS HAVE BEEN OMITTED
                              FROM ALL VALUES
Period of
the Night
    2

    3

    4

    5

    6

    7

    8

    9
Age Group (Years)
21-26
0.20
1.07
1.44
1.64
1.68
1.56
1.79
1.65
1.62
40-45
0.37
1 . 20
1.66
1.75
1.86
1.94
1.80
1.84
1.66
'60-72
0.27
1.35
1.93
1.91
1.89
1.78
1.85
1.74
1.44
Baseline

Boom

Recovery
1.51

1.44

1.26
1.43

1.70

1.57
1.26

1.69

r. 77
                                 Mood States
     Scores derived for 15 mood factors and an overall mood inde?c were evaluated by
 analyses of variance. More detailed treatment of this aspect of the study is reported else-
 where by Smith and Hutto (1972). No boom-related effects were obtained.
                                     554

-------
                                    Table 9
          MEAN BSR LEVELS IN KILOHMS PER SUBJECT DURING NINE PERIODS
          OF THE NIGHT FOR BASELINE, BOOM, AND RECOVERY NIGHTS. EACH
           VALUE REPRESENTS A DIFFERENCE SCORE FROM MEASUREMENTS
                       MADE IN THE FIRST 5-MINUTE EPOCH
Period of
the  Night
    3

    4

    5

    6

    7

    8

    9



Baseline

Boom

Recovery
Age Group (Years)
21-26
1.70
2.61
3.13
3.55
3.90
4.34
4.53
4.82
4.92
4.38
4.44
2.35
40-45
7.83
12.03
13.34
13.57
14.40
16.04
15.45
15.96
16.48
20.32
13.80
7.57
60-72
8.23
9.71
9.72
10.23
10.72
11.00
11.29
11.62
12.12
11.17
11.02
9.36
                             Complex Performance
    Mean scores for the ten measures of performance were calculated for morning and
evening sessions for each age group in the Baseline, Boom, and Recovery phases. A detailed
treatment of < this aspect of the study has  been reported elsewhere by Chiles and West
(1972). No decrement in performance was attributable to the booms.
                                     555

-------
                                     Overview
     There were no significant effects of the simulated sonic booms, presented during sleep,
on overall patterns of sleep in comparing Boom nights with Baseline and Recovery night.
There were also no changes in complex performance  measures or assessed moods which
could be attributed to  the booms. However, individual booms did evoke ECG, EMG, and
BSR responses  in all subjects; average heart rate increased during the minute following
booms (by less  than one beat per minute), EMG responses occurred for 45-50 per cent of
the booms (about three times more often than chance  changes might be expected based on
pseudo-stimulus controls) and BSR changed following about 19 per cent of boom presenta-
tions. The frequency of these occurrences increased with the age of the subject-group.
     That the boom-induced responses were  functionally mild is best  attested  to  by  the
fact that nightly patterns of sleep and physiological activity were not significantly affected;
the booms rarely produced  shifts in stage of sleep (about 14 per cent of the time as
compared with  4 per cent for pseudo-stimulus controls) and even more rarely produced
awakenings (about 5 per cent of the time). However infrequent, the occurrences both of
awakening and  of stage shifts increased from the youngest to the oldest age groups, in
agreement with other findings (Lukas, Dobbs, and Kryter, 1970; Lukas and Kryter, 1968,
1970a, 1970b).
     Present results are also in agreement with other studies of noise effects on  sleep  (cf.
Williams,  1970) in that there were  no significant reductions across Boom nights in  the
physiological changes which occurred following boom presentations, nor did EEG measures
change significantly as a result of repeated exposure to the booms; the latter finding agrees
with data reported by Lukas and Kryter (1970a,  b) for simulated booms, by Thiessen
(1970) for truck noise, by  Kramer  et  al. (1971)  for the  striking of a hammer, and as
summarized by Williams (1970)  for  other acoustic stimuli.  This lack of change  is usually
referred to in the literature on effects of noise as a failure  to obtain "adaptation." However,
such results might better be described as a failure to obtain  "habituation" considering the
nature of the test situation and the relatively brief and infrequent presentations of the
acoustic stimuli. Moreover, it would seem that the lack of apparent habituation to at least
some noises  during sleep may  be a characteristic of only  certain  types of physiological
measures; there appears to be enough anecdotal evidence, and some laboratory data (Lud-
low and Morgan, 1972; Lukas and Dobbs,  1972), to indicate that the frequency  and dura-
tion of awakenings to noises repeated  nightly declines with that repetition.


                                   REFERENCES

  1.  Chiles, W. D.,  and West, G., Residual performance effects of simulated sonic booms
     introduced during sleep.  FAA  Office of Aviation  Medicine Report No. AM-72-19,
     Washington, D.C( 1972).
  2.  Collins, W. E.,  and lampietro,  P.  F., Simulated sonic booms and sleep:  Effects of
     repeated booms of  1.0 psf. FAA Office  of Aviation Medicine Report No. AM-72-35,
     Washington, D.C.( 1972).
                                        556

-------
 3.  Dobbs, M.  E., Behavioral responses to auditory stimulation during sleep. J. Sound &
    Vib., 20,467^76(1970).
 4.  Fern, A., and Schwartz, I. R., Sonic boom generation propagation and minimization.
    AIAA Paper No. 72-194, American  Institute of Aeronautics and Astronautics, New
    York, New  York (1972).
 5.  von Gierke, H.  E., Effects of sonic boom on people: Review and outlook. J. acoust.
    Soc. Amer., 39, 543-550 (1966).
 6.  Kramer, M., Roth, T.,  Trindar, J., and Cohen, A., Noise disturbance and sleep. FAA
    Office of Environmental Quality Report No. NO-70-16, Washington, D. C. (1971).
 7.  Kryter, K. D., The effects of noise on man. New York: Academic Press (1970).
 8.  Ludlow, J.  E., and Morgan,  P. A., Behavioral awakening and subjective reactions to
    indoor sonic booms. /. Sound & Vib., 25, 479-495, (1972).
 9.  Lukas, J. S., Awakening effects of simulated sonic booms and aircraft noise on men
    and women. / Sound & Vib., 20,457-466 (1970).
10.  Lukas, J. S., and Dobbs, M. E., Effects of aircraft noises on the sleep of women. NASA
    Report No. CR-2041, Washington, D. C. (1972).
11.  Lukas, J. S., Dobbs, M. E., and Kryter, K. D., Disturbance of human sleep by subsonic
    jet aircraft  noise and simulated sonic booms. NASA Report No.  CR-1780, Washington,
    D. C. (1971).
12.  Lukas, J.  S., and Kryter, K. D., A preliminary study of the  awakening and startle
    effects of  simulated sonic booms. NASA Report No. CR-1193, Washington,  D. C.
    (1968).
13.  Lukas, J.  S., and Kryter, K.  D., Awakening effects of simulated  sonic booms and
    subsonic aircraft noise.  In: B. L. Welch and A. S. Welch (Eds.), Physiological effects of
    noise. New  York: Plenum (1970).
14.  Lukas, J.  S., and Kryter, K.  D., Awakening effects of simulated  sonic booms and
    subsonic aircraft  noise on  six subjects 7  to 72 years of age.  NASA  Report No.
    CR-1599, Washington, D. C. (1970).
15.  Morgan,  P. A., and  Rice, C. G.,  Behavioral awakening in  response to indoor sonic
    booms.  Institute  of Sound  and  Vibration  Research, University  of Southampton,
    Technical Report No. 41 (1970).
16.  Rechtschaffen, A., Hauri, P., and Zeitlin, M., Auditory awakening thresholds in REM
    and NREM sleep stages. Percept. Mot. Skills,  22, 927-942 (1966).
17.  Rechtschaffen, A., and  Kales, A. (Eds.), A manual of standardized terminology, tech-
    niques and  scoring system for sleep stages of human subjects. Public Health Series, U.
    S. Government Printing Office, Washington, D. C. (1968).
18.  Richards, E. J., and Rylander, R., Sonic boom exposure effects  III. Workshop perspec-
    tive. /. Sound & Vib., 20, 541-544 (1972).
19.  Smith, R. C., and Hutto, G.  L., Sonic booms and sleep: Affect change as a function of
    age. Faa Office of Aviation Medicine Report No. AM-72-24, Washington, D. C. (1972).
20.  Thackray,  R. L, Touchstone, R. M., and Jones, K. N., The effects of simulated sonic
    booms on  tracking performance and  autonomic response. FAA Office of Aviation
    Medicine Report No. AM-71-29, Washington, D. C. (1971).
                                       557

-------
21. Thiessen, G. J., Effects of noise during sleep. In: B. L. Welch and A. S. Welch (Eds.),
    Physiological effects of noise. New York: Plenum (1970).
22. Williams, H. L., Auditory stimulation, sleep loss and the EEC stages of sleep. In: B. L.
    Welch and A. S. Welch (Eds.), Physiological effects of noise.  New York: Plenum
    (1970).
23. Williams, H. L., Hammack, J. T, Daly, R. L., Dement, W. C., and Lubin A., Responses
    to auditory stimulation, sleep  loss, and the EEC stages of sleep. Electroenceph. Clin.
    NeurophysioL, 16, 269-279 (1964).
                                        558

-------
            PROLONGED EXPOSURE TO NOISE AS A SLEEP PROBLEM*

                 Laverne C. Johnson, Richard E. Townsend, Paul Naitoh,
                     Navy Medical Neuropsychiatric Research Unit,
                             San Diego, California, U.S.A.,
                                         and
                                   Alain G. Muzet,
                   Centre d'Etudes Bioclimatiques, Strasbourg, France

     *This study was supported, in part, by a special grant from the Naval Undersea Center, and by
Department of the Navy, Bureau of Medicine and Surgery, under Task No. MF12.524.004-9008DA5G.
     The opinions and assertions contained herein are the private ones of the authors and are not to be
construed as official or as reflecting the views of the Navy Department.

INTRODUCTION
     A number of recent studies have been concerned with the effects of both intermittent
and continuous noise on the human sleep  cycle.  The papers being presented during this
Congress represent the  most recent and, in many  instances, the most systematic and best-
controlled studies in this area. The noise stimuli have ranged from continuous white noise at
sound pressure levels as high as 93 dB (Scott, 1972) to the more  common kinds of sleep-
disturbing stimuli such as truck noise (Thiessen, 1970), traffic noise (Schieber et al., 1968),
sonic booms and  aircraft fly-over noise (Lukas & Kryter, 1970; LeVere et al., 1972), and in
the most recent report Globus et al. (1973) observed the effect of aircraft noise on sleep
recorded in the  home.  The  results of the Globus et al.  study  were presented by Dr.
Friedmann as a part of this symposium.
     Because of the expense  and effort involved in long-term exposure to noise and  physi-
ological monitoring of sleep, most studies have been of relatively short duration (e.g., 3 to
16 nights),  with  small numbers of subjects (6 or less), and have used only nocturnal ex-
posure to the  noise stimuli.  In this paper, we will  discuss the effect of 24-hour exposure to
noise stimuli in two laboratory-type controlled environments and during a routine training
cruise. The  first  laboratory study lasted 15  days and involved 15 men; the duration  of the
second  laboratory study was 55 days and involved 20 men. In a separate study, sleep was
examined during a 7-day submarine training cruise and involved 39 men.  It was expected
that these  three  studies would  provide  both controlled  and  realistic  environments for
determining if sleep disruption would result from long-term  exposure  to noise stimuli.
Changes in sleep were viewed as an important indicator of the  subjects' ability to adapt to
such an environment.

METHOD
     The sleep studies were part of a larger study concerned with the behavioral and physi-
ological effects of tone-like  bursts of sound  (pings), when presented 24  hours a day for
     The authors wish to express their appreciation to Robert S. Gales and his staff at the Naval Undersea
Center and to  George E. Seymour  at the Navy Medical Neuropsychiatric Research Unit, San Diego, for their
cooperation and  assistance.

                                         559

-------
sustained periods. In addition to sleep, measures of temporary threshold shifts, performance
on several cognitive and vigilance-type tests, measures of attitudes, and ratings of affect and
mood were obtained.
     During the two laboratory studies, the subjects  were confined to a two-story barracks
building (approximately 2300 square  feet)  containing sleeping, eating, working, and recrea-
tional areas. Over 100 loud speakers were distributed throughout the building to produce a
reasonably uniform sound field (± 3 dB)  throughout the building.
     Each subject completed a sleep log each day detailing the times and duration of sleep
within the past 24 hours and an estimate as to the quality of the  sleep as reflected by
duration  and  time of sleep, estimate of difficulty in falling asleep, time (minutes) to fall
asleep, number of awakenings during sleep, ratings as to how rested  the subject  felt upon
awakening, and whether he felt he could have used more sleep.
     In addition to the sleep log data,  all-night electrophysiological monitoring was obtained
from  selected  subjects during  the laboratory studies.  These data  included  right and left
electrooculogram  (EOG), electroencephalogram  (EEC) €3 - Aj + A2,  electrocardiogram
(EKG), skin potential (SP), and  finger pulse volume  (FP). These variables plus a time code
and the pings were recorded on both polygraph paper  and  on FM magnetic tape.
     The all-night EEC sleep records  were scored by  a digital computer program (Martin et
al.,  1972) using  the  standardized criteria of the Association for the Psychophysiological
Study of Sleep (APSS) (Rechtschaffen & Kales, 1968).
     As a check on the validity of the computer scoring  of sleep stages,  four records from
the  15-day study and six records from the 55-day  study were scored  manually by  two
trained sleep stage scorers, using the same APSS criteria. The overall agreement between the
computer and human scoring was  84.4%,  with individual agreement ranging from 79% to
88%. These figures are representative of the agreement obtained between two human scorers
and consistent with the 82% agreement reported by Martin et al. (1972).
     The polygraph records  were also scored manually for body movements and autonomic
responses to pings. All-night median  heart rate  was  obtained from a heart rate histogram
analysis of the EKG using  a computer of averaged  transients. During the  55-day study,
auditory-evoked EEC responses were also computed.
                               FIFTEEN-DAY STUDY

Procedure
     Twenty Navy Male volunteers, mean age 20.3, range 17-32, comprised the test popula-
tion. Following three days of baseline, subjects were exposed to pings of 0.75 seconds with
an interstimulus interval (ISI) of 45 seconds over 24 hours for 15 days. The pings were in
the 3 - 4 KHz region with an intensity of 80 dB SPL for the first 5 exposure days and 85 dB
SPL for the remaining 10 exposure days. There were 3 post-exposure recovery days.
     All 20 subjects completed the sleep logs and on four subjects all-night sleep EEGs were
obtained  during baseline, during ping exposure, and during  recovery.  Two subjects were
recorded  each night, which meant each subject was recorded every other night of the 21-day
experimental period.                                                 ;

                                        560

-------
Results
     As the 55-day study involved the longest exposure to the noise and included 80, 85,
and 90 dB levels, a detailed analysis of the 55-day study data will be presented but only a
summary of the results of the  15-day study.  The 15-day study was viewed as a pilot study
for the longer 55-day study.
     There were no significant-either statistical or practical-sleep effects from the 24-hour
exposure to the 80 or 85 dB SPL pings during the 15-day study. The average sleep time over
all subjects and over all days was 6.6 ±1.2 hours. When daytime naps are included, the mean
was 6.8 ± 1.5  hours of sleep over a 24-hour period. There was no significant change in total
sleep time when baseline, ping exposure,  and recovery sleep durations were compared. None
of the  usual sleep measures, e.g.  total sleep time, total movement time, sleep onset latency,
percent time spent in the various sleep stages, or number  of sleep  stage changes, varied
significantly during the 21-day test period in the four subjects from  whom all-night EEGs
were obtained.
     Two subjects showed a decrease in percent-time of stage-4 sleep on  all ping nights
relative to baseline and recovery  sleep, while the other two showed decreases on some ping
nights  but not on others. These latter two subjects had unusually low stage-4 percent during
baseline, suggesting an adaptation problem and a need for a longer pre-ping baseline period.
In support of this hypothesis was the higher stage-4 percent on recovery nights for these two
subjects; a percent higher than that seen  on baseline or on any ping night. Two subjects
showed a decrease in  REM sleep on all ping nights, and the other two showed no consistent
change relative to either baseline or recovery sleep.
     Analysis of changes in heart rate (HR), finger pulse amplitude response (FPR), and
EEC activity  (K-complexes during stage 2) indicated significant responses in all three
measures during sleep. There was no extinction of the  responses during  sleep over the  15
days of ping exposure. As in  previous studies (Johnson & Lubin,  1967),  the autonomic
responses were not  seen  before sleep  onset  while the  subject was awake,  reflecting
habituation during awake to the pings, but on each night the responses returned with sleep
onset.

                           THE FIFTY-FIVE DAY STUDY

Procedure
     Subjects were 20 Navy enlisted men, mean age  20.7, range from 18 to 33 years. As in
the  15-day study,  subjects were medically and  psychologically  examined before the
experiment and found to be normal.
     The  ping was of 660  msec  duration,  approximately  3.5 KHz,  presented every  22
seconds on a 24-hour per day  basis for 30 days. There were 15 days of baseline, 30 days of
ping exposure, and a  10-day recovery period. The 30  days of exposure were divided into 10
days each at 80, 85, and 90 dB SPL of ping intensity. Background sound level was about 70
dB during the daytime and 50 dB at night in the berthing areas.
     Ten  subjects were  randomly selected from the total  population of 20  subjects  for
electrophysiological monitoring.  These subjects were paired randomly and each pair slept in
the  sleep recording room every fifth night. This arrangement maximized  the number of

                                        561

-------
subjects from whom data could be obtained, while still permitting at least two data points
for each subject under each of the experimental conditions.
     To confirm the validity of the sleep log data, a night-by-night comparison was made of
total  bed  time  reported  on   the  sleep logs  with  that  obtained  on  nights   of
electrophysiological sleep recordings  for the 10 subjects with electrophysiological sleep
recordings for the  10 subjects with  electrophysiological sleep data.  As in previous studies
(Naitoh et al., 1971), there was no significant1  difference in  sleep time between the sleep
log data and the electrophysiological recording.
     The subjects  were  required to be in bed  by 2400 and to rise by 0700 on  Sunday
through Friday nights. Most subjects were in bed by 2330. On Saturday nights, ad lib. sleep
was permitted. As  a result, many subjects stayed up most of  Saturday night playing cards,
reading, or talking. Because of these short hours of sleep, unrelated to the noise stimulus,
Saturday nights were excluded from the analysis. (A more detailed description of procedure
and results of the 55-day study can be found in Townsend et al., in press.)

Results
     For the sleep log analysis,  the only significant result  was the  subjective report of
greater difficulty in falling asleep during 85 dB, 90 dB, and post-ping nights compared to
pre-ping baseline nights (see Table 1). The  80, 85, and 90 dB, and post-ping conditions did
not differ significantly from each other.
     Compared to the pre-ping baseline, there were no changes in the percent time for either
awake (W) or stage 1 sleep during ping exposure (see Table 2). The  changes in stages 2, 3,
and REM were not consistent over the  three dB levels, and  those  changes that were
statistically significant were  not considered large enough to be of practical significance.
Under the 85 dB condition, percent time for stage 2 was significantly decreased.  Percent
time for REM sleep was significantly increased for both 80 and 90 dB. The 85  dB  and
post-ping periods did not differ from the pre-ping baseline for REM sleep. For stage 4 sleep,
there was a decrease  in percent  time  under all tone conditions,  and this decrease  was
significant under the 90 dB and post-ping conditions. With the decrease in stage 4, there  was
an increase in stage 3 sleep. The increase in stage 3 was significant during exposure to the 85
dB pings.
     None of the measures of quality of sleep (i.e., number of nocturnal arousals, number of
stage changes, and  sleep cycle  stability) showed any consistent or significant change with
ping intensity (Table 3).

Body Movements during Sleep
     There was a small but significant increase  (p < 0.02) in number of body movements
during  REM sleep,  but total number of body  movements over all sleep stages during  the
night did not increase during exposure nights. Only 3.7% of the pings during stage 2 were
followed by movements and 5% in stage REM. Many of the movements that did occur during
exposure nights, however, were clearly related to ping onset. In stage 2 sleep, 53.7% of all
body movements followed the pings by  < 7 seconds. During REM sleep, 45.3% of all body

1 Significant refers to p < 0.05 using a zero-mu t test, two-tailed.

                                       562

-------
                                                     Table I

                             COMPARISON OF SLEEP MEASURES FROM SLEEP LOG CARDS AND
                              NAP LOG CARDS DURING BASELINE AND EXPOSURE TO PINGS


. Mean
Difficulty falling asleep
S.D.
Mean
Number awakenings per night
S.D.
Mean
Median hours nocturnal sleep
S.D.
Mean
Mean hours of naps
S.D.

Pre-Exposure
Baseline
1.80
0.51
0.55
0.40
7.64
0.60
0.30
0.35

80 dB
2.03
0.59
0.51
0.46
7.65
0.30
0.37
0.48

85 dB
2.10*
0.58
0.54
0.39
7.40
0.06
0.35
0.10

90 dB
2.15**
0.55
0.65
0.41
7.45
0.21
o.so
0.48
Post-
Exposure
Period
2.13**
0.48
0.63
0.48
7.51
0.69
0.31
0.40
U>
     *p_ < 0.05
    **£ < 0.01

     Quantification of difficulty - 1 = none,  2~- = slight,  3 = moderate, 4 = considerable

-------
                                     Table 2
             PERCENT TIME FOR SLEEP STAGES FROM COMPUTER-SCORED
       ELECTROENCEPHALOGRAPHIC AND ELECTRO-OCULAGRAPHIC RECORDINGS

Sleep Stage
Mean
Awake
S.D.
Mean
Stage 1
S.D.
Mean
Stage REM
S.D.
Mean
Stage 2
S.D.
Mean
Stage 3
S.D.
Mean
Stage 4
S.D.

Pre-Exposure
Baseline
3.50

3.19
2.93
2.00
21.75
4.97
59.88
8.10
8.88
3.44
10.37
6.17

80 dB
2.66

2.18
2.19
0.99
26.26*
5.71
56.88
S.34
7.91
3.40
6.77
7.71

85 dB
3.06

2.68
1.83
1.21
24.54
4.48
54.96*
4.21
11.56*
4.69
6.92
5.69

90 dB
5.74

6.30
2.48
1.39
24.66*
4.51
57.03
6.87
11.19
3.73
4.69**
5.90
Post-
Exposure
Period
3.28

4.97
3.16
2.09
23.68
6.93
57.30
5.90
-9.44
4.79
5.97*
6.61
   * p_ < O.OS
  ** p_ < 0.01
movements followed the ping by < 7 seconds. Application of the pseudostimulus technique,
where  pseudostimuli are imposed  on  the  baseline record, indicated  that  there were
significantly more movements  in the first 7  seconds after the stimulus than in  the same
interval after the pseudostimulus. A more detailed analysis of the body movements has been
reported by Muzet et al. (in press).

Median Heart Rate during Sleep
     The average all-night heart rate  under  each  experimental condition did not differ
significantly from the pre-ping baseline.
                                       564

-------
                                     Table 3

          VARIABLES REPRESENTING GOODNESS OF SLEEP AND STABILITY OF
       SLEEP CYCLING FROM COMPUTER-SCORED ELECTROENCEPHALOGRAPHIC
                    AND ELECTRO-OCULOGRAPHIC RECORDINGS

Sleep Parameter
Mean
Number arousal episodes
S.D.
*Mean
Number of stage changes
S.D.
Mean
Time1 to 1st stage 2
S.D.
Mean
Time to 1st stage 3
S.D.
Mean
Time to 1st stage 4
S.D*
Mean
Time to 1st stage REM
S.D.
Average REM-REM interval Mfian
in minutes _ n
o • u •
Pre-
Exposure
Baseline
5.85
6.36
33.85
11.26
14.80
13.34
39.35
39.49
157.50
159.60
96.65
37.52
78.49
20.40

80 dB
3.95
4. 88
33.20
7.07
7.30
6.30
50.50
56.02
153.40
156.36
81.54
34.53
80.31
15.92

85 dB
3.60
4.61
32.30
8.30
11.70
11.41
30.60
11.05
93.70
120.78
95.20
30.69
87.14
16.50

90 dB
3.70
2.62
32.95
6.80
14.60
18.19
43.60
24.72
147.30
112.28
90.80
25.91
83.33
20.97
Post-
Exposure
Period
4.70
5.34
31.65
5.65
13.20
15.92
44.50
28.21
125.10
104.81
88.20
39.05
88.26
24.18
      times are in minutes
Individual Subject Analyses
    The data for each variable were examined to determine if any subjects showed a greater
ping effect  than that suggested by the group mean. For the group of variables associated
with goodness of sleep (sleep onset time, number of stage changes, number of awakenings,
and sleep cycle stability), only one of the 10 monitored subjects showed a consistent,
although  small, decrement in quality of sleep with increases in ping intensity. However, this
subject did  not show any  decrement in waking performance (visual  and auditory vigilance,
                                       565

-------
choice reaction time and recognition memory) (Hershman & Lowe, 1972), or behavior
(mood and anxiety scales).  Changes in median heart rate during sleep were seen in two
subjects, one of whom had an average increase of 5 bpm at 85 and 90 dB, the other had an
average increase of 5 bpm at 90 dB. Again, no waking correlates were noted.
     As in the 15-day study, HR, FPR, and EEC responses occurred to the tones during all
ping-exposure nights. To  further determine the  EEG response to  the  pings, a detailed
analysis of evoked EEG response to the pings during stage 2 and REM was made (Townsend
& House, 1973). The amplitude of the evoked response was consistently larger during stage 2
than stage REM. Williams et al. (1962) have also reported lower-amplitude evoked responses
during REM sleep. Similar to the evoked HR and FPR responses, the EEG evoked response
was not detectable before sleep onset but appeared each night with sleep onset. In order-to
determine whether  any change in response amplitude had  occurred  over the 30-day
exposure period, a comparison was made between the average amplitude of the response of
the 10 subjects during their first recorded night of exposure to ping and during their last
recorded night of ping exposure. These results are presented in Figure 1 for stage 2 and in
Figure 2 for REM sleep.
     The early components (< 200 msec) of the evoked response were unchanged over the
exposure period. The only  significant differences from first to last  recorded night were
similar decreases in the amplitude of the N2~?3 component in both stage 2 and REM sleep
and a decrease in  the amplitude of the ?3~N3 component in REM  sleep. There were no
consistent or significant changes in either the amplitude of the other components or in the
latency of any component. While it appears that there may be a reduction in the amplitude
of some later  components of the auditory evoked response in both stage 2 and REM sleep
during prolonged exposure to pings, there remains a striking similarity  between the first and
last night evoked responses.
                           TRAINING CRUISE STUDY

Procedure
    Speakers were installed in the forward half of a diesel-powered submarine so that a
relatively uniform sound field was present in all areas of the forward compartments. No
speakers were installed in the remainder of the submarine.  This arrangement of speakers
resulted in two groups: 1) The control group, N = 17, who lived and worked in an area free
of pings; 2) The experimental group, N = 22, who lived and worked in the ping area. There
was no significant difference in age or rank between the two groups. The mean age was 25
with a range of 17-43. The subjects were not confined to their respective areas, but they
were encouraged to remain in their assigned quarters. The ping duration was .75 seconds in
the 3-4 KHz region. The intensity was set at 80 dB SPL with a 22 second IS'I and was
presented 24 hours a day for five days. The median background sound  level was 66 dBA
with higher levels in some areas such as the messing areas and lower in the sleeping quarters.
    Sleep logs were completed one day prior to the ship's departure from home port while
the crew members were on leave sleeping in "at home" environments, on the first day of the
cruise before the pings were started, during the five ping-exposure days, and on the first day
after the pings were discontinued.

                                       566

-------
                             Stage 2
                                            First Night
          100    200    300    400     500    600    700    800
                           Time (msec.)
Figure 1:  Changes in ping-evoked EEC response over 30-day exposure period during stage 2 sleep.
                               567

-------
   6
   5
   4
   3
   2
©
a
E
  —rt
  -3
  -4
  -5
  -6
                                 REM
,..% Last Night
            100    200     300    400    500    600    700    800
                             Time (msec.)
 Figure 2:  Changes in ping-evoked EEG response over 30-day exposure period during stage REM sleep.
                                 568

-------
Results
     Compared to "at home" sleep durations, both groups obtained more sleep during the
cruise. The  average "at home" total sleep duration was 6.6 ±1.9 hours. During the first
night of the cruise, it was 9.6 ± 2.6 hours, during ping exposure the average was 7.6 ± 1.5
hours, and on the first night after the ping was discontinued the average total sleep duration
was  6.5 ±  1.9 hours. There  were no  significant differences between  the  two groups in
average total sleep duration for any night or change  in sleep  duration  during the cruise.
Factors other than ping exposure, thus, were  major determinants of sleep  duration. The
total  sleep times for our crew were compatible with those usually seen on a cruise of this
nature. There is usually reduced sleep on the last night in port, an increase on the first night
at sea, with a subsequent  decrease as the operating duty and watch schedules are imposed.
The  shortened sleep duration  on the night after cessation of pings was due to the increased
work schedule to prepare for return to port.
     Even though ping exposure had no  significant influence  on total  sleep duration, the
pings appeared to be  a factor in  reported difficulty in getting to sleep. The experimental
subjects reported more trouble going to sleep, and this was reflected in their increased sleep
latency on ping nights. In Figure 3 are the changes in sleep latency on ping nights. In Figure
3  are the changes in  sleep latency for  the two groups when the time to fall asleep during
ping and the night following  cessation  of pings are compared to that for the first night at
sea. There was  a consistent increase in sleep latency over the ping nights, and this increase
was  significantly (t = 3.88)  longer  than  that for baseline. Thirty-two percent of those
exposed to the pings reported increased sleep latency on the first ping night and 68% on the
fifth  night  of  ping when compared to their  first  night  at sea. Forty percent reported
increased latency on the fifth night when compared to their sleep latency on the first night
of ping. Eleven (50%) of the subjects showed sleep latencies double their  baseline sleep
latencies on two or more of the ping nights.
     In contrast, the control subjects' sleep latencies generally  decreased during the cruise,
except  on ping  nights 3  and 4, but none of  these changes in latency were significantly
different from  baseline. Though the mean change was a decrease in time  of sleep  onset,
some control subjects reported an increase in sleep latency during the cruise though never
exposed to the pings. Compared  to their baseline, 18% of the control subjects reported
increased sleep latency on the first ping night (2nd day of cruise) and 36% said it took
longer to fall asleep on the 6th night of the cruise than on the first night. Compared to the
50% of the experimental subjects  with increased sleep latencies on two or more ping nights,
only  three (18%) of the control subjects had longer sleep  latencies on two  or more nights
during the cruise than on baseline nights.
     The increased sleep latencies on ?3 for the control group were probably due, in part, to
the fact that the submarine returned to port for a couple of hours and many of the control
subjects went ashore on various errands. The experimental subjects were restricted to their
quarters. Similarly, the sleep  latencies  on ping night 4 were also partly influenced by the
day's activities. During the 4th day, the boat anchored off a vacation island and the control
subjects had a  steak-fry  topside  while observing the  vacationers  through glasses.  The
experimental subjects were again restricted to quarters and had to eat their steaks in their
mess area. The  marked increase in  latency for the experimental subjects on ?4 might be due,


                                         569

-------
o
"5
 fl>
 0)
 C
 a
u
 w
 C
 a
 o
£
to
     X
     w
     c
     o
     TJ
     a>
     w>
     a
     0)
      IB
      15
      14
      13
      12
      II
      10
      9
      8
      7
      6
      5
      4
      3
      2
       I
      0
0 X
V> V   n
O C -2
V 0)
L. *-   n
W O ~"
a> Jj
Q   .4

     -5
                                                               •
                                                               »*
                                                                               Experimental
                                                                            -..^Subjects
                                                                                **••
                                         •••
                                        »    *••
                                                      *
                                                      '
                                                     *
                                                                    .VVX'N.XN.NVN.NXXXXX'VWW
                                                                                    Control
                                                                                    Subjects
                Baseline
                                                   P3
                                              Nights
                                                                     Pg     Recovery
     Figure 3:  Changes in sleep latency during 5-day exposure to pings and during recovery compared to sleep
              latency during baseline.

-------
in part, to their "confinement" during the day. Examination of duty and watch assignments
for the two groups revealed no explanation for the difference in sleep latency.
     While  the overall pattern showed longer sleep  latencies  for  the experimental than
control subjects   and within-group  significant  changes  for the  experimental group,
between-group comparisons indicated significant  differences  for only ping nights 2 and 4
and on the night after the pings were turned off.
     The experimental subjects also  reported feeling  less rested upon awakening, but the
difference between the two groups was significant only after the 5th ping night.

DISCUSSION
     Contrary to  the findings of other studies using  different noise stimuli  and only
nocturnal exposure, little effect of noise on sleep duration and number of awakenings was
seen in these investigations. One of the  possible reasons for the difference in results might
have been the pre-sleep exposure  of our subjects to the 24-hour regular pattern of pings. All
subjects were exposed to the pings for a minimum of 10 hours before the first night of
sleep. Sounds with aperiodic rates of occurrences, with variable frequencies and intensities,
have been  reported  as causing alterations  in sleep  which included longer  sleep latencies,
increased wakefulness, and body movements (Lukas & Kryter, 1970; Schieber et al., 1968).
     The relatively younger age of our subjects may have also been a  factor, as Lukas and
Kryter (1970), Thiessen (1970), and Williams (1970) have  reported that more  frequent
awakenings occurred in the older age subjects.
     Our results, however,  are consistent with the general findings that noise results in sleep
onset complaints and a decrease in sleep stage 4.  In contrast to the report by Scott (1972),
we found no decrease in stage REM during ping exposure. A decrease in delta sleep has been
reported by  Roth et al.  (1972) and by Globus et  al. (1973). Subjects exposed to 85-and
95-dB-SPL pings during the 55-day laboratory study  and to 80-dB-SPL pings during the
7-day training cruise reported sleep onset difficulties.
     In both the  55-day and training cruise studies,  the sleep latencies did not return to
baseline values when the pings were turned off, suggesting a carry-over of the sleep onset
difficulties. The average increase  in sleep latency was less than 15  minutes in each phase;
thus, one could reasonably question whether this increase was of practical significance. This
question is particularly pertinent since there was no change in  total sleep duration as a result
of the pings in either phase of the study. Further, the EEC recordings in the 55-day study
showed no EEC change in sleep latency over the 30 days of ping exposure even when the dB
level was 90. We appear,  thus, to  be  dealing with a subjective report which was not verified
by objective  EEC recordings of sleep onset. Similar  discrepancies between reported sleep
problems and EEC  sleep  recordings, particularly  for complaints of insomnia,  are often
found in sleep disorder clinics.
     Before  the subjective report  of sleep onset difficulty can be dismissed  as  of no
significance,  however,  attention must be directed to  the finding that during  the training
cruise 23% of the experimental subjects reported lying  in bed for over an hour, and two
(10%)  for  2 hours,  before sleep onset during one or more ping nights.  Baseline  sleep
latencies for  these subjects were less than 20 minutes. No such marked  increases were found
for  control  subjects. Also, the experimental subjects reported  feeling  less rested upon

                                         571

-------
awakening. Subjects with the longer sleep-onset latencies  invariably reported feeling only
"slightly rested" or "not at all" after morning awakening. In all probability, these long sleep
latencies were due in part to other factors and events of the previous day. When sleep is
difficult, one becomes more aware of noises. The importance of physical and emotional
health in determining the response to noise has been mentioned by Harold Williams in his
summary paper.
     The inability of sleep researchers to identify useful indices for goodness of sleep
contributes to the problem of determining whether a reported change in sleep is significant
(see Johnson, in press).  At present, the report by the subject  to questions relating to sleep
onset problems, and how rested he felt upon awakening and how well he slept, may be our
most adequate measures of goodness of sleep. If we accept these subjective reports as indices
of goodness of sleep, then the pings had  an adverse effect on sleep. Both of the reported
sleep difficulties—sleep onset problems, and awaking less rested-suggest that the presence of
noise necessitates increased effort if sleep is to be  obtained. The subject  feels he has to
"work harder" to go to sleep, and an increased effort may  be necessary throughout  the
night to remain asleep. The finding that there is an autonomic (ANS) and EEC response to
the pings throughout the night without any sign of eventual  extinction indicates that  the
subject is not able to "tune out" the pings during sleep as he is able to do when awake. The
decrease in stage 4 sleep as he is able to do when awake. The decrease in stage 4 sleep during
the 55-day study, and noted by Roth et al. (1972) and Globuset al. (1973),  indicates that
there is  a  decrease in "deep sleep" during noise exposure. Some sleep researchers would
posit a causal relation between amount of stage 4 sleep obtained and the recuperative value
of the sleep, but this relationship has been difficult to demonstrate objectively (Johnson, in
press).
     One  final  point regarding  the significance of  the reported  sleep problems.  No
performance decrement  during the 30-day exposure to pings was found on an extensive
battery of cognitive, reaction time, and vigilance-type tests (Hershman & Lowe, 1972). Also,
no changes in mood orattitudes were found during the 55-day study, except for the increase
in the annoyance value of pings from a rank order of  16 out of 20  factors when the ping
level was at 80 dB to a rating of 1 (the most irritating) when it rose to 90 dB. Similarly, no
significant performance differences between  experimental and control groups were found
during the training cruise. Like the 80-dB ping of the 55-day study, the 80-dB pings during
the training cruise were 17th in a list of 20. Other factors such as lack of showers, boredom,
and lack of exercise were much more irritating to the submarine subjects than the pings. The
pings, however, were not completely innocuous as temporary threshold shifts were found in
some subjects in each study. But even if we are correct in  our hypothesis that subjects
expend more effort to obtain their sleep when exposed to noise, the expenditure of  this
extra effort may not be easily detected in waking performance or behavior. Information as
to other possible areas that might be affected and the changes that might be expected  as a
result of exposure to noise will, perhaps, be one of the contributions of this Congress.

                                    SUMMARY
     In  one 15-day and one  55-day laboratory study  and one operational 7-day training
cruise, the effect on sleep of 24-hour-a-day exposure to pings  of intensities ranging from 80

                                        572

-------
to 90  dB SPL was  examined. The pings were  less than a second in duration with an
interstimulus interval of 45 or 22 seconds, and in the 3-4 KHz frequency range. Maximum
duration of ping exposure was 30 days. In this young adult sample, exposure to the noise
did not produce a decrease in sleep duration or an increase in number of awakenings. There
were, however, reports of sleep onset difficulty and a decrease in percent of sleep stage 4
during ping exposure. No significant changes in waking performance or behavior were found
as a result of the ping exposure during any of the three studies.
                                   REFERENCES

Globus, G., Friedmann, J., Cohen, H., Pearsons, K. S., and Fidell, S., Effect of aircraft noise
     on sleep as recorded in the home. Paper presented at the  13th annual meeting of the
     Association for the Psychophysiological Study of Sleep, San Diego, Calif. (May 1973).
Hershman, R. L.,  and Lowe, T. D., Criteria for airborne noise in submarines: The Project
     PING Phase II human performance tests. Naval  Electronics Laboratory Center, San
     Diego, Technical Note 2045 (1972).
Johnson,  L.  C, Are  stages of sleep related to  waking behavior? American Scientist (In
     press).
Johnson,  L.  C.,  and  Lubin, A., The orienting  reflex  during  waking  and  sleeping.
     Electroenceph. din. Neurophysiol, 22, 11-21 (1967).
LeVere, T. E., Bartus, R. T., and Hart, F. D., Electroencephalographic and behavioral effects
     of nocturnally occurringjet aircraft sounds. AerospaceMed., 43, 384-389 (1972).
Lukas, J.  S.,  and Kryter, K. D., Awakening effects of simulated sonic booms and subsonic
     aircraft  noise. In B.  L. Welch and A. S. Welch (Eds.), Physiological effects of noise.
     New York: Plenum Press, 283-293 (1970).
Martin, W. B., Johnson, L. C., Viglione, S. S., Naitoh, P., Joseph, R. D., and Moses, J. D.,
     Pattern  recognition  of  EEG-EOG as a  technique for all-night sleep stage  scoring.
     Electroenceph. din. Neurophysiol, 32,417-427 (1972).
Muzet, A. G., Naitoh, P., Johnson, L. C., and Townsend, R. E., Body movements in sleep
     during 30-day exposure to tone pulse. Psychophysiology (In press).
Naitoh, P., Johnson,  L.  C., and Austin, M.,  Aquanaut sleep patterns during Tektite I:  A
     60-day habitation under hyperbaric nitrogen saturation.  Aerospace Med.,  42, 69-77
     (1971).
 Rechtschaffen, A., and Kales, A.  (Eds.), A  manual of standardized terminology, techniques
     and  scoring system  for sleep stages of human  subjects. Washington,  D.  C.: U.  S.
     Government Printing Office (1968).
 Roth, T., Kramer, M., and Trinder, J., The effect of noise during sleep on the sleep patterns
     of different age groups. Canad. Psychiat. Ass. J., 17, 197-201 (1972).
 Schieber, J. P., Mery, J., and Muzet, A., Etude analytique en laboratoire de I'influence du
     bruit sur le sommiel. Centre d'Etudes Bioclimatiques du CNRS, Strasbourg, France,
     report (1968).
 Scott, T. D., The effects of continuous, high intensity, white noise on the human sleep
     cycle. Psychophysiology, 9, 227-232 (1972).

                                        573

-------
Thiessen, G.  J., Effects of noise during sleep. In B. L. Welch and A. S. Welch (Eds.),
    Physiological effects of noise. New York: Plenum Press, 27*1-275 (1970).
Townsend,  R. E., and House, J. F., Auditory  evoked potentials in stage 2 and rapid eye
    movement sleep during a 30-day exposure to tone pulse noises. Paper presented at the
    29th annual meeting of the Western EEC Society, San  Diego, Calif. (February 1973).
Townsend,  R. E., Johnson, L. C, and  Muzet,  A., Effects of long term exposure to tone
    pulse noise on human sleep. Psychophysiology (In press).
Williams, H. L., Auditory stimulation, sleep loss and the EEC stages of sleep. In B. L. Welch
    and A.  S. Welch (Eds.),  Physiological effects of noise. New York:  Plenum Press,
    277-281  (1970).
Williams, H.  L., Tepas,  D. I., and Morlock, H. C., Jr., Evoked responses  to clicks and
    electroencephalographic stages of sleep in man. Science,  138, 685-686 (1962).
                                        574

-------
            RELATIONSHIP BETWEEN SUBJECTIVE AND PHYSIOLOGICAL
                    ASSESSMENTS OF NOISE-DISTURBED SLEEP

                       A. Muzet, J.P. Schieber, N. Olivier-Martin
                                 J. Ehrhart & B. Metz

                        Centre d'Etudes Bioclimatiques du CNRS
                       21, rue Becquerel - 67087 Strasbourg Cedex
                                      FRANCE

This study was supported by contract n° 690/523/DGRST - Urbanisation. The authors thank J. P. Lienhard
and B. Korosec for their analysis of data. The authors are grateful to the Centre d'Etudes Bioclimatiques
members who were involved in this experiment.

     The aim of this study was to explore the effects of jet aircraft take-off noises on the
sleep of young adults in good health of both sexes, not habitually exposed to this kind of
noise. Several  studies on noise-disturbed  sleep, and particularly by jet flyover noise, have
been already published (2,  3, 4, 5,  6, 8, 9, 11, 12). Some of them  showed differences
between  male  and female subjects, but the age classes often were different. Furthermore,
the noise effects depend upon the sleep  stage in which the noise occurs.  Finally, authors
disagree in regard to the noise threshold and the noise effect during REM sleep (10, 14).

Methods:
     Eighteen  young adults of both sexes (9 males and 9 females), in the age range between
19  and 24, stayed  permanently  in the laboratory during 4  consecutive nights and days.
During  the third night  of each sequence, 32  jet take-off noises were  presented in a
semi-random schedule.
     The second and the fourth nights were not disturbed by noise, while the first night was
rejected from the analysis.
     Each morning, after awakening, subjects had to respond to a sleep questionnaire used
in order to  explore subjective modifications, if any, in relation with the noises (9).
     Two EEGs (central-mastoid  and parietal-frontal leads), one EOG, one EKG and one
actogramm detected by an original procedure (7) were continuously recorded from 23:00 to
07:00.  Noise  pressure level  in  the experimental  rooms  and a time  code were  also
continuously recorded  during 8 hours. The records were analysed  visually  and a sleep stage
score was given for every 10 seconds section of recording.

Results:
1.     Sleep stage latencies
     The latencies of sleep stage 1, 2, 3,4 and of REM were expressed from the start of the
experiment to the first occurrence of each sleep stage. The sleep stage latencies during the
three experimental  nights (Nl and N3 = nights without noise, and respectively second and
fourth nights  in the laboratory; N2 = night with noises and third night in the laboratory)
were compared by  zero-Mu t^ test for correlated means. Table 1  shows values of zero-Mu t:
test between nights.
                                        575

-------
                                    Table 1

               SLEEP STAGE LATENCIES = VALUES OF ZERO-MU T TEST
           N1 AND N3 = NIGHTS WITHOUT NOISE - N2 = NIGHT WITH NOISES

                 (A) Value  for (Nl  -  N2) difference
                 (B) Value  for (N2  -  N3) difference
                 (C) Value  for (Nl  -  N3) difference
Sleep stage
latencies
Stage 1


Stage 2


Stage 3


Stage 4


REM


* 5 7. or
** 1 % or
Males
(A)

-------
     The average latencies of each sleep stage, for the two groups, during nights Nl, N2 and
N3 are given in the Table 2.
     Subjectively, both males and females estimated that their time to fall asleep was longer
during N2 than during the  two other nights. Zero-Mu t tests  applied to  the time to  fall
asleep as estimated by  the male and female subjects show that this estimated time is
significantly longer during N2 than during Nl and N3. This time estimated by the subjects
and the  stage 1  latency are not significantly different, with the exception of disturbed night
N2 for which male subjects overestimate their time to fall asleep.

2.    Total duration of the sleep stages
     The total duration of  each  sleep stage was obtained  by adding  the 10-sec epochs in
which a same stage was scored. For wake (W) only the time in W after the first stage 1 was
retained. These  total durations during the three nights were compared by zero-Mu ^tests for
correlated means. The results are reported in Table 3.
     For the males, the total time  in W is  significantly longer during the disturbed night
than  during  the  subsequent  undisturbed one N3. Also  the  total  time in stage 3 is
significantly longer during N3 than during Nl.
     Total time spent in REM sleep is significantly longer during N3 than during N2, for the
males, and the same result is found when we consider both groups together.
                                      Table 2

            AVERAGE LATENCIES (IN MINUTES) OF THE SLEEP STAGES DURING
               THE TWO NIGHTS WITHOUT NOISE (Nl, N3) AND THE NIGHT
                                 WITH NOISES (N2)
Sleep
Stage
1

2
3
4
REM
Males
Nl N2
19

23
31
38
117
30

37
60
65
163*
N3
17

21
28
35
119
Females
Nl N2 N3
17

20
30
36
96
23

28
37
42
119
16

20
29
33
111
                  * One  subject exhibited his  first  REM
                    period 363  mn after the  start of the
                    experiment  and another one had it 294
                    mn after  the start.
                                       577

-------
                              Table 3

     TOTAL DURATION OF THE SLEEP STAGES: VALUES OF ZERO-MU T TEST
        N1 AND N3: NIGHTS WITHOUT NOISE - N2: NIGHT WITH NOISES
     (A) Value for (Nl - N2) difference
     (B) Value for (N2 - N3) difference
     (C) Value for (Nl - N3) difference

     W duration is the time spent in wake  after  the  first  stage 1
Sleep stage
(Toi.al duration)
W


Stage 1


Stage 2


Stage 3


Stage 4


REM


(A)
(B)
(C)
(A)
(B)
(C)
(A)
(B)
(C)
(A)
(B)
(C)
(A)
(B)
(0
(A)
(B)
(0
Ma
-2
2
1
-1
1
1
1
-1
0
-1
-0
-2
1
-0
0
0
-4
-1
les
.15
.90
.59
.33
.88
.28
.13
.40
.10
.14
.96
.58
.05
.25
.89
.98
.09
.62
Females
NS
*
NS
NS
NS
NS
NS
NS
NS
NS
NS
*
NS
NS
NS
NS
4*
NS
(A)
(B)
(C)
(A)
(B)
(C)
(A)
(B)
(C)
(A)
(B)
(C)
(A)
(B)
(0
(A)
(B)
(C)
-1
0
-1
-1
0
-0
0
-0
0
0
1
1
1
-0
0
0
-1
-0
.74
.47
.10
.79
.72
.50
.31
.20
.80
.32
.55
.88
.13
.89
.18
.12
.02
.83
NS
NS
NS
NS
NS
NS
NS
NS
NS
NS
NS
NS
NS
NS
NS
NS
NS
NS
Malos -*- Fornl.os
(A)
(B)
(C)
(A)
(B)
(C)
(A)
(B)
(C)
(A)
(B)
(0
(A)
(B)
(0
(A)
(B)
(C)
-2.79
2.49
0.51
-2.42
1.88
0.59
1.01
-0.94
0.07
-0.87
0.50
-0.23
1.59
-0.88
0.71
0.87
-2.99
-1.73
*
*
NS
*
NS
NS
NS
NS
NS
NS
NS
NS
NS
NS
NS
NS
±±
NS
*  5 T, or better
** 1 Z or better
                               578

-------
     For the females, there is no significant difference for the total duration of every sleep
stages in the disturbed and non-disturbed nights.
     The average total durations of each sleep stage during the three experimental nights are
given in Table 4 for the two groups.
     Subjectively, males  and females estimate  that the sleep quality  was worse during the
disturbed night than during the two others. For the males, it seems that N3 was the night
with the best sleep quality, while for the females the best night seems to  be Nl. Such an
estimation should be  compared with the values given in table 4, which show that the total
duration of W and stage  1 is  smaller in night N3 than in night Nl  for the  males, while the
opposite phenomenon is observed in females.


3.   Effects of the jet take-off noises on sleep in night N2
     The 4 types of noise used in this study differed both in peak level and duration:
Noise A:
Noise B:
Noise C:
Noise D:
Peak Sound Level:
 93 PNdB - Duration:    90 sec.
 95 PNdB -             30 sec.
100 PNdB-     -       90 sec.
112 PNdB-     -       30 sec.
     A total  of 32 jet  take-off noises were delivered in  the experimental rooms with a
semi-random  schedule,  the  random  sequences of 8  noises (two  of each type) being
distributed over every two-hour period.
     Thus the noises occurred independently of the sleep stage with the consequence that
the number of noises for each sleep stage differs from subject to subject.

                                       Table 4

           AVERAGE DURATIONS (IN MINUTES) OF THE SLEEP STAGES DURING
                THE TWO NIGHTS WITHOUT NOISE (Nl, N3) AND THE NIGHT
                                  WITH NOISES (N2)
                  Sleep        Males
                 I Stage    Nl   N?    N3
                                 Fema le s     \
                              Nl   N2    N3   !
W
1
2
3
4
REM
20
36
178
38
71
111
34
42
162
45
62
101
9
28
178
50
64
129
4
21
191
46
76
116
12
29
186
46
64
114
10
24
190
36
74
126
                                        579

-------
    Four types of responses to noise were scored:
- Type 0 response:

— Type 1 response:

— Type 2 response:


— Type 3 response:
no change in the EEC during the time of the noise presentation and
during a period of 1 minute following the noise.
occurrence of a phase of transient activation (P.A.T.) (13) during the
two periods described above, but without sleep-stage change.
occurrence of a sleep-stage change in the direction from stage 4 to
stage 1 or from REM sleep to N-REM sleep during the same period as
above but without sleep stage change to W.
during the same delay, occurrence of a sleep-stage change to W.
3.1 Responses of males and females to the noises
    Chi-square tests done on the 4 response types obtained during sleep shows a significant
difference of the noise effects between males and females. Males respond to the noise more
than do females and the number of stage changes to W is larger for the male group, while the
female group exhibits a larger frequency of type 0 and type 2 responses to noise.


3.2 Responses of males and females in three sleep stages
    Chi-square tests done on the responses obtained during stage 2, stage 3+4 and REM
sleep show no significant difference between groups.
    Note in this table that the frequencies of the four types of response differ considerably
according to the sleep stage in which the noises occurred. REM sleep seems to be the one in
which response types 0 and 1 are more frequent and response types 2 and 3 less frequent.

                                     TableS

                 RESPONSE FREQUENCIES OF MALES AND FEMALES TO
                        JET TAKE-OFF NOISES DURING SLEEP
                   (NUMBERS IN PARENTHESES ARE PERCENTAGES)
Response Type
Sex

Hale

Female


0
68
(26)
88
(31)

1
33
(12)
36
(12)

2
103
(40)
130
(45)

3
50
(19)
29
(10)
                     Males versus Females comparison
                     X2 - 9.43   ,  3 df  , p < 0.05
                                      580

-------
                                    Table 6

                RESPONSE FREQUENCIES OF MALES AND FEMALES TO
                JET TAKE-OFF NOISES DURING THREE SLEEP STAGES
                  (NUMBERS IN PARENTHESES ARE PERCENTAGES)
Sleep
Sex
Stage
Male

2
Female

Male
3 + 4
Female

Male


Female



0
19
(17)

26
(22)
22
(27)
26
(27)
27
(40)

41
(53)
Response

1
6
( 5)

9
( 7)
3
( 4)
2
( 2)
24
(35)

25
(32)
Type

2
52
(48)

65
(55)
43
(54)
60
(63)
8
(11)

5
( 6)


3
31
(28)

17
(14)
11
(13)
6
( 6)
8
(11)

6
( 7)
                 Malesversus Females  comparison
                 Stage 2  :    X2 = 7,02  ,  3 df  ,   NS
                 Stages 3+4:  X2 = 3,23  ,  3 df  ,   NS
                 REM :        X2 = 3,19  ,  3 df  ,   NS

3.3 Responses of males and females to the four types of noise
    Only the X2 test  done for the responses to noise C  (100 PNdB, 90 sec) shows a
significant difference between the two groups, males being more disturbed by this type of
noise than are females.
3.4 Responses of males + females to the four types of noise and for three stages of sleep
    In view of the results of 3.2. and of the fact that, considering separately the two groups
of subjects, we would have too small frequencies in each cell, X2  tests were applied to both
groups together, for each type of noise in the three sleep stages.
                                      581

-------
                    Table?

 RESPONSE FREQUENCIES OF MALES AND FEMALES TO
      THE FOUR TYPES OF NOISE DURING SLEEP
   (NUMBERS IN PARENTHESES ARE PERCENTAGES)
Type
of
noise




C

D

I
Sex 0
Male 36
(61)
Female 39
(59)
Male 18
(31)
Female 22
(32)
Male 11
(16)
Female 18
(24)
Male 3
( 4)
Female 9
(12)
lesponsc
1
7
(11)
7
(10)
3
( 5)
10
(14)
9
(13)
8
(10)
14
(19)
11
(14)
i tj
2
13
(22)
17
(25)
31
(54)
30
(44)
28
(42)
43
(58)
31
(43)
40
(53)
rpe
3
3
( 5)
3
( 4)
5
( 8)
6
( 8)
18
(27)
5
( 6)
24
(33)
15
(20)
Males versus Females comparison :
Noise A : X2 - 0,47  ,  3 df  ,  NS
Noise B : x2 « 3,56  ,  3 df  ,  NS
Noise C : x2 =12,00  ,  3 df  ,  0.01> p> 0.001
Noise D : x2 - 6,69  ,  3 df  ,  NS
                     582

-------
                     Table 8

RESPONSE FREQUENCIES OF SUBJECTS MALES & FEMALES TO
 THE FOUR TYPES OF NOISE DURING THREE SLEEP STAGES
     (NUMBERS IN PARENTHESES ARE PERCENTAGES)
Type Sleep
°f
noise Stage
2
3 + 4
A
REM
2
3 + 4
B
REM
2
3 + 4
C
REM
2
3 + 4
D
REM
0
24
(54)
29
(63)
22
(62)
8
(14)
7
(19)
25
(75)
7
(11)
9
(20)
13
(36)
1
(1)
3
( 6)
8
(20)
Response
1
2
( 4)
0
( 0)
12
(34)
7
(12)
0
(0)
6
(18)
3
( 5)
1
( 2)
13
(36)
3
( 5)
4
( 8)
J8
(45)
type
2
13
(29)
16
(34)
1
( 2)
34
(60)
25
(69)
2
( 6)
34
(56)
32
(72)
5
(13)
36
(60)
30
(63)
5
(12)
3
5
(11)
1
( 2)
0
( 0)
7
(12)
4
(11)
0
( 0)
16
<26)
2
( 4)
5
(13)
20
(33)
10
(21)
9
(22)
Noise
Noise
Noise
Noise
A
B
C
D
: X2
: X2
: X2
: X2
=
=
=
=
41,
54,
52,
52,
3
9
1
0
,
»
• ,
»
6
6
6
•6
df ,
df .,
df ,
df ,
, P<
, P <
, P <
, P <
: o.
: o.
CO.
C 0.
001
001
001
001
                       583

-------
     There are significant differences of the response frequencies in the three sleep stages
for the four types of noise. For each noise, REM sleep always shows more type 0 and  1
responses  and less type 2 and 3 responses. Response frequencies  obtained for stages 3 +  4
are generally closer to those in stage 2 than are those in REM sleep.


Discussion:
     Although the two groups of subjects estimated subjectively  their sleep  as being worse
and their time to fall asleep as being longer during the disturbed night N2 than during nights
Nl  and N3, only the males showed significant differences in respect to the corresponding
physiological data.
     These male subjects showed a significant increase of the sleep stage 2, 3 and 4 latencies
during N2  and also, but  less  marked,  an increased total  time  in W  after the first stage 1,
during night N2 when compared with N3, which was considered  by  these subjects as being
the best of the nights spent in the laboratory.
     For the females, the increase of the sleep stage latencies and  of the total time in W and
stage  1, which appear in tables 2 and 4,  are not significant but  they certainly contribute to
the significant differences when the  two groups are tested together.
     REM  latency  was  not  significantly  increased if  considered for each  sex  group
separately, while the significant increase, found  for both groups together,  of the time in
REM  during N3 when compared with N2 resulted mostly from the strong increase exhibited
by the male group. This result has  already been published (8) and could be interpreted as
some kind of "REM rebound" without previous deprivation.
    With  the exception of the difference for the response frequencies to the noises during
the sleep  between males and females (see  table 5), there was no significant difference
between these two groups, neither as for the response frequencies to the noises in the three
sleep stages considered, nor in the responses to the four types of noise, except for noise C
(100 PNdB, 90 sec), in which case sleep-stage changes  to W were more frequent in males
than in females.
     However, the results in Table 8 show clearly the difference between the effects of the
various noises on sleep stages 2, 3,4 and  REM for all subjects. It was during REM sleep that
the greatest frequency of no-response and  that the smallest frequency of types 2 and  3
responses were observed.
    These results seem to disagree with those of  some other studies, especially with  the
recent study of Lukas and Dobbs (6). These authors found that  middle-aged women tend to
be more frequently awakened by noise than do middle-aged men. It is certainly difficult to
compare this study with our own, but perhaps an explanation of this disagreement could be
found in the fact that the two populations were not similar,  especially regarding to  the
subjects' ages.
    In  1962, McGhie and Russel,  investigating the subjective assessment of normal sleep
patterns by subjects of both  sexes, found significant differences between sexes for some
classes of  age but not for others. For young people there were fewer differences between
sexes  than for older people. Furthermore, the sleep pattern of  the females tends usually to
change earlier than that of the  males, that is, during middle age (1).

                                       584

-------
     The difference between our results and those of Lukas and Dobbs, as regards the noise
effects in relation to the sleep stage in which the noise occurs, might be explained partly by
the fact that the  response types  used in the two studies are not identical. However this
aspect needs further study, particularly for the noise effect in REM sleep.

Conclusion:
     The present study does not bring an evidence of a clear-cut difference between noise
effects  according  to   the  sex of the subjects,  either  for subjective  assessment or
neurophysiological criteria,  but it does permit conclusions on the effect of noise on sleep.
Subjective results, such as time to fall asleep, sleep quality, sleep quantity, number of awake
periods, number of movements, and morning tiredness, showed for every subject that the
disturbed night was the worst of the three experimental nights (9).
     Several physiological data tend to show that sleep in night N2 was more disturbed than
during Nl and N3 (sleep-stage latencies, total duration of wake and stage 1, number of
sleep-stage changes) while some variables, such as number of P.A.T., REM periodicity, were
quite stable from night to night (8).
                                   REFERENCES

  1.  McGHIE, A., RUSSEL, S. M., The subjective assessment of normal sleep patterns, /.
     Merit. Sci., 1962, 108: 642-654.
  2.  GLOBUS, G., FRIEDMANN, J.,  COHEN, A., Effect  of  aircraft noise on sleep as
     recorded  in the home,  Oral presentation, 13th annual  meeting of APSS, San Diego,
     May 3-6,  1973.
  3.  LEVERE, T.  E., BARTUS, R. T., Electroencephalographic and behavioral effects of
     nocturnally occurring jet aircraft sounds, Aerospace Med., 1972, 43: 384-389.
  4.  LUKAS,  J. S.,  KRYTER, K. D., Awakening effects of simulated sonic booms  and
     subsonic  aircraft  noises on  six subjects, 7 to 72  years of age,  NASA Report n°
     CR-1599, 1970.
  5.  LUKAS,  J. S.,  DOBBS, M.  E., KRYTER,  K. D.,  Disturbance of human sleep by
     subsonic  jet aircraft noise and simulated sonic booms, NASA Report N° CR-1780,
     1971.
  6.  LUKAS,  J. S., DOBBS, M. E., Effects of aircraft noises on the sleep of women, Final
     report. NASA-CR-2041, 1972.
  7.  MUZET,  A., BECHT,  J.,  JACQUOT, P., KOENIG, P., A technique for  recording
     human body posture during sleep, PsychophysioL, 1972, 9: 660-662.
  8.  MUZET,  A.,  OLIVIER-MARTIN, N., Periodicite du sommeil paradoxal au cours de
     nuits normales ou perturbees par le bruit, Association des Physiologistes de Langue
     Francaise, Creteil, 9-10 fevrier 1973.
  9.  OLIVIER-MARTIN, N., SCHIEBER,  J. P., MUZET, A., Reponse  a un questionnaire
     sur le  sommeil nocturne  et un  questionnaire sur  la forme diurne au cours d'une
     experience  de perturbation du  sommeil par 4 types de bruits d'avions, Bulletin de
     Psychologic, in press.

                                       585

-------
10.  RECHTSCHAFFEN, A., HAURI, P., ZEITLIN, N., Auditory awakening thresholds in
    REM and N-REM sleep stages, Percept. Motor Skills, 1966, 22: 927-942.
11.  SCHIEBER, J.  P., MARBACH, G., Les perturbations du sommeil nocturne par les
    bruits de decollage d'avions. Essai devaluation de leur nuisance, Travail Humain, 1967,
    30: 162-163.
12.  SCHIEBER, J.  P.,  MERY,  J.,  MUZET,  A., Etude  analytique en laboratoire de
    1'influence du bruit sur le sommeil, 64 pp., Centre d'Etudes Bioclimatiques, Strasbourg,
    1968.
13-  SCHIEBER, J.  P., MUZET, A., FERRIERE, P. J., Les phases d'activation transitoire
    spontanees au cours du sommeil normal chez rhomme, Arch. Sci. Physiol,  1971, 25:
    443^65.
14.  WILLIAMS, H. L., HAMMACK, J. T., DALY, R. L., DEMENT, W. C., LUBIN, A.,
    Responses to auditory stimulation, sleep loss and EEC stages of sleep, Electroencephal.
    din. Neurophysiol, 1964,  16: 269-279.
                                      586

-------
      THE EFFECTS OF AIRCRAFT NOISE ON SLEEP ELECTROPHYSIOLOGY

                          AS RECORDED IN THE HOME
                  Gordon Globus, Joyce Friedmann and Harry Cohen
                    Department of Psychiatry and Human Behavior
                                College of Medicine
                         The University of California at Irvine

                                       and

                         Karl S. Pearsons and Sanford Fidell
                           Bolt, Beranek and Newman, Inc.


     The two objects of  this pilot study were first, to demonstrate that in-home sleep
studies could be accomplished readily, reliably, and at reasonable cost, and to elucidate the
nature of the relationship between noise and sleep.
     Two technological advances have permitted the present  experiment to address itself to
these questions:  a compact and highly reliable system for recording electroencephalograph
(EEC),  and electro-oculogram (EOG)  in  the home environment was designed by Jim
Humphries  in our University  of California at Irvine  laboratory.  A portable digital noise
monitoring  unit was developed  by Bolt, Beranek and  Newman, Inc. The two systems
brought an  increased level  of sophistication to the instrumentation required to measure the
effects of noise on sleep.

Method
     Acoustic measurements were made  in a number of neighborhoods in the vicinity of Los
Angeles International Airport.  After overnight records of noise exposure inside these homes
were analyzed, a one-square-mile target area was chosen. Planes came in 500 feet over these
people's homes.
     After  canvassing the  area, six middle-aged married  couples whose mean age was 45
years, and  who showed no substantial  hearing loss in an audiometric screening test were
chosen as experimental subjects. They had lived in the area for an average of 6 years.
     Similar screening techniques led to the selection of two  control neighborhoods,  several
miles from  Los Angeles Airport,  where  five middle-aged couples were selected. The couples
were all paid $10.00 per night for five consecutive recording nights.

Recording Procedure
     Each night, one hour before the couples retired, a technician from each research group
applied  the electrodes,  calibrated and  turned on the equipment. Upon  awakening, the
subjects  removed  the  electrodes,  turned  off the  equipment,  and  filled out a brief
questionnaire about the quality of their sleep.

                                       587

-------
Equipment
     A modified four-track analog tape recorder acquired both the physiological and the
acoustic data. The physiological data consisted of multiplexed FM recordings of EEG and
EOG, recorded on separate channels for husband and wife. The acoustic data consisted of a
continuous direct recording of noise exposure in the sleeping quarters.

Results
     The analysis of the noise data showed that the two neighborhoods differed radically.
The airport received a greater number of high level noise events (28 vs 17 events), as well as
a higher mean dBA level (77 vs 57 dBA). It should be noted that these noise levels, recorded
in close proximity to the subjects' beds, represent close estimates of the noise to which the
subjects were exposed, rather  than  outside noise levels (105 dBA,  approximately  per
flyover).
     The sleep data was then analyzed hi two ways. First, we looked at the differences in
the sleep patterns of the two groups, and then we looked at  how specific noise events
affected the experimental group.
     All of the  data from the last three nights of successful  recording, (the earlier nights
were discarded), were consistent with the general hypothesis: that intense noise exposure
degrades  the quality of sleep. Because of the small  number of subjects, and the large
individual variation, only the  following attained acceptable levels of significance:
1)   experimental subjects spent less absolute  (p  < .005) and percent  time (p < .025) in
     what we called "deep sleep", that is, stages 2, 3, 4 and REM, (45 minutes),
2)   they spent a greater proportion of their time in "light  sleep", that is, stage 1 (p <
     .005), waking and movement time. (Figure 1),
3)   as well as being awakened more often, (p < .025), than the control subjects.
To recapitulate, experimental subjects have less  "deep"  sleep, more  "light"  sleep, and
awaken more often than control subjects, (Table  1).
     These between-group differences  constitute necessary but insufficient evidence for the
hypothesis that aircraft noise exposure, per se, degrades sleep  quality. It is necessary to
show that  individual noise  intrusions  are, in  fact, responsible for disturbances of the
subject's sleep.
     A timeline analysis of sleep  stage-changes was developed. Each experimental subject's
state-change-data served as its own control. Within each experimental subject, three-minute
epochs beginning with a noise event were compared with control three-minute epochs
without a noise event.
     From  this analysis, it  was  determined that  when a flyover  noise occurred, during
"light" sleep, persistence in the same sleep stage or transitions to another "light" sleep stage
tended to occur, whereas transitions to "deep" sleep were lessened,  compared to controls.
When a flyover noise  occurred  during "deep" sleep, there tended to be a transition to
"light" sleep, compared to controls (Figure 2). These findings suggest how noise  exposure
directly effects the observed differences in total amounts of light and deep sleep.
     Since we began this study, the flight pattern  has changed at Los Angeles Airport, and
now there are no flyovers between 11 p.m. and 7 a.m. Therefore we are in the process of
re-examining the sleep patterns of these same subjects, in order to see how they habituate to
the new situation.
                                        588

-------
   400 -
   300
C
-  200
 0)
 E
    100
             Total  Sleep        Deep Sleep
             (M,1,2,3,4,R)     (2,3,4,R)
        Quiet

        Noisy
Light Sleep
        Figure 1: Average amount of sleeping time in "quiet" and "noisy" areas.


                                  589

-------
                     Table I

AVERAGE AMOUNT OF TIME (IN MINUTES) SPENT IN VARIOUS
                  SLEEP STAGES

Sleep
Stages
Waking
Movement
1
2
3
4
REM
Total Sleep
Deep Sleep
Light Sleep
NOISY
Mean
18.58
5.66
57.12
209.20
32.23
10.05
87.03
401.85
339.04
81.35
AREA
Standard
Deviation
26.50
1.93
33.21
35.33
26.06
16.43
9.92
34.02
40.42
51.70
QUIET
Mean
6.57
6.15
31.71
235.84
37.32
13.78
96.70
421.67
383.36
46.53
AREA
Standard
Deviation
7.62
3.32
21.76
52.20
14.46
30.30
22.60
34.42
36.82
29.29

-------
ro
•D

S"
I


a

£
•D

12.

9
•a
i-f
o


f
r+

!£L
u
-Q
           §
           u
           o
   to
                                              34567

                                         Twenty Second Epochs  From Start of Three Minute  Analysis  Period
                                                                                                          8

-------
                  NOISE AND MENTAL HEALTH - AN OVERVIEW

                                     W. Hausman
                                University of Minnesota
                                Minneapolis, Minnesota

     I appreciate the opportunity to address this distinguished Congress and to present an
overview of the work reported to date on the relationship between noise and mental health.
In exploring the literature to examine what has been  done by both mental health workers
and audiologists on this intriguing subject I find myself confronted by a remarkable dearth
of definitive work despite a number of clues that strongly suggest that there is, in fact, a
relationship between  a number  of facets of  the problem of excessive noise and the
developing scientific areas related to psychiatry. My  feeling is that of the optimistic little
boy  who wandered into  a large stable full of horse  manure and who plowed through its
contents for the best part of an hour and then emerged with a  distinctly unpleasant odor
and a large smile on his dirty face. When asked about  his experience he said "I looked and I
looked  and I looked and I couldn't find him but I  know that there is  a pony in  there
somewhere." Similarly I can report that I have looked  and looked through the literature and
that  while I have not found undisputed evidence that noise causes disorders of mental health
I too feel that there is  a pony in there somewhere. There are signs that a clear relationship
between noise  and mental  health will be found when sufficient interest develops in the
communities  of mental  health workers and  those  hi  the various fields related to the
psycho-physiology of audition.
     It may be significant that my  own introduction to this intriguing question has  come
through a series of circumstances which are unique to the experience of a psychiatrist. In
the early  1960's while I  was in military service I was assigned  as Chief of the Behavioral
Sciences Research Branch of the Army Medical Research and Development Command in the
Surgeon General's Office, where, as  part of my duties I monitored in-house and contractual
research on sensory psychophysiology.  During that time I was also assigned to represent the
Army on the Executive Committee of the Committee on Hearing and Biocoustics of the
National Research Council. About two years ago I was asked to participate in a task force of
the American Psychiatric Association on the Environment and Mental Health. At that time
my earlier Army experiences were brought  to  mind and I took  on the task of further
exploration of the literature on the  effects  of noise on mental health. This in turn brought
me to Dixon Ward who  was  very helpful to me in  that study  and who has now further
piqued my interest in asking me to take on this assignment to present an overview of the
field.
     My remarks will be organized into several relevant categories. In the first of these I will
examine the effects of premature deafness which, in our society, appears in large measure to
be the product of a combination of sociacusis,  a term  coined  by Glorig and Nixon15 to
describe the adverse affects on hearing of noise associated with  the general environment in
which we live,  and occupational noise-induced hearing loss which, in a sense, represents a
special  type of  sociacusis  for  those individuals who  work hi a particularly  noisy
environment. Anticaglia points out that about 50% of the machines used in heavy industry

                                         593

-------
produce noise above the level which is critical for potential hearing loss in predisposed
individuals.2 A number of aspects of our society appear to be increasing the probability of
exposure of the average individual to premature deafness including the increasing number of
large aircraft,  the development of rapid transit in our cities, and the increase of highways
running through and near urban centers. The shift of our population from rural to urban
areas and the increasing demand for consumer products as well as the rise in popularity of
discotheques and increased attendance by young people at loud concerts aimed at satisfying
their apparent need for repeated sensory stimulation all contribute to the likelihood that an
increasing part of our population will suffer from varying degrees of premature deafness.11'21
So, too, does the draft with  its  exposure to  the firing range  and the battlefield.
Unfortunately the  consequences of acquired deafness  are  poorly documented in the
literature, but, where examined, suggest  that  they should  be   of concern to  those
investigating constraints on mental health  and the quality of life. I will elaborate on this
issue later in this talk.
     A second area that I will examine involves the non-acoustical effects of noise on mental
health. This will include some of what is known about the direct and indirect effect of noise
on those aspects of human behavior and physiology of concern to the psychiatrist. In this
context I will make some  observations on the  vulnerability of  specific groupings of
individuals within society to noise, and touch on  some work that suggests that just as some
individuals appear to be more sensitive  to the potential deafening effect of acoustic trauma
than others so some parts of the population may well be unusually sensitive and responsive
to the non-acoustical effects of noise.

PSYCHOLOGICAL EFFECTS OF ACQUIRED DEAFNESS
     Although the literature on acquired deafness and on  the hard of hearing is remarkably
meager, as compared with work with the deaf child, several workers have commented on the
significant  psychological  crippling  of individuals  with deafness  that  develops  after
childhood.12*13 Perhaps the most specific examination of the psychological consequences of
acquired deafness has been that of Myklebust in which he describes  a variety of studies on
the hard of hearing including psychological testing of, and descriptive statements by, such
subjects.2 4  He describes a sense of isolation in these individuals which is borne out by their
patterns of living. Virtually all of the individuals with acquired deafness describe a shift in
occupation and  a  change in  their social patterns after they had lost their hearing. The
pattern of shifting jobs is also described by Dorman14 and Anticaglia2 and they point out
the great difficulties placed in the way of the deafened  by their sensory handicap. This
disorder also affects their  pattern of  leisure time activities so that most facets of their
quality of life are profoundly affected by the deafening. Myklebust notes  that the deafened
appear to marry  later than those who hear and  they remain more dependent on their
families. The extent of the psychological consequences of deafening may be measured by
their scores on the Minnesota Multiphasic Psychiatric Index where 64% of males and 56% of
females score high on the depression scale,  about 54% of both sexes high on the paranoia
scale, 64% of males and 55% of females high on the schizophrenia scale and 57% of males
and 50% of females high on the hypomania scale. In general the hard  of hearing  males
appear more emotionally maladjusted than is true of normals and the females are similar

                                        594

-------
although their scores are somewhat less dramatically abnormal. These findings are consistent
with the experimental work of Hebb and others on sensory isolation effects16  and clearly
suggest an acquired secondary emotional disorder in these individuals. Depending on the age
of acquisition of deafness the effects vary but in all groups the impact on mental health
appears to be significant and profound.

NON-ACOUSTIC EFFECTS
     Studies  on  the  effect of  constant  noise  in the  environment by Atherly and his
co-workers suggest that exposure to constant noise may lead to a mild depressive syndrome
which appears to be similar to the neurotic depressive reaction.5 Studies on performance by
Wilkinson,  Broadbent and others indicate that  the noisy environment interferes with work
accuracy although not particularly with speed of performance.2 8 The specific relationship of
environmental noise to mental illness per se has been investigated only by Abey-Wickrama et
al in the famous London airport studies which do not appear to be replicated elsewhere.1
Although the Abey-Wickrama study  suggests an increase in admissions to mental hospitals
for  those  maximally exposed  to  airport noise, other  workers have  challenged  the
epidemiological aspects of this  study, pointing out that the population under study was
over-represented  by  older unattached women who  tend to have a higher incidence of
psychiatric disorders under any circumstances.10  We  will hear more about  the follow-up of
the London airport studies later in this program.
     Another line of clues on the non-acoustic effects of noise comes from the studies in
Italy and Czechoslavakia on  the effect  of chronic noise exposure on factory workers?6'19
While  these  studies  do not specifically exclude the  interactive affects of such factors as
vibration and monotony of work, they  are striking in that both studies found a very high
incidence of gastrointestinal complaints among the workers exposed to  a high  level of
factory noise for 15  years or longer. In  one study 65% of the subjects were found to have
identifiable  gastrointestinal lesions  through  x-ray  examination.26  To the  best  of my
knowledge neither of these very  provocative  observations have been followed up in the
United States or in other countries than those in which the original studies were made.
     Some of  the non-acoustical pathways  through which noise appears to affect the
organism are outlined by Anticaglia and Cohen3 and are elaborated in a recent symposium
edited by the Welches.27 In this symposium several interesting studies suggest neurochemical
pathways mediating  the effect  of noise  in specially  sensitive breeds of animals23'8 and in
man.18'4 An  interesting question here is whether the physiological response to noise is  a
specific and direct result of the sound stimuli22 or whether it operates through an avoidance
mechanism as elaborated by Brady, Mason and  others.6'7

VULNERABILITY OF SPECIAL GROUPS
     Several  studies  suggest differential responsiveness of special classes of individuals to
loud noise stimuli. In one well  documented study Hunter observed increased physiological
response and decreased performance of dyslexic children as compared with normals, during
studies performed in an area located under the  air lanes leading into the San Diego airport.17
Similarly,  in experimental  studies  with groups of normal, hypertensive and psychotic
patients  Arguelles  et  al. demonstrate  significant increase  in  the  epinephrine  and

                                         595

-------
norepinephrine  responses of hypertensive  patients exposed  to  noise and catecholamine
responses  in  a  similar direction in a group of schizophrenic  and depressive psychotic
patients.4  The  latter group of patients  is unfortunately small  in  number and Arguelles
indicates that they were medicated at the time of the study, thus leaving largely unanswered
questions  about  the  differential effects  on  the  non-medicated  mentally  ill  of noise
annoyance. In both pathological groups the noise exposure was acute, thus leaving open the
question of the effect on these subjects of chronic or repeated noise exposure.
     In  view  of recent  developments  in  the  study  of  catecholamines  and  other
neurochemicals  in depressive and schizophrenic patients9'20'25 it is striking that among the
drugs most effective in protecting vulnerable  mice from acoustically triggered seizures were
a group of substances widely used as antidepressants, all of which effect norepinephrine
metabolism. Accordingly to Lehman "it appears that all drugs protecting mice or decreasing
the severity of the seizure increase the norepinephrine (NE) level into the receptor sites, and
that all drugs increasing the  severity of  the  seizure decrease  the NE  level."23  This is
particularly interesting in  light of the  work of the Arguelles group noted above. Obviously
one cannot make direct application from the study of mice to the study of man but it
appears  that  there  is  an  increasing body  of knowledge  that bears consistently  and
suggestively on  the issue  of the relationship between noise and  the induction of certain
types of psychochemical  responses. At  the same time it should be emphasized that the
definitive work on this intriguing topic with vulnerable humans is still to be done.

IMPLICATIONS AND CONCLUSIONS
     In summary it must  be noted that in the areas both of emotional disorders related to
and following the acquisition of noise-related deafness, and of the mental health effects of
noise other than through  deafness, relatively little specific work has been done. At the same
time there are sufficient clues to suggest  that those  who do involve  themselves in studying
these relationships may well make important  contributions to the body of knowledge both
in psychiatry  and in  the  acoustic sciences. Obviously the implementation of these studies
will  involve a range  of other workers interested in the environmental aspects of public
health. In  my view the most valuable clues point to the need for identification of vulnerable
populations and for the study of catecholamine and other neurochemical responses in those
individuals who appear by history or genetic  predisposition to have  a high vulnerability for
schizophrenia, endogenous depression or other  mental disorders. I suspect that the sort of
large scale epidemiological studies typified by  the  Abey-Wickrama  London airport study
may represent an approach  that is too broad in scope for our purposes. Similarly, the
criterion of hospital admissions may  not be  sufficiently sensitive to pick up  the acute or
chronic  effects of noise on the members  at risk of a large target population. Whatever work
is done, it is apparent that the problem of noise in society will not go away. There is a need
for greater recognition of the  potential  of work in this area and for working alliances of
audiologists, physiologists, neurochemists, epidemologists, psychologists and psychiatrists.
There may well be need  for new techniques  of collaborative  study  sensitive to the special
problems of eco-psychiatry so that the relevant parameters can be examined. Under these
circumstances it is hoped  that the real pony can be located somewhere in that malodorous
barn.

                                        596

-------
                                  REFERENCES

 1.  Abey-Wickrama,  I.A.,  Brook,  M.F.,  Gattoni,  F.E.G.,  and Herridge,  C.F., Mental
    hospital admissions and aircraft noise. Lancet, 2(7633) 1275-1277 (1969).
 2.  Anticaglia, J.R., Introduction: Noise in our overpolluted environment, in Welch, B.L.
    and Welch, A.S., Physiological Effects of Noise, Plenum Press, N.Y., pp. 1-3 (1970).
 3.  Anticaglia, J.R. and Cohen, A., Extra-auditory effects of noise as a health hazard.
    American Industrial Hygiene Association Journal, 31:277-281 (1970).
 4.  Arguelles,  A.E., Martinez, M.A., Pucciarelli, E., and Disisto, M.V., Endocrine and
    Metabolic effects of noise in normal, hypertensive and psychotic subjects, in Welch,
    B.L. and Welch, A.S. (Eds.) Physiological Effects of Noise, Plenum  Press, N.Y., pp.
    43-56 (1970).
 5.  Atherley, G.R.G., Gibbons, S.L.  and Powell, J.A., Moderate acoustic stimuli: The
    interrelation  of subjective importance and certain physiological changes. Ergonomics,
    13:536-545(1970).
 6.  Brady,  J.V.,  Operant  methodology  and  the  experimental production of altered
    physiological states in Operant Behavior Areas of Research and Application, Honig,
    W.K. (Ed.), Apple ton-Century Crofts, N.Y. (1966).
 7.  Brady, J.V., Porter, R.W., Conrad, D.G. and Mason, J.W., Avoidance behavior and the
    development of gastro-duodenal ulcers. Journal of Experimental Analysis of Behavior,
    1:69-72(1958).
 8.  Buckley, J.P., and Smookler, H.H.,  Cardiovascular and biochemical effects of chronic
    intermittent noise stress, in Welch, B.L. and Welch, A.S., (Eds.), Physiological Effects
    of Noise, Plenum Press, N.Y., pp. 75-84 (1970).
 9.  Bunney, W.E., Jr., and Davis, J.M., Norepinephrine in depressive reactions. Archives of
    General Psychiatry 13:483-494 (1965).
10.  Chowns, R.H., Mental hospital admissions and  aircraft noise. Lancet,  1(7644): 467
    (1970).
11.  Cohen, A., Anticaglia, J.R., and Jones, H.H., Noise induced hearing loss. Exposures to
    steady-state noise. Archives of Environmental Health, 20:614-623 (1970).
12.  Cornforth A.R.T., and Woods, M.M., Progressive or sudden hearing loss. Nursing Times,
    68:205-207, (1972).
13.  Denmark,  J.C.,  Management  of  severe deafness  in adults:  The  psychiatrist's
    contribution. Proceedings of Royal Society of Medicine, 62:965-967 (1969).
14.  Dorman, G.D., The sixth American Medical Association  congress on environmental
    health, Chicago, April 28-29, 1969. Welcoming Remarks, Archives of Environmental
    Health, 20:610-611 (1970).
15.  Glorig,  A.,  and  Nixon,  J., Hearing  loss  as  a  function  of  age. Laryngoscope
    72:1596-1610(1962).
16.  Hebb, D.O., A Textbook of Psychology, Saunders, Philadelphia (1958).
17.  Hunter, E.J., Autonomic responses to aircraft noise in dyslexic children. Psychology in
    the Schools, 8:362-367 (1971).
18.  Jansen, G., Relation between  temporary threshold shift and  peripheral circulatory
    effects of sound, in Welch, B.L., and Welch, A.S. (Eds.), Physiological Effects of Noise,
    Plenum Press, N.Y., pp. 67-74 (1970).
                                        597

-------
19. Jirkova, H.,  and Kromarova, B.,  Sledovani vlivu hluku  na celkoviy  zdravotni  stan
    pracujicich ve velkych strojirenskych zavodech. (Studies of the influence of noise on
    the general health of workers in large engineering works; an attempt at evaluation).
    Procovni Lekorstor (Prague) 17:147-148 (1965).
20. Kety; S.S., Toward hypotheses  for a biochemical component in the vulnerability to
    schizophrenia. Seminars in Psychiatry, 4:233-238 (1972).
21. Kryter,  K.D., Evaluation of exposures to impulse noise. Archives  of Environmental
    Health,20:624-635 (1970).
22. Leake, Chauncey, D., Foreword, in Welch, B.L. and Welch, A.S. (Eds.), Physiological
    Effects of Noise, Plenum Press, N.Y. (1970).
23. Lehmann, A.G., Psychopharmacology of the response to noise, with special reference
    to audiogenic seizure in mice, in Welch, B.L., and Welch, A.S., Physiological Effects of
    Noise, Plenum Press, N.Y., pp. 227-258 (1970).
24. Myklebust, H. R., The psychology of deafness. Grune and Stratton, N.Y. (1964).
25. Schildkraut,  J. J., The  catecholamine hypothesis  of affective disorders:  a review  of
    supporting evidence. American Journal of Psychiatry, 122:509-522 (1965).
26. Tarantola, A., Grignani, A., Lalli, M., and Santarnelli, G., Syndrome  dispectica  in
    lavoratori  esposti  a vibrazioni  ed elevata rumorosita.  (The dyspeptic syndrome  in
    workers exposed to vibrations and noise). Lavoro Umano (Naples) 20:245-265 (1968).
27. Welch, B. L., and Welch, A.S. (Eds.), Physiological Effects of Noise, Plenum Press, N.Y.
    (1970).
28. Wilkinson, R., Some factors influencing the effect of environmental  stressors upon
    performance. Psychological Bulletin, 72:260-272 (1969).
                                        598

-------
             OBSERVATIONS OF THE EFFECTS OF AIRCRAFT NOISE
                NEAR HEATHROW AIRPORT ON MENTAL HEALTH
                                  C.F. Herridge
                                      and
                                   L. Low-Beer
                             Sutton, Surrey, England

Introduction and Results:
     Interest in this subject was aroused eight years ago when one of us (C.F.H.) became
Community Psychiatrist to a population of about 150,000 people who live in the London
Borough of Hounslow. The Borough lies immediately to the East of Heathrow Airport, and
inevitably lies under the landing approaches of all aircraft approaching East-West. It also lies
under the  take-off routes  of some  aircraft leaving West-East. As  prevailing winds are
westerly, landing noise is the major problem. Figure 1 shows the location of the area and
and the  two general hospitals at which we have worked.
     In  1969 (1) a research project was carried out in an endeavour to see if aircraft noise
affected mental health. This was a retrospective study which measured only admissions to
the area psychiatric hospital during a two year period.  Admissions were studied from the
Maximum Noise Area (MNA), shown shaded in Figure  1, and compared with admissions
from the less noisy remainder of the Borough, though it must be remembered that the less
noisy part,  which includes the West Middlesex Hospital, is hardly quiet. MNA was defined as
an area  where the NNI (Noise and Number Index) was over 55, or where the PNdB from an
approaching Boeing 707 was over 100.
     The results are shown in Tables 1 and 2. Table 1 shows sex, age and social status, whilst
Table 2 shows  diagnostic  categories. From them it was concluded that Psychiatric
Admission  Rates were higher from the MNA than from outside it, and the person most at
risk was the single, widowed or separated woman, in the older age group, suffering from
organic or neurotic mental illness.
     The work was challenged by Chowns (2) who questioned the demographic analysis and
the use  of a combined PNDb and  NNI maximum  noise area. We believe the demographic
analysis to be correct, but the combined noise values could be regarded as suspect.
     Last year, therefore, members of the British  Medical Research  Council carried out a
similar smaller study, using the same methodology but taking the 55-and-over NNI contour.
This is shown in Figure 2. This work is not to be published, and constitutes  a pilot survey
for a much larger intended study. We are very grateful, therefore, to Dr. Tarnapolski and Mr.
Gattoni for allowing us  to show their summary of results in Table 3. This is a smaller
sample, but the results show a similar trend for  all first admissions and for female first
admissions, though they do not quite reach the 5% significance level.
     An important new  factor has appeared in the last year,  however, which has  to be
mentioned. Housing prices in the very  noisy areas have not kept pace with national average
rises, according to local Real Estate Agents. Probably not coincidentally, a large number of
Asian immigrants, many from Africa, have settled in the very noisy areas. It is  obviously not

                                       599

-------
                  LONDON  BOROUGH  OF HOUNSLOW
                 Maximum  noise area and study  area
                                Hounslow Hospital
 Heathrow
  Airport
                                                    West Middlesex Hospital
                                                           1   mile
                                             maximum noise area
                                               area studied  ( Borough of
                                                                Hounslow )
            Figure 1: London Borough Hounslow maximum noise area and study area.

only aircraft noise which is responsible, but it is perhaps not too fanciful to suggest that
"Noise Ghettos" are being created, with all the  implications of such a problem, including its
own effect  on mental  health  as shown  by  Schneider (3). "Permanent Noise Slums" are
already described by tally and  Holmes (4). This seems to be an extension of the concept.

Discussion:
    To obtain really reliable data of the effects of Aircraft Noise on Mental Health around
Heathrow is notoriously difficult, as Atherley (5) has  pointed out. and the relatively 'hard'
data which  we have presented is very scarce. Social Surveys may be illuminating  and  two
reports by the British Board of Trade (now  Department of Trade and Industry). (6.7) are
valuable, but in the eyes of some, including ourselves, inadequate.
    As a result of recent work with the Department of the  Environment, however, an
interesting hypothesis has occurred to us and  may help  to  explain, and  perhaps be used
prophylactically against, the adverse effect that aircraft noise  appears  to have on mental
health.
                                     600

-------
Table 1

Category
Ail admissions:
Total
First
Female admissions:
Total
First
Male admissions:
Total
First
Females aged > 45:
Total
First
Females (married) :
Total
First
Females { single, widow,
separated, divorced):
Total
First
Females (widow,
separated, divorced )-.
Total
First
Females ( widows h
Total
First
M.N.A.
0
245
96
147
54
98
42
96
36

68
24
79
30
44
21
36
19
E
212.9
78.1
125.7
41.2
87.1
36.9
75.4
22.8

61.5
19.8
62.3
20.75
38.5
15.5
27.5
13.0
Non-M.N.A.
0
243
83
143
41
100
42
79
17

69
• 20
74
21
33
10
19
7
E
275.1
100.9
164.3
53.8
110.9
47.1
99.6
30.2

75.5
24.2
90.7
30.25
38.5
15.5
27.5
13.0

*2
8.58
7.28
6.37
7.02
2.44
1.30
9.98
a 42

1.30
1.70
7.55
6.92
1.57
3.90
5.26
5.60
Significance
0.005>P>0.001
0.01>P>0.005
0.025>P>0.01
0.01>P>0.005
N.S.
N.S.
0.005>P>0.001
P< 0.0005

N.S.
N.S.
0.01>P>0.005
0.01>P>0.005
IV. S.
0.05>P>0.01
0.05>P>0.01
0.05>P>0.01
 601

-------
                                       Table 2
                       DIAGNOSTIC CATEGORIES (FEMALES ONLY)

Category
11 Neurotic "
Total
First
Organic mental illness
Total
First
MM. A.
0

26
9

35
22
E

18.6
6.9

24.3
12.1
Non-WLN.A.
0 E

17 24.4
7 9.1

21 31.7
6 15.9

X2

5.19
1.02

8.33
13.94
Significance

0.05>P>0.01
N.S.

0.005>P>0.001
P<0.0005
     *  Affective, schizophrenic, "psychopathic", and"epileptic, alcoholic, & etc." groups were
        analysed for females and all six diagnostic categories were analysed for males, but in
        all instances  either P was > 0.05 or {he numbers were too few for analysis.

     In  1971  (8)  Minimum  Noise  Routes  for take-off  from  British  Airports  were
recommended, and these were implemented in July, 1972. Some of these routes are shown
in Figure 3—the idea being that  aircraft taking off from Heathrow should be routed over
areas of minimum population density, though it must be  realised that this is a very relative
term in over-crowded South East England.
     The flood of protests, together with the setting up of highly articulate, well-managed
and  well-financed  protest groups, was immediate. Expensively-produced and extensively-
researched protest documents have been produced. The areas mostly affected are social class
1  and 2 residential, and their  psychology as organized protesters has been  described by
a'Brook (9).
     Compare  this, however, with the much noisier but far less affluent Hounslow area. The
Chief Public Health  Inspector for the Borough reports (10): "During the year (1971) the
department received nine complaints concerning noise from aircraft. This figure is extremely
low and cannot reflect the nuisance and inconvenience suffered  by people living under the
various approach and take-off paths. It can only be concluded  that the residents of the
Borough are sceptical as to any reductions in noise  being achieved."
     We  have  tried  to  portray  this  tremendous difference  in attitude in Figure  4.
Admittedly, organized  campaigns can collect signatures and therefore magnify results but,
even so, this difference is extremely startling.
     Elsewhere (11)  one of us  has described more journalistically clinical cases of mental
illness where aircraft noise appeared to be an important aetiological factor, and we see new
cases each week. It was suggested that there were many instances where aircraft noise might
be the "last straw" in precipitating breakdown.
                                       602

-------
              LONDON  BOROUGH OF  HOUNSLOW
                     M.R.C.  Pilot  Study

                          Hounslow Hospital
Heathrow
Airport
West Middlesex Hospital
                                         0      1  mile
                                          I
                                     55 and over N.N. I.  contour
                                flw*'iw Area studied  ( Borough of
                                                      Hounslow )
               Figure 2: London Borough of Hounslow M.R.C. Pilot Study.

-------
                                               Table 3
                              IN PATIENT PSYCHIATRIC HOSPITAL ADMISSION RATES
                                ( Gattoni F. & Tarnopolski A. - unpublished work)


BOTH All admissions
SEXES First "
MAire All admissions
First "
FEMALES Al1 admissions
First "
High Noise
Area Rate
per 1000
3.46
2.01
3.15
1.71
3.73
2.28
Numbers
182
106
79
43
103
63
Lower Noise
Area Rate
per 1000
3.22
1.54
2.81
1.45
3.57
1.61
Numbers
213
102
87
45
126
57

.Increased Risk for
55+ Zone %
7.2
26.4
11.5
16.4
4.4
34.4

xz
0.50N.S.
3. 75 TREND
0.54N.S.
0.60N.S,
0.10N.S,.
3. 59 TREND
I

-------
             TAPLOW
        AIRCRA FT HEIGHT
                 JOOO
'
                                     37°/o of West Takeoffs
                        WINDSOR
o/o of West Takeoffs
                         AIRCRAFT HEIGHT
                          JOOO FEETAPPPOX.
     TAKEOFF  ROUTES
     ("MINIMUM NOISE ROUTES")
                                                                         LANDING HEIGHT
                                                                         400 FEET
                                                                           PROPPING..
                                                                  33% of East Takeoffs
                                               AIRCRAFT HEIGHT JOOO FEETff iSH ER
                               Figure 3: Take-off routes ("Minimum noise routej")

-------
                 AIRCRAFT NOISE COMPLAINTS
            4500 r-
            3500
Total number
of          '2500
complaints
            1500
             500
                        complaints from residents of Hounslow
               1960  '61 '62  '63  '64 '65  '66 '67  '68 '69  70 71

                                    Year


                   Figure 4: Aircraft noise complaints.
                               606

-------
     We now propose to forward a further theory based on the results shown in Figure 4.
Whilst people are aggressive, articulate and active, and whilst they feel that their actions are
producing some  effect which may eventually  improve their lot, they may remain well.
When, however, they despair of results, try to live with aircraft noise and become externally
apathetic, then they may be turning their aggression inwards and producing, by accepted
psychological  mechanisms,  psychiatric  breakdown.  This  can be  called  "The Four A's
Hypothesis" - Aggression, Articulation and Action versus Apathy.
     If this theory  is correct, then psychological prophylaxis is possible. It does not remove
the necessity for cutting down aircraft movements at Heathrow, producing quieter aircraft,
and  ultimately removing  major airports  from heavily populated areas, but an active
Heathrow Advisory Council,  with genuinely easy access for complaints  by the ordinary
person,  and an emphasis on  popular communication,  would be a real help in reducing
morbidity. All aggrieved people should be able  to see, either in Newsletters or in the local
press, that their personal complaint has actually been noted, and real concern and practical
action be seen to be shown by, at present, apparently uncaring authority. This would act as
palliative  therapy  until  the real  treatment-removal of intensive aircraft  movement from
Heathrow to the projected and much-disputed coastal airport at Maplin-can be completed.
The fact  that Heathrow  may  still have 12% spare  runway capacity by 1985  (12) is
immaterial. The population around the airport has already run out of psychological reserves.
                                   REFERENCES

  1.  Abey-Wickrama, I., a'Brook, M.F., Gattoni, F., and  Herridge, C.F. "Mental Hospital
     Admissions and Aircraft Noise". Lancet ii 1275-7 (1969).
  2.  Chowns, R.H. Lancet i 467 (1970).
  3.  Lally, G., and Holmes, G. "Aircraft Noise and Flight Routeing". Local Authorities
     Aircraft Noise Council — London (1972).
  4.  Schneider,  E.V.  "Inter-Relations between  Social  Environment  and  Psychiatric
     Disorders". New York  (1953).
  5.  Atherley, G.R.C. Documenta Geigy. "Noise";  p. 4. Manchester (1968).
  6.  McKennell, A.C. "Aircraft Noise Annoyance around London (Heathrow) Airport - the
     Social Survey". Central Office of Information. London SS.337 (1963).
  7.  "Second Survey of Aircraft Noise Annoyance around  London (Heathrow) Airport". H.
     M. Stationery Office, London (1971).
  8.  The Noise  Advisory Council. "Aircraft Noise: Flight Routing Near Airports". H. M.
     Stationery Office, London (1971).
  9.  a'Brook,  M.F. "A Brief Examination of the  Psychology  of Associations formed to
     promote control of Aircraft Noise". Sound 6.2 (1972).
 10.  "The Health  Services of Hounslow,  1971"  (p.64) Hounslow  Health Department.
     London, (1973).
 11.  Herridge, C.F. "Aircraft Noise and Mental Illness". Sound 6.32-6(1972).
 12.  Reed, A.  The Times. London, April 23rd, (1973).
                                       607

-------
       SESSION 7


 COMMUNITY RESPONSE I
Chairmen: G. Thiessen, Canada
     P.N. Borsky, USA
          609

-------
                   METHODOLOGICAL ASPECTS OF STUDIES OF
                        COMMUNITY RESPONSE TO NOISE

                   Erland Jonsson, Ola Arvidsson, Kenneth Berglund
                                        and
                                   Anders Kajland
               From The National Swedish Environment Protection Board,
             Department of Environment Hygiene, The Karolinska Institute,
                       Department of Environmental Hygiene,
                                Stockholm, Sweden

     According to a generally accepted definition, noise consists of all undesirable sounds.
For the most part this definition has been used in investigations concerning the effects of
community noise on individuals.
     This presentation will  give  an overview of the methodological problems  associated
with  studies of community  response to  noise. Community  response can  mean  the
coordinated actions from  the community where the individuals live to change conditions
that  are considered to be undesirable,  for example attempts to remove the source of noise
from  the area;  but community response can also mean individuals' experiences with  the
effects of noise and their reactions to those experiences. As will later be shown, a great
many irrelevant factors influence the individuals' willingness to join coordinated actions, for
example, against a source  of  noise.  For  this  and other reasons,  the  investigations of
community  noise and  its  effects have  been  mostly concerned  with the individuals'
annoyance reactions.
     In hitherto-reported investigations of the prevalence of annoyance reactions caused by
noise, a type of dose-response thinking has often been used. This means that attempts have
been made to correlate the subjective  reports of noise disturbance to an objective measure
of sound dosage (some combination of sound level and duration). When establishing norms,
it  is necessary  to estimate  the  highest permissible dose  on the basis of what reactions
different doses  give rise to. There are  two possible approaches. One either tries  to protect
the most sensitive group or one  starts from  the "normal person's" reactions. In the first
approach the search for the dose-response  relationship begins  with  the "most sensitive
group's" reaction to different doses,  which also  means that less sensitive individuals  are
automatically protected. In the other approach the search for the dose-response relationship
begins with the normal population's responses. It is the  latter approach that has been most
often used.
     In  attempts to map  the  effects of community  noise, survey investigations have
traditionally been used. These, as a rule, have studied the disturbing effects of noise. Several
definitions  have  been used for "disturbances", which has  often  resulted in differently
phrased questions; consequently individuals have  been  classified  as  "disturbed"  from
different starting points. In all  cases, the assessment has been based on the individual's
report on his reactions to noise. This means that the  individual not only has to try to
describe his own experiences but also to evaluate them.

                                       611

-------
     The following will first describe how the effects of noise exposure have been studied in
several investigations. Different methodological problems will then be considered, followed
by a discussion on the possibilities for measuring the relationship between noise exposure
and different disturbance reactions.
     The reactions that have been mainly used for evaluating individuals' experience of the
exposure conditions are the so-called "disturbance reactions". Empirical studies of what a
respondent really means when he states that he is "disturbed", "irritated" or "bothered"
have not been done. This means that when a person at an interview or with a questionnaire
declares  that he is disturbed, it is not possible to know  more than that the exposure
concerned gives the respondent an experience of displeasure, that is to say, something that is
not desirable.
     The presence of disturbance in the above meaning can be said to be the first dimension
of disturbance observed in the investigations that have been done. Other dimensions are the
intensity of disturbance stated by the respondent,  the duration of disturbance—that is, the
frequency and length of the occasions he is disturbed-and finally, the manifestations of the
disturbance, such as psychosomatic  symptoms, actual  activity disturbances, attempts by
various means to bring about a change in the  exposure situation.
     Besides the aspects of disturbance reactions already mentioned there are also others
that should be considered. One of importance is  how  conscious the individual is of the
disturbance, which is possible to determine by studying the respondent's readiness to report
the existence of annoyance.
     The importance of the disturbance reactions is decided partly by how much attention
the individual places on the disturbance problems, and partly by how great the disturbance
is seen in  a relative sense—that is to  say, seen in relation to disturbances from  other
environmental factors.  Both these dimensions  of "importance" must be considered so that
the classification of what disturbance reactions really mean will be more complete.
     Investigations that have been published have mainly studied the following reactions to
noise exposure:
     — reports on general disturbing experiences caused by noise exposure (both intensity
as well as frequency)
     — reports on disturbances of the exposed individual's daily activities
     — reports on psychosomatic symptoms as a result of noise exposure
     — reports of complaints to some authority
     — report of different types of activities the exposed individual has undertaken, such as
using noise protecting measures, to diminish  the actual noise exposure.
     Very  briefly we shall now recount a couple of investigations executed in the United
States, Great Britain and Sweden  where different reactions to noise exposure  have been
studied. The first of these investigations is Borsky's study "Community Aspects  of Aircraft
Annoyance", 1954. Quite often, inconsistent answers appear in the investigation; that is to
say, a given respondent may spontaneously state  at  one time that aircraft  noise  is a
disturbing  factor in the environment, but in response to a direct question concerning the
occurrence of aircraft noise it has sometimes been the case that  the respondent denies any
inconvenience  caused  by  aircraft noise.  It has  therefore been necessary   to have a
summarized conclusion of all the statements expressed in interviews  concerning aircraft

                                       612

-------
noise. In accordance with this principle two judges, independently of each other, classified
the respondents into different disturbance classes,  such as "not bothered", "somewhat
bothered", and "greatly bothered".
     A Swedish aircraft-noise investigation of 1958-59 studied the applicability of Borsky's
results to Swedish conditions. When summarizing the individual's reactions to aircraft noise,
however, answers to seven questions were used,  of which the first five were open questions
dealing with the individual's attitudes toward  his home  and his community.  Here the
respondent had the  opportunity  to mention aircraft noise. Of the last two  questions, one
dealt with noise in general, the other with aircraft specifically.
     An investigation by McKennel in the Heathrow Airport area, done in the early  1960's,
attempted  to construct a disturbance scale according to the  so-called Guttman principle.
This scale  includes  a question concerning  intensity of disturbances as well  as questions
concerning other activity disturbances.
     The  Swedish  traffic noise  investigation  performed  in  the years  1967-68  used  a
combination of disturbance intensity  and disturbance frequency to construct a disturbance
index with six index values.
     The Scandinavian aircraft noise investigation of 1972 used questions similar to those of
the  above-mentioned  traffic noise  investigations.  Considering questions  dealing  with
disturbance intensity,  individuals were classified into  four different disturbance  classes:
individuals who are not bothered by noise, individuals who are not especially bothered by
noise, individuals who are rather bothered, and individuals who are very bothered.
     Reports of  disturbance of  the  individual's daily activities in the  above-mentioned
investigations   have   mainly  concerned  disturbances  of  sleep,  rest,  relaxation,   and
communication  difficulties.  Different  social  epidemiological investigations  of  the
individual's experience  with community  noise have worked with different definitions of
disturbance, which has made it difficult to compare the results. Other factors that have also
diminished the comparability have  to do with the question's formulation and the way the
purpose of the investigation is presented.
     The following problems are involved in formulating questions:
     — which verbal expression for disturbance should be used?
     — should  general or specific  questions concerning disturbance experiences be asked?
     — should  the questions be open or structured?
     The problem of which verbal expressions should be used for disturbance experiences
has been studied in a number of investigations, like that performed by Jonsson (1963a) and
Sorensen and Jonsson (1973). These studies aimed,  among other things, at clarifying the
connection, from a linguistic point  of view, between different "noise and air pollution
stimuli"  on  the  one  hand, and different verbal expressions for the  reactions  on the
other,—so-called response expressions. Stimuli were, for example, "aircraft noise" and "dust
or soot from industries". Response expressions were "unpleasant", "painful", "disturbing",
"irritating" and "troublesome".
     The purpose was also to study what degree of injurious  effect the different response
expressions implied.
     The results indicate that different verbal  expressions used in annoyance studies can
express different  degrees of injurious effect; that is to say, the percentage  of interviewed

                                         613

-------
individuals indicating an injurious effect from exposure can vary with the choice of the
verbal expression in the questionnaire.
     Questions that directly name a source of disturbance (specific questions) produce a
greater  number of disturbed respondents than questions dealing  with  noise  in general
(general questions) (Chapman, 1948). It has also been shown that questions dealing with the
disturbance that individuals suffer from noise exposure give a lower disturbance frequency if
no fixed  alternatives are given (open  questions)  than  if there  are fixed alternatives
(structured questions), (for example, Arvidsson, 1972).
     Another investigation by Jonsson (1963b) studied how disguising the investigation's
purpose affected the number who reported that they were bothered. The results proved that
many more individuals  report that they are bothered by  noise when the purpose of the
investigation is clarified than when it is not.
     The investigations cited show that many more individuals claim to be disturbed or very
disturbed  when  the investigation's purpose  is  not disguised,  when direct annoyance
questions  are used instead of general annoyance questions, and when the questions are
structured instead of being open. What type of questionnaire gives the "true"  value is at
present impossible to decide, but it is conceivable that the measurement methods used have
not overestimated the frequency of disturbed individuals.
     What importance the above-mentioned sources of error have  for the assessment of
individual  disturbance reactions is difficult to judge because both the construction of the
questionnaires and the control of the measure's relevance have been derived from the
individual's subjective reports. An example of this method of procedure is the control that
was used  in the latest Scandinavian aircraft noise investigation (Arvidsson 1972). A clear
connection was  found  there  between  degree  of  disturbance  and number of  medical
symptoms, and the  number of reported activity disturbances, respectively.
     Another related problem is that it cannot be assumed that individuals have the same
frame of reference when judging the disturbing effect of a certain exposure. This means that
different individuals can assess the same exposure differently, even though the exposure,
objectively viewed,  causes the same reactions in the individuals. This can be due  to the fact
that the importance of noise problems in relation to other problems—for example, economic
problems—are experienced differently by different individuals.
     The purpose of the investigations that have been done has, on the whole, been to give a
basis for an assessment of whether a certain noise exposure is acceptable or not. This has led
to the problem  of  dividing, in a relevant way, the respondents into different disturbance
categories.   When  designing  the  dose-response   relationship,  the  earlier-mentioned
investigations and other published investigations  of community  noise have all  proceeded
from  the principles  that have been used in earlier investigations of a similar type, and as a
rule the boundary between bothered and not-bothered individuals has been set up  without
careful consideration of the total effect of noise upon a person.
     It has long been known that reports of disturbing experiences used  for mapping the
individual's reactions to noise exposure have been influenced not only by the characteristics
of the stimuli and  the exposure  situation, but also  by the individual's own attitudes, for
example, toward the source of noise. Thus we work with a measure of noise reaction that is
easily influenced by extraneous factors. When assessing the relevance of a statement on

                                        614

-------
noise annoyance it is therefore necessary to try to connect the report of the subjective
disturbance experience with some other part of the individual's life situation.
     This connection can be made in several ways. One can give the report a  partial social
anchorage, that is to say, assess the importance  of noise problems in  relation to the
occurrence of other problems in the environment. The other type of connection that can be
given to a subjective report of noise disturbance is a so-called functional anchorage. This
means that one studies how noise may  cause reactions that change the individual's ability to
function in a normal way in his  environment.
     Finally, one can give the report a physiological anchorage  and in this way study how
noise stimulation bring about changes  in the individual's physiological status.  This type of
reaction can of course mean poor health, but it may not necessarily be the case.
     When controlling the relevance of the  measure  methods based  on the reports of
subjective experiences, the last two mentioned types of connections should be used.
     If the functional or physiological  connections are  to be used it is necessary  to clearly
know what effects noise  can have upon a person. In different contexts attempts have been
made to divide the effects of  noise upon the individual into different groups, taking into
consideration, among other things, the type of the reaction and the measurement methods.
It can  be  said that  the  different classification proposals are mainly  answered by  a
classification of the effects into  the following four types:
     1.  Physiological effects
     2.  Effects upon different activities
     3.  Psychosomatic effects
     4.  Experienced psychiatric effects
     The last two types  of effects have above  all been  studied by using  survey methods
while effects upon other activities have been studied both with laboratory experiments and
with traditional survey investigations. Only laboratory experiments have been used to study
physiological effects.
     In  the  different  laboratory  experiments  it  has been  found  that  noise causes
physiological and performance-reduction effects, but only a few attempts have  been done to
correlate simultaneously  these  objectively measurable effects to the individual's subjective
experiences of noise  exposure.  In  some investigations, however, a  similar method of
procedure has been  used, among others, by Miller (1957) who did  not find any relation
between performance  changes  and subjective annoyance experiences. On the other hand
Moreau and Nordberg (1967) and Arvidsson (1972) have found a relation.
     To conclude, a problem will be discussed which has not been considered  very often in
reports dealing with investigations of the reaction to community noise, that is, the problem
of how much the differences between individuals' reactions are explained by the differences
in the exposure situation.
     In published investigations, the  emphasis has been mainly on studying the relationship
between dose and response at the aggregate level. This means that a number of areas have
been  chosen with different exposures; within  each of these  has then been studied the
percentage of respondents who  claimed to be bothered or very bothered. This means that all
the individuals within an  area  are assigned to the same  dose  group regardless of the
individual's length of stay in the living area or the number of hours spent daily in  that area.

                                        615

-------
A requirement for the composite dose measure to be comparable between different areas is
that the inhabitants' characteristics should be similar in relation to the variables that can
influence how much the individual is actually present in his living area or his residence.
(Examples  of such variables are sex, age, civil status and type of work.) Regarding the
response measures used  in  this  type  of studies, one must object to the  fact that
consideration often is not given to the distribution of the response on categories other than
"very bothered". Correlations obtained in studies at the aggregate level are as a rule based
on  few  observation pairs (10-20).  In this  context  it should be  stated  that ecological
correlations usually are high, which is due to the fact that consideration is not taken of the
individual variations that appear with regard to both dose and response within the respective
chosen areas.
     Several  traffic and  aircraft  noise investigations show, however,  that  not more than
10-20% of the total variance (variance among individuals independent of the exposure level)
can be explained by the difference in dose level; the remaining variance must originate from
individual differences.  Approximately the same ratio for the explained variance  has been
obtained from analysis of data that handles dose-response relation of smell.
     To summarize, it can be emphasized that even if worthwhile results have been obtained
from the up-to-now published investigations of people's reactions to community noise, there
is still a lot to do from the methodological point of view.
     The methods for measuring and describing both exposure and individual response no
doubt have to be refined.
                                   REFERENCES

  1.  Borsky, P.N., Community aspects of aircraft annoyance. National Opinion Research
     Center, University of Chicago, Report No 54 (1954).
  2.  Carlsson,  G.,  Ekdahl,  A.,  och  Ronge, H., Intervjuundersokning i Halmstad och
     Bromma. (Survey investigation in the Swedish cities Halmstad and Bromma.) Stencil
     (1959).
  3.  McKennel, A.C., Aircraft noise annoyance around London (Heathrow) airport, Central
     Office of Information (1963).
  4.  The National Swedish Institute for Building Research and the National Swedish Insti-
     tute of Public Health, Traffic Noise in Residential Areas. The National Swedish Insti-
     tute for Building Research, Report No 36 E  (1968).
  5.  Rylander,  R., Sorensen, S.  and Kajland  A.,  Storningsreaktioner vid  flygbullerex-
     ponering. (Annoyance reaction from aircraft noise exposure.) Omgivningshygieniska
     avdelningen,  Statens Naturvardsverk och  Hygieniska  Institutionen,  Karolinska In-
     stitutet (1972).
  6.  Jonsson, E., On the formulation  of questions in Medicohygienic interview investiga-
     tions. Preprint. Acta Sociologica, vol.  7(1963a).
  7.  Jonsson,   E.,  and  Sorensen,   S.,  Om   undersokningsresultatens  beroende  av
     frageformuleringar vid  hygienisk-sociologiska frageundersokningar.  (The effects  of
     different formulations of the questions in hygienic-sociological survey investigations.)

                                         616

-------
    Hygieniska Institutionen, Karolinska Institute!, Omgivningshygieniska  avdelningen,
    Statens Naturvardsverk och Sociologist Institutionen, Stockholms Universitet, Stencil
    (1973).
 8,  Chapman, D., A survey of noise in British Homes. Building Studies Techn. Paper Nr. 2,
    His Majesty's State Office, London (1948).
 9.  Arvidsson, O., Om manifestationer av bullerstorning. (On effects of noise distrubance.)
    Sociologiska Institutionen, Stockholms Universitet, Omgivningshygieniska avdelningen,
    Statens Naturvardsverk  och Hygieniska Institutionen, Karolinska Institutet, Stencil
    (1972).
10.  Jonsson, E., Om olika metoder for studier av yttre miljofaktorers storande effekt. (On
    different methods for studying the disturbing effect of environmental factors.) Nordisk
    Psykologi, vol. 15, 1 (1963b).
11.  Miller,  H-G.,  Effects of high  intensity noise on  retention. Journal  of Applied
    Psychology,  Vol41,No.  6. 370-372(1951).
12.  Moreau, C-E,, och Norberg, J.O., Effekter av auditiv stimulation pa vissa intellektuella
    funktioner. (Effects of  auditory stimulation on certain  intellectual functions.) MPI,
    Rapport Nr.  60, Stockholm (1967).
                                         617

-------
        DECISION CRITERIA BASED ON SPATIO-TEMPORAL COMPARISONS
                        OF SURVEYS ON AIRCRAFT NOISE

                                Ariel ALEXANDRE
                          Environment Directorate, OECD*

I.    SYNTHESIS OF SURVEY FINDINGS
     The following synthesis mainly combines the findings of the United Kingdom (1961
and  1967), French (1965 and 1972) and Dutch (1964) surveys, which all use Guttman scale
analysis for measuring annoyance and correlating it with physical noise parameters.
     In the  matter  of  annoyance scores, the number  of steps used in constructing the
annoyance scale differed from country to country (5 categories for France, 6 for the United
Kingdom and  7 for the Netherlands), so that for purposes of valid comparison the raw
values  of annoyance scores have  had to be  converted into relative values-or "annoyance
indices", i.e. into percentages of maximum 5, 6 and 7 scores. Thus the maximum annoyance
score obtained in France is regarded as equivalent to the maximum score obtained in the
United  Kingdom  and in the Netherlands, while  of course all intermediate scores have
likewise been converted into "degrees of relative annoyance".
     At first the annoyance scores were correlated with separate indices, since the precise
purpose of the surveys conducted in each country was to calculate an index  which would
best  represent annoyance.  Calculations  undertaken  after the surveys  were conducted,
however, showed  that all indices were strongly intercorrelated. To make the consolidated
findings intelligible, a single index therefore had to be  chosen. The one index used in
common by the British, French  and Dutch researchers  was the NNI. This allowed
international comparison of the findings to be made.
     The combined findings of the 5 surveys carried out in the United Kingdom, in France
and in  the Netherlands appear in Figure 1 herewith. These are the average annoyance scores
classified by noise stratum for the two British surveys, and by community surveyed for the
two French surveys and the Netherlands one (1).
     (1)  Sources of combined data in the figure:
         -    page 207, "Noise,  Final Report", HMSO, London,  1963 (London findings);
         —    page  192, "Second Survey of Aircraft Noise Annoyance Around London
              (Heathrow) Airport", HMSO,  London, 1968 (London findings): we have
              used the findings concerning
              (a)  the zone within a 10-mile radius of London-Heathrow;
              (b)  the total mode of operation of the airport, for otherwise it would have
                  been impossible to make comparisons with the 1961 survey;
         —    pages  32 and  142   of  Report  No.  AAA   -  16/67  of  Association
              d'Anthropologie Appliquee:  "Enquete sur le bruit autour des aeroports",
              Paris 1967 (Paris findings);
     This paper is being presented as a private contribution; it does not necessarily reflect the views of
the OECD on this question.

                                       619

-------
            Indies de gene
            Annoyance index
 Gene maximum 100 f*'
 Maximum annoyance

             90  •
             70

             60

             50

             40

             30

             20

             10

 Gene null*    n
 No annoyance
  RELATION ENTRE  LA GENE  ET LE  BRUIT  DES AVIONS
                  (Resultats de 5 enquetes)
RELATIONSHIP  BETWEEN ANNOYANCE  AND AIRCRAFT NOISE
                   (Results of  5 surveys)
                         r = 0,%
           Loot/on 7967
           London 7967

           Amsterdam 7965

           Paris 7965
           Paris 7972
Gene
Annoyance «
                                                                                   2/3
                   10
        15   20    25    30    35    40
45    50
55
60    65   70
 Indie* c/e bruit NN1
   Noise index NNI
                   Figure 1:  Relationship between annoyance and aircraft noise.
          —   page 92, No. 10-1970 of Revue Francaise d'Acoustique, op. cit. (Netherlands
              findings);
          —   unpublished findings of the French 1972 survey.
     It will be noted that the average annoyance scores are closely linked to values for the
NNI and that differences between countries  are rather small  (the difference within each
noise class does not exceed ± 5 per cent).
     No systematic variation is found between airports  or between  time  periods  except,
perhaps, that annoyance  increases faster today between 35 and 45 NNI than was the case 8
or 10 years ago.
     Although all  the enquiries  show a very clear-cut correlation between aircraft noise
indices  and average annoyance scores by noise class or location of the survey correlation
                                          620

-------
(correlation coefficients above .90 in France, the Netherlands and the United Kingdom),
they show relatively low correlation between the noise indices and individual annoyance
scores (correlation coefficients under .50). It is likely that personal factors, and doubtlessly
also those related  to the kind of life one  leads (children or none, all-day or  occasional
presence at home, windows habitually opened or closed) as well as certain acoustical factors
(apartment exposure, window size, etc.) play a role  in the  creation of the  feeling of
annoyance. Recent research^) has shown that  by including personal factors such as fear of
aircrashes, sensitivity to noise in general, attitudes towards the  usefulness of aviation, etc.,
the correlation with annoyance increases. Could we then say that "appropriate educational
campaigns" would reduce annoyance? This is not at all sure because the elimination of fear
and hostility with respect to an airport can only modify reactions up to a certain threshold,
i.e. where sleep or conversation are objectively disrupted, irrespective of the individual's
attitude  towards aviation.  Furthermore, we can't even be sure that  annoyance is really
influenced by fear or hostility, at least when these feelings are  acknowledged orally in  the
course of the interview. It could be that annoyance itself unleashes a whole set of defensive
or aggressive reactions. This hypothesis seems in fact to be confirmed  by  the Tracer Study
itself, which shows  that the  probable  sequential relationship of  "stages in response to
aircraft  noise  proceeds  from  hearing aircraft  to disturbance of activities,  thence to
annoyance,  and culminating  in  the formation  of a  negative  attitude  toward aircraft
noise"(2 \ We should not forget that when we are dealing with problems of this kind, even if
we come up with correlations, these do not clearly indicate the direction of the causality.
Thus we have to be very careful in interpreting the results. Which is first, the hen or the egg?
All we know is  that fear and hostile attitudes increase with noise. In the example of the hen
and the egg, it doesn't really matter whether annoyance is taken to be hen or egg. But noise
is  the rooster,  that's for sure. No rooster, no  egg. To return to more serious matters, we
should add that the public is increasingly seeking a better quality of life and fewer daily
disturbances.  It is  therefore highly  probable  that  campaigns  to  promote  the easier
acceptance of aircraft noise would be offset by campaigns for environmental improvement
led by associations for the protection of residents in the vicinity of airports.
     Knowledge of some  personal  variables therefore  seems to be  of limited value  for
decision   makers.  However,  the following  important conclusion  emerged  from  all  the
surveys:
     — even at high noise levels, a small number of people suffer little or no annoyance.
     — even at low noise levels, a small number of people are always annoyed'3 \
    (^In  particular,  see "Community Reaction to Airport Noise,  Final Report",  TRACOR  No.
T.70-AU-7454-U. Austin (Texas), United States, 1970.
    (2) "Community Reaction to Airport Noise, Final Report", op. cit., pp. 78 and 86.
    ^3^lt has often been claimed that  10% of the  population will always be annoyed, whatever the
conditions, while 20 to 30% will remain imperturbable. In fact, however, in truly quiet surroundings (below
15 NNI), only 5% are annoyed; and in extremely noisy surroundings (above 65 NNI), only 10 to  15%
remain relatively unaffected, of whom only 5% are not at all annoyed. This has been found in both the
second French survey and the second British one.

                                          621

-------
     The  existence  of "imperturbable"  just  as  "externally dissatisfied"  subjects is a
relatively stable factor in all problems and for decision-making purposes must therefore be
taken into account.
     It is  obvious, however, that a dissatisfied group in any population will create more
problems for the public authorities than a group of people who are always satisfied with
their environment. Cpmplaints made in the vicinity of airports  are indeed likely to come
primarily from a dissatisfied group whose general discontent will be crystallized around the
noise question. If these complaints increase in number they will finally set in motion a true
wave of public protest which could then be joined by those people who are truly annoyed
but who don't necessarily belong to the "externally dissatisfied" group.
     But it is by no  means certain that complaints are a good indicator of annoyance'4', as
the formulation of a complaint depends on many exogenous factors of a practical (to whom
should the  complaint  be  addressed?); political (to  what  extent does one  believe the
complaint will be taken into consideration?) and economic nature (according to the United
Kingdom  and United  States surveys,  complaints depend on  levels  of education,  family
income and  the value  of  dwellings exposed  to noise^5' ; from this  it may moreover be
deduced  that  as the standard of living rises, complaints  will increase or be formulated by
people at lower noise exposure levels than in the past).
     Furthermore, while average annoyance in a given area can be predicted on the basis of
the  level of noise exposure in that area (see the previous synthesis of European surveys),
complaints can only be predicted by considering all the sociological, political and economic
variables characterizing the population exposed to noise.
     Complaints do  of course annoy the  people to whom they  are addressed, that is, the
airports,  the municipal government, etc. But the action taken  by the public authorities
should be  based on the  annoyance to which the population is subjected and not the
annoyance to which the administration is subjected.
     In view of all the foregoing reasons, it seems advisable to take account of annoyance
rather than simply to count complaints.

H.   PRACTICAL USE OF SURVEY FINDINGS FOR DECISION MAKING
     We have already seen  that a certain segment of the population is hardly ever annoyed
even at very high noise levels, whereas even at very low noise levels a small section of the
population will feel annoyed.
     To circumvent this problem of individual variations of attitude with regard to noise, it
would thus be enough to consider the  percentage of people annoyed according t,o locality
(instead of considering the average score, which conceals individual variations). On what
     ^^Nonetheless complaints do precipitate an awareness of the problem and it quite often happens that
surveys are organized only after complaints have been repeatedly made.
     (s)"Aircraft Noise Annoyance Around London (Heathrow) Airport", by A. C. McKennell, HMSO,
1963, pp. 7.1 to 7.4) and "Community Reaction to Airport Noise, Final Report", by TRACOR, op. cit., p.
84.

                                         622

-------
criterion would decisions then have to be based—when 30 per cent of people are annoyed or
90 per cent? The need to satisfy 95 per cent  of the population rather than 50 per cent
cannot be established scientifically (not to mention that the satisfaction of 95 per cent of
the inhabitants of a community numbering 500 people is radically different from trying to
satisfy 50 per cent of the people in a town with a population of 20,000).
     Although this problem cannot be resolved, it is still possible to use the ratio established
between  the percentage of people annoyed and  the noise level for predicting the results of
any  decisions  which  may be made, "people annoyed" being defined by their  individual
annoyance scores.
     As an outcome of the first United Kingdom survey,  the investigators estimated those
people to be seriously annoyed whose annoyance score equalled a minimum of 4 (on a 6
point scale the relative score comes to 65%) because above  this score aircraft annoyance
prevailed over all other causes of dissatisfaction with living conditions. In France, anyone
who  scored 3  or more (on  a  5  point scale) was considered seriously annoyed. In the
Netherlands, anyone who scored 4 or more (on a 7 point scale) was also considered seriously
annoyed.
     The reasons employed by investigators in determining  annoyance thresholds are of
course more complete and numerous than those here summarized (for details, reference may
be made to the researchers' reports). But the survey findings agree at least on two major
points: 1) an annoyance threshold is exceeded when the annoyance score indicates frequent
disturbance  of speech communication (conversation and radio or TV reception); 2) this
annoyance threshold  proves to be much the same for all the surveys when expressed in
relative  terms. Unquestionably, whenever the spoken word—man's principal  means of
contact with the outside world—can no longer be perceived on account of noise, annoyance
reaches a limit of tolerability which if exceeded interferes with the normal pattern of daily
living.
     Figure 2 combines the findings of the European surveys by taking the above annoyance
criteria into account.
     As  these results expressed in NNI have been found to correspond closely with those
obtained in the United States by K.  D. Kryter using the CNR and NEF indices, the NNI,
CNR and NEF indices may be regarded as virtually interchangeable, subject to the margin of
error which must necessarily be expected in any such evaluation^6'.
     This method provides a valid tool for predicting the probability of annoyance: if, for
example, in areas currently  subjected to 60 NNI, the noise of each aircraft were to be
reduced  by 10 dBA  or the air traffic were to be cut by three-quarters, the proportion of
people annoyed would fall from 75  to 50 per cent. It would thus be enough to know the
level of noise  exposure for each area (in terms of NNI or CNR) to ascertain and predict how
many  people  will  be objectively   annoyed (according  to the  meaning  attributed to
"annoyance"  during  the surveys, which implies the  disruption of a  number of activities).
     (6)Kryter's estimate (which corresponds to the lower part of the curve in our figure) is quoted from
"The Effects of Noise on Man", by K. D. Kryter, Academic Press, New York, 1970 (figure 238 a).

                                         623

-------
               10 L

                       -
                       -.
-
-.
                                    .•
100
 -.
 K
110
 •.
 •
120
102
 •
130
112
 70
140  CNR iUS>
122  ;. •
 80  NN
                   Figure 2:   Percentage of annoyed persons in relation to noise.
The percentage of annoyed people in each area can be predicted  in simple terms by  using
one of the following two formulas^7':
percentage of annoyed people
   2 (noise level in NN1 - 25) or
   2 (noise level in CNR - 85) or
   2 (noise level in NEF - 15)
     Individual  reactions, which vary one  from the other and cannot be forecast, thus form
;i statistical whole  that can be  forecast when all the individual reactions in a given popula-
tion exposed  to the same noise level are taken into consideration. (We are referring here to
the sum of individual reactions to annoyance and not to the reaction of the community as a
whole which is expressed only  in the form  of petitions, defense associations, etc. and which.
as we have previously seen, depends on many other factors than noise alone.)
     To sum  up.  the surveys so far conducted have made it  possible to determine  noise
indices which are  closely correlated  with average reactions of annoyance in a population to
           shall  here  consider  the  indices  to  be  interchangeable  because they show much  greater
intercorrelation than correlation with the annoyance experienced by residents around airports. In  view of
probable errors of estimation (approximately  ±  5 per cent in relation to the values obtained from the
preceding figure),  the multiple of the logarithm for the number of peaks (15 in the case of NNI and lOfor
CNR and NEF) has but a secondary effect. The  proposed formulas may be applied when the noise level is at
least equivalent to 30 NNI (or 90 CNR. or 20 NEF).
                                           624

-------
 establish limits of noise acceptability based on the number and scope of activities disrupted
 by noise, and  to estimate the proportion of persons annoyed (i.e. whose annoyance goes
 beyond the threshold of tolerance) among a population in terms of the latter's exposure to
 noise.
     These conclusions drawn from the surveys provide practical decision-making criteria
since they provide an evaluation of annoyance in probabilistic terms, i.e. by allowing for
likely  individual variations (which would  not be taken into account if average annoyance
were merely ascertained in each area).
     This criterion thus makes it possible  to  calculate  the number of people annoyed in
some area in  absolute terms:  the percentage  of people annoyed-which can  be predicted
whenever the noise to which the area is (or will be) exposed is known-need thus only be set
against the population density for the area. Actual "maps of annoyance" might hence  be
plotted around airports, and  used  for reaching certain decisions concerning landing and
take-off paths, noise  abatement procedures, the preferential use of certain runways, etc. so
as to subject the smallest possible number of people to
III.  SHORTCOMINGS OF SOCIAL SURVEYS AND NOISE INDICES
     However, the proposals here formulated should be considered cautiously, as the social
surveys and noise indices are guilty of a number of shortcomings which cannot be ignored.
     It was pointed out  for instance  that  as reactions to  noise vary so greatly from one
person to the next, predictions can only be considered as relating to an entire population.
     Is this a major shortcoming of social surveys? True, the effectiveness of some noise-
abatement measure for a community's benefit may well be questioned if overall annoyance
diminishes while a few individuals remain  dissatisfied  with the steps taken. Actually this
problem would assume major importance if noise were the only issue revealing comparable
individual variations. But there are many fields where major variations in behavior occur. It
may even be asked whether there is any single field unaffected by such variations and where
all decisions result in unanimous approval.
     Since variations in individual behavior are universal and by no means peculiar to noise,
there is no reason why they should create any specific problem for decision makers.
     The real drawback of the social surveys is that they provide a good correlation between
annoyance and noise only for the daytime. In regard to nightime noise it is not yet possible
to define  an  objective threshold of disturbance.  It would appear  that in-depth studies
concerning the effects of aircraft noise on sleep are urgently needed. These studies  should
include the measurement of secondary effects of the disturbance of sleep: fatigue, consump-
tion of sleeping pills, job  alertness and performance, etc. No overall study of this kind has
           other criteria used at present as a basis for decisions concerning aircraft noise protection are
criteria of an economic nature (drop in housing values because of noise exposure and direct social costs of
noise); comparisons and evaluations of psychological and social criteria as opposed to economic criteria are
found in "Le temps du bruit", by A.  Alexandre and J.-Ph. Barde, published recently in France (Flam-
marion, Paris, 1973).

                                         625

-------
yet been  undertaken,  yet  one is needed if any standards of night-time noise which the
majority of the population can accept are to be proposed.
     The noise indices also have undoubted drawbacks. To begin with, a common index
would be  desirable so that  noise abatement measures taken at the international level could
be based on  coherent conclusions. At present the conversion of one index into another is
complicated by the fact that  the weighting for numbers of aircraft heard as well as units
adopted to measure  noise  vary from one country to another,  whereas  these units are so
closely intercorrelated that the use of dBA, recorded directly on a sonometer, proves fully
adequate.
     Furthermore, noise indices give rise to problems of application. The contours of noise
equivalence  plotted  around  airports  should be calculated generously  rather than over-
accurately, since  the paths followed by the various types  of aircraft are shown  to be  far
from precise, and levels of noise transmitted to the ground depend on various meteor-
ological and  other factors. Great care  seems to be  called for in this connection, since
otherwise  the plotting  of some index  values  may fail to match  the  annoyance which is
actually felt. Because of the drawbacks inherent in  the use of calculated noise contours,
some  airports propose  more comprehensive  noise  monitoring  systems  (for  example
London-Heathrow). All these problems are important and should not be overlooked on the
premise that they have not yet been sufficiently investigated.

                                         o
                                       o    o

     The social surveys have however already contributed a great deal. They have made it
possible to determine  noise tolerance limits  based on simple criteria (the disturbance of
essential daily activities) and also to calculate noise indices which, notwithstanding certain
shortcomings, are simple, handy tools  for predicting which areas  around  airports will be
exposed to noise causing some given degree of annoyance.
                                        626

-------
           PSYCHO-SOCIAL FACTORS IN AIRCRAFT NOISE ANNOYANCE

                                 Aubrey McKennell,
                             University of Southampton,
                                      England.

1.   Acoustics, Noise Control Policy and Community Response
     For administrators, the ultimate criterion of what constitutes a noise problem lies not
so much in the physical or even acoustical characteristics of the noise source as in the nature
and  extent of the public protest that is generated.  If people in communities responded to
noise in the same fashion  as sound level meters, or even like subjects in laboratory experi-
ments, then policy decisions for noise control would be that much simpler. The facts about
community response to noise, as discovered by social surveys, and reviewed in this paper,
raise awkward  problems for administrators and legislative bodies, but need  nevertheless to
be taken into account. The evidence reviewed comes mainly from studies of aircraft noise
since this happens to be the field in which community response has been most thoroughly
investigated. The main implications drawn, however, will be seen to apply to other sources
of environmental noise disturbance.
2.   The Exposure-Annoyance-Complaint Paradigm
     To understand  community reactions to noise it might seem that all we have to consider
is a stimulus-response relation between, on "the one hand, the degree of noise exposure in a
locality, and on the other, the volume of complaint generated. Unfortunately, the truth is
not so simple. Much of what I  have to say  concerns the psycho-social variables that inter-
vene in and attenuate this relationship. As a result, we shall see that complaint activity can
be a misleading index to the effect of noise.
     A distinction has to be made between  'complaint' and 'annoyance' reactions to noise.
The term 'complaint' will be reserved for any kind of formal public action directed against
the noise nuisance-writing to or telephoning an official, signing a petition, joining a noise
protest organization, and so on. Those who take such action we shall term  'complainants'.
The term 'annoyance' on  the  other hand  will be reserved for the  subjective feeling or
attitudinal reaction aroused by the noise. The annoyed may express such feelings verbally to
their immediate friends, family  or neighbours, but most of them do not go on to any other
kind of action. In fact, unless  we provide  the opportunity for a sample cross-section to
express such feelings to interviewers in a social survey, we will not normally  learn about the
degree of 'silent annoyance' that  exists in a community. To this extent annoyance can be
termed a latent reaction. We cannot simply infer the volume of annoyance from the volume
of complaint. Of course, it is true that almost all complainants are  annoyed-I say 'almost
all' because there are important exceptions even  to this (see Figure  5c). But the reverse
certainty does  not  hold. All  the evidence  indicates that complainants are  a very small
fraction of those equally annoyed  who do not complain.

                                        627

-------
     So instead of the simple relation

                         Exposure  	"-Complaint

we have to posit a process such as

                         Exposure,             Complaint
                                   ^Annoyance

     We have already noted that the passage from annoyance to complaint is by no means
straightforward or inevitable, and we will return to  the influence at work here. But now
consider the exposure-annoyance arm of the total process.
3.    The Low Correlation between Exposure and Annoyance
     A measure of annoyance can be  obtained from social surveys. In the Heathrow survey
(1), for reasons of comparability, we  followed Paul Borsky's earlier work (2) by construct-
ing a Guttman scale  of  annoyance-caused-through-activities-disturbed by aircraft. Subse-
quent work (5) has established that noise annoyance can be measured much more simply,
and adequately enough for most practical purposes, by a straightforward self-rating ques-
tion.
     Having established our measure of annoyance let us turn to the relationship between it
and noise exposure. Figure 1 shows the regression line through the means of the individual
annoyance scores in each noise exposure stratum. Data are taken from the 1961 Heathrow
survey (1). The fit to  the means is very close. In fact it could hardly be bettered. But it is
important to note that what has been achieved is the prediction of the central tendency of
response only.  The prediction of individual annoyance reaction is poor. The correlation
coefficient between PNdB values and  individual scores on the annoyance scale was only .46.
In other words, in any one noise exposure stratum there is a great range of variation in the
annoyance of individuals. This variation mostly reflects differences in noise susceptibility
between neighbors, or individuals at the  same exposure level.  It can be  accounted  for
statistically by measures of psycho-social variables which affect annoyance independently of
noise exposure. The correlations and  partial correlations in Table 1 (see especially the third
column) show the results for some of the more important psycho-social variables in  the
1961 Heathrow survey (1). Figure 7 shows how the results for one of these variables look
when  presented  graphically. The  same basic findings  emerged in the Second Heathrow
Survey (9) and in the Tracor study in the U.S.A., which are discussed further below.
    These statistical facts correspond to the experience of field workers in a noise-exposed
community. In the same street one can find many people who scarcely notice the noise even
though  some of their neighbors are  severely troubled  by it. Generalizing,  and taking an
inferential  step which I  shall  attempt to justify shortly, one can expect  to find whole
communities reacting  quite differently to  noise even though subjected to much the same
physical conditions of  exposure.

                                        628

-------
                                       Table 1

            VARIABLES WHICH OFFSET THE DEGREE OF AIRCRAFT ANNOYANCE
                      INDEPENDENTLY OF NOISE EXPOSURE LEVEL
Partial Correlations



5.

12.
16.

18.

22.

25.


34.




Variable
Number of things
disliked
Preventability
Aircraft held to
affect health.
Fear of aircraft
crashing.
General attitude
to noise.
Annoyance scale
for noise other
than aircraft.
Reported feelings
and activities of
neighbours.
Correlation
with
Annoyance
(1)
.30

.35
.38

.52

.50

.25


.45


Correlation
with
Exposure
(2)
.07

.11
.21

.28

.15

-.03


.33


Variables x
Annoyance
(Exposure
constant)
(3)
.23

.34
.33

.45

.43

.25


.27


Exposure x
Annoyance
(Variable
constant)
(4)
.42

.42
.38

.30

.41

.43


.35


4.   Acoustical Laws as Central Tendencies in an Attitudinal Mix
     Returning now to the regression line in Figure 1, the importance of establishing law-
like  relationships—such as the Noise and Number Index-which will predict the long-term
average community response  to noise, needs no  urging.  For convenience we will refer to
such a law as the 'Central Tendency' or 'Acoustical' law. However, if our object is to predict
the community response at a given point in time or in a particular locality it is important to
recognise the limitations of the Acoustical law. The very  good fit to the means in Figure 1
was  obtained because all the  various factors which lead  individuals at the same exposure
level to be  more annoyed or less annoyed happened  to counteract one another, in this
particular study, so that they balanced out around the central tendency. The Acoustical law
tells us nothing whatsoever about  the nature and  extent of the variation about  the central
tendency.

                                        629

-------
     Figure 8, based on combined data (24) from the Second Heathrow Survey and the
American Tracer study shows the effect on this variation of the joint operation of two
particular psycho-social  variables, namely the attitudes of Misfeasance (labelled Prevent-
ability in the British study) and Fear. The selected combinations of Fear and Misfeasance
yield three annoyance curves. Those persons included in the lower curve represent the group
most favorably predisposed to accommodate the noise in respect to the particular psycho-
social variables  under consideration. The top line of the figure  represents the opposite
extreme, or the  most psychologically hostile group. What the results bring  out is that we
have not one annoyance curve but a series of annoyance curves. In the more conventional
analysis these curves are  collapsed to show  only the central tendency of annoyance with
variation in exposure, as in Figure 1. But this central tendency is the resultant of a particular
combination of psychological predispositions that existed at a particular point of time in a
particular locality.  Generalizations from the central tendency results to other localities or to
the future in the same locality implicitly assume that the attitude structure in these other
situations would replicate that found in the particular survey. It is evident that this particu-
lar assumption will not hold up if there is any change  in the  mix of relevant psycho-social
attitudes in  the  population  or if the mix is different between the populations being com-
pared.
     Note also that in designing surveys to establish the central tendency relationship it is
important that a specific level of noise exposure is not represented by  just one locality. In
addition to gross individual differences in psychological predisposition there may well be
factors making  for average differences between subregions. It follows that surveys which
sample  a large  number of localities are likely to provide the best  basis for  an unbiased
central tendency result. Samples drawn from  a small  number of  subregions on the other
hand run the risk of regional biases which if extreme enough could distort,  weaken, elimi-
nate or even reverse the expected central tendency result (e.g. reversal could occur if the
bias  ran counter to the exposure so that the noise sensitive in the sample were drawn from
the quieter regions  and the less sensitive from the more exposed regions).
5.   Consequences for the Spread of Annoyance
     The individuals who balance out around the statistical average do not on that account
disappear even from communities for which the Central Tendency law holds good. To see
the practical consequences of this look at Figure 2. Column (b) of the table shows the
percentage in each  noise exposure stratum who are  'annoyed'. The great majority of the
annoyed people are not to be found at the highest levels of noise exposure, but at several
levels below. The  point  being  made is that these  administratively awkward and  over-
whelming facts of the  situation  are not predictable at all from the Central Tendency law.
They arise completely out of the departures from such a law.
6.   Consequences for the Prediction of Community Response
     Turning now to a second main consequence of dispersion about the central tendency,
we have already noted  that the very close fit of the means to the regression line in Figure 1

                                        630

-------
indicates that the individual differences in noise annoyance were sufficiently varied as to
approximate randomness in this study. This was a nice result  to have obtained at the time,
but with hindsight I think it is not one that can be relied upon to occur. The Central Limits
Theorem happened to work for us on this occasion. But there  is no reason to think that the
psycho-social variables will always or inevitably balance out. Under the influence of, say,
anti-noise publicity,  they could well aggregate in one direction. A prediction solely on the
Central Tendency or Acoustical law could then be misleading.
     How misleading? The answer  turns on  the relative weight to be attached to noise
exposure and psycho-social factors in predicting annoyance. Let us look at this. As noted,
the correlation coefficient  corresponding to the regression fit in Figure 1 was only .46. That
is, less than one quarter of the variance in individual annoyance reactions can be attributed
to physical  noise exposure as measured here. But before we can attribute the remaining
variance to psycho-social factors, we have to show that it is not simply measurement error.*
7.   The Relative Importance of Acoustical and Psycho-social Factors
     How far can  the  correlation  between  annoyance  and noise exposure be raised by
improving the index of noise exposure? In Figure 1, PNdB measures are used, but the data
plot changes little when NNI values are substituted. Recalculating and putting in an objec-
tive  measure corresponding to the  Noise and Number Index makes little difference to the
regression fit. The correlation coefficient is raised from .46 to a mere .48.
     These are the data from which the Wilson Committee calculated the Noise and Number
Index (3). At the time  they noted  the imperfections in the data. Two recent major surveys,
in the U.K.  (9) and the U.S.A. (10), made vigorous attempts to remove these imperfections.
In neither survey was a correlation greater than .5 achieved  between  noise annoyance and
any  of the combinations of noise exposure parameter that were tried. In both surveys, the
collective independent contribution of psycho-social variables to variance in annoyance was
approximately twice as great  as the  contribution due  solely to noise exposure. Similar
results appear to have been found in community reaction to motor vehicle noise. In a study
(11) reviewed in reference (12), the  analysts were able to predict whether an individual
would voluntarily express annoyance with freeway noise or not in 64 percent of the cases,
using attitudinal data from the interviews. A measure of freeway noise used alone yielded
only a 0.23  correlation. There is also the survey work  (13) leading to the Traffic Noise
Index (TNI). In developing this index Langdon and Griffiths unfortunately did not examine
the role of attitudinal factors. It is  notable, however, that the correlation between TNI and
individual annoyance scores  falls to .29 from the .88 using median  values, reflecting the
wide range of individual reactions.
     *In a survey of Noise Annoyance in Central London (8) no correlation was found between annoyance
and measured noise exposure. The reasons for lack of correlation here are thought to be the restricted range
of noise exposure considered, compared with that occurring in the aircraft noise situation, and the uncon-
trolled distance of the measuring point from the informant's home.

                                         631

-------
                        83    86    89   92   95    98   101   103-r
                                    Average pert loadness (pNdt)
 Etch point is  tht arithmetic mean  of the scores on the annoyance  scale (or
 informants  within the  particular noise  stratum.
The  dotted lines indicate  the  re|ion  on either  side of the  avenge within which
two-thirds  of  the individuals  can  he expected  to fall.

Although  the  strata  are designated in PNdl  the  average  annoyance  caused is
actually  the joint result of  the numier a*  well as  the fondness  of the  aircraft.

        Figure 1:  Relationship between annoyance scores and noise exposure strata.

                                       632

-------
     It begins to appear, therefore, that the scope for improving the objective characteriza-
tion of the stimulus for annoyance amounts to no more than a few percentage points in
amount of annoyance  variance explained. Even when all available acoustical knowledge is
utilized to the  full, by far the greatest proportion of variance in  annoyance will remain
attributable to the psycho-social factors independent of exposure.

8.   Implication for Noise Exposure Indices
     The low statistical correlation found between noise exposure  and annoyance has un-
fortunate  implications  for attempts or claims to base the validity of particular noise ex-
posure indices on multiple regression equations fitted to survey data. When the frequency of
flights and their average loudness are used jointly as predictors, for example, the amount of
annoyance variance explained is only marginally higher  than it is for either of these para-
meters used separately. It can be shown (see statistical note on pages  186-7 of reference 9)
that the weights attached to separate parameters in  a composite index like the NNI can vary
enormously without having much effect on the value of the multiple correlation coefficient.
Accordingly,  the weight  3 5 for the number parameter in the NNI is not critical. Neither are
the departures from 15 in the various alternative noise indices that have been proposed. The
notion of some  kind of "trade off in annoyance between the number of flights over an area
and their average loudness may survive conceptually, but its quantification in an empirically
validated equation still eludes us. The required study methodology here may be difficult to
arrange in practice. What seems to be called for is before-and-after panel surveys of the same
community subjected to "natural" changes in exposure parameters.
     It should  be noted that the difficulty of fixing the weights for the parameters in a
composite noise index  is not avoided  by using the  Central Tendency  law. When individual
differences are  balanced  out by using median- or average annoyance as the dependent vari-
able in regression, multiple  correlations of the order of .9 are achieved. This leaves little
room for improvement. There has been insufficient secondary analysis of airport surveys to
be emphatic  on the point, but on a priori statistical grounds it does seem that the various
measures that have been developed to deal with aircraft noise (NNI, CNR, NEF, etc.) can be
expected  to  do about  as well in  prediction, despite the rather sophisticated differences
between them.
     The difficulty here in choosing between objective  measures of community noise ex-
posure on statistical grounds parallels that found for the relationship between single noise
measures  and measures  of individual's reactions  under laboratory conditions. Literature
surveys by Botsford  (14) and Parking (15), Young and Peterson (16) and a study of the
varieties of PNdB by Ollerhead (17) imply that, statistically, there is no significant differ-
ence between the various measures. Young and Peterson (16) conclude:
   "For simple  noise reporting and comparisons, the public will benefit,  at no loss of
   precision, if  sound  level A (A-weighted sound level in dBA) is employed for aircraft
   noise as well  as for other kinds of noise in the community."
Endorsing this conclusion, the Serendipity report (12) comments:
   "The point  has been  reached where researchers are striving for a  level of precision
   which exceeds their capability to collect, process  and analyze data."
It seems that this comment, though applied to laboratory studies, may apply equally to the
methodology based on the single cross-sectional survey.
                                         633

-------
9.   Percentage Annoyed and Population Density in Planning
     The above are some of the facts which are not easy to accommodate in planning and
policy decisions for noise control. However,  towards this end, a  helpful notion is the
criterion of "percentage annoyed". We may note that  the percentages in column (b)  of
Figure 2 rise in a  lawful way with noise exposure. The prediction from noise exposure  to
annoyance in percentage or probability terms appears more lawful than the prediction  of
individual annoyance. Percentage annoyance calculated from exposure values can be com-
bined with population density figures to give the absolute numbers annoyed in any region.
Areas in a map can then be shaded accordingly, for such purposes as land-use decisions in
regional planning,  the alignment of runways, and the determination of flight routes. For a
discussion of the application of such community noise indices see reference (21), and for a
critique of the concept reference (22).

10.  Criteria of 'Serious Annoyance'
     On  the percentage-annoyed  criterion, annoyance is treated like an attribute. Since
annoyance is in fact a continuum permitting many degrees, the use is implied of a cut-off
point or fence value below which annoyance is not counted and above which it is treated as
serious.
     Operational meaning can be given to  alternative fence values,  and hence to "annoy-
ance", by describing each score on the annoyance scale in terms of specific types of reaction
which the average  (or median) person reports at that level of annoyance (see Figure 3). The
Wilson Committee (3) chose a score of between 3 and 4,  since this is the point beyond
which informants, unprompted, named aircraft as more disturbing than any other  features
of local living conditions (i.e. the point at which noise disturbance becomes salient, Figure
3c) and tend to mention sleep disturbance (Figure 3d). It is also the point at which people
rate themselves consciously as moderately  to very  much annoyed (Figure 3b), and corre-
sponds to  PNdB values of 103 and an NNI of 50-60. Similar results were obtained in the
second Heathrow survey (9).

11.  Problems for International Standards based on Community Reaction
     The point at which noise becomes salient over other inconveniences of living in an area
may not vary excessively between communities with comparable standards of living. How-
ever, even if agreement can be reached on a cut-off point for  "serious  annoyance", other
difficulties should  be noted. A percentage of noise-susceptible people are seriously annoyed
even at very low levels of exposure (see Figure 2, column (b)). Moreover, for administrators
10 percent of a population of 1,000 is a very different matter than the same percentage of a
population of one million (e.g. around Heathrow).  In general, the setting of standards  of
acceptable living conditions in the light of survey data is a matter of balancing percentages
and absolute numbers against economic, administrative and political considerations. These
considerations  can, of course, vary with both the nature of the planning problem and
differences between countries in planning criteria. A notable attempt to assess the  costs of
noise disamenity in a total  cost-benefit exercise for planning purposes was made by the
Roskill Commission on the Third  London Airport (18). This cost-benefit exercise, however,
has attracted severe criticism on both sociological (19) and economic (20) grounds, not least

                                        634

-------
o\
u>
        Up to  85
     In the  figure, the width  of  the  column represents
     the total number in that PNdB  stratum.
     The shaded section represents the  number annoyed11
     in  that  stratum.
     Total  in  all strata is  1,400,000
     Total  in  shaded  section U  27%  of this.or 378,000
PNdl
Stratum
103 +
100-102
97- 99
94- 96
91-93
88- 90
85-87
Up lo 85
Annoyed #
Remainder
Total
PERCENTAGES
%of total
population
in stratum
a
3
6
7
13
27
22
II
II
% of stratum
annoyed*
ft
68
SI
48
36
24
23
16
10
% of total
population
annoyed*
2
3
3
5
6
S
2
1
27
73
100

mo
ABSOLUTE
No. of peoplr
annoyed *
d
28/100
42,000
42,000
70,000
84,000
70JDOO
28,000
14.000
NUMBERS
No.ol people
in stratum
e
42,000
84,000
98,000
(82/)00
378,000
308,000
154,000
154.000
378POO
1,022.000
1.400.000
1.400.000
For these  data the 'annoyed1 refers to those  having  a score on the annoyance scale
of 3-5 or above.
The population is  that  within a ten mile  radius  of  London Airport: 1,400,000 adults.
Entries  tn column  (c) are derived from those in  columns (o) and (b)  e.g. 68°/«  al
3'/« aivei 2Vo(roun4ed)
Tic numbers in column (e; corresponds to  the percentages in  column (a)  and  those in
column (d)to  column  (c).
                            Figure 2:  Data showing the distribution over noise  levels of the total population and the
                                      annoyed* in the population.

-------
        CO til terms  of  loudness  exposure  Uvtls  (PNdEp
                      nrr.nrr
                      I          2          3
   61234          5-r
The arrows  indicate  the average  annoyance  score of informants  exposed
at that  PNd5 level.
Ot) In terms of self-ratings of annoyance

                (Main   i
     at all       sample) I A ,-mil  Moderately
          Not at all
                                                      Very
                                                      much
                                                      I
                                                              (Special
                                                        complainants
                                                      1 sample)
           6123          4         3+
The arrows indicate the average annoyance score  of informants who said they were 'not
at  all".'a little', "moderately' or'very  much' annoyed  by  aircraft.  Also shown are the
overage  scores for the 20mile area, and  for  the  special  complainants  sample

        (c) In relation  to  other inconveniences of living  in the area
                                          A/C  dislike
                                                     A/C the one thing most
                                                     disliked
                                                           Area  liked less no*
                                                           because of aircraft
                                                                wishes to move
                                                                from area
                                                                because of A/C
           0          I23         4J+
The arrows  indicate the point on the  annoyance  scale at which  replies spontaneously
mentioning aircraft exceed mentions  of any other Item causing inconvenience in  living  in
the  area.
        (0 In relation to the specific typet of  disturbance experienced
           T.V. picture  flickers
                               Interferes  with ,
                               listening to T.V.
                                  Wakes  up
                                  iHousc  vibrates
                                      Interferes  with
                                      conversation      Disturbs rest  or relaxation
                                        	I Prevents from going  to sleep
          6l2          3          4          5+
    Arrows show the point on  the annoyance  scale  at which  50°/o of  people report
    the  specific  type  of disturbance.

            Figure 3: Meaning of scores on the aircraft annoyance scale.

                                     636

-------
for its treatment of the noise  factor.  At present  it remains an open question how far
monetarization of the noise problem can provide a rational basis for decision making.

12.  Complaints as a Criterion: Characteristics of Complainants
     So far we  have been discussing only the relation between noise exposure and annoy-
ance. Now let us look at annoyance reactions in relation to complaint activity. As a group,
numerically, those who register complaints constitute only a fraction of one percent both of
the total exposed population, and even of the annoyed  in this  population who do not
complain. In the Heathrow survey I drew a special sample of complaints and analyzed their
characteristics. Figure 5 shows some  of the results. The main points to note are that, as
expected, complainants, by and  large, are recruited from those who are highly annoyed
(Figure 5c). Like the highly annoyed  who do not complain, and for the same reasons, most
complainants are not found at the highest noise levels, but are spread over all the strata of
noise exposure (compare Figures 5e and 5f).
         60
         20
                 Up to  84    85    88    91     94     97     100
                          Average  peak  loudntss (PN4B)

                  Figure 4: The one thing residents would change in their area.

                                        637
103+

-------
OJ
00
 Figure ?».
 All non-complamcpti
 ToUl WormanU to ZOmile
 area.
 Representing  l.4mllllon
 adults.
                                            Mian
31%




1
1
20%

|

.84

IAOL
10 ft

fe
• »

13%

»=I73I
00%)


9%
Inform



9%
ants



2 /o
        Figure 3b.
        Non-complainants to  the highest
 Representing  3% of 42.000 out of
 l.4million  adults
(Base- (4 8 informants
Ihest
33 + P


out of
oOt
870

NdB)




5%
Mean
score


12%



3-58
1
I
J—
t!4%

t

24%





31%












6%
  Ffgure  »•
  All non-complolnanU
  Total Mormants 'n ZOraili  area.
  Representing 1.4 million adults.
                                                                                                                      Mean PNdB
s area.
dults.
11%
H%
»i
22%
^1
*7%
1
1
1
1
1
1
(pate. 100%

13%
- 1731
7 «/o
adults)
6%
3%
                                                                                                        63    86    80    92
                                                                                                                               98   101    103+
.Figure   »•                                        Mwn ^
  The  most highli  annoyed  non-complainants            95(-2
  Non-complainants scoring 5or6 on the annoyance  scale
  Representing  11% or 154,000 out of
  1.4million adults
 (Base -  100%
  • 243 informants)
                                                                                                        83
                                                                                                                               98   101     105+
        Figure   jc.
        Special complainants sample

       (Bait=l73 informants
         - I00°/j)
                                   0     1     2     3     4
                                             Annoytnct  tetlt

              DISTRIBUTIONS  OVER  ANNOYANCE  SCALE
                                                                         Figure   Sf •
                                                                         Special  complainants sampla

                                                                        (B«»-IOO°/o
                                                                          • 178 informants)
                                                                                                                                   ! Mean PNdB
jpja



.
Ar^


.
n%



15%




29%





94.9
17%







10% 9%


                                                                                                 83    86    89    92    93    98    101    103 +
                                                                                                            Average peak  PNdb

                                                                                            DISTRIBUTIONS  OVER  STRATA   OF  NOISE EXPOSURE
                                Figure 5: Distributions of non-complainants and complainants on the annoyance scale and
                                          over noise exposure strata.

-------
     We ran an elementary discriminant function analysis to see what distinguished com-
plainants from those equally  annoyed, and equally exposed to the noise, who took no
complaint action. Briefly, what we found was that variables of occupational class, educa-
tional level and value of house, which were quite unrelated to degree of annoyance in 1961
now stood  out as important. So did membership  of  organizations and political  activ-
ity. Complainants, in short,  come  from that section of the politically active, articulate
middle class who are sensitive to noise. There was no evidence that they were any more
neurotic than  the equally annoyed  non-complainants, but they did tend to be even more
convinced that the noise could be prevented and that it was affecting their health.
     In general, attempts to understand community response to noise need to take into
account the fact that noise exposure information can describe only the central tendency.
What has been said about annoyance in this respect applies with extra force to complaints.
The relationship between noise exposure and complaint activity is attenuated by all the
psycho-social factors in annoyance plus the further  factors  which determine whether the
annoyed person will complain.

13.  Political and Organizational Factors in Complaint
     Complaints have, so far, been discussed as if each complaint constituted a discrete act
carrying unit weight. To leave the  discussion here would be  to commit "the atomistic
fallacy". In practice, of course, the weight carried by a complainant will be proportional to
his influence in the political hierarchy — and in this respect the ordinary citizen cannot be
equated with his local community leader, still less with his senator or member of parliament.
     The volume and vigor of the complaints received from any community will also depend
on its level of organization for protests. In Bauer's terms "we seem bound to find both some
communities whose organized response is disproportionate to the discomfort of its  mem-
bers, and some whose members suffer in silence because they do not have the disposition or
capacity to organize for protest" (23). Bauer also points out that airport officials and others
responsible for such problems have accumulated  a great  deal of informal,  and largely unre-
corded, knowledge in the course of their, for the most part sensitive and sensible,  dealings
with protesting communities. There has, however, been little social research into the condi-
tions under which individual noise annoyance becomes translated into social action.
     The study of further intervening variables describing the social structure and dynamics
of a community  would  seem to be called for here. For example, take a variable like
"attitude to the contribution the airport makes to local prosperity". In the Heathrow survey
(1), this correlated hardly at all with annoyance. It did, however, correlate—negatively—with
complaint activity. Yet clearly the effect of such a variable on complaint would depend on
the degree of identification of local residents with their community. Many features of the
community  structure could influence this. A town where most people derived their income
from the airport, for instance, would be influenced  differently from a dormitory  town in
which most men commute  to work. More generally, over and above economic factors, there
are many features of the  social  and  political  structure, including the methods  used by
community  leaders to handle problems in the past, which could influence the course of
development of a local noise protest movement.  Such factors will vary between communi-
ties, and vary even more between communities in  different nations.

                                         639

-------
    "
£-520
I!
                   Never felt like moviitf
                   Felt like moving for reuons
                        other then A/C
                               Felt  like novinf beceusc of
            under 90         91-96      97-102
                  Average  peak fetidness (PNdl)
     I5r
  3
  7
  w

  0>
  e
  °»
  0*

  r 10
  2
  Z   5
           Under  90         91-96      97-102
                   Average  peek  loudniis (PNdl)
                                                                 103+
                                    r. For  change ol  scenery.
                                    f. To be merer work.
                                    g. To get eway fro* A/C
 Retso* for wishing to move!
       •.  To go where  climetc  better.
       b.  To go to better  living eccomodetion.
       c.  To get mj  fron smoke/dirt smells.
       i.  To get wcy  from midcsiribfe people.

Question  9A ui  B. Hove  you  ever  felt like  moving ewey  from  this  eree (if
why  did you feel like moving?

Figure 6:  The reasons why some residents have felt like moving away from their area.
                         640

-------
         ±  A
         u
                    Under  91          91-96        97-102
                            Average  peak  loudness   (PNe"&)

                Figure 7: Opinions on aircraft annoyance and the effect on health.
105 *
14.  Long-term Sociological Changes which could Influence Community Response to Noise
     Interesting speculations  on these  changes have  been presented by Bauer (23). It is
by now a cliche that our attitudes towards noise and other forms of pollution are in  a
state of profound transformation, and that we are in the midst of the first serious attempts
by Governments to manage the quality of our environment, and of planners to look at the
full range of consequences of existing  or new technology. The conservationist movement
itself is partly a product of rising levels of affluence which are bringing about improvements
in people's living conditions. As the background level of other environmental inconveniences
diminishes, then the point at which the noise nuisance stands out as salient may be expected
                                        641

-------
    100



     •


     •:
     .
                                                                         90
                                     7 1

•
'




 •
                          '•
                                    4.4
                                -
                                                                               •
                                           i;

                                                      	  NO Fear - No Misfeasance

                                                            Mod Fear  Mod Misfeasance

                                                      o-o-o  High Fear - High Misfeasance

                                                            Bated upon: 4647  British &
                                                                          2912 American
                                                                              interview*
                   96
                        96 99      100 104      106  109

                                           CNR
   110 114       115*

Columbia Universily -  Noise Research Unir
                      Dec  1971
          Figure 8:  Reported high annoyance with aircraft noise by CNR FEAR 8 misfeasance.
to occur at lower levels of exposure.  Further, in many countries there are indications that
large  segments of the  public are coming  increasingly  to challenge  the  technicians' and
administrators' definition of their interest and to want direct representation in the decision
process. "// we are moving  into a period in which individual citizens increasingly expect to
be free from various forms of environmental nuisance and //"citizen groups are tending more
and more  to  take an active role  in  the decision-making process, then  it  is probable that
complaints and effective organized protest will occur at lower levels and frequency rates of
noise exposure than in the past" (23). These speculations run counter to the hypothesis that
                                           642

-------
people will learn to adapt to noise. There is little definitive information either way. Some
relevant data, not held to be definitive, is obtained by comparing the results of the 1961 and
1967 surveys around Heathrow. In the intervening years between the two surveys there was
a considerable increase in the number of aircraft heard (but not in their average loudness).
Complaints increased during this period, but the survey data reveals no  increase for the
general population in either the proportion spontaneously mentioning aircraft noise as a
disturbing feature or the average score on the annoyance scale (see reference 9, chapter 4).
                                  REFERENCES

  1.  McKennell, A.C. Aircraft Noise Annoyance Around London (Heathrow) Airport. U.K.
     Government Social Survey Report SS. 337, 1963.
  2.  Borsky,  P.N.  Community Reactions to Air Force Noise. WADD Technical Report
     60-689, 1961.
  3.  Noise. Wilson  Report, H.M.S.O. Cmnd. 2056, 1963.
  5.  McKennell, A.C.  Methodological Problems in a Survey of Aircraft Noise Annoyance.
     The Statistician, 1969, 19, 1-29.
  8.  McKennell, A.C., and Hunt, E.A. Noise Annoyance in Central London. U.K. Govern-
     ment Social Survey Report SS. 332, 1966.
  9.  Second  Survey of  Aircraft Noise around London (Heathrow) Airport,  London,
     H.M.S.O., 1971.
 10.  Community Reaction to  Airport Noise. Final Tracor Report, NASA: NASW 1549,
     1970.
 11.  Galloway, et  al. NCHRP Report 78. Study  for Highway Research Board, reviewed in
     pages 52-56 in reference 12.
 12.  Serendipity Report OST-ONA-71-1, prepared for the U.S. Department of Transporta-
     tion. A  Study of the Magnitude of Transportation Noise Generation and  Potential
     Abatement. Volume 2, Measurement Criterion, 1970.
 13.  Griffiths, I.D. and Langdon, F.J. Subjective Response to Road Traffic Noise. Journal
     of Sound and  Vibration, Vol. 8, pp. 16-32, 1968.
 14.  Botsford, J.H. Using Sound Levels to Gauge  Human Response to Noise. Journal of
     Sound and Vibration, Vol. 3, No. 10, 1969, pp.  16-28.
 15.  Parkin, P.H. On the Accuracy of Simple Weighting Networks for Loudness Estimates of
     Some Urban Noises. Journal of Sound Vibration, No. 1, Vol. 2, 1964, pp. 86-88.
 16.  Young, R.W.  and Peterson, A. On Estimating Noisiness of Aircraft Sounds. Journal of
     the Acoustical Society of America, Vol. 45, No. 4, April 1969, pp. 834-838.
 17.  Ollerhead, J.B.  Subjective Evaluation of General Aviation Aircraft  Noise. FAA No.
     68-35, April 1968.
 18.  Roskitt Commission on The Third London Airport. Vol. 7, H.M.S.O.,  1970.
 19.  Self, P. Nonsense on Stilts: The Futility of Roskill. Political Quarterly, July, 1970.
 20.  Pearce, D.W.,  Chapman, C.B.,  and Wise J.,  Southampton  University.  Evidence submit-
     ted to the Roskill Commission on the Siting of the Third London Airport, 1970.
                                        643

-------
21.  Richards, E.J. Noise  and the Design of Airports. World  Airports the Way  Ahead,
     Proceedings of the Conference held  in London, September 1969  (Published  by the
     Institution of Civil Engineers,  1970), Paper 15.
22.  McKennell, A.C.  Population  Density and Indices of Community  Noise Annoyance.
     Memorandum  No. 446, Institute of Sound & Vibration Research,  Southampton Uni-
     versity, 1970.
23.  Bauer, R. A. Predicting the Future: Evaluating the Noises of Transportation, Chalupnik
     (Ed.); University of Washington Press, 1970.
24.  Data from an unpublished comparative  analysis of the Second Heathrow Survey (9)
     and the American Tracor study (10), carried out in collaboration with Paul Borsky.
                                       644

-------
                A SURVEY OF AIRCRAFT NOISE IN SWITZERLAND

                     Etienne Grandjean, Peter Graf, Anselm Lauber,
                           Hans Peter Meier,  Richard Muller

              Department of Hygiene and Applied Physiology, Swiss Federal
           Institute of Technology, Clausiusstrasse 25,8006 Zurich, Switzerland
1    Program and Methods
     Noise measurements and social surveys were carried out at the three civil airports of
Zurich,  Basle and Geneva in  1971/1972. The whole investigation included the  following
steps:

—    Noise contours for 30,40 and 50 NNI were calculated on the base of data given by the
     airport authorities and the PNdB contours of the I.C.A.O.
—    We chose randomly 400 households in each noise contour zone of each airport for the
     interviews.
—    A fourth zone without aircraft noise was chosen randomly in Zurich, Geneva and Basle
     as a control group, which consisted of 400 households in Zurich and Geneva, and 200
     in  Basle. The choice was made in  such a way that these control  groups were com-
     parable to the other noise-exposed groups.
-    Aircraft noise was measured around each airport at approximately 140 points. These
     measuring points were located in such a way that representative results for a group of
     interviewed households could be  obtained.  This made it possible for us to determine a
     relatively precise measurement of aircraft noise exposure  for each household.
—    The noise was recorded for 24 hours at each point. We calculated the mean annual
     values in NNI from these records  for day and nighttime (06:00 - 22:00 hours and
     22:00 -  06:00 hours,  respectively). In addition,  the  following noise ratings were
     determined with a statistical analysis:
     cumulative noise levels 1,99,1.59, L\ and LQ \
     equivalent noise level Leq
     noise pollution level Lj^p
     traffic noise index TNI.
—    First analysis of the results showed a low correlation for Zurich between traffic noise
     and individual annoyance. This led us to increase the number of measuring points in
     Basle in order to get a better representativeness of the acoustical data for traffic noise,
—    A questionnaire, based on the O.E.C.D. procedure, was developed for the social survey.
     A  French translation  was  used  for Geneva, which  of  course caused some semantic
     problems.
     Results

                                        645

-------
21   Noise measurements
     Comparison between the previously-calculated noise contours and the measured noise
values showed characteristic differences: the measured values near the take-off runway are
smaller compared with the calculated ones; they  are larger at a greater distance from the
airport. These differences are due respectively to the greater ground attenuation near the
airport and the large deviation of flight passes from the route prescribed by the airport.

22  Noise and Annoyance
     In the interview  the subjects were asked, by a direct question, to rate the degree of
annoyance due to  aircraft noise  on a thermometer-like scale. This direct measurement of
annoyance showed a better  coincidence with noise exposure than the indirect procedure,
using a Guttman scale.
     The correlations of the individual values (Pearson's coefficient) between different noise
rating procedures are given in Table I.
                                        Table 1

         CORRELATIONS BETWEEN 4 NOISE RATINGS AND THE INDIVIDUAL DEGREE
           OF ANNOYANCE (SELF-RATING TEST ON A THERMOMETER-LIKE SCALE).
                      All values are related to daytime noise (06:00-22:00).



Noise ratings
Noise and number index NNI
Noise pollution level L^p
Equivalent noise level Leq
Cumulative noise level L0 ,

Zurich
1471
subjects
0.53
0.44
0.46
0.45
Airport regions
Geneva
1524
subjects
0.68
0.27
0.30
0.35

Basle
944
subjects
0.53
0.16
0.13
0.25
     Differences between the NNI correlations and the other three are significant.
     The differences between the three airports are due to  the  fact that aircraft noise is
dominant in Zurich while in Geneva and Basle, a greater part of traffic noise interfered with
the assessment of annoyance. We conclude, from these noise ratings, that the NNI procedure
is to be preferred, especially when traffic and aircraft noise are mixed.

                                         646

-------
     Figure 1 shows the  mean annoyance values for each airport in relation to the noise
exposure measured in NNI.
     There is a clear difference between Basle and the other two airports. We had reason to
assume that the lower number of overflights in Basle led to an underestimation of the NNI.
In fact, regression analysis showed a better correlation between noise and annoyance if we
put less emphasis on the number of overflights in the NNI formula. Results of this analysis
are shown in Figure 2. Actually, the best regression was found when we used 6.6 x log N for
the number of overflights.
     This result  confirms similar observations  made in a French  (1), and  a British (3)
investigation. We use, below, the corrected NNI for statistical purposes, but the results are
given in traditional NNI when the data are to be compared with other results. It is not our
intention to  prejudice any necessary future internationally-accepted noise exposure assess-
ment with our corrected NNI.
     One of the specific aims of this study was to compare the effects of aircraft noise with
traffic noise. To  this end, we used the same self-rating procedure by asking a direct question
about traffic noise annoyance.  Table 2 shows the  correlations obtained in Basle:  we had
satisfactory acoustical data of traffic noise only there.
     tu
     o
     <
     o
     Q
     UJ
     oc
     U-
     UJ
     CO
        10  -
         8  -
         6  -
          2 -
          0
•   ZURICH
•   GENEVA

A   BASLE
             0
  10
20
30
50      60
70   tINI
       Figure 1: Mean annoyance values for each airport.
               (On the self-rating scale: 10 = intolerable annoyance and 0 = not at all annoyed)
               NNI  = LNP • 15 log N - K
               rind
                    •• 0.559
                rm = 0.914
               Interviewed subjects: Zurich = 1471; Geneva = 1524; Basle = 944.

                                          647

-------

10-


8-


UJ f
o b-
^^
^
o
2
< £l
Q
in
1—
^
CC
S 2-
CO



o

• ZURICH

• GENEVA
A BASLE
x
X
• X
A X •
A X •
• X
x
X _
Ax "
x
X*
X
X
X
•r'"*
x A
x
x'»
•x
X
1 1 1 1 III
0 10 20 30 40 50 CO 70 NNI
                                                                               CORR

 Figure 2: Mean annoyance values for each airport related to a corrected NNI with less emphasis on over-
         flights.
         NNIco,.,. = LNP + 6.6 • log N - K
         rind. = 0.587           rm = 0.947
         Interviewed subjects: Zurich = 1471; Geneva = 1524; Basle = 944.
     The individual  correlation with  the 1,50 value was significantly higher than all the
 others. Further analysis showed a relatively high independency of 1,50 from  aircraft noise,
 while the other ratings were strongly influenced by both types of noise.
     The mean annoyance values in relation to traffic noise expressed in 1.50 are shown in
 Figure 3.
     Another interesting comparison between aircraft  and traffic noise is apparent if we
 consider the spontaneous answers to an open question  concerning disturbing factors in the
 surroundings. The results are given in Figure 4.
     The  most striking point appears in the three histograms farthest to the right: there is a
 decrease in complaints  about  aircraft or aircraft  noise when traffic noise  increases. We
 conclude that the surrounding noise is relevant to  the disturbing effect of aircraft noise.

 23   Noise and disturbed Activities
     The effect of noise on various activities was also studied with the aid of the O.E.C.D.
questionnaire (direct  questions). The results from Zurich and Geneva are shown in Figure 5.
 (We left out the results from Basle because of its small aircraft-noise range.)

                                          648

-------
                                  Table 2

     CORRELATIONS BETWEEN 6 NOISE RATINGS AND INDIVIDUAL DEGREES
              OF ANNOYANCE DUE TO TRAFFIC NOISE IN BASLE.
                All values are related to daytime hours (06:00-18:00).
Noise ratings
      8  -
      6  -
  
-------
DISTURBING FACTORS:
SOCIAL    INSUFFICIENT   TRAFFIC AND      TRAFFIC NOISE
FACTORS   INFRASTRUCTURE  OTHER
        AND         I MM I SS IONS
        SURROUND INGS
                                                           AIRCRAFT NOISE.
                                                           AIRCRAFT
        C.  10 20  0 t 10.20  30 Q  10 20 30 /H)  0 10 20  30 10 50 ^ '0  10 20  30 10 50 ^60 70 Z
    FBI

    < 20
    20-29
    30-39
    40-49
                                                                      LOW TRAFFIC
                                                                      NOISE
                                                                      L  - 10-52
    < 20
    20-29
    30-39
    10-19
    • 20
    20-29
    30-39
    10-10
                                                                      foDERATE
                                                                      TRAFFIC
                                                                      L  - 52-60
                                                                      HIGH TRAFFIC
                                                                      NOISE
                                                                      L  - CO-72
 Figure 4: Disturbing factors related to aircraft and traffic noise in the interviewed subjects' surroundings.
         Ordinate: % of interviewed subjects giving narrnpnnrinj answers. 100% = number of subjects
         answering in each noise category.
             values are expressed in dB(A).
      At first glance, the figures are similar to those of the French and British studies (1,2):
 the communicative activities (conversation and television) are primarily disturbed. A more
 thorough analysis shows marked differences between the two Swiss airports. We must point
 out further that the patterns of night flights are different at the two airports. It is therefore
 doubtful whether we should relate the sleep disturbances to the day NNI values.
      In the Basle region we asked the subjects about disturbances due  to traffic noise. The
 disturbed activities in relation to the traffic noise L5Q are shown in Figure 6.
     There  is an interesting difference in the comparison between aircraft and traffic noise:
the rank  order of disturbed activities is changed. Traffic noise disturbs primarily rest and
sleep while  conversation is less disturbed. We assume  that this is due  to the different
characters of both noises.
     The  disturbed activities recorded at the two Swiss airports were  compared with the
French and British studies, and put together in Table 3.
      Table 3 compares the disturbed activities of respondents near Swiss, French and British
 airports under similar noise exposure. The differences between the three populations are
 remarkable. However,  semantic differences probably intervene so that inferences about dif-
 ferences in reaction levels should be drawn very cautiously.
                                           650

-------
                  z
                  GO-
                  50.
                  40 -
                  30 .
                  20-
                  10-
                   0
  ZURICH
\
  \
   \
                     0
   20
30
50      GO     mi
                  90-
                  80.
                  70-
                  60-
                  50-
                  40-
                  30-
                  20.
                  10-
                   0-
  GENEVA
                                        CONVERSATION

                                        TV AND RADIO
                                        VIBRATION
                                        REST
                                     _  SLEEP
                                     .-  STARTLED

                                        WORK
                     0
   20      30
        40
50
CO
UN I
Figure 5: Disturbed activities du«.
         Geneva: 807 subjects answers
         Ordinate: % of answers in each
         noise in each category.
                      '•irich and Geneva.
                            ~*s answered.
                                  Dumber of subjects annoyed by aircraft
                                             651

-------
      40-
      30-
      20-
      10-
                BASLE
                                                                  REST
                                                               „  SLEEP
                                   STARTLED
                                   TV,  VIBRATION
                                   CONVERSATION
                                   WORK
             46       50

                  L50
54
58
C2
68    oB(A)
Figure 6: Traffic noise and disturbed activities in the Basle region. Subjects were questioned on disturb-
        ances due to traffic noise.
        Ordinate: %  of answers in each  noise category. 507 subjects answered.  100% = number of
        subjects annoyed by aircraft noise in each category.
                                         Table 3
            DISTURBED ACTIVITIES AT SWISS, FRENCH AND BRITISH AIRPORTS.
Kind of
disturbance
French study
84-89 R
British study
47-52 NNI
Swiss study
47-52 NNI
Disturbed subjects
(occasionally - very often)
Startled
Asleep
Awake
Rest and recreation
Radio and television
House vibrating
Conversation
29
20
37
50
75
69
61
59
47
64
44
76
75
'73
51
Us
)
57
68
60
71
                                          652

-------
                                      Table 4

               ACTIVITIES DISTURBED BY AIRCRAFT AND TRAFFIC NOISE.
                                 (Cumulative frequency)
     Kind  Of
     disturbance
Noise exposure values  at  which
50  %  of  subjects  are disturbed
                                    Airports of Geneva
                                    Zurich  and Basle
                                    NNI
                             Region  of
                             Basle

                             L50
Startled
Sleeping
Rest and recreation
Radio and television
House vibrating
Conversation
43
42
44
44
43
45
55
56
56
56
57
57
     Table 4 shows mean NNI and 1,50 values of disturbed respondents for different activi-
ties.  The  mean values  are relatively stable for aircraft as well as traffic noise. It is also
interesting to note that a mean value of 44 NNI for aircraft noise corresponds to the one of
56 dB(A) for traffic noise.
24   Noise, Behavior and Weil-Being
     The answers to direct questions about the effect of aircraft noise on the use of ear
protective devices, consumption of sleeping pills and the need to consult a doctor are given
in Figure 7.
     There is a clear relation between these behavioral patterns and noise exposure.  From
the  medical point of view, we consider the strong increase of subjects taking sleeping pills
because of aircraft noise as  a severe health-hazard.
     The answers to direct questions about social habits are shown in Figure 8.  Here, too, is
a clear effect  of noise. We  consider these restricting effects on social habits as a serious loss
of life quality.
     Noise has also a strong effect on people closing their windows (see Figure 9).  Open
windows and fresh air indoors are  conditions for healthy living. The necessity to close
windows because of noise can also be evaluated as a loss of life quality.

                                       653

-------
         CONSULTS DOCTOR
   'CORR  0
                10
20
  < 8
  8-14
  15-22
  23-29
  30-37
  58-44
  45-52
  53-GO
            462
            357
            585
           1091
            730
            5C2
             73
                                                    UNI
                                                       CORR
                                     USES EAR PROTECTION

                                     0     1C     20
< s
8-14
15-22
23-29
30-37
38-44
45-52
53-60
140
] 462
357
n 585
_J 1091
I 730
1 SD?
1
                                                            TAKES SLEEPING PILLS
                                                      'CORR

                                                    < G
                                                     8-14
                                                    15-22
                                                    23-29
                                                    30-37
                                                    38-44
                                                    45-52
                                                    53-60

Figure 7: Effect of aircraft noise on use of ear protection, consumption of sleeping pills and consultation
        of doctor. 100% = number of interviewed subjects in each noise category.
     Like the French and British studies, we also included some direct questions, for those
who are annoyed by aircraft noise, on their willingness to do something about it. The results
are given in Figure 10.

     As expected, the real protests are less frequent than the intended ones. But both show
a clear connection to noise exposure.
     Figure 11 shows that the annoyed subjects primarily consider the political authorities
and technology as responsible for the insufficient noise abatement. It is striking that, with
increasing noise exposure, the responsibility for insufficient abatement is shifted from politi-
cal authorities to technology.
                                           654

-------
                                                   [ml
                                                          WISHES TO MOVE AWAY
                                                      CORR
0
                           10
                                                                       20
                                                             30
REMAINS LESS OUTDOORS
lliUrr.Dt.
CORR o jg 2g jg j.
< 8
8-14
15-22
23-29
30-37
38-44
45-52
53-60
k
"I
1

 N

 140
 462
 357
 585
1091
 730
< 8
8-14
15-22
23-29
30-37
38-44
53-60
140
L 462
1 	 357
. . 1,, „ HJK
T
1 1091
1 7*f'
1 ^02
_ 	 	 , 1 73
                                                   Niil
                                                          GOES OUT MORE  FREQUENTLY
                                                      CORR
                                        0
      10
                                  20
                                                                              30
< 8
8-14
15-22
23-29
30-37
58-44
45-52
53-00
140
] 4G2
357
11 585
_1 	 , 1091
1 7^n
1 w
I
1 75
Figure 8:  Effect of aircraft noise on the wish  to  move away, frequency of going out and remaining
         outdoors. 100% = number of interviewed subjects in each noise category.
      NNI
          CORR
CLOSES  WINDOWS

0        10        20
                                               30
                    40
         50
\
1

1

140
462
357
585
1091
730
502
1 73
       < 8
       8-14
       15-22
       23-29
       30-37
       38-44
       45-52
       53-60
 Figure 9: Effect of aircraft noise on windows being closed more often than desired. 100% = number of
         interviewed subjects in each noise category.
                                             655

-------
        HAVE ALREADY COMPLAINED OFFICIALLY
   'CORR
              20
10
CO
80
« 8
8-11
15-22
23-29
30-37
38-11
15-52
53-60
110
1C2
357
_l 585
~ 1091
' 730
1 
-------
                  z

                  GO -


                  50 -


                  10 -


                  30 _


                  20 -


                  10 -
.  AIRPORT AUTHORITIES

•  POLITICAL AUTHORITIES
.  INTERESTED GROUPS
x  NOISE EXPOSED PEOPLE
o  TECHNOLOGY

A  FOREIGN AIRLINES
                         |<15  !l5-22|23-29|3C-37|38-H<6-52|53-eo| NNI
                     N -  53    GO   250   MO   185   351   C5
            CORR
Figure 11: Judgement on responsibility for insufficient noise abatement 100% = number of subjects who
considered actual abatement as insufficient in each noise category.
25   Influence of social and psychic Factors on individual Hyperreactivity to Noise
     Hyperreactivity only  stands for reported annoyance in this study. In fact, we define
hyperreactivity as a deviation of the reported annoyance from the mean values.
     A detailed discussion  of all results derived from this approach would exceed the limits
of this paper. We shall restrict ourselves only to the compilation shown in Figure  13.
     All characteristics given in the Figure, though highly significant, have a relatively small
explanatory power, the clearest being found in the  characteristic of strong  fear  of air
crashes.
     On the other  hand, we were surprised that income  and education had no direct in-
fluence on  hyperreactivity; but these factors have an indirect influence on "attitudes" and
"flight experience". Furthermore, the influence of age on hyperreactivity is eliminated if
"residence duration" is considered.
                                           657

-------
          RESIDENCE SATISFACTION
          0    20    40    60    I
40-45
44-47
48-51
52-55
56-59
60-63
C4-72
1

|
|
|
]
I
tJlll
     RESIDENCE SATISFACTION
CORK  o     20     40     60
                                               <  o
                                               8-11
                                               15-22
                                               23-29
                                               30-37
                                               38-44
                                               45-52
                                               53-CO
                              80 Z
                                    N
                                  " 140
                                    462
                                    357
                                    585
                                   1091
                                    730
                                    502
                                    73
          SUPPOSED EXTERNAL JUDGEMENT

          0     20     40     CO
40-43
44-47
48-51
52-55
S6-59
60-C3
64-72


1
1

|
I
I


HillCORR

< 8
 8-14
15-22
23-29
30-37
38-44
45-52
53-60
     SUPPOSED EXTERNAL JUDGEMENT

     0    20    40     60
80

                                                                                   N
                                                                                   140

                                                                                   357
                                                                                   585
                                                                                  1091
                                                                                   730
                                                                                   502
                                                                                    73
Figure 12:  Relation between residence satisfaction and supposed external judgement to aircraft and traffic
          noise. 100% = number of interviewed subjects (N) in each noise category.
Summary
     Noise measurements and social surveys with 3939 interviews were carried out at 3 civil
airports in Switzerland.
—   The  self-rated  annoyance gave the best correlation with aircraft noise if this was ex-
     pressed in NNI or in a corrected NNI (with less emphasis on the number of flights).
-   The  self-rated  annoyance related  to traffic noise  gave the best correlation  with  the
     cumulative noise level L$Q (rjncj  = 0.43).
—   There is a decrease in complaints about aircraft noise with increasing traffic  noise for
     the same aircraft noise exposure.
-   Aircraft noise disturbs primarily communicative activities, while traffic noise interferes
     more with rest and sleep.
—   Use of protective devices, consumption of sleeping pills, consultation of doctors, wish
     to move away,  real and intended political protests, tendency to spend less spare time at
     home, to go less outdoors or to keep windows closed, and residence dissatisfaction are
     increased with higher aircraft noise exposure.
-   A factor analysis showed  no significant  relation between a "mental  health status-
     factor" and noise exposure.
                                           658

-------
                                     MARITAL STATUS:
                                     HARRIED; WIDOWED
STRONG FEAR OF
AIR CRASHES
»,

CHARACTERISTICS OF HYPER-
REACTIVE INDIVIDUALS


NEGATIVE PROJECTED INTER-
PRETATION OF NOISE SOURCES
     No HABITUATION

     TO NOISE
                  LOW PERSONAL BENEFITS
                  FROM AIRPORT
NEGATIVE ATTITUDES TO
MODERN TECHNOLOGY. RETRO-
SPECTIVE ORIENTATION
Figure 13:  Characteristics of hyperreactive individuals concerning aircraft noise. Diagram shows individual
          characteristics which are significantly over-rated by the group of hyperreactive individuals.
     A group of subjects hyperreactive to aircraft noise revealed several significant charac-
     teristics  such as fear, long residence duration, negative  attitudes and no flight ex-
     perience.
3.

4.
                                 REFERENCES

Centre Scientifique et Technique du Batiment, Paris: La gene causee par le bruit autour
des aeroports. Rapport de fin d'etude ler mars 1968.
Committee on  the problem of noise:  Noise, final report. London H.M. Stationery
Office,  1963.
Office of population censuses and surveys: second survey of aircraft noise annoyance
around  London (Heathrow) airport.  London H.M. Stationery Office, 1971.
Rylander R., Sorensen S.  & Kajland A.: Annoyance reactions from aircraft noise
exposure. J.  Sound Vibration 24, 419-444, 1972.
                                           659

-------
             AIRCRAFT NOISE DETERMINANTS FOR THE EXTENT OF
                            ANNOYANCE REACTIONS

                        Ragnar Rylander and Stefan Sorensen

            Institute of Social and Preventive Medicine, University of Geneva,
                    and the Department of Environmental Hygiene,
                  National Environment Protection Board, Stockholm.

1.   Introduction
     From  the public health point of view, a satisfactory  control of environmental agents
can only be achieved in light of detailed information concerning dose-response relationships
(WHO, 1972). This presentation proposes to  review various indices to express aircraft noise
in relation  to the  extent of annoyance  in exposed communities, and to discuss the dose-
response relationships found.

2.   Analysis
     Aircraft noise exposure is  a combination of several  physical factors. The number of
overflights, their duration and frequency spectra are some of the most important physical
characteristics. It has been assumed that several of these contribute to the exposure reaction
and  thus should be included in  an index to express the noise exposure.  These indices are
usually  constructed according to the acoustical  principle of "equal energy"  and contain
expressions for the number of exposures as well as some mean noise level.
     The NNI index was derived as a result of the first  Heathrow study (HMSO, 1963). The
concept of the index is that both the noise exposure level and the number of overflights are
significant for expressing the exposure.
     The exact formulation of the index was obtained by adapting it to the data from the
field investigations using weighting factors. It should be noted that the investigation  areas
with different exposure levels were not distinctly separated  and that some combinations of
overflights/noise levels were poorly represented in the material.
     Another equal-energy index was developed according to results from an investigation
around  Schiphol (Kosten et al., 1967).  Interviews were performed  in 6 areas around the
airport and in addition to constants for the number of aircraft movements, arbitrary weight-
ing factors were included for the distribution of the flights over different hours.
     The R. index was  derived from results  from a  French  investigation performed in
1965-66 (CSTB, 1968). Social surveys as well as noise measurements were performed  in 20
areas around the airports of LeBourget, Orly, Marseilles and Lyon. The results on the extent
of annoyance in the various areas were used to develop the French aircraft exposure index
R. Again a technical index was fitted to the results from the field study with an adjustment
of weighting factors to  give the best adaptation.
     A  major study has  been reported from the United States (Hazard,  1971). Interviews
were performed around the airports of Atlanta, Dallas,  Denver and Los Angeles. An analysis
of the correlation between  exposure to noise and the extent of annoyance demonstrated
that the CNR index was about  equal to NNI. However,  the CNR index was found  to be
more stable in relation  to variations in socio-psychological factors.

                                        661

-------
     Various international bodies have then further refined the indices by adding different
weighting factors or modifying the original values. Much of this work is based upon assump-
tions; sufficient information on the importance of these adjustments for the development of
the exposure reactions is not present.
     Table 1 summarizes some of the noise-exposure indices that have been related to the
extent of annoyance reactions in the community.

                                       Table 1

          EXAMPLES OF INDICES USED TO EXPRESS AIRCRAFT NOISE EXPOSURE.

     NNI = PNdB max + 15 logN - 80                              (England)
     B = 20 log 2 n - 10LA/I5 .157                               (Holland)
     R = PNdB + 10 logN - 30                                     (France)
     CNR = PNdB max + 10 logN + Kj                                (U.S.)
     LPNeq = 10 log 1 QEPNdB/10 + i Q log ^                        (ISO)
     In a review of then-available results, Alexandre (1970) was able to demonstrate a very
close correlation between the various types of noise indices. By standardizing the expres-
sions for the annoyance reaction, he was also able to show a close agreement among various
investigations concerning  the  dose-response relationship. The  correlation coefficient be-
tween noise exposure expressed as NNI and annoyance was r = 0.92.
     From a scientific point  of view, when an index is developed by making a best fit to
existing data, it will be necessarily valid  only for that particular set of data and cannot be
generalized. This is illustrated in the second Heathrow study (HMSO, 1971) in which the
weighting factor for "log N" was found to be relatively arbitrary. As was pointed out in the
report, the previously suggested value 15 can only be looked upon as one of several weight-
ing factors that can  be  used with the  same degree  of  accuracy. This  conclusion is also
supported by findings in the recent Swiss investigation, in which the weighting factor for the
NNI index was found to be about 7,  again by adapting the NNI index against existing data
from the investigation.
     Also  in  some investigations mere coincidence  will make the  data  support the equal
energy concept. This is particularly true for the data from the French investigation 1965-66.
In the 4 areas with the highest exposure, an increasing number of flight movements was
always accompanied by a higher noise level,  although  these  factors are naturally inde-
pendent from each other.
     In summary, the method of adapting data to a best fit relationship for the development
of noise exposure indices fails to "prove" that the so-developed index is generally valid. This
is demonstrated by the variation in the  mathematical formulation of indices found in the
different field investigations.
     In a recent Scandinavian  investigation on annoyance due to aircraft noise  exposure, a
different method was used to  analyze the noise exposure. In the experimental design, the
number of exposures and the  noise levels were kept as independent variables. Variation in

                                        662

-------
noise levels expressed in dB(A) were obtained by choosing areas at different distances from
the runways. Variation in the overflight frequency was obtained by studying areas around
different airports. The number of exposures was expressed as the total aircraft noise events
(landings and take offs) exceeding 70 dB(A).
     When the results were evaluated against the equal-energy indices used in other investi-
gations (NNI, CNR, etc.) the correlation between the extent of mean annoyance and noise
exposure was found to be similar to that demonstrated in earlier studies (correlation coeffi-
cients .5 7-. 74).
     However, when  the  analysis considered the  number of overflights and noise levels
separately, a new reaction pattern  developed. It was found that if the areas were divided
into such exposed to  a low (< 35) and high (> 50) number of exposures, the extent of the
annoyance in the  areas was  determined by the nominal noise contour from the  noisiest
aircraft in the area, expressed in dB(A).
     For high exposure areas with exposure frequencies from 50 to 180, a linear correlation
was obtained between dB(A) and annoyance (r = 0.99). In low exposure areas, the extent of
annoyance was low, up to noise levels of 90 dB(A), whereafter an increase was found.
     The  dose-response  relationship for the high/low exposure area principle versus dB(A)
levels and corresponding NNI values is illustrated in Figure 1.
     It is demonstrated in the figure that  the correlation  between noise exposure and the
extent of annoyance was fair, when the noise exposure was expressed as NNI, but improved
if a division was made into high and low exposure areas.
     The conclusion drawn from this analysis is that both the number of overflights and the
noise level are of importance in determining the extent of annoyance caused by exposure to
aircraft  noise. The number of overflights  is only  used to divide the areas into exposure
categories, after which the noise contour from the noisiest aircraft will determine the extent
of annoyance.
     This finding represents a new concept  in the relation between aircraft noise and human
reactions. If it is to be proclaimed  generally valid,  it should be present in investigations
performed earlier, provided the same analysis principle is applied. To evaluate this hypothe-
sis, a cooperation was established with the researchers who had participated in the French
investigation from 1965-66. The areas involved in their studies were classified into high-or
low-exposure categories by using the number of overflights recorded in 1965-66. The noise
level in each area  was determined by plotting the nominal noise contours for the noisiest
aircraft  at the time,  along the flight paths from the different airports. The  results are
illustrated in Figure 2.
     It is seen in the figure that the same type of dose-response relationship as demonstrated
in the Scandinavian investigation could be found in the  French study.  This finding was
verified in the analysis of 5 additional areas from a French investigation performed in 1971.
     A similar re-analysis is underway concerning the 1963  Schiphol study, the Japanese
Yokota airbase study, the Swiss study from 1971-73 and the German study from Munich
1969.
     In  summary,  an analysis of the now  available  data on  the relation between aircraft
noise exposure and the extent of annoyance has demonstrated the following:
     1.    A relatively good correlation is found  between aircraft noise exposure expressed
          as different equal-energy indices and the extent of annoyance reactions.

                                         663

-------
                                       /o very annoyed
                                        40
                                        10
                                              10    20    30    40    SO
                                                                      NNI
  v«fy annoyed
               HIGH EXPOSURE
40
30
20
10
                                                                      LOW EXPOSURE
                                                       /c v*ry annoyed
       eft
                                                      20
                                                      to
       TO    SO    90    tOO
                           dB(A)
                                                            TO   80   »0    100
                                                                                    dB(A)
       Figure 1: Relation between aircraft noise exposure and annoyance for Scandinavian material.
      2.   An improved correlation can be demonstrated if the overflight frequency is used
           only to classify  areas into exposure categories, whereafter the extent of annoy-
           ance is related to the noise level in dB(A) from the noisiest aircraft type.
      For the time being these conclusions are only valid  for exposure frequencies up; to
 about  350/24 hours and for a diurnal traffic pattern that comprises about 10% night traffic.

 3.   Comments
      The  evaluation of the relation between aircraft noise exposure and  the extent of
 annoyance according to the principles developed in the Scandinavian investigation, by and
 large represents  an increase of the correlation between exposure and effect from about 0.85
 to 0.99. At first glance, this increase in correlation might seem to be of academic interest
 only. There are, however,  some practical consequences which become important from a
 public health point of view.
                                           664

-------
     If the NNI principle is used as illustrated for the Scandinavian material in Figure  1, a
specific NNI value could represent  a  large variation of the extent of annoyance in  the
exposed area (30 NNI = 3-40% very annoyed). This variation around a planning norm such
as NNI is not acceptable. The same applies to CNR where for the value 90 one can find
between 3 and 20% very annoyed.
     If the noise contours based  upon the NNI concept are compared to those based upon
the max dB(A) concept, important differences are found in the extension of the runways.
Figure 3 illustrates NNI values and dB(A) contours from the  noisiest aircraft type at an
existing European medium-large airport. The NNI calculations are based  upon noise meas-
urements. It is  seen in the figure that  the critical noise contours based upon the NNI
concept will end sooner  than those based  upon the dB(A) max concept. A  risk of  not
protecting  certain areas in the  airport surroundings might therefore be present  if equal
energy noise indices are used. The new contours are in certain cases slightly narrower at the
°/
Xo very
M
                            SO
                            30
                            10
                                 annoyed
                                         m* '   •
                                         ^•*«
                                  60    70   BO    tO   100
  very annoyed
                HIGH EXPOSURE
 60
 30
                                                     So very a
                                                           innoyed
                                                                   LOW EXPOSURE
                                                      30
                                                      20
      • 0
           TO   »0    M   100
                            d*r«>
                                                           fO    70   10    M   100

         Figure 2: Relation between aircraft noise exposure and annoyance for French material.


                                          665
                                                                                   dICAl

-------
Figure 3: Noise contours from max dB(A) level (solid line) and NNI (dotted line) around an airport.
                                          666

-------
sides of the runway, since the new principle does not take into account the number of
aircraft movements once this exceeds 50 per 24 hours.
     According to the  new principles, attention should be focused on the noisiest type of
aircraft using an airport a certain number of times per day. The critical zones used should be
based upon the noise contours of this aircraft and will not be affected by a decrease or an
increase in the number of aircraft movements. A change of the critical noise zone due to the
number of aircraft movements occurs only when a specific area is transformed from the high
exposure to a low exposure category,  e.g. when flight diversion  is applied after take-off. If
the noisiest aircraft type is changed to a quieter aircraft, the critical area will decrease
corresponding to the difference between the two noise contours.
     From a practical point of view, the new principles thus offer interesting possibilities for
traffic control  in order  to diminish  the annoyance  caused by aircraft. For each traffic
situation it is possible  to plot  the  critical  noise contours based upon the noisiest type of
aircraft. Noisy  aircraft could be routed out of and into the  airport along special  zones
outside populated areas of towns. An alternative would be to  apply noise abatement pro-
cedures, which would  not have to  be enforced upon the less noisy  aircraft providing that
their noise contours stay within those caused by the noisiest aircraft abatement.
     Considerable effort is presently being devoted to continuing the evaluation of the two
aircraft noise exposure  concepts:  the equal-energy  principle and  the maximum-noise-
type-level principle. The re-analysis of previous investigations  is still underway and field
experiments  are being designed to  test, at the  same location, the validity of the two con-
cepts for expressing aircraft noise. Only when this work is finished can it be decided if the
new principles are generally valid to establish dose-response relationships for aircraft noise.
In view of the important consequences for the formulation of criteria and standards, this
work is given a high priority.
                                    REFERENCES

WHO: Environmental Health Criteria and Standards. Report from a Committee, November,
      1972.
H.  M. Stationary Office, London, CMND 2056, 1963, Final Report: Committee on the
     problem of noise.
C.W. Kosten, G.W. de Zwann, M.H. Steenbergen, C.A.F. Falkenhagen, J. A. C. de Jonge and
     G. J. van  Os, 1967, T.N.H.  Report 1-119. Geluidhinder door vliegtnigen.  (To be
     published).
Centre Scientifique et Technique du Bailments, Paris  1968. La gene causee par le bruit des
     aeroports.
W. R. Hazard, 1971, Journal of Sound and Vibration, 15: 425-455. Predictions of noise
     disturbance near large airports.
A.  Alexandra,  1970, Anthropologie Appliquee 28/70, 1-151. PreVision de la gene due au
     bruit autour des aeroports et prespectives sur les moyens d'y remedier.
H.  M. Stationary Office, London 1971. Second survey of aircraft noise annoyance around
      London (Heathrow).

                                         667

-------
     REACTION PATTERNS IN ANNOYANCE RESPONSE TO AIRCRAFT NOISE
                Stefan Sorensen, Kenneth Bergtund and Ragnar Rylander

                        National Environmental Protection Board
                                  Stockholm, Sweden

1.    Introduction
     The relation between exposure to aircraft noise and the extent of annoyance reactions
has been studied in several investigations, and various factors influencing the development of
annoyance have been  evaluated (1).  The purpose  of this presentation is to analyze the
annoyance reaction  with reference to the  dose-response relationship  found in a recent
Scandinavian investigation (2). In the  presentation, the importance of a precise description
of the duration of the noise exposure will be estimated. Different expressions for annoyance
will  be  evaluated, as well  as reactions among persons with  different  individual charac-
teristics. The  different  factors comprising the annoyance will be evaluated using factor
analysis.
     Figure  1 demonstrates  the dose-response relationship obtained in the Scandinavian
investigation for high exposure areas.
            Very annoyed
           kO -
           30 -
           20 -
           10 -
                    —\—
                    70
—1—
 80
	T
 90
                                                                      dfl(A)
                           1OO
                 Figure 1. The dose-response relationship for high exposure areas.
                                         669

-------
     The subsequent analysis will be made with this dose-response relationship as the experi-
mental model.

2.    The noise dose
     The exposure to noise is usually determined by measuring or estimating the noise level
in  the investigation area.  In order to  improve the  dose estimation, the individual dose was
calculated for each respondent by determining the number of overflights to which each
individual was exposed. The respondents were then divided into two groups, those exposed
to 50-120 and to more  than  120  overflights per day. The  results showed that in both
exposure  groups  the same original  dose-response relationship is  found. The results thus
obtained support the conclusion from the earlier  analysis that the number of exposures is
not a determinant of annoyance, once these exceed 50 per 24 hours.

3.    Different intensity of annoyance
     The Scandinavian investigation uses the expression "very annoyed" as a measure of the
extent of annoyance in the exposed area. If instead the expression "rather annoyed" is used,
the dose-response relationship  becomes  less  pronounced, and for the expression "little
annoyed" the extent of the reaction is independent of the noise level.
     The  dose-response relations for different levels of annoyance indicate that the  lower
the annoyance intensity,  the lower the correlation to  the  dose.  Annoyance scales which
include  expressions for low annoyance thus seem to be less precise than those applying
strong expressions for annoyance only.

4.    Individual reactions
     From a practical point of view it is desirable to define the annoyance reaction in an
individual rather than work with means from groups of respondents. This will allow a more
exact prediction, especially in populations which for various reasons differ from the average.
     Several  attempts have been made to increase the relatively poor correlation between
noise exposure and individual annoyance  by  construction of annoyance scales or by analyz-
ing individual factors such as socio-economic conditions. In the following,  the influence of
some individual characteristics for  the extent  of annoyance will be analyzed.  Figure  2
illustrates the relation between  the dB(A>level and the extent of "very annoyed" for men
and women.
     It  is seen in the  figure  that the same dose-response  relation exists for both  sexes.
Women show a slightly lower annoyance, but this difference was not statistically significant.
     The  dose-response relation for individuals remaining in  the  area during the day  and
those working outside is shown in Fig. 3.
     A difference is found between the two groups, and this increases with increasing dB(A)
levels. At the 90-dB(A) level the difference in annoyance is about 15%.
     The dose-response relation for different age groups is illustrated in Fig. 4.
     The age groups 31-50 and 5 1-70 show similar functions but  the  20-30-year group  was
found to be different. At  70 dB(A)  the extent of annoyance was the same as for the other
groups but at 90 dB(A)  the  young age group reports about 25% less annoyance than  the
other groups. This pattern  is found for both men and women (Fig.  5).

                                        670

-------
                    % Very annoyed
                        A
                    50  -


                    40  -


                    30  -


                    2O  -


                    10  -
                                   70

                             • Hen
                             ii Women
                                                 —I—
                                                  80
—I—
 90
dB(A)
                          Figure 2.  Dose-response relation for men and women.
                     Very annoyed
                   50  -
                   30  -
                   20  -
                   10
                                                                                
-------
         % Very annoyed

         50  -
         30  -
         20
         10  .
                        70
                                      80
                                                     90
                                                                 -> dB(A)
                   — Age i  20-30
                   --- Age i  31-50
                   ------- Age:  51-70
              Figure 4.  Dose-response relation for different age groups.
          50 -
          30 -
          20 -
          10 -
                                                             Age: 31-70
                                                             Age! 20-30
                                                                      dB(A)
                        70
                                      80
                                                     90
                   '  • ..... Women
                  --- Men
Figure 5. Dose-response relation for different age groups divided into men and women.
                                      672

-------
     The difference in reaction pattern between age groups is found also when the respond-
 ents remaining in the area during the day are compared to those working outside (Fig. 6).
     The results from these analyses show that different reaction patterns exist for different
 individual characteristics but that they are present only at higher noise levels.

 5.   Expressions for annoyance
     An analysis was made of the relation between noise exposure and different expressions
 for annoyance. The results are presented in Table 1.
     It is seen in the table that television flicker was poorly correlated to the dB(A) level,
 which indicates that it is not a relevant expression for annoyance. A high correlation exists
 between the dB(A) level and disturbance of telephone  conversation, normal conversation,
 and listening to  radio/TV. This indicates  that annoyance due to aircraft noise exposure to a
 great extent can  be defined as communication interference.
     A factor analysis was performed for the various components of annoyance at different
 noise levels. The results are shown in Table 2.
     It is seen in the table that at the lower noise levels the factor that explained most  of
 the variance consisted of fear, nervousness and awakening. At  the higher noise level, the
 factor that explained most of the variance was based on interference with relaxation and
 sleep.
                 % Very annoyed
                     A*

                 50  -




                 kO  -




                 30




                 20




                 10  -
Age: 31-70
Age: 20-30
                              70
                                          80
                                                                  ->  dB(A)
                         _^_ Remaining in the community during the day
                         -• —— Working outside the conniunity
Figure 6. Dose-response relationship for different age groups divided into those who remain in the area
during the day and those working outside.
                                           673

-------
                                       Table 1
                 CORRELATION BETWEEN NOISE LEVEL AND DIFFERENT
                           EXPRESSIONS FOR ANNOYANCE.
Type of activities/disturbances
Conversation
Radio/TV
Telephone
Sleep disturbance
Awakened
Rest/relaxation
Vibration
Startle
Flickering TV-picture
rxy
0.96
0.98
0.99
0.63
0.78
0.57
0.96
0.42
0.14
                                       Table 2
                RESULTS OF FACTOR ANALYSIS AT DIFFERENT NOISE LEVELS.
70 dB(A)
Fl
Frightened

Nervous
Awakened

F2
Radio/TV

Telephone
Conversation

90 dB(A)
Fl
Sleep
disturbance
Awakened
Rest/relax-
ation
F2
Radio/TV

Telephone
Conversation

    The second most important factor for the variance was found to be communication
disturbances which was the same for both noise exposure levels. The factor analysis further
demonstrates that house vibrations, although closely correlated to the dB(A) level, did not
form part of the annoyance pattern.
    The results from the factor analysis thus demonstrate that reaction patterns for annoy-
ance are different at different dB(A) levels. In view of results presented above, the reason
for this could  be the influence  of age; i.e., older people are more easily disturbed during
sleep.
                                        674

-------
     A further analysis was  therefore made for different age  groups and for respondents
who did not remain in the area during the day. The results for the 90 dB(A) level are shown
in Fig. 7.
                  Mean factor  score
                  1.0 -
                  0.5 -
                      Age:  20-30
31-70
                             ii- Remaining in  the community during the day

                         — — — Working outside the community


                Figure 7.  Mean factor score (factor 1: sleeping problems) for those
                who remain in the area during the day and those who work outside
                the area, divided into two age groups.
     The figure shows the "mean factor score" for sleep disturbance according to age and
stay in the area during the day. It is seen  that younger respondents reported  less sleep
disturbance due to aircraft noise and that respondents who work outside the area during the
day report  less sleep disturbance than those remaining in the area during the day.
     A different  pattern is obtained if the second factor for annoyance-communication
interference—is analyzed in a similar way. These results are illustrated in Fig. 8.
     It is seen  in the  figure that the importance of communication interference is greater
among those who work outside the community during the day. However, communication
interference did not vary with age.

                                         675

-------
                 Mean factor score
                  1.0  -
                  0.5
                         Ape:  20-
31-70
                         ^• Remaining in the community during the day

                         — —•— Working outside the  community

                 Figure 8. Mean factor score (factor 2: communication  interfer-
                 ence) for those who remain in the area during the day and those
                 who work outside the area, divided into different age groups.
6.   Comments
     In summary, the following conclusions can be drawn from the analysis performed on
respondents exposed  to noise consisting of more than 50 flight exposures/day.
     1. The  expression  "very  annoyed" shows  the  best dose-response  relationship with
noise exposure.
     2. In the group  of respondents, the age and the time spent in the area during the day
are determinants of the extent of annoyance.
     3. The various components in the overall annoyance are different  at different noise
levels.
     4. At higher noise levels, the difference in reaction between young and old age groups
can be explained  by differences in sleep disturbance. Communication interference was equal
for different ages.
     Several of the components of annoyance analyzed here have earlier been  used  to
construct annoyance scores, e.g. the Guttman scale.
     Against the results from the present study, based upon the dose-response relationship
found  in high exposure  areas in the Scandinavian investigation, several of the factors con-
tained  in earlier scores  were found  to  be less important for annoyance. The relative im-

                                         676

-------
portance of the various components was also found to vary with the noise levels and with
certain individual characteristics.
     The Guttman scales are therefore not  ideal to measure the extent  of annoyance in
exposed  communities,  either to determine individual reactions or to establish mean reac-
tions in a group of individuals.
     With the techniques available today it is  questionable if meaningful dose-response
relationships  can be established  for individuals. It  will thus not be possible to use the
reaction  of individuals  to establish criteria, and consequently some kind of a mean reaction
has to be used. Studies are underway to further refine the measurement, with the objective
of finally establishing techniques which will  make the measurement of individual reactions
possible.
                                   REFERENCES

     Hazard, W.R.  1971. Predictions of noise disturbance near large airports. Journal of
     Sound and Vibration. 15:425-445.
     Rylander, R.,  Sorensen, S. and Kajland, A.  1972. Annoyance reactions from aircraft
     noise exposure. Journal of Sound and Vibration. 24:419-444.
                                         677

-------
            THE REDUCTION OF AIRCRAFT NOISE IMPACT THROUGH
                 A DYNAMIC PREFERENTIAL RUNWAY SYSTEM*
                                    Martin Gach
                              Freeport, New York 11520

          the exception of Sections 1  and 2, significant extracts are included in this paper from
References 2 and 3.

INTRODUCTION
     One of the unfortunate results of the growth of a highly mechanized and urban society
has been the gradual increase of noise levels to which the ordinary citizen is constantly
exposed. Power-driven machines make noise, and since machines serve people and are cen-
tered in  urban centers, the  impact of such emission on  concentrated populations which
continue to grow  becomes  a matter of increasing concern. Accordingly,  the  problem of
mass transportation associated with the development of larger and more powerful aircraft,
notwithstanding that noise  complaints existed  prior to  the  advent of such aircraft, has
become particularly severe,  especially in  communities surrounding airports.  The  intro-
duction of jet transport aircraft has raised the problem to a critical level.
     This  problem  of  aircraft noise presently,  and in the foreseeable future, is directly
related to airport operations since an airport produces a high concentration of low-flying
aircraft at its boundaries, with increased engine  settings, and in narrow flight paths.  These
factors tend to emphasize and increase the exposure of communities to noise.
     The concern with  this  problem develops from two  different but complementary re-
quirements. First, as a matter of public responsibility it is necessary, to the extent possible,
to protect communities from the effects of aircraft noise. This is directly  related to the
second requirement, which is the  orderly development and effective  functioning of the air
transport system and its protection from undue harassment. The resulting combination of
requirements,  when rationally  considered,  provides a  better basis for understanding the
relationship between the airport's operation  and the community's reaction.
     In  order to  improve this relationship, the Federal Aviation Administration Eastern
Region Noise Abatement Office conceived,  for the summer of 1970, the development and
trial of a modification to the existent preferential runway system at John F. Kennedy
International Airport.
     The motivation behind  this change stemmed from  the Noise Abatement Officer's con-
cern that one of the primary sources of public reaction to aircraft noise was the extended
periods of noise exposure experienced in the same communities, which resulted from the
use of the then existent preferential runway system. This operational usage, during previous
summers, resulted in as much as 24 to 32 hours of continuous overflight on  one community,
with resultant volatile public reaction.
     It was suggested  that  whenever practicable, and  when  runway usage  permitted, its
operational utilization not extend beyond an average eight-hour period. Subsequently, after
considerable study, this concept  was  found to  be acceptable  by the airlines, the Port Au-

                                         679

-------
thority and  the Federal Aviation Administration Air Traffic representative responsible for
the control of traffic utilizing these runways.

INITIAL RUNWAY USE MODIFICATION
     Based upon  these  determinations, a  four-month experimental program was imple-
mented on 1  June 1970 which considered the average weather conditions that prevailed for
the past eleven summers, square miles affected, numbers of people who resided under the
flight paths of the airlines, and scheduling demand correlated with hours of the day both for
arrivals and  departures.  From these data a new runway index in lieu of the preferential
system previously in  use was  introduced  by the Kennedy Control Tower, whiph index
outlined a pattern designed  to limit the use of any particular pair of runways to not more
than eight hours per day, when wind and weather conditions permitted.
     Upon completion of an evaluation, the data were analyzed and for this paper, specifi-
cally  related to the noise complaint  pattern. It was found that when runway usage was
confined to  an average eight-hour period, noise  complaints maintained a low level. Further,
the investigation revealed that when the chronic complaints were removed from numerical
tables, during those times when the runways  were used within the eight-hour factor, noise
complaints as received were  insignificant, averaging two to four a day.
     Of greater significance was the sudden  rise in noise  complaint reception by  official
channels when the runway  usage exceeded  eight hours. The proportionate rise was as much
as  100 percent on many occasions when the runway usage exceeded the eight-hour time
frame.
     The adjustment of the above described  state of relative compatibility between neigh-
boring communities and the airport through the action of changes to the runway system
usage indicated the practicability of the application of system solutions and the  further
requirement for additional  methodology to be explored, cost  effectiveness correlated, and
further adaptations of the principles involved to be considered for further sophistication of
the System.

COMPUTERIZATION OF NEW SYSTEM
     The next appropriate step was a Dynamic Preferential Runway System (DPRS) using a
small computer as a part of the System. This involved hardware and considerable prepara-
tion. The Federal  Aviation Administration,  the Port of New York  and New Jersey Author-
ity, and  the  Aviation Development Council jointly sponsored an effort  to have such a
system operating at J.F.K. in the summer of 1971. This System was developed and installed
on an experimental basis. Because of contractual delays, it did  not begin its operational test
ui.til 2 August 1971.  (Therefore only two months of experience during the high complaint
season were obtained.)
     The purpose  of the System was to determine an optimum  mode of airport operation at
any time from the standpoint of community noise exposure. By mode of operation is meant
the combination of runways used for arrival and departures. Runway use systems are basi-
cally  noise abatement priority  listings of available operating  modes and  are widely used
throughout the world. The use  of the word  "dynamic" in the DPRS reflects a distinction in
that there is  no fixed priority, but rather an order of preference which changes according to
past and probable future  community exposure conditions.
                                      680

-------
     Design Goals
     The following imperatives,  which form  an  intuitive basis for the DPRS, were con-
sidered design goals for the new system on the assumption that their achievement would
reduce community annoyance to aircraft operations.
     1. Avoid excessive dwell (periods of continued overflight).
     2. Avoid exposing the same community in the same time period (particularly evening
and night) on successive days.
     3. Recognize the need for efficient airport  operation to maintain  the air service re-
quirements of the community.
     4. Provide sufficient information to allow cogent selection of alternate runways.
     5. Incorporate reasonable time of day and day of week sensitivity corrections.
     There are two essential  steps in minimization of community noise disturbance by a
DPRS. The first step is to calculate the current disturbance of each community, based upon
a complete history of exposure in that  community. The second step in the process is to
calculate  the further effect of the future use of each runway combination in  terms of
community disturbance. Runway selections made on the basis of the latter results can then
control and minimize disturbance.
     The technical approach to calculation and utilization of community disturbance data is
described in the following sections.

     Community Disturbance Model
     The basic element of  the DPRS is a model  for the disturbance in each area of the
community, incorporating all of the factors known to be related to general disturbance.
These factors, which will be discussed individually, include for every flyover the time of day
and week, the noise levels produced and the number of persons exposed to those levels, and
the disturbance  caused by  previous  flyovers. If all these factors are  properly taken  into
account,  the best  choice of runways from the noise standpoint  can  be determined.  The
DPRS does this  and also simplifies the  selection process by taking into account probable
traffic loads. The DPRS computer thus serves as a specialized accounting and  decision-
making device, the purpose of which is to indicate the optimum choice of runways and at
the same time relieve the controller of unnecessary work.
     Figure  1 is a diagram  of the  community disturbance model. In the JFK DPRS this
model is applied to each of four principal  community zones lying under the flight paths of
the four major runways. The input to  the model is a FLYOVER EVENT affecting the
community of concern; this is specified as to time of occurrence and type, i.e., approach or
departure. If operations are frequent then the input rate is high. The TIME OF OCCUR-
RENCE FACTOR reflects the fact that people are  more sensitive to aircraft flyover noise at
certain times of the day or  week. The weightings used in the present DPRS are given in the
following table:
Hours
0700-1859
1900-2159
2200-0659
Weekends and
holidays
3
3
10
Other
days
1
3
10
                                       681

-------
                                 FLYOVER
                                   EVENT
                                 TIME  OF
                                OCCURRENCE
                                 FACTOR
                               POPULATION/
                                EXPOSURE  '
                                 FACTOR
                                 MEMORY I
                                 FACTOR
                                COMMUNITY
                              DISTURBANCE
                                 RATING
                  Figure 1: Elements of the community disturbance model.
    The POPULATION/EXPOSURE FACTOR is present because different operations af-
fect different numbers of people. The effect of a flyover on a given community is propor-
tional  to the disturbance  of the prototype individual times the community population.
Since not every person in the community is exposed to the same noise level, it is desirable to
use  a community weighting which reflects the composite disturbance of the community.
The established EPNL contours for typical aircraft provide a basis for evaluating this weight-
ing. Each person exposed  to an individual EPNL of 100 EPNdB perceives twice as much
                                    682

-------
noise as a person exposed to 90 EPNdB and four times as much as a person exposed to 80
EPNdB. From this property of EPNL we can define the community sensitivity weighting as

                             W = fn p(x, y) N(x, y) dxdy
where p(x, y) = population density at latitude x, longitude y
      N(x, y) = effective noy value of a flyover at (x, y)
           d = antilog [(EPNdB -40)/33.2]
           S = area covered by the community

     In practice this weighting is calculated in the following way:

         W = P90 * 1+P100  *  2 + P110 * 4 + P120 * 8 + P]30 *  16+  ...

where  PL = population within the (L) EPNdB contour but not within the (L + 10) EPNdB
contour. There  are separate weightings for arrivals and departures.
     The final operative element, the MEMORY FACTOR, is of particular importance. The
effect of a particular flyover is dependent upon preceding flyovers, i.e., upon past exposure.
Disturbance potential is  heightened after an uninterrupted series of flyovers but decreases
during a respite  period. Thus  the total effect of a given number of  flyover events is a
function of the  temporal  pattern  of exposure. The significance of this fact is that, by
optimizing this pattern, community disturbance may be decreased by decreasing the dura-
tion of the exposure.  In  order to do this, however, it is necessary to incorporate a kind of
memory into the system which simulates  the hypothesized human reactions. The DPRS
provides this in terms of the temporal function W = 2"(T/24).  Each flyover event is weighted
by  this function according to the time (T) in  hours that has elapsed since it occurred.
Remote events  carry less weight as they are "forgotten". On the other hand, a succession of
recent events tends to maximize the weighted sum of flyover events. The time period over
which  such a continuing succession occurs  is  called the "dwell".  It has  been  observed that
long dwells are  strongly associated with community disturbance.
     The community disturbance model thus provides a means of continuously computing
the disturbance in each community around the airport. It also is the basis for assessing the
effect  of continuing operations depending upon which runways are used. In the DPRS an
overall rating of disturbance for all four communities is computed using the  criterion that
the disturbance in any one community should  not  greatly  exceed that in another. This
rating is evaluated by the DPRS for each possible airport operating mode  for the present and
for probable future conditions. The  latter are based upon wind predictions for the next 3, 6,
9, and  12 hours, each successive set of predictions being discounted by half in the overall
disturbance  rating. This rating, proportional to the variance of the separate community
ratings  summed  over the  prediction  period, is  the  basis for rank ordering the possible
operating modes  for present use. An additional  function in the DPRS monitors traffic load
versus  time of day and causes operating modes with insufficient capacity to be denoted by
the symbol I rather than X in the printout for controller use.

                                        683

-------
     Fundamentals of Operation
     The DPRS is physically embodied in a small computer with Teletype located in the
 control tower. Data on aircraft operations, wind conditions, and wind predictions are read
 into the  system periodically. At regular intervals, or upon interrogation, the system prints
 out a currently optimum listing of operating modes for various wind conditions. This listing
 is delivered to the controller, who can then readily  determine the best choice of runways.
 Operations data from the existing CATER system (used for tower record keeping) include
 (for the purposes of the DPRS) time of day, runway  used, and type of operation, i.e., arrival
 or departure, for each aircraft operation  at JFK. These are transferred by punched paper
 tape. Weather data,  which are transcribed remotely to the tower location, consist of the
 predicted wind direction and speed for 3, 6, 9, and 12 hours in the future. The DPRS
 interprets these data in  terms of community noise disturbance,  according  to the model
 discussed earlier, and ranks the existing possible choices of airport operating mode in order
 of increasing probable disturbance.
     Table 1  is a typewritten condensed  version of a DPRS printout. Listed under RUN-
 WAY are the eight feasible modes of operation. No distinction is  made between left and

                                     Table 1

                             SAMPLE DPRS PRINTOUT

                COMPUTERIZED  NOISE  REDUCTION SYSTEM

   DATE    08/02/71        TIME    2004        LAST OPERATION   1902
2211
 311
  1

  X
           WIND FACTORS
           5-15 KNOTS
                                   0-4
                                   KTS
311/
 041
041 /
 131
131 /
 221

  X

  I
  X


  X

  X
                                    X

                                     684
                        X
                        i
                        i
                        x
                        i
                        X
                        X
                        X
RUNWAY
 AR DP

 13 13

 13 22

 31 22

 31 31

   4 13

   4 31

   4   4

 22 22
VALUE

 0336

 0355

 0355

 0365

 0396

 0434

 1165

 1796

-------
right runways, although all runways exist as parallels. Constraints on choice of runways
include both traffic load and wind. Those modes which are not capable of handling the
normal traffic for the particular time of day of the printout are shown by Fs. Usable modes
are represented by X's in the WIND FACTORS columns for which they are appropriate. The
modes are listed  in order of preference and the figures of merit on which the order is based
are given in the VALUE column. To determine the current best choice, one simply finds the
highest X under the prevailing  wind condition. If for some  extraneous reason  the corre-
sponding runways are not available, then one would go to the next highest X.

RESULTS AND PERFORMANCE
    The operation of the DPRS under "real input" conditions is extremely complex. The
community disturbance model is intuitively reasonable, but it is perhaps difficult to see how
the model achieves the stated design goals. It was discovered early in the trial program that
the system users could not always predict the DPRS recommended runway  usage. Close
examination, however, revealed  that in every case the DPRS was simply able to consider far
more information in its decisions than any air traffic controller could reasonably hope to be
able to consider.
    The effect of the system, however, can be illustrated by examination of records for a
few days of use.  It is easiest to see the improvement offered by the DPRS by comparing
what actually occurred operationally  during those days to what would have occurred had
the previous preferential runway system (described in JFK Tower Bulletin 69-1) been used.
Figure 2 shows the overflight periods in  each of four geographic sectors surrounding the
airport. The upper set of lines illustrates what would have occurred had the previous method
of operation  been used during  four particular days, while the  lower set illustrates actual
operation under  the DPRS. A solid line represents a period of continuous arrivals or depart-
ures, or  both. The dotted lines indicate periods of extremely light traffic peculiar to Sector
A.
    Weather conditions during  this period would have caused excessive dwell for communi-
ties in Sectors B and D, had the old system been in use. In contrast, the DPRS used the
available flexibility  of the airport  to more equitably distribute the operations over all
communities, while avoiding excessive dwell in any sector.
    As  described, the DPRS does not expressly consider dwell. Rather the system records
each individual overflight, giving weight to increased sensitivity to evening and night opera-
tions, and to overflight  of densely  populated  areas.  Thus Figure 2 represents only the
resultant patterns from an intricate series of mathematical processes.
    Table 2 is a compilation of some of the aspects of Figure 2. Comparison of actual
operation under the DPRS with operation  under the old system shows the following:
     1)  The actual dwell periods  were more evenly distributed among the sectors, and total
dwell was better  distributed.
     2)  There were more dwell periods, indicating a "breaking up" of the exposure.
     3)  The  mean dwell period was considerably shortened for Sectors B and D, without
increasing dwell in Sectors A and C beyond reasonable limits.
    4)  The respite periods are also better distributed, more frequent, and more uniform.
     5)  The night exposure (10:00 p.m. to 7:00 a.m.) is more uniformly distributed.

                                       685

-------
                            Exposure Periods under Previous System (Estimated)
Sector
A
B
C
D
mum mum mini iiua

u


> ii i i i i i
                 0700         2200      0700         2200     0700         2200      0700          2200
                       1  Sept                2 Sept                 3 Sepc                 4 Sept
00
                                  Exposure Periods under DPRS (Actual)
Sector
A
B

B k-— 3 1 	 III F


	 , 	 i 	 , 	 _j 	 ' 	 1 	 1 	 	 1 	
                 0700         2200
                     1 Sept, 1971

            t -^  Normal Traffic
            lilllll  Light Traffic
0700         2200
     2 Sept. 1971
0700         2200

     3 Sept. 1971
0700        2200

     4 Sept. 1971
                                    Figure 2:  DWELL patterns for previous systems and OPRS.

-------
                                                  Table 2
OS
00
-J
                            COMPARISON BY GEOGRAPHIC SECTORS OF PREVIOUS SYSTEM
                            OPERATIONS (ESTIMATED) WITH DPRS OPERATIONS (ACTUAL)
                                FOR 1 SEPTEMBER 1971 THROUGH 4 SEPTEMBER 1971
PERFORMANCE
MEASURES
Total hours dwell
Percentage of total
dwell
Dwell periods
Mean dwell
Total hours respite
Respite periods
Mean respite
Total hours of-
night operations
Percentage of dwell
occurring at night
Previous System
(Estimated).
A B .
16,0
7.9
4
4.0
80.0
4
20.0
0.
0.
68.2
33,5.
4
17.1
27.8
4
7.0
20.2
29.6

C
35.4
17.4
t
6
5.9
60.6
5
12.1
27.8
78.5

D
83.8
41.2
4
21.0
12.2
3
4.1
23.4
27.9
Trial DPRS
(Actual)
A
35.4
18.6
7
5.1
60.6
7
8.7
15.0
42.4
B
37.9
19.9
6
6.3
48.1
6
9.7
10.2
26.9
C
65.3
34.2
6
10.9
30.7
5
6.1
29.5
45.2
D
52.2
27.4
7
7.4
43.8
7
6.3
16.8
32.2
                *Xhese data are approximate.
                       Area A - Communities NE of airport  boundary  )
                       Area B -      "      SB  "    "       "      ).
                       Area C -      •«      SW  «    "       "      )
                       Area D -      '•      NW  "    "       "      )
See Chart 1

-------
     It is very important to note that the PRS does not "discriminate" against Sectors B and
D, as might appear from Figure 2. Previous analyses demonstrate that the former system was
a good preferential runway system when viewed on a long-term basis but the application of
the DPRS equalizes the exposure on a short-term basis. The addition of a computer pro-
vided the ability to anticipate and to compensate for the short-term effects of weather and
other uncontrolled factors.
     It can be seen that the DPRS not only breaks up exposure, resulting in shorter dwell
periods, but that the system distributes exposure among the various sectors in a way which
is more equitable. In these respects, the  DPRS may be expected to out-perform a static
preferential runway system.

LIMITATIONS OF PRESENT SYSTEM
     There are three types of constraint  involved in the  present operation of the DPRS:
external limitations, functional limitations, and lack of conceptual validation.

     External Limitations
     External  limitations are temporary conditions which are  outside the control of the
DPRS. The following constraints are all within this category:
     1) Adverse weather conditions.
     2) Nonavailability of runways resulting from construction, maintenance, or other con-
ditions.
     3) Delays in input of the CATER punched paper tapes.
     Since the DPRS can better operate with maximum runway options and current infor-
mation, these  constraints should be minimized whenever possible (particularly during the
complaint season).

     Functional Limitations
     Functional limitations are caused  by lack  of equipment or by airport geometry. Such
constraints are:
     1) Elimination  of some useful runway  combinations  because of operational safety
requirements or taxiway congestion.
     2) Inability to account for variance in the noise levels of individual operations.
     3) Inability to determine precise departure or arrival paths, and exact community
areas affected.
     4) Inability to effect more frequent runway changes when  needed.
     These constraints are more  severe than those in the preceding category, but are  well
within the present conceptual framework of the DPRS. Removal of any of these restrictions
could improve the performance of the system.
     Conceptual Validation
     The experience thus far with the DPRS has demonstrated the feasibility and practi-
cality of the system. While the results of implementation of the trial system were favorable,
the concepts and weightings involved in formulation of the community disturbance model

                                       688

-------
had little or no experimental confirmation. This model is the heart of the DPRS, so that
concern over validation is appropriate. It is therefore hoped to conduct such studies in the
future.
                                  REFERENCES

1.  Gach, M., "Appraisal  of Community Response to Aircraft Noise", Society of Auto-
    motive Engineers, Doc. No. P-37, Paper No. 710317.
2.  Edmiston, R., "A Comparison of the Experimental DPRS with the Manual Preferential
    System at JFK International Airport", Tracer Inc. Federal Aviation Administration,
    Jamaica, New York, 31 July 1972.
3.  Edmiston, R., and Connor, W., "Installation & Operation of a Dynamic Preferential
    Runway System  at JFK International Airport, Jamaica, N.Y." Tracer Inc. Federal
    Aviation Administration, Jamaica, New York, 12 March 1972.
                                       689

-------
          A CAUSAL MODEL FOR RELATING NOISE EXPOSURE, PSYCHO-
             SOCIAL VARIABLES AND AIRCRAFT NOISE ANNOYANCE

                                  Skipton Leonard
                                   Paul N. Borsky

                                 Columbia University
                               School of Public Health
                                New York, New York


Supported by
     DoT Research Grant   OS-20190
     NASA Research Grant  NGL 33-008-118
                                    ABSTRACT

     In an extension of research methodology and strategies begun by earlier researchers,
data from  a community noise survey were analyzed in an attempt to causally relate noise
exposure, psycho-social variables and aircraft noise annoyance.  Results from these analyses
indicated that noise exposure has a relatively small direct impact upon aircraft noise annoy-
ance. Intervening variables such as fear of aircraft operation and concern for harmful effects
upon health are  required in the present causal schema relating exposure and annoyance. In
other words, annoyance significantly increases or decreases with noise exposure only to the
extent that fear and health concern also increase or decrease.
     The results indicated  that health concern also served as a mediating variable in explain-
ing the relationship between beliefs of negligence (misfeasance) on the part  of the aircraft
industry and aircraft noise  annoyance.
     Implications for the mediating nature of aircraft fear and health concern are discussed.
It is reasonable to expect that the quieter, though larger, aircraft currently being introduced
to international  aircraft fleets  will be effective in  reducing  aircraft annoyance in most
situations.
     There can be little question that persons exposed to aircraft noise experience annoy-
ance, irritation, and at times fear. Many must wonder whether intensive exposure to aircraft
operations  might also be  physically  and  psychologically  damaging. While we are sure  of
the experiential quality of life in high aircraft noise areas, a phenomenological methodology
does not supply the kinds  of  evidence  necessary for an adequate  understanding of the
dynamics of aircraft noise  annoyance. One person cannot speak for a population, and even
if he could, his  mind would become hopelessly overloaded with confusing and often con-
flicting information.
     By judicious use of intuition and sophisticated data analysis techniques, however, one
can construct  logical and  consistent causal networks of variables which  will increase our
understanding  of aircraft annoyance phenomena.

                                        691

-------
     This  paper will be  focused  upon  an understanding of the aircraft noise-annoyance
relationships rather than simple prediction. It is unfortunate that analyses are often stopped
when adequate  prediction has been attained of a phenomenon  whether or not  these
empirically based "equations" make any conceptual sense. Without giving our punchline
prematurely, we believe that one will find surprising and  interesting differences between
simple prediction approaches to the data in this paper and a more extensive causal analysis
of the aircraft annoyance problem.

Previous Research:
     Early social survey  studies in  the  U.S.A. and  Britain (Borsky, 1952; Borsky,  1961;
McKennell, 1963) reported relationships between measures of aircraft exposure and annoy-
ance. The annoyance variables used in  these studies were based upon reports of activities
disturbed  by aircraft noise and the degree to which this disturbance caused annoyance.
     In  addition, it was reported that the respondents' reported fear of aircraft operations
was also related to annoyance. Also associated with annoyance was the respondents' belief
that aircraft noise was "preventable" (McKennell, 1963) and the respondents' perception of
"considerateness" of aircraft officials, pilots,  etc. in keeping aircraft noise at a minimum
(Borsky, 1961).
     In  these  studies there was little evidence to suggest anything but a simple positive
relationship between  noise  exposure and annoyance since this pattern was consistently
obtained.  Indeed, Van Os (1967) in a Dutch  study reported a correlation of .95 between
exposure  and annoyance. This correlation, however, is the relationship between  mean
annoyance ratings and exposure for each survey location and would be considerably reduced
if individual annoyance ratings were  used in computing the correlation.
     TRACOR (1970), on the basis of survey data obtained in seven big city airports in the
U.S.A.,  reported the first test of the simple stimulus-response or exposure-annoyance model
implicitly assumed in  the earlier studies. The relationship  between noise exposure and
annoyance, it  was noted, dropped markedly when  the effects  of other variables, such as
respondents' reported fear of aircraft operations and feelings that aircraft noise could be
reduced and for some insufficient reason was not (feelings of misfeasance), were controlled
or held constant. In fact, some of these variables (fear, noise susceptibility, adaptability) had
a greater independent effect upon annoyance than did exposure.
     As our sophistication in the analysis and  measurement of the relevant variables grows,
so will our understanding of the dynamics of aircraft noise annoyance. This growth is seen
as a natural extension of the research methods and strategies begun by Borsky, McKennell
and TRACOR.

                                   THE SURVEY

     The data reported here were obtained from a community noise survey conducted  by
Columbia  University Noise Research Unit in February-March, and August-October  1972.
The  survey was conducted  in order to provide data for an evaluation  of the Dynamic
Preferential Runway System (a computerized method for assigning runway use) at John F.
Kennedy International Airport (JFK), New York City, U.S.A.

                                        692

-------
Sampling Design
     The sampling procedure was designed so as to maximize the homogeneity of noise
exposure within each surveyed area. Since noise levels from  aircraft drop rapidly  as one
moves laterally away from landing and take-off flight paths, and as one moves farther from
the end of a  runway, it was necessary to intensively sample areas only a few blocks in
diameter. Survey tracts were located 1.1, 2.5 and 5.2 miles j from the end of the various
runways at JFK. These sampling sites are presented in Figure 1.
     All interviewers were given predesignated addresses in the sample areas, each consisting
of small clusters of adjacent blocks. In some assignment locations  where the number of
dwellings was limited, every household was contacted. In other areas, every nth dwelling was
selected. Respondents were required to be over 18 years old,  a permanent resident  of that
dwelling and not in employment at that residence.
     Aircraft noise  annoyance  data for the  months  of June and  July  exclusively were
directly obtained from those respondents (795) interviewed in August.  Respondents inter-
viewed in February and March 1972, (670) however, were contacted by telephone at the
start of August in order to obtain annoyance data for June and July.
     All respondents (those interviewed  in February, March  and August 1972) were con-
tacted  by telephone at the  start  of October to obtain annoyance data for the months of
August and September.
     Of 1740  assignments for the three distance areas, 1465 face-to-face interviews were
completed (85%). Interviewers were able to complete  1103 telephone follow-up interviews
in order to obtain June-July, August-September annoyance ratings.

Community Questionnaire
     The questionnaires used for the face-to-face interviews and for the telephone interviews
are similar in many  ways to instruments used in previous noise studies. Many items related
to aircraft  noise annoyance, fear of aircraft operations, beliefs in the negligence (mis-
feasance) of those  connected with  aircraft operations, are very similar  to items used  in
earlier  questionnaires (Borsky, 1961; McKennell,  1963; TRACOR, 1970). Items related  to
aircraft noise were interspersed among items asking about other kinds of disturbance in the
community.
         PSYCHO-SOCIAL AND AIRCRAFT NOISE EXPOSURE VARIABLES

Aircraft Noise Annoyance
     Previous  researchers have used  the aircraft noise disturbance model as a method for
measuring an individual's positive or negative feelings towards aircraft noise. The rationale
has been to measure the number of disturbances and the degree of annoyance caused by the
disturbance. This is the basic model that has been employed in the current study.

     3Some  locations  1.1,  2.5, 5.2 miles from a runway  were not sampled because of 1) practical
considerations (too few residents, local opposition) or, 2) the tract was unsuitable for purposes of the study
from which this data was obtained.

                                        693

-------
                                                                 :
         LONG  ISLAND,  NEW YORK
                        ATLANTIC     OCEAN
A
n
O
1
2
5
,
5
.2
M
M
'"
ILE
IL
IL
E
;
SAMPLE
SAMPLE
SAMPLE
AREAS
AREAS
AREAS
                  Figure 1:  Sample areas for 1972 community noise survey.
    TRAC'OR I 1(>70) has differentiated between  the event of activity disturbance and the
onset  of  negative  attitude. TRACOR has presented evidence that  suuuests that the two
concepts ma\  not  be synonymous ami that the causal Relationship between disturbance ami
.mno\ ance ma> v.iry with  the noise stimulus. For instance, where sonic boom is the stimu-
lus, evidence was  presented that a negative affect state preceded activity disturbance. KM
normal aircraft operations, however, the sequence of effects  was  nist the reverse. 'Flic
TRACOR distinction senes as a reminder of possible differences in interpretation for the
present aircraft annoyance  scale.
                                      . ...:

-------
     Table  1  presents the aircraft annoyance items considered for inclusion in the present
annoyance scale.2
     Based  upon a factor analysis (Principal Components, varimax rotation) it was determ-
ined that all items except sleeping pill use formed a general factor. The annoyance scale was,
therefore, constructed  by summating annoyance ratings for the eleven remaining activity
disturbance items. TRACOR (1970) had previously demonstrated that an unequal weighting
system based upon factor loadings contributed little to improvement in the prediction of
annoyance  by predictor variables similar to those used  in the present study.  A measure of
internal  consistency or reliability (coefficient alpha, cf. Nunnally (1967) p. 196) yielded
values of .91  and  .93  for the aircraft noise annoyance scales for June-July and August-
September.

                                         Table 1

                       AIRCRAFT NOISE ANNOYANCE SCALE ITEMS

           1 .  Interferes with listening to radio or TV.
           2.  Makes the TV picture flicker.
           3.  Startles or  frightens anyone in family.
           4.  Disturbs family's sleep,
           5.  Makes house rattle or shake.
           6.  Interferes with family's rest and relaxation.
           7.  Interferes with conversation.
           8.  Makes you keep your windows shut during the day.
           9.  Makes you keep your windows shut during the night.
          10.  Makes you feel tense and edgy.
          1 1 .  Gives you a headache.

Noise Exposure
     CNR (Composite Noise Rating, (Galloway and Bishop, 1970) ) was used as the primary
measure of community aircraft noise exposure. CNR  was  calculated from  known  EPNL
values for existing aircraft and operations data at JFK for the periods June-July and August-
September 1972. The following equations were used in the computation;
                   5   = EPNLj + 1 0 1 ogi o (NDj + NNj ) - 1 2

               CNRj   = lOlogio ? antilog (CNRj / 10 ) ,
where j refers  to  a  particular class of aircraft operation  and NDJ and N^j are the mean
number of occurrences during day and night respectively.
     Although  there are a number of objections to the use of this scale, it seems to be
related to aircraft noise annoyance (as measured by the activities disturbance model) as well
as any of the other conventional measures of exposure (TRACOR, 1970).
     2Figure IA and 2 A in the appendix presents the distribution of the two annoyance scale scores. Each item was
scored 0-4 with 4 representing highest annoyance. This produced a possible range of annoyance scores of 0-44.

                                         695

-------
 Fear
      The fear scale used in the present study consisted of a summation of four items from
 the community questionnaire. Respondents were asked to  rate 1) their dislike of unsafe
 low-flying airplanes, 2) how much  the noise from airplanes startles or frightens them, 3)
 how often they felt airplanes were flying  too low for the safety of the residents, 4) how
 often they felt there was some  danger that they might crash nearby.
      These items have strong face validity as well as high item intercorrelation. In addition,
 a number of the items have been shown to be related to annoyance  in previous research
 (Borsky, 1961; McKennell, 1963; TRACOR, 1970). The coefficient of reliability (alpha) for
 the fear scale is .83.3

 Misfeasance
      The  concept  of misfeasance (TRACOR, 1970) is an outgrowth  of Borsky's  (1961)
 concept of "considerateness" and  McKennelPs (1963) concept of  "preventability". This
 scale  was  intended to  measure the  respondents' belief that various agents connected with
 aircraft noise production are capable of reducing the noise but for some insufficient reason
 are not. The agents  in the present  scale  include  "the  people who run the  airlines", "the
 airport officials",  "the other governmental officials", "the  pilots",  "the  designers and
 makers of airplanes", and  "the community leaders". Each item was scored 0-4 with 4 being
 highest misfeasance. This produced a possible range of misfeasance scores of 0-24. The
 coefficient of reliability (alpha) for the misfeasance scale is .76.

 Health Attitudes
      McKennell (1963) reported a  strong relationship  between the belief that aircraft ex-
 posure effected the respondent's health  and annoyance. In the present questionnaire, re-
 spondents were asked "how harmful do you feel the airplane noise is to your health?" This
 item was scored 0-4 with 4 being very much.

 Importance of Aircraft
      A small relationship (r =  .12) was reported by McKennell (1963)  between an aircraft
 importance  scale  and annoyance. In  the present study respondents were asked  how  im-
 portant they felt commercial airplanes were to national welfare, the community and their
 own family. Each  item was scored  0-4 with 4 meaning very important. The sum of these
 three items was termed respondents'  feelings of aircraft importance.
     The scales and items described thus far were included in the present study because they
 had demonstrated theoretical promise in previous research. The relationship  between many
 other items in the  present  questionnaire and annoyance were computed. The increase in the
 understanding of annoyance, however, was  minimal and will be discussed only briefly in a
 later section.
     Figure 3A in the appendix presents the distribution of fear scores in the present study. Each item was scored 0-4
with highest fear being 4. This produced a possible range of fear scores of 0-16.

                                         696

-------
                    RESULTS AND CAUSAL INTERPRETATIONS

Zero-order (Pearson) Correlations
     Table 2 presents the zero-order or simple correlation coefficients between June-July
aircraft  noise annoyance (June annoyance), August-September aircraft annoyance (August
annoyance), CNR computed  for June-July  (June CNR), CNR computed for August-
September (August  CNR), fear,  misfeasance, health  attitudes, aircraft  importance (A/C
importance), and sex.
     Of immediate interest is the fact that the  correlations between aircraft noise annoyance
and the psycho-social variables decrease uniformly from June-July to August-September. An
explanation for this phenomenon is not immediately apparent  and this writer will not
speculate as to  possible  reasons for this drop. More detailed analyses are underway and it is
hoped that the  reasons will be determined. The directions of correlation in all cases remain
the same with most variables (except sex) remaining significantly related to annoyance for
the months  of August-September. Since the patterns of relationship have not been changed
but only weakened in the August data, it is reasonable to assume that the same processes in
relation to annoyance are involved.
     The  fear and health-attitudes variables  are highly  correlated with  June  and August
annoyance,  indicating once again  their importance in relation to annoyance. The greater a
respondent's expressed  fear of aircraft operations and his belief in  the harmful effects of
aircraft noise, the greater is his annoyance with aircraft noise.
     It  would  seem from these correlations  that  the misfeasance and CNR variables are
moderately  related  to both June and August annoyance. Small correlations exist between
annoyance and the A/C importance and sex  variables. In the case of A/C importance, the
more important a respondent believed aircraft to be, the less annoying  he rated aircraft
noise. Males reported less annoyance with aircraft noise than women.
     Variables which were minimally related  to aircraft noise annoyance were respondent's
age, education,  ownership of dwelling, income, length of residence in area and installation of
air-conditioning in dwelling. In addition, noise and general sensitivity scales were not related
to annoyance.

Multiple Regression Prediction Model
     Multiple regression analyses were  computed with June annoyance and August annoy-
ance as dependent variables and June CNR, August  CNR, fear, health attitudes, misfeasance,
A/C importance, and sex as predictor variables.
     The  analysis for the  June-July  period  suggested that the fear and health  attitudes
variables explained significant  amounts of annoyance variance independent of the effects of
other predictor variables. Other variables contributed little to annoyance prediction when
the other predictors were held constant.
     While other variables had little value as  independent predictors of annoyance for the
June data, it seemed quite possible that they  might be related to other predictors and thus
be related indirectly to annoyance. Subsequent multiple regression analyses with fear, health
attitudes and misfeasance as dependent variables indicated that CNR (June) and misfeasance
were independently related to fear and health attitudes respectively, while A/C importance

                                        697

-------
00
June Annoyance

August Annoyance

CNR June

CNR August

Pear

Misfeasance

Health Attitudes

A/C Importance

Sex
                                          Table 2

                       CORRELATION COEFFICIENTS FOR 1972 SURVEY DATA
                    June     August    CNR     CNR
                   .Annoy.   Annoy.    June    August
                                                            Fear
Misf.
Health   A/C
At ti t.   Import.   Sex
1100
.74
.32
.33
.72
.32
.63
•r?2
-.09
1.00
.38
.38
.62
.27
,55
-.18
-.08

1.00
.98
.41
.10
.24
-.14
-.02


7..00
.40 1.00
.09 *.30 1.00
,24 .64 .31 1.00
-.13 -.19 -.14 -.18 1.00
-.07 -.11 .00 -.07 -.04 1,00

-------
and  sex seemed to have little independent relationship to any of the other predictor vari-
ables.
     The  August regression analysis substantiated these findings with the exception that a
small to moderate degree of independent relationship was indicated between August CNR
and  August annoyance. The final set  of predictor variables included CNR, fear, health
attitudes and misfeasance. The A/C importance and sex variables added little to an under-
standing of the dynamics  of aircraft noise annoyance and were therefore eliminated from
further analyses and from subsequent theoretical models. Table  3 presents the results of
multiple regression analyses with a reduced set of predictor variables. These analyses again
indicated that the fear and health attitudes have far greater standardized prediction weights
than the CNR or misfeasance variables.
     Fear, health attitudes, misfeasance  and CNR as a predictor set explained 58 per cent of
the annoyance variance for the months of June and July. For the months of August  and
September,  45 per  cent of the variance was accounted for by these four variables. These
percentages represent a considerable increase  in  the amounts of annoyance explained by
CNR alone (10 per cent in June, 14 per cent in August).
     The  results presented here  compare favorably  with  data presented by TRACOR
(1970). A multiple regression analysis conducted by TRACOR with annoyance as depend-
ent variable yielded the  following standardized regression weights: fear, .36; CNR,  .16;
misfeasance, .06; importance of Airport, .05.
     On the basis of the similarity of  these  sets of data, it  would seem that misfeasance
attitudes are not related directly to aircraft noise annoyance. The observed similarity of the
August data and the TRACOR data suggest that a small to moderate independent relation-
ship exists between noise exposure and aircraft  noise annoyance although the June data
would bring into question  the stability  of this relationship. It  would be a mistake, however,
to assume that exposure and misfeasance are unimportant in understanding the dynamics of
aircraft noise annoyance.

Causal Model
     June annoyance and June CNR data were used in constructing a  causal model between
CNR,  fear,  misfeasance  and health attitudes since the  relationships for this period are
stronger than  for the August period. Patterns for the August data will be compared later
with the June model.
     The first step in  developing a causal model is to identify those relationships in which
the causal direction can clearly be inferred on  an a priori basis. Our model assumes that
aircraft noise annoyance is caused by some combination of antecedent variables such as
CNR, fear, health attitudes and  misfeasance,  and not  vice versa.  It is also clear that CNR
cannot be caused by  any combination  of the other variables.  It is  not so clear as to the
"natural" causal directions between the other variables.
     Figure  2 presents the possible causal relationships based upon the above assumptions
and the simple or /ero-order correlations. One-way arrows represent possible causation from
the variable at the tail of the arrow to the variable at the point of the arrow. Double-ended
arrows indicate indeterminate causal relationships at this stage of the analysis.
     The goal of further analyses is to validate  each of these possible causal relations, i.e., to
identify each as real or spurious. This validation can be accomplished by the use of partial
                                        699

-------
                                                 Table 3

                      BETA WEIGHTS (STANDARDIZED REGRESSION COEFFICIENTS) AND B WEIGHTS
                      (UNSTANDARDIZED REGRESSION COEFFICIENTS) FOR REGRESSION ANALYSES
                   WITH JUNE AND AUGUST AIRCRAFT NOISE ANNOYANCE AS DEPENDENT VARIABLES
                                            TIME PERIOD
o
o
            Dependent
            Variable
Annoyance
June-July
   Annoyance
August-September

Fear
Health
Attitudes
Misfeasance
CNR
Regression
Constant
Beta Weights
(Standardized)
.50
.28
.08
.04

B Weights
(Unstandardized)
1.22
2.23
.16
.06
-4.74
Beta Weights
.38
.25
.06
.17

B Weights
.96
2.08
.14
.25
-26.94

-------
   Fear
                                 Health Attitudes
Misfeasance
                             Annoyance  (June)
Figure 2: Possible causal relationships based upon a priori assumptions and zero-order (Pearson) correlation
        coefficients.
   Fear
                             CNR  (June)
Misfeasance
                                 Health Attitudes
                             Annoyance  (June)
                     Figure 3: Empirically established causal relationships.
                                         701

-------
correlations. A partial correlation is the relationship between two variables that exists when
the effects of one or more other variables have been held constant or "partialled" out.
     For instance, if a direct causal link exists between CNR  and annoyance, the partial
correlation between  CNR and annoyance should not be  near zero when the effects of fear,
health attitudes and  misfeasance are partialled out.  In fact, the correlation between CNR
and annoyance shrinks from  .32 to .06. We  conclude, therefore, that  there is little direct
causal effect of CNR upon annoyance and remove that arrow from the model.
      Similar procedures will  produce a simplified model as presented in Figure 3. Three
causal links have been substantiated; CNR -»• fear, fear -> annoyance, health attitudes -*
annoyance.  The  two  remaining double arrow relationships, fear    health attitudes and
health attitudes ** misfeasance,  require  further clarification in order to establish causal
direction.
     Double-ended arrows in a causal model imply reciprocal causation, i.e. that the pair of
variables are causally dependent upon each other. This situation does not lend itself easily to
a simple "path analysis" of causation (Simon, 1960).  If possible, one would like to establish
whether these relationships are really bi-directional since the  implications  for  reciprocal
causation situations are quite different than for the unidirectional case.
     In the absence  of a priori knowledge as  to causal direction between two variables, one
may analyze the data to see if both points of the arrow are required to explain the data. If
not,  that point not required  will  be dropped from the model.  For instance, since  the
zero-order correlation of .24 between CNR and health attitudes is reduced to .01 when the
effects of fear are partialled out, we infer that a fear -»•  health attitudes link exists. In  the
absence of other variables not measured, there is no other way to explain the zero-order
correlation between  CNR and  health  attitudes without  including fear as an intervening
variable.
     In a  similar fashion it is observed that the zero-order correlation between misfeasance
and annoyance of .32 is reduced to .10 when the  effects of health attitudes are partialled
out. This  indicates that health attitudes are an intervening variable between misfeasance and
annoyance and that a misfeasance -*• health attitudes link  exists.
     The reduction of the zero-order correlation between fear and misfeasance of .30 to . 14
when the  effects of health attitudes are partialled out necessitates  the existence  of at least
one more  arrow of causation. If health attitudes are an intervening variable between fear and
misfeasance  either a  health attitudes -*• fear link or a health attitudes ->• misfeasance link, or
both, must exist.
     It would seem quite likely that concern  with  the health of one's family could increase
one's fear concerning aircraft operations.  On  a common  sense basis it also would be logical
to expect a reciprocal relationship to exist between  health attitudes and misfeasance. The
sequence might be as follows: "I'm concerned about aircraft noise and my family's health; I
wish something would be done about the noise; If they can put a man on the moon they  can
reduce the noise; therefore, the aircraft people aren't doing enough to reduce the noise (are
being misfeasant)." These proposed links are pure conjecture, however, and the most one
can say is that at least one of these links must exist in order to explain the simple relation-
ships among this set of variables.
                                         702

-------
      Figure 4 presents  a  schema based upon  the causal  inferences made so far in the
  discussion. The dotted arrows represent reciprocal relationships which this investigator be-
  lieves exist but which must remain tentative at this time.
                                                   Tentative Relationship — — - —
                        CNR (June)
Fear
Health Attitudes
*             ^Misfeasance
                           Annoyance (June)
                Figure 4: Final causal model for the June CNR and annoyance data.
                                                    Weakened Relationship*
                                                    Tentative Relationship
                          CNR  (August)
 Fear
            ^ Health Attitudes^
        -•^     .               *•*.-
                                                                        isfeasance
                            Annoyance (August)
                 Figure 5: Final causal model for the August CNR and annoyance data.
                                         703

-------
     A causal model for the August data includes a causal arrow from CNR to annoyance.
Although the simple correlation of .38 between CNR and  annoyance is considerably re-
duced by partialling out the effects of fear, the relationship remains great enough (.21) to
warrant inclusion in the model. Figure 5 presents the causal model for the August data.
     We are unsure  of the explanations for this discrepancy between the  causal models for
June and August. We are currently investigating sources of variation between the two time
periods which may have contributed to these differences in pattern.
                    IMPLICATIONS FROM THE CAUSAL MODEL

     Fear of aircraft operations and concern with the harmful effects of aircraft noise are
the major intervening variables in the proposed causal model. Increases in noise exposure or
misfeasance beliefs  can be expected to result in significant increases in aircraft annoyance
only to the extent that fear and health concerns are also increased. The present data and the
TRACOR data suggest that exposure has a small independent effect upon annoyance. The
reliability of this relationship, however, is suspect.
     What aspects of aircraft noise are related to fear of aircraft operations? Noise level is
one cue to the altitude of an aircraft. To a resident  inside his house, a louder aircraft may
indicate that there is  a greater danger of that craft crashing near his home. If this is so, the
new generation of aircraft with quieter engines should yield important pay-offs in terms of
annoyance since residents would believe the aircraft to be flying higher than they really are.
Interestingly, one might expect the effects of the quieter engines to be neutralized outdoors
since visual cues would indicate the giant wide-body aircraft to be flying lower than they
actually are. A test  of the indoor-outdoor differences in annoyance would be an interesting
investigation of the noise-as-danger cue thesis.
      It has been proposed that reciprocal relationships exist between fear *- health attitudes
and health attitudes •«* misfeasance beliefs. If so, these three variables comprise a feedback
system in which an increase in the first variable causes an increase in a second variable which
"feeds back" or causes additional increase in  the first variable, etc. Systems of this kind are
unstable (Turner and  Stevens, 1959). The reverberation of causation ceases only because of
the internal frictions of personality dynamics.
     The three psycho-social variables in the proposed causal system, therefore, should not
be considered  separately but as a general or central factor in the  causal model. One would
expect that a decrease in any of the three variables would result in diminution of the effects
of the other psycho-social  variables as well.
     While this prediction requires further  validation,  it should offer encouragement to
those investigators and administrators who are developing techniques for attenuating atti-
tudes and feelings of fear, health hazards and misfeasance. The effects of intervention and
change in any of these variables  would be amplified  by  the proposed system of reciprocal
relations.
     In conclusion,  we emphasize the tentative nature of the causal model presented here.
An attempt has been made to make causal inferences from correlational data. The model,
however,  has not been subjected to the rigors of experimentation on  a more controlled

                                        704

-------
basis.  This model should be considered an intermediate step between descriptive, correla-
tional analysis and hard-nosed experimental validation. In addition, one should keep in mind
that these data were obtained from respondents intensively exposed to aircraft noise. Causal
models  based  upon  this population may not be entirely generalizable to less-exposed
samples.

                                  REFERENCES

Borsky, P.  N.  Community  Aspects  of Aircraft Noise. National Advisory Committee for
    Aeronautics, 1952.
Borsky, P. N. Community Reactions to Air Force Noise, Part 1. WADD Technical Report
    60-689(1), 1961.
McKennell, A. C. Aircraft  Noise Annoyance Around London Airport. Central Office of
    Information. London, 1963.
Nunnally, J. Psychometric Theory. New York: McGraw-Hill, 1967.
Simon, H. A. Models of Man. New York: John Wiley and Sons, Inc., 1957.
TRACOR. Community Reaction to Aircraft Noise,  Vol.  1  and 2.  TRACOR Document
    T-70-AU-7454-U, Austin, Texas, 1970.
Turner, M. E. and Stevens, C. D. The regression analysis of causal paths. Biometrics, 1959,
    75, pp. 236-258.
Van Os,  G. J.  Advisory Committee on Aircraft Noise Abatement, Ministry of Traffic and
    Public Works. Delft, Netherlands, 1967.
                                       705

-------
          COMMUNITY RESPONSES TO AIRCRAFT NOISE IN LARGE AND
                           SMALL CITIES IN THE U.S.A.1

                       Harrold P. Patterson and William K. Connor
                                     Tracor, Inc.
                                  Austin, Texas 78721

     From 1967 through 1969 communities around the airports of seven large cities in the
U.S.A. were  surveyed (both socially and acoustically) for response to aircraft noise.  Pro-
cedures, methods, results and  recommendations can be found in Tracor (1970). The project
was  conducted in two phases. During Phase I, Dallas, Texas; Denver, Colorado; Chicago,
Illinois; and Los Angeles, California, airports were surveyed and 3,590 interviews collected.
Phase II involved Boston, Massachusetts; Miami, Florida; and New York, New York, airports;
2,912 interviews were collected for this phase.
     In order to  extend our knowledge about the relation between lower levels of noise
exposure (primarily involving  lower numbers of aircraft) and community response, a second
project was initiated  in  1970. In this study two  small cities, Reno, Nevada, and  Chatta-
nooga, Tennessee, were  surveyed using techniques and procedures similar to those of the
previous  project. In Chattanooga 1,114 interviews were collected; 846 were obtained in
Reno (For details, see Connor and Patterson, 1972.)
     Comparing responses  between  large and  small cities,  one notices immediately  that
responses in the smaller cities are of a different order. For almost all "reaction" variables the
intensity of response is  less. This paper examines some of these differences, discusses the
relation between annoyance and noise  exposure,  and suggests several alternative explana-
tions for these differences.
     Table 1 shows the distribution of annoyance2  and the mean CNR (Composite Noise
Rating) among the large and small cities. In  terms of high annoyance, New York is the most
disturbed, followed by Los Angeles and Boston.  Chicago  is about midway in the distribu-
tion. Clustered after Chicago are Dallas, Denver, and Miami. Reno and Chattanooga are
located at the very bottom  of  the distribution. The spread  of  low, medium, and  high
annoyance in the small cities  is most similar to that of Denver. Mean CNR values are  also
generally distributed in the same manner. The not-unexpected conclusion is that the amount
of high annoyance varies directly with the mean level of CNR.
     However,  if we  examine the percentage of  high annoyance by each noise exposure
category  for the large and small cities (Figure  1),  we find that this percentage is generally
less in each noise category for the small cities.  Only at the extremes (the 80-84, 85-89, and
125-129 categories) do the small-city and large-city responses compare. This would indicate
that  the relation between annoyance and noise exposure is of a different type in the small
cities.

     !This paper was developed  from work performed under the auspices of the National Space and
Aeronautics Administration-Contracts  NASw-1549 and NAS1-10216 by Tracor, Inc., Austin, Texas, USA.
       Annoyance" refers to the Annoyance-G scale developed during Phase I of the large-city study and subsequently
used throughout Phase II and also the small-city study. For details see Tracor, 1970 and Connor and Patterson, 1972.

                                         707

-------
                                            Table 1



                           DISTRIBUTION OF ANNOYANCE AND MEAN CNR AMONG

                                LARGE AND SMALL CITIES IN THE U.S.A.
-o
o
00
City
Large
Chicago
Dallas
Denver
Los Angeles
Boston
Miami
New York
Small
Chattanooga
Reno
Year

1967
1967
1967
1967
1969
1969
1969

1970
1970
Annoyance
Low

43%
52
62
31
29
56
14

74
65
Moderate

23%
23
17
22
28
23
23

18
21
High

34%
25
21
47
43
21
63

9
14
Mean CNR

107.3
110.1
99.5
110.7
107.5
106.3
114.6

101.2
101.1
Sample Size

872
923
1009
786
1166
676
1070

1114
846
                   v'C
>lfFrom Connor and Patterson,  1972, p.  24 and Appendix D,

 p. 141, and Tracer, 1970,  Appendix B,  p.  B-9.

-------
o
VO
           Q
           w
  100
   90
   80

   70

   60
o
3  50
H  40

   30

g   20
    10

     0
                 80
                90
100       110

     CNR
                                                                     LARGE CITIES
                                                                       (N  = 6502)
                                                            SMALL CITIES
                                                              (N = 1960)
120
130
                            Figure 1: Percentage of high annoyance by noise exposure*
                  *From Connor and Patterson, 1972,  Figures  6  and 7,
                   pp. 25-26.

-------
     A clearer picture of this annoyance differential is shown in Figure 2, where annoyance
is regressed on CNR. In this particular analysis, the Phase II sample of the large cities, being
closer temporally and methodologically to the small-city sample, is the basis of comparison.
The two regression lines are given by

     Large city - Phase II:  Annoyance = -35.3 + 0.497 CNR
     Small city         :  Annoyance =  -9.2 + 0.190 CNR

     The ratio of the two slopes is 2.6, indicating a substantial difference between the two
relationships.
     Other variables show the same kind of pattern. Table 2 shows a comparison of six
variables which  were found to be related to annoyance.  "Fear" refers to the anxiety  over
the  possibility of aircraft crashing  in the neighborhood. "Susceptibility" deals with  the
sensitivity to everyday neighborhood noises. "Adaptability" means the degree to which the
respondent is willing to tolerate more noise. "Misfeasance" taps the respondent's belief that
public officials are not doing a proper job in preventing or reducing the noise.  "Importance"
refers  to  the respondent's  evaluation of air  transportation.  "Discussion"  is the  rate of
discussing aircraft noise with friends, relatives or neighbors. Except for "Susceptibility," and
possibly "Importance," there  is a clearcut division between small and large cities. However,
there are similarities between Miami and  the small cities. We also  noted above that the
distribution  of  annoyance  in  the two  small  cities  was similar to  that in Denver. These
similarities indicate that differences in response are not necessarily related to differences in
city size per se.
     The  question is to what these differences may  be attributed. Field and measurement
techniques and  procedures  remained the same over the two projects. Two possibilities are
differences in the formulation of the noise exposure parameter  or  differences in sample
characteristics.
     Exposure values were computed according to the following formulae
     CNR    =      101og10?antilog(CNRj/10)

     CNRj     =     PNLj= l01ogio(ND. + 20NN.)- 12

where j  is a single class of operation producing a particular  noise characteristic at some
particular reference point,  NDJ and NN: are the number of occurrences in that class during
the periods 0600-2100 and 2100-0600, respectively, and PNLj is the energy-mean maximum
perceived noise level for that class. The PNL for each measured  flyover was determined by
adding seven  units to  the  maximum N-weighted level, this  being  the  correction factor
established  from  the  large-city  data  between   the   latter  and  the  non-discrete-
frequency-corrected PNL calculated from band analysis.
     There are two ways in which the above equations can be  altered so as to produce a
shift  in  the  noise exposure. One  is to have a drastic difference between nighttime and
daytime  operations. The other is to use a different number correction than 10 logjo N.

                                        710

-------
  r   45
   PC
   o
      30
w
C_>

1
>•
o
!S
13
       15
        0
         80       90
                                       i
100       110

     CNR
                     LARGE  CITY -  PHASE II
                            (N = 2912)
                                                 SMALL CITY

                                                 (N = 1960)
120      130
                  Figure 2: Regressions of annoyance on noise exposure1
        *From Connor and Patterson,  1972, Figure 9, p.  36.

-------
                                           Table 2
                                 ANNOYANCE-RELATED VARIABLES
(O
Variable
High fear
High susceptibility
High Adaptability
High misfeasance
Low importance
High discussion
Small City
Chattanooga
18%
5
57
5
1
19
Reno
13%
9
61
8
1
14
Large City -
Boston Miami
44% 16%
10 4
29 50
16 9
4 2
46 29
Phase II
New York
51%
7
19
19
6
54
               *From Connor and  Patterson 1972,  Table 4, p. 14,  and
                Table 1, 'p. 11.

-------
     A  difference between day-night operations among the large and small cities can be
quickly dismissed. In Figure 3, a graph of the percentage of operations against the time of
day shows that the pattern in the small cities is the same as that in the large cities.
     The use of 15 log N term, as used by the British NNI measure (McKennell, 1963), or of
a 13 log N term, as in the German Q index (Burck et al, 1969), instead of the 10 log N term
would emphasize the differences in operation counts between the large and small cities. This
could reduce  apparent  annoyance differences. The  range of daily operation counts in  the
small city study was 50 to  54  (mean  = 52). In the large-city study the range was 353 to
1.573 (mean =  834). If a 15 log N term were  used, an effective shift in 6 units could be
expected.
     Table 3 shows what would happen to the correlation between annoyance ami CNR if
the constant K is altered in the term K log N in the noise exposure formulation.
Comparing the correlations across, one finds little effect in changing the constant K.
     The possibility  that differences in  the small city and large city samples are responsible
for the differing relation between annoyance and noise exposure must also be discounted.
Table 4 shows a comparison of sample  characteristics between the small and large cities. All
of the values for the small cities are within the range of values for the large cities, except for
residential mobility in Reno.
     Three other possibilities exist for explaining the annoyance differential:  1) a seasonal
effect, 2) differential response to takeoff vis-a-vis landing noise, and 3) varying amounts of
social interaction.
     All of the  seven large cities were  studied  in the summer, which is known to be  the
season for heightened reaction. For example, as is shown in Table 5, the "complaint season"
at Kennedy International Airport is June, July,  and August, which together account for 65
percent of the total  complaints in the year. On the other hand, the small cities were studied
in the winter.  The mean monthly complaint at Kennedy during the small-city survey period
(October through January) was  1.9 percent. For the large-city survey period (May through
September), it was 16.6 percent.
     The hypothesis being  offered is that reaction to noise exposure is affected by the locus
of normal living activities. During inclement weather, individual activity tends to be con-
strained toward  the indoors, thus providing  insulation from aircraft noise exposure. Under
better  weather  conditions the locus of individual  activity  is much wider, thus  providing
greater exposure to noise. When these ideas are  applied to the small-city data, we speculate
that individuals  living under the same objectively measured  noise conditions as in the large
cities did not react in the same manner because the effective noise exposure was less.
     An alternative explanation for the annoyance differential is that individuals react  dif-
ferently to takeoffs vis-a-vis landings. If we construct CNR measures separately for takeoffs
and for landings, and then correlate these with the combined CNR measures and annoyance,
we  obtain  the results found in  Table  6. Here we see that indeed there are differences
between the small and large cities. The large-city  noise exposure is based mostly on landings;
the  small-city  noise  exposure is based mainly on takeoffs. One sees this  same effect when
annoyance is correlated with takeoffs and landings separately.
     The landing-doninated  CNR measures for the large cities and the takeoff-dominated
CNR measures for the small cities are probably due  to a combination of geography and size.

                                        713

-------
          CO
          g
          o
          H
          Oi
              30
          %  20
              10
              •LU
• Dallas
• Denver
A Los  Angeles
                                              o Chattanooga
                                              n Reno
                0000       0600       1200
                                   TIME OF DAY
                              1800
2400
*From  Connor  and Patterson,  1972, Figure  4,  p.  18.

                  Figure 3: Percentage of total daily operations by time of day.

The three Phase II airports (Boston, Miami, and New York) are located by the ocean or next
to largely unpopulated land  areas and all have multiple-runway  systems. The options for
flight operations are thus greater. Where options exist, noise exposure from  takeoffs  is
normally minimized, since takeoffs are considered noisier than landings. The airports at the
small cities had only one main runway and were inland. Very few options existed at these
locations.
     There is an indication that these data show a greater sensitivity to landing noise than to
takeoff noise. Note that although the CNR for large cities is composed mostly of landing
noise (landings correlate 0.61 with the combined CNR in Table 6) and  the CNR for small
cities is composed mainly  to takeoff noise (takeoffs correlate 0.83 with the combined CNR
in Table 6), annoyance correlates 0.42 with  landings for the  large cities but only 0.21 with
takeoffs in the small cities. If sensitivity were the same, one would expect a higher correla-
tion between takeoffs and  annoyance in the small cities.
     A  third possibility for the explanation of the  annoyance  differential is the lesser
amount of social interaction focused on aircraft noise  in the  small cities. For example, the
variable "discussion," which measures  the number of times in  an average week aircraft noise
was  discussed  with friends, relatives,  or neighbors, was a significant predictor of annoy-
ance.3 This  was not the case in the large-city study. Yet, the  percent with high discussion

      Discussion" was ranked 6th among six predictors of annoyance in a multiple regression scheme. It correlated 0.20
with annoyance. See Connor and Patterson, 1972: 47-50 and Table 14.
                                       714

-------
                          Table 3
        CORRELATION OF ANNOYANCE WITH CNR-TYPE VARIABLES

CNR -Type
Variable

Landing
Takeoff
Combined
Large

5
0.42
0.04
0.43
City -
Constant
10
0.42
0.03
0.41
Phase
(K)
15
0.42
0.02
0.40
II

20
0.41
0.02
0.40
Small
City

Constant (K)
5 10
0.14 0.13
0.23 0.21
0.27 0.25
15
0.13
0.19
0.24
20
0.12
0.17
0.22
*From  Connor and Patterson, 1972,  Table 7,  p.  38.

-------
                            Table 4
               COMPARISON OF SAMPLE CHARACTERISTICS
Characteristic
Percent High
Occupational Rating
Percent Income
$10, 0004-
Percent Education
More Than High School
Percent Age 604-
Percent Homeowners
Percent High
Visitation
Median Residential
Mobility
Small City
Chattanooga
25
39
32
24
81
39
0.59
Reno
19
50
34
16
75
39
2.09
Large
Boston
22
37
29
24
63
50
0.08
City -
Miami
20
36
43
33
74
35
0.44
Phase 11
New York
37
56
38
13
82
54
0.13
*From  Connor and Patterson, 1972,  Table  1,  p.  11.

-------
                          Table 5

      PERCENTAGE OF ANNUAL COMPLAINTS RECEIVED BY MONTH,
           1959-1967 KENNEDY INTERNATIONAL AIRPORT
Month
Jan
Feb
Mar
Apr
May
Jun
Jul
Aug
Sept
Oct
Nov
Dec
Percent of
Annual Complaints
20.0
1.5
2.5
4.0
7.5
18.0
24.5
23.0
10.0
3.5
2.0
1.5
*From  Connor  and Patterson, 1972, Table  6,  p. 32


                          Table 6

     CORRELATION OF CNR-ELEMENTS WITH CNR AND ANNOYANCE
Partial
CNR
Landings
Takeoff s
Large Cities - Phase II
CNR Annoyance
0.61 0.42
0.29 0.03
Small Cities
CNR Annoyance
0.35 0.13
0.93 0.21
   "From Connor and Patterson, 1972,  Tables  7
    and 8,  p. 38.

                           717

-------
rates are much lower in the small cities than in the large cities (see Table 2). We suggest that
a certain level of social reinforcement is necessary for a feeling of annoyance to develop, and
that the small cities  fall below this level. Of course, the idea of social reinforcement may
also be connected with seasonal variations in social activities,  indicating operation of an
interaction-type variable.
     We do  not insist that the three possibilities for the explanation of the annoyance
differential adduced above are the sole causes of the varying patterns of annoyance reactions
between the small and large cities. We do believe,  however, that they are worth serious
consideration and that research into the effects of aircraft noise should begin  to take into
account such variables.

                                     SUMMARY

     In 1967 and  1969, 6,502 interviews were collected around the airports of seven large
cities in the  U.S.A.  An additional  1,960 interviews  were collected around airports of two
small cities  in 1970.  Acoustical surveys were conducted parallel to the interviewing. A
comparison of responses to aircraft noise between the two-city  types shows that the inten-
sity  of response is different in small  cities than  in large cities.  In almost all "reaction"
variables the level of intensity is much lower in  the small cities.  These differences, which
increased as noise levels increased (except at levels above CNR 130) could not be attributed
to differences  in  the  formulation  of noise exposure measures nor in the sample charac-
teristics. Possible reasons for the differences are 1) a seasonal effect, 2) differential  response
to takeoff vis-a-vis landing noise, and 3) varying amounts of social interaction.
                                    REFERENCES

Burck. W.. M.  Griitzmacher. and  F. J. Meister. Fluglarm: Seine Messung und Berwertung,
     seine  Beriicksichtigung  bei der Siedlungsplanung,  Massnahmen zu seiner Minderung.
     Bundesministeriums fiir Gesundheitswesen, Gottingen, 1969.
Connor.  William K.. and  Harrold  P.  Patterson, Community Reaction to Aircraft Noise
     Around Smaller City Airports, NASA CR-2104, Washington, B.C., August 1972.
McKcnncll. A.  C., Aircraft Noise  Annoyance Around London (Heathrow) Airport. Social
     Survey Report S.S. 337. Central Office of Information, April 1963.
Tracer, Inc., Community  Reaction to Airport Noise, Vol. I, NASA CR-1761; Vol. II, NASA
     CR-1 11316. September 1970.
                                         718

-------
        SESSION 8
                     Vr-V
 COMMUNITY RESPONSE II
Chairman: R. Rylander, Sweden
           719

-------
       MEASUREMENTS OF STREET NOISE IN WARSAW AND EVALUATION
          OF ITS EFFECT ON THE ACOUSTIC CLIMATE OF DWELLINGS,
       SCHOOLS, OFFICES, HOSPITALS, HOTELS AND PARKS; THE DEGREE
   OF OFFENSIVENESS TO  INHABITANTS IN THE LIGHT OF A QUESTIONNAIRE

                              Aleksander Brodniewicz
                               Department of Hygiene
                            Institute of Biological Sciences,
                            Academy of Physical Culture
                          Warszawa, ul. Marymoncka, Poland

     The progress of technical civilization, urbanization and trend towards development of
motor transport are paralleled by steady intensification of noise, owing to multiplication of
its sources and emission of very large amounts of acoustic energy. Noise is man's companion
in industrial  plants,  workshops, offices,  schools  etc., markedly  affecting the efficiency,
quality  and  safety of work.  Noise is  unavoidable in streets, means of transport, shops,
catering or recreation places and even in apartment houses expected to ensure silence and
repose. Thus, the  disturbances in rest by day and especially during night sleep (both being a
prerequisite  of regeneration of strength and health) are greatly harmful and entail major
bio-physical risk, in my opinion.
     Noise most  severely affects the urban population exposed to 24 hours' loud street
noise and dwelling in many-storied residential buildings without adequate sound-proofing;
present-day building materials being less massive than in the past,  the insulating  properties
of walls and ceilings are much inferior. Street noise reaching the apartments, together with
sounds from  neighboring quarters, e.g. conversation, children's games and shouts, inappro-
priately-utilized musical instruments, radio and television sets, household appliances, eleva-
tors  and chutes,  banging  of  doors,  chair moving, etc., blend into a perpetual acoustic
inferno.
     People  subjected during  conscious state and sleep to uninterrupted reception-against
then" will—of acoustic disturbances reaching them from all sides  are in no position to
effectively isolate themselves  from the environmental noise. Thus, in urban population the
import of audible  sensations definitely outweighs that of all  other ones, even visual. The fact
of a constant increase in the percentage of urban population subjected to  excessive city
noise gradually and imperceptibly impairing its comfort, stamina and health, has been  and
continues to be of concern to  the community and authorities.
     In  Poland, the first studies of noise in the cities of Warsaw, Cracow and Wilno had been
initiated in  1933 in  the Department of Hygiene of the University of Warsaw and then
reported by  prof. W. Gadzikiewicz (1936). Their results are greatly  useful for comparative
purposes.
     Rapid reconstruction of Poland as well as its intense economic and industrial develop-
ment after World War II have cleared a way for technical progress in all fields of activities,
together with noise as its inherently-present stigma and nuisance.
     Initially, between 1945  and 1956, despite warnings of hygienists, this problem  had
been disregarded  and  unperceived from the standpoint of both health and economics. Only

                                        721

-------
                                                         Table 1



                                 RESULTS OF THE FIRST INVESTIGATIONS ON STREET NOISE

                                                       IN WARS AW
                     Result* of the first investigations on street noise in Warsaw


                     on 7th llaroh 19J6 /15-16h/  after Prof.W.Oadslkiewio*.
                                                                                            Table nr 1
to
to
Place
Square Zbawlciela
Corner of HaraaaZkowska
and Koszykowa Streets
Central Hallway Station
Corner of Towarowa and
Srebrna Streets
Corner of Topolowa and
Nowowiejska Streets
dorner of Polna and fco-
kotowska Streets
Square Unli Lubelskiejj
Palat Street
Kind of pavement
Gobble smooth stone
„«_ -«_
Cobble smooth stone
Rou£h cobble stone
Cobble stone
Cobble smooth stone
— "« -"—
_«- -"-
Source of noiae
traiway and motorcar
moving slowly
motorcar moving quickly,
tramway noving quickly,
waggon spring noving walking pace
tramway movinc slowly,
horse drawn cab,
waggon without rubber tyres,
autobus,
signal of a ringing tracivay bell
distance ]to.
Coal cart moving a uolkirv; pace,
Waggon spring moving at a trot
Waggon spring moving a walking
pace
2 waggons spring aovinc at a trot
Coal cart and traoway
tf lying plane
Noise level
in phones
50
60
65
55
35-10
^5
55
55 - 60
75
6C - 65
80
70 - 75
60 - 65
75
85

-------
the potential public health hazard from noise, the multitude of complaints voiced by city
inhabitants, the increased number of accidents augmenting economic damage, and the over-
all social importance of the problem of noise made of the latter an object of many-sided
investigations and a source of concern to the authorities, with respect to measures to be
taken to prevent and  control noise. Most studies have dealt with evaluation of street and
apartment  building noise in the cities of Poznan, Lodz, Cracow and, especially, Warsaw; the
capital was particularly suitable  for investigating the problem of noise, on account of its
rapid transformation into a modern industrial, trade and culture center.
     The results of studies performed by various medical and technical research centers
permitted characterization  of the biophysical parameters of noise, their assessment from the
standpoint of hygiene and, in parallel, exploration of the opinion of inhabitants by means of
questionnaires.
     My own investigations involved noise both in streets and railway network, both being
in Warsaw  an integral part of the city's means of transport. The main sources of noise were
studied and its  physical properties (intensity, spectrum, distribution in time) were charac-
terized; analysis was made  of its spatial range, and especially of its effect on the acoustic
climate of  apartments, schools, hotels, hospitals, offices, railway stations and parks both in
the city center  and suburban districts. Also, studies were made of the additional acoustic
load to which passengers are exposed furing rides on a bus, tramway, trolley bus and electric
train; recently,  the problems  of noise in sport centers (gymnasia, swimming pools) and
students' dormitories were investigated as well.
     To accurately record  noise, which is a physical  phenomenon rapidly  changing  as a
function of time  and site, application  was made of Bruel-Kjaer's microphones and  self-
recording electro-acoustic equipment, Philips-Professional and Emi  tape recorders as well as
of sound-level  meters of Polish make.  Measurements were taken on windless  or nearly
windless days at 173 selected sites in Warsaw, under conditions standardized with respect to
localization of equipment etc.
     The overall results of the investigations on street noise in Warsaw, including tables,
graphs and  figures, were published as a monograph in 1963.
     In this brief report, only the most important results will be summed up in the form of
figures  with a  short commentary. Sources of sounds most  contributing to  generation of
street noise include motor  and rail vehicles as well as trolley buses (starting, braking and
signalling sounds),  aircraft,  loud-speaker broadcasting programs, street works and building
operations, sounds from railway stations, bridges and industrial plants.
     In the relatively quiet suburban districts, in addition to vehicular traffic noise, there is
a predominance of sounds resulting from children's shouting and games, voices of passers-
by, sounds  from apartment buildings, sports fields, dogs' barking etc.
     The highest intensity of noise was observed along the  main traffic arteries and their
crossings where  the recorded maxima varied  from 95 to 103 dB, with average intensity of
85-90 dB. The highest noise level,  112 dB, was noted in the tunnel of the East-West route,
during simultaneous passage of tramways, cars and of a tractor.
     Other streets  and squares with intense  traffic also showed a very high average noise
level of 80-85 dB.
                                         723

-------
     Motor vehicles, especially tractors, tramways, buses, cars and motor cycles are a source
of most intense and annoying street noise. These relationships are clearly visuali/ed in I;igs.
2. 3. and  4. showing  in parallel to noise records  the intensity of vehicular traffic at a busy
crossing  i.\l.  Jero/.olimskie  and Nowy Swiat) and on one of the central  bridges (Slasko
-Dabrovvski Bridge i.
     In the central trade districts,  intense noise prevails also on side-streets,  especially those
with  trolley bus  lines.  These streets, being narrow  though of small traffic capacity, are
jammed  with vehicles and passers-by; chiefly with one-way intense traffic, they are lined
with a continuous row of many-storied buildings
     It is  stressed  that in big cities, street noise uninterruptedly prevails day and night, as
well as throughout the whole year. The  former phenomenon is illustrated in Fig. 6. which
shows a noise  level record taken in a room on the IVth floor of an apartment house situated
near one of the central bridges.
                              •   \


                                                    
-------
 BUSES
 TRUCKS
 CARS

                .x-vvvv
M01OK-
TKAMWA¥S


TRO1AEV-
BUSES


TRACTORS
^^^>^sN^x^^xN^NX^^sN^:s^3 '»
                         XsXXX^J 63
                          3C1BB 6»
               ..••.•*..••.••..••.*•.••..*•.
DRIVEN
CARS

SUBURBAN
                                           es
TRA1NS
€l£CTRlC

              io
                        5O   6O  'O   8O   9O   «OO  no
dB
           1NA«*RT*A£NTS            ON THE STREET
          /SHUT WINDOW/
          /VllNIVt.  0BB WAX.      5HJ MINIM.  •• /V1AX.
  Figure 2: No«e caused by vehicles as heard in Warsaw streets and apartments.
                                 725

-------
                      POZKDMU QtOSMO*Ci  UALA«U UUC1NEQO OffA? POCwu MOtO*tQO
                  MA  8WU2V2OWAWIU  Al. 3rPO7OUM8HICM I MOttFCO  4WUTU
                                                                        MOM llvll.
                                                                     WICMUU MOi«f LIVU
                                                                              HVf 1
                                           COOl'M*  BOUM
Figure  3: Dependence of street  noise of traffic intensity in one of Warsaw's central crossings during 24
hours.
          -•
          :
          .
              300
              -
          - a
          :
          i
,00
 60

 3O -
                       KRZVWE POUOMU  QhOSNOSCl MAbASU
                         NA  WO8CIE  3LASKO-DA8ROKJSNIM
                                                               MAI NOISE  Llvtl
                                                                          llvll
                        JO»24?i68O
      Figure 4: Dependence of street noise of traffic intensity of ttie Silesian Bridge during 24 hours.
                                            726

-------
85
70-
                                                       SS-7O
                   Figure 5: Noise level in a section of the central part of Warsaw.
     Owing to a substantial drop in vehicle traffic at night, at this time the mean noise level
decreases; however,  at night the amplitude of noise fluctuations (40-44 dB) is greater than
the amplitude  in  the daytime (22-25 dtU So high  a level of fluctuations  of the audible
disturbances at night,  their irregularity and  frequent  recurrence  at  brief  time intervals,
seem  from the standpoint ol hygiene even more objectionable.
     The  railway  network, the main  lines of which cross  the most densely  populated dis-
tricts of Warsaw, is also an integral part  of the city's communication system and a source of
intense  noise. I he penetration range and degree of nuisance of railway  noise  mainly depend
on the kind and course of tracks along  bridges and embankments; the intensity of day and
night  railway traffic, rapid passenger transportation and transport of goods; lack of protec-
tive zones and noise-breaking barriers: speed and direction of winds; etc.
     Small passenger stations and central railway stations, as well as apartment buildings
situated near them, are most exposed to  railway noise.
                                          727

-------
           Noise  level  records during  15 min  in the same room
                / speed of  recording 0,1 mm/sec./
      08

      JO

      to

      M
                  open  window
                  1010- 1025h
shut window
1046- 11  h
Figure 6: Noise level records (in decibels) during 24 hours in a room, with open window, on the 4th floor
of an apartment house situated near one of central bridges (speed of recording 0, 003 mm/sec).
                                     Table 2

       NOISE LEVEL MEASURED AT ELECTRIC RAILWAY STATIONS IN WARSAW (dBA)
No.
1
2
y
4
5
6
Railway station
Ochota
Srodmie6cie
Powisle
Stadion
Wscnodnl
B K D
Platforms on arrival
of train
68-86
b1 - bb
VG - 56
70 - 7?
72 - 84
66 - 95
Platforms during
slit£t traffic
54 - 5&
55 - 57
GO
48 - 50
52 - 55
52 - 58
Loud speaker
aanounc ement s
76 - 80
7i - 05
-
74 - 80
62-84
60 - 65
                                      728

-------
                                                     Table 3
                            NOISE LEVEL MEASURED AT MAIN RAILWAY STATIONS IN WARSAW (dBA)
K)
Ho.
1
2
3
4
5
Llain railway
station
G16wnj
Central ny
WBchodni
Wilencki
Gdansk!
Station
hall
61 - 80
54 - 61
61 - 65
62 - 69
62 - 76
Waiting-
-roorn
58 - 66
54-61
54-72
5& - 62
60 - 76
Station
restaurant
66 - 72
	
65 - 83
63 - 70

Platforjas
55 - 78
62-66
00-84
60 - 79
59 - 89
Loud speaker
axuiounc ement s
65 - 80

	
70 - 71
64-68

-------
     Noise levels are excessively high  also in the neighborhood of many hotels and other
 public utility buildings. On these grounds, the location of some of them is greatly inappr-
 opriate.
     Special consideration was given to the acoustic climate inside hotels, hospitals, research
 institutes and schools, as well as to the disturbances in this climate caused by external noise
 prevailing in their environment.
     Biophysical risk  entailed by noise is highest in case of patients and  convalescents,
 interfering  with treatment  and often  causing its  prolongation,  on account  of their
 augmented susceptibility to any kind of stress.
     The high degree of nuisance resulting from street noise is confirmed by complaints of
 patients, opinions of physicians and results of measurements presented in Table 4. (Fig. 7).
     A modem hotel is expected, aside from providing conditions of hygiene, comfort and
 esthetics, to insure silence as a fundamental biological requirement which is a prerequisite of
 good general feeling, rest and sleep.  In  what measure large hotels in Warsaw meet these
 requirements is illustrated in Table 5 and  Fig. 3, 8 and 9.
     The level of noise  reaching the surroundings of research institutes depends whether
 they face the street or are situated either inside a park or else at the  rear of front buildings
 shielding them from street noise.
     Noise levels measured in the neighborhood of several research institutes varied within
 the following ranges:
      — minimum   45-73 dBA
      - medium    50-82 dBA
      - maximum   58-98 dBA
     The amount and intensity of audible stimuli reaching an individual during a definite
time  interval  constitute a measure of psychophysical noise load; some  relevant data are
shown in Tables 6 and 7.
     Special attention has to be given to school noise. Modern school is,  irrespective of its
character or specialization, an institution devoted to intense systematic mental work requir-
ing a calm  and quiet climate.  Creation  of an appropriate climate in schools is a major
pedagogical problem decisive of the intelligibility of speech as a universal means of commun-
ication permitting attainment of the intended educational aims.
     In Poland, the  problem of school noise has come into prominence,  and studies of its
various  aspects  (acoustic, pedagogical, hygienic and social) have been undertaken in the
fifties. To illustrate  the acoustic climate  in an elementary school in Warsaw (TPD No. 27,
situated at a distance of 50 and 100 m, respectively, from two streets), it is worthwhile to
quote  the results of measurements—paralleled by an inquiry—performed in winter when
most of the  time the windows were closed and both pupils and staff mainly stayed indoors.
     Loud ness level in class  in the course  of a lesson carried on fairly quietly amounted to
59-92 phons; during a lesson of physical exercise, 59-72 phons; in the recreation hall during
breakfast break, 57-80 phons; in the teachers' room during break, 57-72 phons.

                                        730

-------
                                                   Table 4

                                   STREET NOISE LEVEL MEASUREMENTS IN THE
                                       ENVIRONMENT OF HOSPITALS, in dBA
OJ
Ho
1
2
3
4
5
6
7
8
9
10
Hncrrvt +-ol

Henatologioal H.
Blubietanski H.
Hospital of Infe-
ctions Diseases
No. 1.
Clinical Hospi-
tal Ho. 3.
Clinical Hospi-
tal Ho. 1.
Institut for Sta-
te erculosis Re-
search
Municipal Hospi-
tal Ho. 4.
Grochow Hospital
Pediatric Clinic
Hospital of Pe-
diatric Surgery
Forenoon
max
»9-93
65
100-102
86
88
96
92
88-95
96-97
90
inin
53
49
69
64
65-67
65
6S
58
68-72
70-72

medium
65
53-55
75-77
73-75
73-76
72-75
75-78
65
90
78-80
Afternoon
max
93
63
92
63
93-97
91-92
85
' 61
92-99
93
rti'ln
52
48-50
68
67
68
55-56
68
65
65-66
63-64
medium
65-68
54-56
72-74
74-76
75-78
63-65
74-76
70-72
73-75
70-72
nignt
Ifljfty
53
44-46
86-87
72-75
72-76
83
91
79
92
82
din
43
36
48-50
50
48-50
40
56
45
62-64
56
medium
48-50
42
55-58
52-55
55
42-45
65-68
52-55
72-73
70

-------
                                             10- -
                                  •pern wind«w
                                                    ~ io*»A.M.
                              •hat window
Figure 7: Street noise level in a ward of Pediatric Hospital situated in one of central streets in dbA.
                                      732

-------
                                                    Table 5
-j
u>
u>
                                   STREET NOISE LEVEL IN THE ENVIRONMENT OF

                                       SOME HOTELS MEASURED 1958 in dBA
•o

1
S
3
4
Hotel

Oread Hotel
Buropejski
Hotel
WareMM
Hotel
MOM Hotel
lore
MZ
86-89
98
95-96
87
noon
•in
68
73
73
73-7*

Bedim
77-79
80-82
80-82
78-80
Attei
«*
87-88
98
89-91
90-91
•BOOO
Bin
69-70
73
63
73

•ediuB
75-78
82-65
75-76
78-60
1
—
85
73
70
82
Ight
•in
45-50
48
42-44
o^^oo

•edliiB
65-68
52-55
54-56
55
                                              •e«mired 196? - la dB/V
1
2
3
4
5
Oread Hotel
Buropejaki
Hotel
W«MM*« Hotel
M D M Hotel
•etaopol
Hotel
65 - 80
68-83
63 - 78
64-84
75 - 82
72-80
73-87
67-87
65-80
70-83
59-75
75 - 89
58-80
66-84
68-77

-------
                              HOTEL
                                                  Z A
                                                                   "1. VaxecJci
                                   Potc6J  227.  IT  ptr.  od ul'.Gwi'tglcrzyekiej
                                                                   zapls przy zemknlgtyr: oknie
                                           1 »OW1
1045  -  11   zapie  priy  ofcnie otwartym     g.11°5_ 11*

 Front  room Nr.227.  2nd floo^- - directed north to the 3wi^tolcrzyEk« 3troet,
                          with op*n, a:id clos»rt window.

                                       rf'oa
                        ?okoj 1504.  XV ptr.  nti-oinik  '1.  u>*cklcgo  1  fwi?tokrzy»kl«J
             ;.114° - 1155   okno otwtu-ta               g.12  -  12^5   Zapi»  priy oluilc zaraknlrtyn
                        Corner roon;  ,Tr.1S04  -  15th floor  with op«n  and closed window.
                                                          "~
                                   PolcrfJ  1502.  ::v  ptr.   fftr.zachodnla  (..a
              J.I2?0  -  I?15   zapis  przy  oknl* otwartyn        g.1250  -  13°5   okno  i«mknlc;t»

                               Mr. 1502 -  15th floor  facing  west  to th»  :.'aj-szaikowsk«  Ltreat.
                                   1% rain,  registration  with  open and  closed window.
Figure 8: Noise level measured in different apartments of Hotel Warsaw with open and closed windows.
                                                   734

-------
                                HOT " L   '.I T) R
                                                    Varszawa,  Fl.  Konstytuojt nr 1
                                    oj  326.  Ill  ptr.  od ul.
                         zapis  pr?y  "toiie  ntvu-ty-       1C*0 - 10^5 z»pis przy oV-nj^ seller.!•=,tym
                     Front  room !Ir.3i8  -  3rd  floor situated on the r.:arsialltow3)ca street
                                       w?th  op*n and  cloeej  window.
                                                                 cf; OB
                                         319.  Ill  ptr.  od  PI.  Konatytucji
            1355 . 141(^  inpis przy  otoiie otwartynt        H1"?- 143^ znple przy olcnie zamkni^tyc
                 Front  roon: '.'n the  onetltution T.quare  with open
                                            and closed  window.
                                                                  0 D8
                                         okoj 338. Ill ptr. od podworta
             1240 _ 1255 Zap.is przy otoiie otwartys:         V}00 - 1315 zapis pr^y oknie za-il-Jiivt
               Dact-roor. '(r, 330 - 3rd floor. 15 rain, resisirnt! on with open and closed window.
Figure 9: Hotel MDN—Constitution Square, noise level measured in different apartments of the MDN-Hotel
          with  open and closed windows.

-------
                         TableG

NUMBER OF ACOUSTIC IMPULSES OF STREET NOISE AUDIBLE INSIDE
           DIFFERENT ROOMS DURING 15 MINUTES
Ho

1

2
3
4

5

6
7


8

9

10

f-lace of investigations

Dept. of Pediatrics
Clinical Hospital Ko.4
Hospital of Pediatric Surgery
Solec Hospital
Apartneut, IV floor, on one of
central Street
Apartaent, IV floor, another
central Street
Warszava Hotel
U D L Hotel
a/ lU floor, back roou
,, 3|
b/ III floor, facing one
of central public squares
c/ IIIrd floor, facing one
of central streets
Office of Polish Press Agency,
one of central street crossings
Bditorial Office, one of
central streets
Trade Onion Office, near one
of central bridges
number of impulses over
noise level of
95 dB A

22
-
-

7

-
21
8
1
7

2

17

51
90 dB A

129
16
6

106

26
98
177
52
135

84

165

560
:
85 dB A

382
111
79

332

156
380
151
410
205

562

198

174
                          736

-------
     The results of an inquiry on noise, carried out in the same school (respondents: 487
pupils, 7-14 years of age, and  19 school staff members), are presented in Table 7.
                                       Table 7
Perception of
noise as:
indifferent
pleasing
troublesome
^ __ 	 !
Pupils
during playtime
73
44
370 (75%)
during mental
work or rest
46X
—
441
School staff
members
1
—
11
     "including 12 pupils with uni- or bilateral hearing impairment

     Parks and public gardens, much attended for their recreation value, are reputed to be
oases of tranquility. Measurements taken in 11 parks showed that on the whole only about
half of them, especially the hilly ones, come up to expectation (Table 8).
     Noise levels in parks varied within the following ranges:
     — minimum
     — medium
     — maximum
35 - 60 dBA
38 - 68 dBA
42 - 84 dBA
     Rapid urbanization of Polish cities, involving predominance of many-storied buildings,
brings about marked concentration  of inhabitants in apartment houses and recreation areas
attached to residential settlements.  The resulting intensification of noise level is, however,
much less disturbing than that arising from the development of motor transport. The num-
ber of cars being parked along quiet streets within residential settlements and even on the
pavement next to buildings continually increases, adding to noise inside the apartments.
These facts often  find expression in critical  articles published in the local newspapers, as
well as in complaints lodged with the authorities.
     The number of complaints yearly pouring in to the Warsaw Sanitation and Epidemiol-
ogy Station was as follows:
                                     1967-  74
                                     1968 - 145
                                     1969-206
                                     1970-212
                                     1971-216
                                     1972-205
     To explore the  opinion of inhabitants on the degree of nuisance caused by city noise,
an inquiry was carried out. Its greatly simplified results are recorded in Table 9.
                                         737

-------
                       Table 8

NOISE LEVEL MEASUREMENTS IN WARSAW GARDENS AND PARKS
                       INdBA
iio
1
2
y
4
5
6
7
8
9
10
Park or
Public Garden
Krasiuski Park
Saski Garden
Palace of Culture
and Science Garden
Skaryszewski Park
Powisle Park
Botanical Garden
Lazionki Part

Bielanjr Park
Citadel Garden
Hakowiec Park
Noise level
Lax
70
66-70
72
61
67
62
46
45
61
56
Uin
56
55
58-60
15
44
yf>
57
59
46
44
Medium
62-64
60-62
66-67
50-55
49 - 51
45 - 44
40 - 42
40
52
47 - 48
                         738

-------
                                       Table 9
         RESULTS OF SURVEY ON NOISE IN APARTMENTS AND ON THE DEGREE OF
               ITS ANNOYANCE TO RESPONDENT INHABITANTS OF WARSAW
Ho
1
2
3
4
5
6
7
Sources of noise invading
the apartments
street vehicles
planes passing overhead
passere - by
children's shouts and
games
street loud speakers
railway stations,
industrial plants,
recreation places etc.
others
Interference with sleep,
rest and mental work
75 %
41 %
19,1 %
36,5 %
9 ji
17, 4 %
12 %
     After systematic studies of city noise between 195 8 and I960, the results of which had
been  published as a monograph in 1963, further pertinent noise measurements were per-
formed during  1968-1969. As a result of collaborative work of the Warsaw Sanitation and
Epidemiology Station and Institute of Building Technique an "Acoustic map of Warsaw"
characterizing traffic noise (cars, tramways, airplanes) according to  the criteria of loudness
level and as a function of frequency was prepared. It provides city-planners and architects
with valuable information, being of assistance in prospective planning of the development of
Warsaw.
     In conformity with  the "Acoustic map of Warsaw", the Warsaw Sanitary Service keeps
carrying out-at 2-year intervals-supplementary control measurements; the latest series of
noise determinations was performed in 1971.
    The results showed  that owing to improved condition of street surfaces, better techni-
cal state of vehicles as well as to standards restricting the noise of passenger cars to 80 dB(A)
and that of buses and trucks to 85 dB(A), maximal noise levels at central street crossings
exhibit a pronounced decrease, compared with the past years.
    The present brief survey of the results  of investigations on the acoustic climate of
Warsaw proves  that in  Poland, city noise commands continually increasing  interest: its
social, sanitary,  economic and work safety aspects.
                                        739

-------
     Many years' activities of several research institutes, an increasing number of specialists,
as well as vivid co-operation of the public opinion and of the press has contributed to rapid
progress  and popularization of bioacoustics of cities and towns, especially with respect to
noise prevention and control.  In  Poland the most important achievements  in this field
consist of acoustic standards that are obligatory in the building trade, bio-requirements in
city  and  regional planning, setting up of new research  centers attached to the  Polish Acad-
emy of Sciences, organization of the Polish Acoustic Society, preparation of a draft of the
Noise Control Act and  calling  into being the  nation-wide "League for Noise Control" in
1970.
     The effects of all these practical steps and progress in noise control measures can best
be observed  in case  of  the constantly expanding city  of Warsaw as a test model for other
Polish cities.
     In Warsaw, practical realization of the principles of noise prevention is exemplified by
zoning of the city, modernization of its road network, as well as by building of new railway
stations,  hotels and  schools with consideration given to noise-absorbing material  and to
bio-requirements in city-planning.


                                  BIBLIOGRAPHY

  1.  Bell, A., Noise:  An occupational hazard and public nuisance. Publ. Health Paper No 30.
     World Health Organization /1966/.
  2.  Beranek, L., Revised criteria for noise in buildings. Noise Control, 3, /1957/.
  3.  Brodniewicz, A., Sikorski, St., Badanie halasu ulicznego w Poznaniu. /Investigations of
     street noise at Poznari. Nowiny Lekarskie 23/24, 329-342, /1950/.
  4.  Brodniewicz, A., Zagadnienie halasu w oddziale leczenia snem. /A problem of noise in
     the ward of treatment with sleep. Pol Arch. Med  Wewn. 3, 245 - 256, /1954/.
  5.  Brodniewicz, A., Halas uliczny Warszawy i jego wplyw na klimat akustyczny mieszkan
     i biur /Street noise in Warsaw and its influrence on the acoustic climat of apartments
     and offices/. Gaz, Woda i Technika Sanitarna 9, 347-354, /1960/.
  6.  Brodniewicz, A., Badanie  halasu ulicznego na terenie m.st. Warszawy. /Investigations
     on street noise in Warsaw/. Warszawa 1963. 1-177.
  7.  Brodniewicz, A., Halas w swietle bad an ankietowych prowadzonych na  terenie m.st.
     Warszawy. /Noise in the light of a questionnaire/. Miasto, 11,13-17, /1963/.
  8.  Brodniewicz, A.,  Badanie warunkow  akustycznych w krytych plywalniach warsz-
     awskich./Examination  of  acoustic  conditions in  Warsaw indoor swimming  pools/.
     Roczniki Naukowe A WF Warszawa 6, 289 - 302, /1966/.
  9.  Brodniewicz, A., Komunalhygienische Untersuchungen uber das Larmproblem in War-
     schau. Zeitschr.f.d.  gesamte Hygiene u.i. Grenzgebiete. 10, 760 - 764, /1967/.
10.  Brodniewicz, A., Ergebnisse  von larmuntersuchungen  in Warschauer Hotels. Larm-
     bekampfung 2-3, 26- 32,/l968/.
11.  Brodniewicz, A., Klimat akustyczny obiektow szkolnych i jego znaczenie higieniczna.
     /Acoustic  climate  in schools and its hygienic importance/.  Papers  of the II. Polish
     Symposium on school hygiene. Zielona Gora 1968. 229 - 240.


                                        740

-------
12.  Brodniewicz, A.,  Srodowisko akustyczne niekt6rych hotel! warszawskich w ocenie
    higienicznej. /Acoustic environment of some Warsaw hotels/. Roczniki Naukowe A WF
    - Warszawa, 265 - 288, XI. 1969.
13.  Dominik-Ziemborakowa,   M.,   Halas  szkolny.  /School  noise/.  Roczniki
    P. Z.H.-Warszawa. 2-3,/1958/.
14.  Gadzikiewicz, W., Badanie halasu ulicznego w miastach polskich. /Street noise measure-
    ments in polish towns /. Zdrowie Publiczne 7, /1936/.
15.  Harris, C. M., Handbook of Noise Control.New York 1957.
16.  Maleckie,  J., Wplyw wymagan akustycznych na ksztaltowanie miasta. /Influence of
    acoustic criteria on the town development/. Miasto 3, III. /1952/
17.  Puzyna, Cz., Zwalczanie halasu w miastach. /Noise control in towns/. Ochrona Pracy 6,
    /1960/.
18.  Sadowski, J.,  Akustyka w  urbanistyce, architekturze  i budownictwie. /Acoustic in
    town planning, architecture  and building construction/ Warszawa 1971.
19.  Sadowski,  J.,  Postep  badari  w  zakresie akustyki urbanistycznej  na  przykladzie
    Warszawy. /Progress in town  planning on the example  at the city of Warsaw/. Archi-
    wum Akustyki 1, 21-52, /1970/.
20.  Siemienski, M., Kultura  a srodowisko akustyczne czlowieka./Culture and  the acoustic
    environment of human/ Warszawa 1967.
21.  Szudrowicz, B., Zuchowicz, I., Wyniki badan i wytyczne projektowania akustycznego
    w budownictwie szkolnym./Results of investigations and directives on design in school
    construction/.
    Warszawa 1962 - Unpublished paper.
22.  Zuchowicz,  I., Wyniki  badania  halasow w nowych budynkach szkolnych /Results of
    noise investigation in new school buildings/. Biuletyn Informacji Naukowo-Technicznej
    I.T.B. 19,/1965/.
23.  Tomaszewski, L.,  Protection des habitations  centre  le bruit de la rue an moyen
    d'ecrans.
    Revne Generate de Routes et des Aerodromes 439,11969/.
24.  Tomaszewski, L., Prevision  et etudes de developpement du traffic routier en Pologne.
    Rev. Gener.de Routes et des Aerodromes 435, /1968/.
                                        741

-------
               A NEW FIELD SURVEY-LABORATORY METHODOLOGY
                    FOR STUDYING HUMAN RESPONSE TO NOISE

                                    Paul N. Borsky
                                  H. Skipton Leonard
                                 Columbia University
                                School of Public Health

                                     SUMMARY

     1. The noise from the Boeing 727 airplane with engine treatment is judged significantly
less annoying than the standard  untreated 727  in  landing approaches for the three noise
levels found at residential areas 1.1. 2.5 and 3.5 miles from landing touchdown.
     2.  An additional simulated engine treatment landing noise, about double the attenua-
tion of the actual  Boeing modified airplane was also judged significantly less annoying than
the actually retrofitted  plane for  the  noise levels at 1.1 and 2.5 miles from touchdown. For
the more distant area at  3.5  miles, annoyance judgments for the  two types of treatment
were about the  same, but the additional noise reduction at this distance was less than 3
EPNdB.
     3. All three groups of subjects from the different distance areas reported these signifi-
cant reductions in  annoyance for the two types of engine treatments. Since the definition of
the annoyance  unit in the rating scale was left  to  each subject, however, it cannot  be
assumed  that an average numerical difference can be interpreted in terms of a percentage
change in annoyance.
     4. In general,  it was found that a reduction of 6 EPNdB produced in landing operations
by the Boeing retrofit airplane resulted in about a 0.7 reduction in the average annoyance
score, on a scale where  "0" represents no annoyance and "4" means very much annoyance.
     5.  At  the  indoor noise levels  heard  at  1.1  miles from touchdown (95.9  EPNdB
untreated and 89.6  EPNdB for the Boeing retrofit engine), average reported annoyance is
reduced from a  score of 3.58  to 2.95. Of even greater possible importance, however, is the
drop in  the  highest annoyance "4" ratings from  72% of  all subjects for the untreated
airplane to only  34% of all subjects for the retrofit airplane.
     6. At  the somewhat more distant 3.5 mile area, the indoor noise is reduced from 82.3
to 75.0  EPNdB and average  annoyance score drops from 1.55 to only 1.03, with "0"
annoyance reports increasing from 18% of all subjects for the untreated 727 to 40%- for the
retrofit airplane.
     7. These positive findings of reduced annoyance  for the  727 retrofit  package are valid
for the conditions tested  - indoor noise levels interfering with communications activities
engaged in by only moderately fearful residents.  The effects of higher outdoor noise levels
on other  types of residents engaged in  different activities cannot be predicted without actual
study.
    8. The new methodology developed by Columbia  University of an  integrated field-
survey-laboratory study has been  successfully used in  an investigation of the retrofit noise
problem. A representative sample  of previously interviewed residents, classified according to

                                        743

-------
selected  psychological characteristics,  participated  in a realistic .controlled  laboratory
experiment. Their generally relaxed behavior, observed through a one-way mirror, and their
voluntary comments in debriefing sessions indicated that they felt they were hearing real
airplanes as experienced  in  their homes. Many subjects in the discussions spontaneously
compared  their own usual  home noise reactions to those  reported in the laboratory.
Another technical accomplishment was  the development of the experimental sound tapes
from engineering data. This capability will enable testing human responses to fly-overs of
proposed airplanes  that exist only on engineers' drawing boards. It also demonstrates the
ability to  test  for meaningful annoyance responses to the great variety of variables that
describe  the real noise environment.

            A NEW FIELD SURVEY-LABORATORY METHODOLOGY FOR
                     STUDYING HUMAN RESPONSE TO NOISE

1.   Introduction

     This is the first major  study using a new methodology  developed  at Columbia Uni-
versity to  study human  response to noise.  Our approach differs  from standard psycho-
acoustic  laboratory  procedure* in six major aspects.

          1.   Subjects were  randomly  selected from populations  actually exposed to
              environmental noise.

     Most  psychoacousticians  use students or other volunteers as subjects. While this is a
readily available convenient  procedure, the  representativeness or bias of such a sample is
unknown.  At Columbia, a random sample of almost 1700 residents in the vicinity of JFK
Airport,  who are exposed to known noise environments, were interviewed in their homes as
part of a  regular hour-long  community study. These respondents are  used  as a pool of
eligible laboratory subjects.

         2.   Sub-samples  of Respondents,  with known psychological predispositional
              characteristics were invited to participate in the laboratory  experiment.

     Previous community surveys in  the  United States  (1,2), England  (3), Netherlands (4),
Sweden (5), France (6) and elsewhere, have clearly identified the importance of socio-
psychological variables in explaining variance in  annoyance responses. Figure  1,  which
summarizes recent British and American  surveys, clearly indicates how feelings of fear and
misfeasance together with noise  exposure levels, differentiate annoyance responses. From
our survey  questionnaires, a subsample of over 500 residents were classified as moderately
fearful and  invited to participate in this study.

         3.   The laboratory is a realistic replication of a typical middle class living room.

     Most laboratories are small  acoustic chambers with bare walls and floors made of
acoustic absorbent materials. As Figure 2 shows, our laboratory looks like and provides the
atmosphere of a real living room.

                                       744

-------
          4.    /'//<• /7\ orrr.v /mixed by llic subjects sound like real airplanes moving across
               the room from left to right

     Most previous studies have used  ;i  monophonic  or stereo speaker system, where the
sound  starts  from a  fixed  direction, peaks and recedes hack into the same direction  Our
laboratory uses a quadraphonic system  which  provides the illusion of directionality and
movement overhead.
                                                                           --

                     •
                   *
                      NO Ff cr - Ho y  •     :'.ce
                             . ' - "Ci. Mi«T Cf ? C^CC  •


                                   .: .  .          ft

                                   20  ','  A -. C 11 C C T '
                                                '
                                                                    _1	I
                 <9C
• 9   1OO- 10-J  i O'- - 109    HO
          CNS

             COIumt a Urnvi
                                                                - NOSC fiececrcn U-.
                                                                           Oe.c  I97i
           Figure 1. Reported high annoyance with aircraft noise by CNR-FEAR8 misfeasance

                                           745

-------


                                                 ••
                          Figure 2. Picture of interior of laboratory

          5.   The sound tapes used in this experiment were generated electronically from
              engineering data, rather than simple field recordings.

          6.   Subjects were engaged in a real task of watching a color TV program, so that
              the airplane flyovers were actually unwanted

     Most past studies ask subjects to judge different noises, where the primary task is the
listening to sounds and making judgements. Since noise, by definition, is unwanted sound.
these sounds in  past  studies wen.1 not  unwanted, but  the focal point of the studs  In our
project,  the  sounds interfered  with  the  desired task of watching and listening to a TV
program. Consequently, real annoyance  responses were  possible and could be recorded

II.   A Comparative Study of  Annoyance Judgements of Three 121 Airplanes in Landing
     Approaches - One a  Standard Aircraft and Two with Acoustical!) Treated Nacelles

     Since the issue of whether or not to  require older aircraft to be retrofitted is so timely
and important, it was decided to use this question as the basis for the first substantive study
at the new laboratory.
                                         746

-------
     A.  Experimental Design

         1.   Acoustic Characteristics to be Tested

     Since  prior experience indicates that the maximum duration of a laboratory session
should normally not exceed  \l/2 - 2 hours, the number of physical variables that could be
included in this experiment was limited to the following:

              a.    Type of aircraft - 727 (JT8D engine)

     While  the 707 and DC-8 are larger and  noisier aircraft, there is general agreement that
most of them will probably be phased out of the active  fleets in the next 5-10 years. The
727, however, is expected to continue to be an  important short and intermediate range
aircraft well into the '80s, and thus was selected for this initial test.

              b.    Type of operation - landing

     Boeing Aircraft Company has developed and certified a retrofit package for the 727
that produced a measured noise reduction of about 6  EPNdB in landing noise at 1.1 miles
from touchdown.  The measured reduction in  take-off noise  levels was much less,  so it
appeared logical to test first the meaningfulness of landing noise reduction.

              c.    Number of noise levels tested - three

     The following three noise levels were tested: The levels correspond to noise produced
at the following altitudes along the glide slope: level A  at 370', level B at 750' and level C at
about 1000'.  These altitudes correspond to the following lateral distances from touchdown:
1.1 miles, 2.5 miles and about 3.5 miles from touchdown.

              d.    Number of retrofit treatments tested  - two

     The untreated 727 landing noise was compared to the actual Boeing measured reduc-
tion of about EPNdB and a theoretical noise with about a 12 EPNdB reduction. These three
noise groups will be referred to as:

                         U - Untreated
                       Tl - Low goal attenuation - 6
                       T2 - High goal attenuation - 12

              e.    Rate of operations - 20 per hour

     A flyover was programmed, on  the  average, every  three minutes, which corresponds
approximately to the average daytime rate of operations at JFK Airport.
                                        747

-------
              f.    Time of day - afternoon or early evening

     It was decided that during this time period TV viewing normally occurs.

              g.    Location of subject - inside a living room - windows open

     The outside  noise spectra and levels were adjusted in accordance with suggested SAE
values for northern climate, inside room, open window conditions (7).

              h.    Ambient noise level in room - 60 dBA

     The average  ambient noise level was about 60 dBA and was provided principally by a
color TV program which the subjects watched.

         2.     Experimen tal En vironmen t

              a.    Acoustic environment

     All tests were conducted in a triple-wall sound-proof IAC chamber (Model 400-A),
 18'X14', with an 8* ceiling, furnished as a typical living room  in a middle class house. The
drawing in  Figure 3 shows a schematic of the interior of the room and its furnishings, with
the location  of a couch  comfortably seating three persons, a low cocktail table and two
chairs facing a 23" color Setchell-Carlson (Model S EC 904)  television set, and  simulated
windows in two of the walls. Four Klipschorn loudspeakers were located in the corners of
the room, and a one-way mirror in the wall alongside the television set permitted observation of
the subjects  from the  control room located adjacent to  the  acoustic chamber.  The floor
was covered by a rug, and all interior surfaces had pictures and drapes of the types used in
the average home, so that the interior appearances and sound conditions were as realistic as
possible. Figure 3 presents a schematic drawing of the room.
     The aircraft  sounds  in the chamber were produced by the four Klipschorn corner-horn
speakers to provide an accurate replication of a fly-over as heard under actual conditions in
an average  home. The airplane was heard flying directly over the room from left to right, at
the sound pressure levels which are heard in a typical northeastern United States house with
the windows  open. Our previous studies have shown that the use of the four-speaker system
gives a true sensation of overhead flight in the direction of the phasing of the speakers. They
have also shown  that listeners inside a  room judge a direction of motion of the outside
aircraft  and,  therefore, the sense of directionality must be provided to fulfill the subject's
expectations (8).

              b.    Sound reproduction system

     The aircraft  flyovers were reproduced by the following sound system. The recording of
the flight was played back by a Crown model 800 tape recorder. The left and right channels
were connected to two calibrated variable attenuators (Daven  T-730G) which were used to

                                        748

-------
                      0 U C
     CHAIR
                                FALSE WINDOWS
                                (LIGHT BOX )
                     TABLE
                                            TABLE
                                              *•—«
                                              AM
               LABORATORY
               LIVING  ROOM
~\ /
                              INTERCOMM
                               ONEWAY MIRROR
                Figure 3. Laboratory living room

-------
obtain accurate repeatable settings of the reproduced sound pressure level in the chamber.
The electrical signals through the attenuators were amplified by two Crown Model DC 300
power amplifiers having an output power rating of 150 watts per channel, which powered
the four loudspeakers.
     The system is capable of producing a sound pressure level of over  120 dB in the
chamber. The lowest ambient noise level in the chamber is 14  dBA, and therefore, the
available dynamic range is 105 dB. When the subjects were in the room, with the heating or
airconditioning system in  operation, the ambient noise level averaged about 30 dBA. The
sound of the television set was adjusted to a mean level of 60 dBA during the tests.
     Sound pressure levels of the flyovers in the chamber were calibrated prior to each
session with a B&K model 2204 Sound Level Meter.  Rudmose ARJ-6 audiometers were used
for testing the subjects' hearing.

              c.    TV programs watched

     A  comparison of national Nielsen ratings indicated that "All in the Family" was one of
the most popular half hour TV programs and that "Ironsides" was one of the most fre-
quently-watched hour-long programs. A small telephone survey  of Long Island residents
confirmed  these national ratings* so it was decided to video tape these two programs for use
in the experiment.

              d.   Order of fly o vers presented

     Subjects judged three noise levels - A,  B and C  and three comparison flights at each
level  - U, Tl and  T2. To counterbalance completely  these nine types of flights was not
feasible, but 36 random order combinations did succeed in eliminating possible order effects.
Since there were 36 different orders of stimulus  presentation, it was necessary to have a
minimum of 36 subjects from each of the three distance areas being tested, or a total of 108
subjects in all.

              e.    Subjects to be Tested

     A  group of 108 subjects  were  selected from a pool of 1651 persons previously inter-
viewed  by the Columbia University  Noise Research Unit in March and August  1972. These
respondents resided in  13  sample survey areas which were selected so as to include persons
living about I.I, 2.5 and 5.2 miles away  from various runways at JFK International Airport
and located directly under primary landing  and take-off flight paths as designated by the
FAA. A highly concentrated random sampling procedure was employed which maximized
the uniformity of aircraft  noise exposure within sampling areas and between sampling areas
of comparable distance from JFK runways. Respondents for the surveys were required to be
permanent  residents of an assigned  block and at least  18 years old. In addition, only one
respondent  from each household  was interviewed.  No  domestics  or hired household
employees  were interviewed,  nor  were persons  with a poor  command  of  the English
language.

                                        750

-------
     The interviews averaged about an hour in length and proceeded from general questions
about likes and dislikes about neighborhood environments to more specific perceptions and
reactions  to general noise and  finally to aircraft noise exposures. Since previous survey
research had  clearly demonstrated  that annoyance was related  to psychological  and atti-
tudinal  variables  as well as to the noise  stimulus, it was decided to select a moderately
predisposed group of residents as the most average group for this first experiment, and test
the extremely favorable or unfavorable groups in other experiments.
     Each  survey  respondent was  classified as to the extent to which he or she feared
aircraft  operations around his or her home and  the extent to which he or she  believed
various manufacturing, airport and  community organizations to  be misfeasant with respect
to controlling aircraft noise. Only 531 respondents were classified as moderately fearful and
eligible to participate in the present study. No attempt was made to select a subsample of
respondents with  respect to  the misfeasance variable. It was decided to use a statistical
co-variance analysis for this variable.

          3.   Procedures Used

     Respondents classified  as moderately fearful were telephoned by  a member of the
Noise Research Unit Staff and invited to the research facility in the following manner:
     "Hello:  I  am	, a supervisor from  Columbia University  Research
Center.  May I speak to (the person who was interviewed earlier)! I want  to thank you for
helping  us in  our study of community problems  by answering all of our questions on the
interview. As you probably know, we found that aircraft noise is one of the major concerns
in your area. For  this reason, city planners, airplane manufacturers and interested commu-
nity and environmental groups  have asked us to conduct an intensive study into aircraft
noise specifically.
     "While we know that almost everyone wants less noise, we don't know how much
aircraft  noise must be  reduced in order to be acceptable to the public. Columbia University
has  constructed  a special  research center, nearby, in Franklin Square,  to which we are
inviting citizens,  like yourself, to help in this vital, and we hope interesting research. Our
participants will relax in a living room, watching popular TV shows while different types of
aircraft  fly over. The  participants  are simply asked to judge the annoying qualities of the
various aircraft.
     "You will receive $6 as a small token of thanks for your cooperation and the study will
take from  l'/2 to 2 hours. We will also provide door-to-door transportation and refreshments.
We have  a number of alternative times  and  dates  for our study and  would appreciate
knowing when it would be best for you to come. First, could you come	?"
     Three subjects were scheduled for each session. One subject lived in one of the sample
areas 1.1  miles from a JFK runway, and the other two subjects lived 2.5 and 5.2 miles from
a JFK runway. Thus, all three types of subjects received each order of stimulus presentation.
Upon arrival at the research facility, the  three subjects were escorted into the living room
and asked to sit on the couch in  a specified location.
     In the event that a "subject failed to keep his appointment or it was not possible to
schedule  three subjects at the same time, a staff member who  was not known to the real

                                          751

-------
subjects substituted for the  absent subject, so that three persons were always present for
each session. Actually 18 additional repeat sessions had to be scheduled with  real subjects
for the stimulus sequences that had used substitute subjects. The subjects were then given
the following instructions:
     "Please go into the living room and be seated  over here (indicate position). As you
know, Columbia University has an extensive environmental research program, of which our
group  is a part. We are interested in learning more about how people respond to different
noises, especially those from airplane flyovers.
     "We are going to have a TV show for you to watch and we hope you enjoy it. From time to
time you will hear airplanes  flying over here; some may  appear louder; other quieter.
Occasionally you will hear a voice from this speaker (point to front over TV), asking you to
record your responses to the airplanes which you have just heard here.
     "This is your reaction sheet. (See Figure 4). In the first column, I would like you to
indicate the extent to which the aircraft flyovers you hear here interfere with your watching
and listening to the TV  program. In the second column, I would like you to indicate the
extent to which they bothered  or annoyed you.
     "There is no right or wrong answer — We just want to know how you feel. You will
notice on the right liand side of the sheet, a  thermometer with numbers from 0 to 4. 0
means that the airplanes did not interfere at all or that you were not annoyed at all. 4 means
that the interference or annoyance was very much. Any number in between would indicate
that your feelings were something greater than 0 but less than the top category of 4.
     "Please also notice that there are 9 lines. There will be 9 different times when a voice
will ask you to record your responses. You will not be required to do this after each aircraft
flyover, but only when you hear a voice from the speaker. After each time you hear the
voice asking you for your response, you will enter two numbers on each line; one to indicate
how you feel about the amount of interference and the other to express the extent of your
annoyance with the aircraft which  you heard  since the previous time you recorded your
responses.
     "I would like  you to remain seated until the end of the first session, which will be
about  30" minutes. Then, we  will have a brief  coffee-break. In  all, there will be three
30-minute sessions. If at any  time during the session you want to talk to one  of us for
example; if  the TV picture or sound goes off, you can do so by pressing the button on top
of the TV speaker and then you will be able to talk.
     "Please try to record your own personal feelings about the airplanes flying here. Try
not to influence,each  other by avoiding any discussion or indication of how you, yourself,
feel about  them. Of course, if you want to talk about the TV program, as you  would at
home, feel free to do so. OK?"
     At this point the TV monitor was activated and the interior and exterior chamber
doors were closed by the departing experimenter.
     The first segment  of the session consisted of  a  27-minute video-taped "All in the
Family" program which had previously been rated as one of the most interesting and most
watched TV programs.  Coincident  with  activation of  the TV monitor, a Crown 800
quadraphonic tape deck was engaged which produced simulated  aircraft flyovers  with a
mean inter-flight interval of about three minutes. Nine such simulated flyovers occurred in

                                        752

-------
DATE:
NAME:
ADDRESS :
(Street)
INTERFERENCE
1
2
3
4
5
6
7
8
9 	 	
(Town)
ANNOYANCE








\
It
3
2
1
f
ZERO
VERY MUCH
/

V NOT
f
AT_ALL.
V or
1 NONE
FOR OFFICE USE



No.
Condition
                  Figure 4. Survey 101 Columbia University Sept. 6,1972
                                        753

-------
the living room during this segment of the session. After the third, sixth and ninth flyovers
the subjects were  requested, via  a separate voice channel, to make judgements as to the
annoying and  interfering quality of the flyovers since the previous request for judgments. In
a previous methodological study (8), it  was found  that  annoyance judgments seem to
stabilize after presentation of three stimuli.
     At the end of the "All in the Family" program, the experimenter re-entered the living
room and asked if the subjects wished to  stretch, use the bathroom or would like some tea
or coffee.
     In the second and third sessions, nine flyovers were also presented. The TV program
for these sessions consisted of an "Ironsides" series episode.
     At the end of the third session, the experimenter re-entered the living room along with
an audio-technician and audiome.try records were obtained via two Rudmose ARJ-6 Clinical
Bekesy audiometers. Since only two subjects could be tested at a time, the third subject was
asked to wait in the reception room until the first two subjects had been tested.
     The subjects  were then thanked  and  debriefed, given $6 for participating in the study
and driven home if they had been provided with transportation to the facility.

          4.   Summary of Analytical Design

     Three principal hypotheses were investigated -

              a.   Each  retrofit treatment (T.I and T2) would be judged significantly less
annoying than the  standard untreated (U) 727 landing.
              b.   Each  retrofit treatment would be judged less annoying  than the un-
treated 727 at each of the three levels of noise tested (A, B & C).
              c.   The  type of subject's normal noise environment (residence) would be
related to annoyance judgments. More specifically, it was expected that mean  annoyance
ratings, in general, would have the rank order from greatest to least for 5.2 mile, 2.5 mile
and 1.1 mile distance subjects.
     These predictions were  based  on the concept that each person has a  "comparison
level" (9) based upon previous experience against which he judges new experiences.
     For instance, 5.2-mile-distant subjects should perceive simulated flyovers in the A tape
series to be more annoying than would subjects living 2.5 or 1.1 miles from JFK since these
flyovers, in general, are relatively louder in relation to their normal experience than for the
other residential groups.  By the same token, C series tapes should  be less annoying for 1.1
mile subjects than  for the 2.5 or 5.2  mile subjects, since they are relatively quieter than the
actual exposure levels for the other two groups of subjects.

     B.   Findings

          1.   Representativeness of respondents in field survey

     All  interviewers  were given  predesignated addresses in thirteen primary sample areas,
each consisting of  small clusters of adjacent blocks. In  some assignments where the number

                                         754

-------
of dwellings in a sample  area was limited, every household was contacted. In other areas,
every n'th dwelling was randomly selected. Over 83% of all assignments were interviewed,
5% were not contacted, and only 12% refused an interview. In general, this completion rate
compares very  favorably  with  similar surveys in major metropolitan areas, and the 1651
respondents can be considered  fully representative of the populations in the areas surveyed.
     As indicated in the description of the experimental design, 531 respondents, or about a
third of the total survey sample, were  classified as  expressing moderate fear of airplane
operations and, thus,  became eligible for the laboratory study. Since this is one of the first
attempts to use a representative population sample in a major psychophysical laboratory
study, the outcome of the invitations to participate  in the laboratory is of some interest.
About one-third of all persons who were contacted actually participated in  the laboratory
tests. An almost equal number were judged not physically able to cooperate within the time
limits set for the study. These respondents indicated that their work or home responsibilities
(infants, multiple jobs,  etc.)  made it  very difficult for them to meet our laboratory
schedules.  Some of  these  persons might have  been convinced to cooperate if the lab
schedules were changed  or adult baby sitters were provided. The  other major reason for
non-availability was poor health reported mostly by the  elderly and few of these could be
expected  to travel to the laboratory. Only 77% of those invited  were considered "hard
refusals", while the remaining  15% were busy at the  time of our initial contacts and were
not called back because the required number of subjects had been obtained.
     The question arises about the representativeness of the subjects who were tested in the
laboratory, since they constituted only one-third of those invited to participate. Most lab-
oratory studies cannot evaluate the representativeness of their subjects, since they  rely on
readily  available volunteers. A comparison of selected responses obtained from the initial
field survey enables such  an evaluation. These data indicate that the laboratory sub-sample
was generally representative of the full sample in all  aspects considered most significant to
this study.

          2.   Description of Airplane Flyovers

     The aircraft  flyovers which were  reproduced in the test chamber were those of a
standard untreated 727, a low-goal treatment 727, and a high-goal treatment 727 landing, at
distances of 1.1 miles, 2.5 miles and 3.5 miles from touchdown.
     The test tapes were  based on actual Columbia University field recordings of standard
727 flights at these distances, with modifications for the low-goal treated engine according
to  information  provided by  Boeing. The  high-goal  treatment assumes the same spectral
changes as the  low-goal, with more attenuation.  Since actual recordings of the low-goal and
high-goal treated  aircraft were unavailable, it  was necessary to introduce the measured
spectral and time history effects of these treatments by electronically modifying the record-
ings of the standard aircraft.
     Since the  Boeing data were for outdoor sound levels, the modification of the Columbia
test tapes  to provide  for  the various engine treatments had, therefore, outdoor sound levels.
The  final test  tapes, however, incorporated  outdoor-indoor sound  pressure level  and
frequency response corrections (18 dBA at 1000 Hz) as given by SAE recommendations for
cold-climate houses with windows open (7).
                                          755

-------
   60-



    75-



£  70-


u
    «-
u   oo
c:
c.

o   50-
r
a
O
(O   45-
    40-
          J
         50  !   80  !  125  I  200
315    500
                     4	^'TREATMENT  I
                     o	CTREATMENT2
600  !  I25K 1   2K    3 J5K    5K
                                     \
 I
8K
                    tOO    160    250    400   630     I K
                          i6K   2.5K   4K     63K    JOK
                   Figure 5. Indoor noiu spectra for 727 landings at 1.1 miles from touchdown

-------
     Figure 5 presents the actual indoor Columbia University noise spectra for the one-mile
distance noise levels used in the experiment. The untreated 727 noise is compared with the
low  goal Tl  and  high goal T2 noises. As  previously noted, the high goal spectrum was
assumed to be similar to the low goal Tl engine treatment with additional attenuation. This
assumption was necessary since the high goal test engine had not yet been completed by
Boeing at the time of this experiment.
     Table 1  presents some selected acoustic summary measures of the flyovers actually
heard indoors and judged by the subjects.

                                       TABLE 1

           INDOOR NOISE LEVELS OF FLYOVERS PRESENTED TO THE SUBJECTS

                                            dBA                 EPNL
                                       Number    Changes      Number    Changes

            Level A (I.I miles)
            Untreated (U)                 80                  95.9
            Low Goal (Tl)                 73       -7         89.6       -6.3
            High Goal (T2)                 68       -5         84.1       -5.5

            Level B (2.5 miles)
            Untreated (U)                 72                  88.2
            Low Goal (Tl)                 65       -7         81.9       -6.3
            High Goal (T2)                 59       -6         74.8       -7.1

            Level C (3.5 miles)
            Untreated (U)                 66                  82.3
            Low Goal (Tl)                 60       -6         75.0      -7.3
            High Goal (T2)                57       -3         72.2      -2.8

     Each test  tape consisted of a set of 9 flyovers at one specific distance for the three
 types  of  aircraft.  Each type of aircraft  flight is repeated three times at approximately
 three-minute intervals.

               3.    Judgtnents of ''Annoyance"

                    a.    Summary of effects

     The main  analytical scheme for evaluating reported annoyance and interference was an
 Analysis of Covariance. All subjects judged the same 27  flyovers, which consisted of com-
 binations of three noise levels (A, B and C), and three types of engine treatment (untreated,
 treatment 1  and  treatment 2).  In this type  of repeated  measures  design,  attitudes of
 misfeasance  could have a possible effect  only upon subject residence differences,  since as
 noted, the same subjects judge all noise levels and treatments. Table 2 presents a summary of
 the covariance analysis.

                                          757

-------
                                         TABLE 2

                   SUMMARY OF COVARIANCE ANALYSIS OF ANNOYANCE

       Sources of Variation
                             Sums of     Degrees of       Mean
                             Squares       Freedom       Square         F Value

       TOTAL                1841.59        971
       Between Subjects         542.70        107

       Subject residence (A)       35.00          2          17.50         3.62 p< .05
       Error (A)                507.70        105           4.83

       Subject residence
       Adjusted for Misfeasance    30,05          2          15.03         3.14p<.05

       Adjusted error (A)        501.46        105           4.7S

       Within subjects           1298.89        864

       Level of Noise (B)         529.64          2         264.82       257.11p<.0l
       Subjects X level            77.11           4          19.28        18.72p<.01
       Error (B)                216.81         210           1.03

       Treat ments (C)           211.57          2         105.79       179.31 p<.01
       Subjects X Treatment         .94          4            .24          .41  n.s.
       Error (C)                123.71         2iO            .59
       Level X Treatment         14.14          4           3.54         7.87 p <  .01
       Subj. X level X Treatment     7.8           8            .98         2.18 p <  .05
       Error (D)                187.14        420            .45


     As can be seen, annoyance judgements for different levels of noise and engine treat-
ments were very significantly different. The analysis indicates that the differences reported
could  have occurred  by chance in less than one case out of 100. (p < .OH The differences in
judgements attributed  to  the residence  types were also statistically  significant and could
have occurred by chance in less than five cases out of 100 (p <  .05).
     The  effect  of  misfeasance  on  between-subject differences was of relatively minor
importance.  The following interactions of the  main variables were also significantly related
to annoyance judgements:  a. subjects and level of noise; b. level of noise and engine treat-
ments; c. subjects, levels of noise and engine treatments.
     The interaction  of subject differences  and engine treatments, however, was not signifi-
cant.  Likewise, unreported analyses indicated that the varied order of presenting the levels
of noise and engine  treatments succeeded in eliminating any significant order of presenta-
tion effects,  in summary, the main and interaction effects, combined, explained about 44'?
of all the reported variations in annoyance responses.

                                           758

-------
              b.   Effects of noise level ami engine treatment
         e A Kraphiejlly presents the  different mean annoyance ratines In subjects for
varying noise levels and engine treatments. It should be noted that subjects were free to rate
annoyance  from "0"  meaning "not  at all" to "4". defined  as "very much". It is quite
evident that there were stable differences  in annoyance between untreated and treatments
for each level of noise. It can also be noted that there is a consistent reduction in annoyance
with lower  level of noise. Hypothesis  I  and 2 have been confirmed by these results. As can
be seen, the differences in annoyance between treatments at level (' are smaller than at the
other noise levels. This pattern is reflected in the significant interaction of noise  level  and
treatments  reported in Table 2.  (F-3.54. df-4.  420.  p  <  .01)  In fact, a "t" test of the
difference between the means of annoyance for Tl and  T2  treatments at level ('indicated
no significant difference.
      4 0
   (0
   o
   *  3 0
   l-
   BC
   o
   z
   <
       I 0
                                                              r-
                                                               UNTREATED
TREATMENT  l_
TREATMENT  2
                \                     B                     C

                            ENGINE  NOISE  LEVEL

                 Figure 6. Mean annoyance for engine noise levels and treatments
                                        759

-------
     This is not an unexpected finding, if one considers the EPNL levels. The actual EPNL
 reductions between treated and untreated A & B level noises are about 6-7 EPNdB, while
 the EPNdB difference between  treatment 1 & 2 at level C  is only  2.8, a much smaller
 reduction. Furthermore, the absolute level of these noises was close to the TV sound level
 and represented minimum C group masking.
     Table 3 presents the  mean annoyance values for each level of noise and type of
 treatment as well as the frequency distribution of annoyance judgements. As can be seen,
 when annoyance judgements for untreated 727s are compared to treatments 1 & 2 noises,
 the drop in higher annoyance (4 & 3 ratings) is quite sharp in noise levels A & B. Corre-
 spondingly, the number of no annoyance answers increases in these comparisons.
                                     TABLE 3

          ANNOYANCE RESPONSES BY LEVEL OF NOISE AND ENGINE TREATMENT

                                                   Annoyance  Scores
          Level         Engine
         Of Noise       Treatment       Mean       43210

           A           U          3.58
                        Tl          2.95
                        T2          2.23

           B           U          2.56
                        Tl          1.74
                        T2          1.23

           C           U          1.55
                        Tl          1.03
                        T2           .80
72%
34
16
18%
36
29
6%
22
29
3%
7
16
\%
1
10
23
8
3
32
19
15
27
32
19
12
22
29
6
19
34
2
1
0
12
9
8
42
21
15
26
29
25
18
40
52
              c.    The effects of subject differences on annoyance

     Figure  7 presents the relationships between average annoyance ratings by different
subject groups for the three noise levels. As can be seen, while each subject group rates noise
level A>level B >level C, the highest average annoyance is reported by the 2.5 mile group
(X=2.5) which is only a little higher  than average annoyance for the 5.2 mile residents
(X=2.03). The closest 1.1  mile group  reported an  average annoyance of only  1.70. These
findings partially confirm our third hypothesis. The pattern of results, however, does not
correspond entirely to our predictions. While the mean annoyance for subjects at 2.5 miles
was greater than that for subjects at 1.1 miles,  the mean for the 5.2 mile group was not
greater than the 2.5 mile means.

                                      760

-------
             d.   Relationships hetween reported annoyance and i'.T\'dH noi\e level

     Figure 8 presents a summary of the averse annoyance judgments for the nine aircraft
flyovers expressed in EPNdB levels. The same  noise  level and treatment differences may be
noted,  but since the acoustic  stimulus is now expressed in common EPNL units, a more
general relationship may be observed. A  least squares regression line has  been plotted in
Hgure  6 for all  1 OS subject judgments for the nine noise stimuli. The corresponding correla-
tion  coefficient  uvtv .62, significant at the  p  < .005  level. The  correlation coefficient be-
tween EPNL and only  the nine mean annoyance values was .971. From  the plotted regres-
sion  line, it appears that below 75 EPNdB. reported annoyance is/ess than  I O. and that an
increase of JO /:7'.\V/# result* in an average increase of / 1 7 in rated annoyance. It should be
•» \J
to
0
z
l_ I A.
EAN ANNOYANCE RA
fo <
> 0 <
2 '
O-










~v
x^
XN
^
"^






•x ^"^i
X.
**s
*X






\.
X S
X
*. V
^v^







^s
s^
•x
\^ -
*s







«,J^







>^












: T
-------
emphasized that these  arc  the reported annoyance relationships found in this particular
experiment and should not be assumed to be valid for other types of aircraft in other imxlcs
of  operation. Additional experiments will be  needed to arrive at possibly more general
relationships.

     C.   Overall Strategy for Further Research

     The general objectives of our laboratory- program arc to disentangle the complex inter-
actions of variations in noise environments and summated community annoyance responses
By obtaining greater control over both the physical and  psychological variables  with our
new field-laboratory  methodology, it  is hoped that  empirical data can be developed to
substitute  for the "best  judgements" that  now constitute the weights used in composite
noise indexes. This information is needed to provide the criteria  required for development
of standards for noise regulations
   40
-
-'
(T
..
u
r
<
   30
20
   1.0







-1C —






*J






°.
X






\^r






/
~







1






^







X






ir


LEAS!



J3J

LEGEND'
X - LEVEL V NOIGE '
O-i f vr i 'n" uoiRF
A -LEVEL "C" NOISE
U -UHTRC ATfD -72 7
I-TREATMTNT i
2 -TREA 1 ME NT 2
\ 60UARLS R{:GR£SSIO^
V onX
Y- 117-771

           700
60£>
NOISE LEVEL IN
                                                       90.0
IOO O
                Figure 8. Indoor noise level in relation to mean annoyance ratings
                                        762

-------
     It is our analytical strategy to accomplish this general objective in the following three
stages:
          1.   Establish  the  relationships  between  annoyance  and acceptability  and
different aircraft (spectra and level differences) in landing and take-off operations. This will
establish the range  of variability and meaningfulness of  annoyance and  acceptability
responses.
          2.   Establish  the  relationships  between  annoyance  and acceptability  and
different combinations (operation mixes) of aircraft in landing and take-off operations. This
will  provide an  empirical basis for combining  different  physical  stimuli into  meaningful
annoyance units.
          3.   Establish the relationships  between annoyance and acceptability and varia-
tions in frequency of operations over time of different mixes of aircraft. This last phase will
replicate a realistic complex community exposure to aircraft noise.
     It is hoped  that insights gained at each stage of research will help simplify and combine
variables in  further research projects. For example, a single measurement unit like EPNL
may prove fully descriptive of the many types of aircraft and operations.

                             SELECTED BIBLIOGRAPHY

(1)  Borsky, Paul  N..  "Community Aspects of Aircraft Noise", - National Advisory Com-
     mittee for Aeronautics, 1952
(2)  TRACOR. "Community Reaction to Aircraft Noise", Vol. 1 & 2, TRACOR Document
     T-70-AU-7454-U, Austin, Texas, September 4, 1970
(3)  Office of Population Censuses & Surveys, Social Survey Division  - "Second Survey of
     Aircraft Noise Annoyance around  London Airport", London 1971
(4)  Van  Os, G.J., Report by  Advisory Committee on Aircraft Noise Abatement, Ministry
     of Traffic & Public Works, Delft, Netherlands, 1967
(5)  Rylander, R.. Sorensen, S. and Kajland, A., "Annoyance Reactions from Aircraft Noise
     Exposure". Karolinska  Institute, Stockholm, 1972
(6)  Alexandre,  A.,  "Perception of Annoyance  due to Airport Noise and Suggestions for
     Ways of Reducing it", Anthropologie  Appliquee, 28/70
(7)  Society of Automotive  Engineers,  Proposal AIR 1087
(8)  Borsky. Paul N., "A New Field-Laboratory Methodology for Assessing Human  Re-
     sponse to Noise", Columbia University Report, October 1972
(9)  Thibant. J.W. and Kelley, H.H., "The Social Psychology of Groups", New York, N.Y.
     Wiley 1959
                                        763

-------
              AN INTERDISCIPLINARY STUDY ON THE EFFECTS OF
                            AIRCRAFT NOISE ON MAN

                    B. Rohrmann, R. Schumer, A. Schumer-Kohrs,
                                R. Guski, H.-O. Finke
                              University of Mannheim
                                Mannheim, Germany

1.  INTRODUCTION

     1.1  Steadily increasing air traffic exposes more and more individuals to noise;expan-
sion  of towns intensifies this  problem. Thus the Deutsche Forschungsgemeinschaft (DFG)
initiated  an interdisciplinary project enabling  scientific  research on  the effects of aircraft
noise on man  to be conducted.
     1.2  The following primary questions were to be clarified:
     — Which sociological, psychological and physiological effects of aircraft noise are
ascertainable? Under what conditions do they occur?
     — In what  way are reactions to aircraft noise determined by influences of the social
environment or by psychic and somatic attributes of the individual concerned?
     - To what extent does the acoustical characterization of noise exposure covary with
the ascertained effects of noise on man? (See Fig. 1).
     1.3  The team of the project was composed of 6 sections:
     - "acoustical section"                  (Dipl.-Ing. H.-O. Finke,
                                            Dir. u. Prof. Dr. R. Martin)
     - "medical section"                    (Prof. Dr. A.W. v. Eiff, Prof. Dr.
                                            L. Horbach, PD Dr. H. Jorgens)
     - "organizational section"               (Dipl.-Psych. B. Rohrmann)
     - "psychological section"               (Dipl.-Psych. R. Guski,
                                            Prof. Dr. H. Hermann)
     - "social-scientific section"              (Prof. Dr. M. Irle, Dr. R. Schumer,
                                            Dipl.-Psych. A. Schumer-Kohrs)
     - "work-physiological section"           (Prof. Dr. Dr. G. Jansen)

2.  PLANNING AND DESIGN

     2.1  It was  agreed to survey  and test highly populated communities (i.e. large cities)
situated  in close proximity to  a large international airport. After considerable method-
ological preparation and a jointly  conducted preliminary study (Hamburg 1966), the main
study was performed in Munchen (1969); each section executed its investigations on the
same respondents (at the same place and during the same period).
     2.2  In the  preliminary study, a sample of noise-exposed respondents was contrasted
with a control group not  exposed to aircraft noise (contrast group design). For the main
study, that area  in which aircraft noise dominated over all other noise sources was selected.
The  area defined  (after previous acoustical measurements) by a noise  curve around the
airport Munchen-Riem covers about 30 km2 and contains more than 100,000 inhabitants

                                       765

-------
      AIRCRAFT NOISE: SOME  POSSIBLE INTERDEPENDENCES,
                 REACTION
      STIMULUS
      number,
      level,
      duration,
      timing
      of flights
   PERSONALITY
 psychological and
physiological traits
                   ENVIRONMENT
                  acoustical and
               sociological situation
                         INTERVENING" FACTORS
                 psychic,
                 social and
                 somatic,
                 perceptions
                 and effects
                        Figure 1. Some possible interdependenties


(1969). Near the airport, the number of daily flyovers was about 80, in the outer parts
about 20; the mean A-weighted sound levels ranged from 75 to 107, Aircraft noise exposure
in this  area was divided into 32 levels; each level was equally considered in the sample
(quasi-continuous approach).
     2.3 In  order to provide a  close  association between respondents and  noise data, a
group of respondents, clustered at one place (i.e. respondents from houses close together)
was selected for each of the 32 defined noise levels. In each of these sample clusters one
locality for sound measurements  was designated. The allocation of the 32 clusters complied
with two (hierarchical) principles: firstly, to represent the aircraft noise levels equally (one
cluster per noise level); secondly,  to represent the demographic structure (stratified sample).
Drawing of  clusters and persons was randomized. The survey was based on preselected
addresses  (about 30  per cluster, 952 altogether). Figure 2 shows the resulting area and
clusters.
     2.4  The general, interdisciplinary investigated,  sample was drawn  randomly from
inhabitants ranging from  21 to 60 years of age. But the interviews were extended to persons
of 15 to 70 years, as well as to former inhabitants of the 32 clusters who had moved within
Munchen or  left it during the last 12 months before the study (additional samples 'youths'
and 'old people', 'migrators within Munchen', 'migrators out of  Munchen'). Part of the
respondents took part in a re test.
                                      766

-------
                           JUKVFY  AREA  un.1  CLUSTER 1.  32  in MUNICH (-  -Ikr
                                  Figure 2. Survey area
     2.5  The data collection program consisted of the following steps:
       Social scientific interview based upon standardized questionnaires (at the respond-
ents homes.  1 hour);
       psychological and physiological experiments and  tests (at  the test station, 2 hours);
       medical case history, examination, and experiments (at the  test station. 2 hours);
       acoustical  measurements (1 control point per cluster).
     2.6  The survey yielded a total of hoO usable social-scientific interviews (general sample
phis '\oulhs' .nul  'old people'): further, 152 interviews with migrators and  1 15 retests: 400
psychological/physiological and 400 medical tests. 357 individuals went through  the entire
program.
          of the  eligible individuals were interviewed. 72 ''r of the individuals invited to the
test station were induced to take part in the entire program.

3.  ACOUSTICAL MEASUREMENT

    3.1  Noise recording was done by portable  measuring instruments  which could record
the noise events automatically  at each measurement point for 24 hours. The measurements

                                        767

-------
were carried out over a period of 7 weeks. Flyovers and background noise were separately
recorded by two tracks of a tape recorder and evaluated thereafter in the lab.
     3.2 From these data, values for the A-weighted flyover and background levels, for the
duration of single overflights and for the number of the flyover events were defined. The
usual rating criteria, as they are proposed in various countries for the judgment of aircraft
noise (Q, NN1, CNR, NEF, R, etc.), were calculated.
     Figure 3 shows the classification of sample clusters according to number and  mean
noise level of flyover events.
     3.3 The mentioned measures are highly correlated with each other (r=.97) and there-
fore can be seen as equivalent  with regard to the data of this study. Correlations between
these measures and variables  of disturbance  and annoyance by aircraft noise show a
tendency for measures involving frequency of flyovers to be more highly correlated with the
annoyance variables than measures invoking the noise level of flyovers.
   100
   95
   90
   85
   BO
                DISTRIBUTION OF SURVEY AREAS
                (CLUSTER  1....32)
   =  Mean of peak  levels
N  -  Number of  overflights
                                                 N
20       30
                              40       50       60
                            Figure 3. Distribution of Clusters

                                      768
                                             70
80

-------
     3.4 After systematic comparison and optimization of the weighting of the different
shares for level and frequency in the calculation formulas, a rating criterion, called aircraft
noise rating criterion FBI ("Huglarmbewertungsmass 1") was derived from the data of the
study:



                 FB1  =10log£lOLAi/10+10logN-50
                                -— A- weighted flyover level
                            N   — number of overflights  per day
                            50  — constant
     3.5 FBI  was developed with the goal of considering only components which exhibit
meaningful relations to the annoyance variables and being simple in composition and also in
measurement necessities. This measure - FB1 - was used for the analyses reported below.

4  SOCIAL SURVEY

     4.1  The  main purpose  of  the  social-scientific part of the study was to clarify the
following questions:
     - To what extent does the stimulus 'aircraft noise' determine reactions to  aircraft
noise (such as dissatisfaction or annoyance)?
     — Which (moderator) variables are apt to explain differing individual reactions under
the same condition of aircraft noise?
     4.2  Topics and variable construction technique  were mainly in the frame of other
aircraft  noise  surveys (e.g., the Tracer studies). The  stability coefficients (computed on
N= 115 retest subjects) indicate a retest-stability of the survey variables that is on the whole
satisfactory.
     4.3  According to  intercorrelations of the variables with the  stimulus, the individual
variables were classified in two groups:
     — Reaction variables, consisting of all those variables significantly related to one of the
stimulus variables.
     — Moderator variables,  consisting  of all those variables  showing  very slight or no
correlation with the stimulus  variables but which correlated with the reaction variables—i.e.,
variables which contributed to the prediction of reactions independently of stimuli.
     The relationships  between  the  variables were analyzed by multivariate procedures
including regression-, factor- and discriminant-analyses. The relative contribution of stimulus
and moderator variables to the prediction of reactions was determined.
     4.4  The analyses show that the greater the aircraft noise the greater
     — the rated loudness of aircraft noise (r=.30 with FBI)
     — the perceived number of aircraft noise events (r=.47)

                                        769

-------
     - disturbances of communication (e.g. disturbances in conversation, in listening to
radio/TV) (r=.56)
     — disturbances of tranquillity and relaxation (r=.39)
     — the sensation of pain (r=.28)
     — the perceived physical consequences (such as walls trembling) (r=.34)
     — the number of subjects spontaneously naming aircraft noise when asked for incon-
veniences (r=.35)
     — the number of subjects spontaneously naming aircraft noise when asked for condi-
tions impairing health and life (r=.34)
     — dissatisfaction with the neighbourhood (especially dissatisfaction with its recreation
value (r=.51)
     — the rated intolerableness of aircraft noise (r=.39)
     — the frequency of taking part in social action (such as participating in protest demon-
strations against aircraft noise) (r=.23)
     - the frequency of taking physical action  (such as installing double windows) (r=. 16).
     These relations are linear; curvilinear determination coefficients lead only to an insig-
nificant increase over linear determination coefficients. Figure 4 shows means and standard
deviations of the variable, "disturbances in communication", for each level of aircraft noise.
     The described relationships are  by no means perfect ones: The  highest correlation
found  amounts to only .58, i.e., only approximately 1/3 of the variability in reactions can
be predicted by means of one stimulus variable alone. Even when correlating more than one
stimulus variable with each reaction variable (multiple correlation) or with more than one
reaction variable (canonical correlation)  a considerable amount of the variability remains
un predicted.
     4.5 The most efficient moderator variables are those referring directly to noise in
general (such as sensitivity to noise or indifference or adaptability to noise) or to air traffic
or aircraft noise (such as the belief that aircraft noise is health impairing, or attributed value
of air traffic, or knowledge about aircraft noise received by mass media), i.e., apart from the
stimulus variables, the above variables contribute most to the prediction of reactions.
     4.6 When analyzing the  relationships between variables in different subgroups  it
appears that they  differ in some respects - e.g.: Analyzing the  relationship  between the
'noise adaptability  factor score' (or 'indifference to noise1) and the 'global reaction' (factor
score extracted from all reaction variables) for various degrees of exposure to aircraft noise,
it appears that the  greater the exposure the closer the relationship between adaptability to
noise and the global reaction.  A simple multiplicative model, or the conception that such
moderators as noise adaptability or noise sensitivity work like an amplifier of the stimulus,
is suitable for describing the data.
     4.7 A survey  of people who had moved from the research area either to other parts of
Munchen or outside of Munchen did not indicate that the sample of the main study would
be biased by selective migration of those subjects who were especially sensitive and/or who
felt particularly affected by aircraft noise.

5 PSYCHOPHYSIOLOGICAL EXPERIMENTS

     5.1 The investigations, which psychologists and work-physiologists have done together
in the  laboratory,  were concerned with information processing behavior, and they were
designed to test  two alternative hypotheses: on  one hand the hypothesis  of  'adaptive
coping' with aircraft noise, which assumes the learning of techniques for disturbance-free

                                        770

-------
    c
   4
   3
   2
        DISTURBANCE  OF  COMMUNICATION
        BY AIRCRAFT  NOISE
                                                                   SD
                                                                   Mean
                                                                   SO
                                                 C=.073[FB1-78.7)+3.4
       60
65
70
75
85
90  FBI
     [811
               (85|
                      (90)
                       (95]  (98) LA
     120)
     (40)
         (60)
                             (80)N/24h
                       Figure 4. Disturbance of communication
information processing in spite of the noise, and a decrement of physiological responses
towards the presentation of noise. On  the other hand stood the hypothesis of 'defensive
blocking', which assumes an interruption of information processing and a physiological state
of defense against noise  as a consequence of frequent and intense day-by-day aircraft noise.
    5.2 The investigations had three aspects:
    — the general activation theory and its possible splitting up into 'orienting* and 'defen-
sive' components,
                                     771

-------
     — the distraction theory, which assumes a damping or disturbance of the information
input in one sense modality in the case of simultaneous stimulation of several modalities,
and
     — a possible change in the connection between aircraft noise stimulus and aircraft noise
reaction by personality characteristics.
     5.3 Besides the manifold  personality tests, recognition, memory, and signal tracking
tasks,  the behaviour of vasomotoric and  muscular activity was continuously recorded in
experimental situations with quiet and noise interchanging.
     5.4 The results, especially  those of the psychophysiological experiments,  do  not
confirm the hypothesis of 'adaptive coping' with aircraft noise: with increasing day-by-day
aircraft noise exposure, the physiological response to the onset of noise  in the laboratory
increases. This response consists of a constriction of the blood vessels at the finger and at
the temple, an increase in the electrical muscle activity, a decrease of the heart rate and an
increase in the tracking error rate. This complex reaction was called "defensive reaction"
following SOKOLOV and the interpretation goes towards a blocking of information recep-
tion processes. The reaction correlates positively both with intensity and  frequency of
aircraft movements  (r=.21), and it occurs especially with persons of low mobility, strong
conservative tendencies and very high blood pressure.
     5.5 It should be  mentioned that hearing ability, measured at 5  tone frequencies,
decreases with increasing aircraft noise exposure as a (statistically insignificant) tendency.
     5.6 Other aspects of human  behavior, such as information processing in complex
stimulus situations,  are not so much affected by aircraft noise as  such, but are affected
indirectly via negative attitudes or annoyance  related to aircraft noise, especially the per-
formance requiring attention in noisy conditions. For instance, the discrimination of optical
signs during concomitant presentation of white noise or spoken numbers  is impaired  by
aircraft noise or by the annoyance caused by aircraft noise.

6.  MEDICAL INVESTIGATIONS

     6.1  In the medical laboratory investigation program, the clinical status was assessed by
means of anamnesis and examination of the body; analyses of blood and  urine, and  experi-
mental tests of vegetative functions were performed in order to check whether the day-by-
day aircraft noise was associated with illness of the blood circulatory system, with diabetes
mellitus and states  of  nervous irritability, or also whether  major functional  changes in
situations of greater load could be seen as first steps of some illness.
     6.2  During the physiological experiments, systolic  and diastolic blood pressure, heart
rate, respiration rate, and electrical muscle activity were recorded for 34 minutes, and the
subjects were submitted to quiet, mental arithmetic, and continuous and discontinuous
white noise.
     6.3  The analyses done with the medical data demonstrate that aircraft noise does not
cause manifest illness, but that it contributes as a tendency to changes in vegetative func-
tions, especially the blood pressure.

7. INTERDISCIPLINARY INTERPRETATIONS

     7.1  In  interdisciplinary statistical analyses, the different data decks were  integrated
(N3357). Some results:
     — The sociological, the psychological, and the physiological variables of aircraft noise
effects show very low intercorrelations (r about .15).
                                         772

-------
     —  Using  an interdisciplinary  set of  sociological, psychological,  and  physiological
moderators, one third of the variability of sociopsychological annoyance is determined (as it
is done by the acoustical stimulus variables).
     Other analyses (also within the sections) concerned special subgroups and cross valida-
tion attempts by  split-half techniques.
     7.2 Some ideas and analyses concerned the mechanism of moderators. In the present
data a  moderator  has  mostly  a "regulating' effect (attenuating or intensifying) on  the
reaction to aircraft noise, but other impact models can be differentiated, as well, such as a
'switch  on' effect  (reaction appears  only in a certain subgroup), or 'switch over' effect
(depending on the level of the moderator different reactions to aircraft noise result) or a
'mediating' effect (an initial reaction acts as the moderator of a secondary reaction).
     7.3 Such chains of reactions are  being described in various path models. One of these
models for the effects of aircraft noise  (using nine variables), which reproduces the empirical
intercorrelation matrix, shows (see Fig. 5) that within the surveyed data set there are two
'direct'  effects of  aircraft  noise, firstly, an intensification of verbalized  annoyance and
impairment by aircraft noise ("R1U") and secondly, (weaker) the  intensification  of the
physiological defense reaction to (laboratory) noise ("DBF").  The annoyance  by aircraft
noise furthermore  causes impacts on  three 'indirect'  effects of aircraft noise, namely fear
associations ("FAF"),  lessening of attention  ("AUF"), and  increase  in  blood pressure
("RRD").
     In  addition, the model shows  the impact of the considered moderators on the reac-
tions.
     Indifference to noise ("ROB") moderates annoyance by aircraft noise and fear associa-
tions concerning  aircraft; age and sex influence all three 'indirect' effects of aircraft noise.
     7.4 With reference to noise protection zones (as they are defined in USA, GB or BRD)
the data demonstrate that outside of these areas considerable portions of the population are
affected by aircraft noise and its consequences (see Fig.  6). In highly noise-exposed areas
more than 3/4 of the inhabitants feel that aircraft noise interferes with their living condi-
tions.
     7.5 It is important in this connection to  take into account the distribution of popula-
tion (see Tab. 1 below). The absolute number of individuals increases considerably toward
the  less noise-exposed  clusters (on account  of their larger extension). This population,
feeling disturbed in their communication and impaired in  their rest and recreation, is less
numerous in relative terms, but is considerably larger in absolute terms.
     This fact allows for predictions of effects of aircraft noise to be expected on behalf of
the further increase of air traffic as well as of the greater density of population.
     7.6 Concluding from the results of this study one can say:
     — Based upon the WHO-definition of health 'as a state of an optimal physical, psycho-
logical and  social  well-being', exposure  to  aircraft  noise  is a serious impairment  to  the
population.
     —  The reduction  of aircraft noise is a problem for those producing the  noise  (e.g.
airlines, aircraft industry) and also for those distributing the noise (e.g. city planners). That
is, it is a problem  involving aspects of engineering as well as of policy.
     For  a  detailed  report on methods,  results,  and  consequences  of  the  study  see:
"DFG-Forschungsbericht: Fluglarmwirkungen - eine interdisziplinare Untersuchung liber die
Auswirkungen des  Fluglarms auf den  Menschen"  (with English  summaries; Bonn, Summer
1973).
     References will be  found there.

                                          773

-------
              FB1 aircraft  noise  exposure
              ROB indifference  to noise
              R1U annoyance and disturbance
                  by aircrafts
              DEF physiol. defense  reaction
              FAF fear associations
              RRD diastolic blood  pressure
              AUF attention performance
Figure 5. Path model of the effects of aircraft noise
              774

-------
                                                                0
7
© "Disturbance of communication"
    "Disturbance of rest and recreation"
     Aircrafts as "Disturbing factor"
      (spontaneously  mentioned)
                                                               60
                                                                20
                                          (least square fits  of
                                           the percentage data)
                                                                100
 59  61  63  65  67 69  71   73  75  77  79  81  83  85  87  89  91    rD1
 .  Al A2 - ^ . AA -6A  ?P ]?. Ik . 7.A .7A '. ?? . ?? . A4. .8J> . ?? . .9A . ??  . : . . .
|20)           (40)           (60)       (70)              (80)  HR/Tag
 (8lT .................... (85)" ........ |90V ....... i95)""(ldoTfj""
 Fiugtdrmgesetz der  BRD        >       2
 Wilson- Committee |GBI
 NEF- Criterion I USA)   >        II       >         \
                 Figure 6. Percentage of people annoyed by aircraft noise
                                 775

-------
TABLE I NUMBER OF PEOPLE ANNOYED BY AIRCRAFT NOISE
Noise Zone              A        B        C        D
Population           45.000    44.000    15.000     2.000
FBl(dB(A))             65,1     74,6      82,7      90,8
disturbed in    rel.      21%     43%      56%      70%
communication
and relaxation   abs.    9.400    18.900     8.400      1.400
rel.: percentage in the sample
abs.:  estimated part of population
                           776

-------
                    RATING THE TOTAL NOISE ENVIRONMENT
                        IDEAL OR PRAGMATIC APPROACH?

                                   D.W. Robinson
                        National Physical Laboratory, England

     Opinions differ on the emphasis to be placed  on the quantitative aspect of environ-
mental noise rating. Human reactions are so numerous and so variable  that it may seem
possible to make only broad  qualitative statements about them. However, those who are
concerned with  noise as a public health problem must adopt a more positive attitude to
these difficulties. In this paper we take the position that quantitative methods are necessary
as a basis of planning and for  comparing data on a uniform footing throughout the world.
The question  we have to consider, therefore, is this: If noise assessment is not an exact
science to what  extent is it possible and useful to treat the problem in a strictly quantitative
way?
     We can readily dispose of one part of this problem. Every sound that enters a listener's
ear first  undergoes a process  of "aural  transduction", the details of which are well estab-
lished  with  the  possible exception  of  sounds having very  low frequencies or impulsive
character. The transduction process can be studied in the laboratory, and the results are to a
large extent already enshrined in standard procedures such as those for calculating loudness
(ISO, 1966). As a general rule, the more faithfully one wishes to mirror the workings of the
ear  as a transducer, the more complicated the formula one will  have  to choose. In the
context  of  total noise environment rating,  however,  the  differences between the  various
measures are merely marginal, being overshadowed by  other factors (Botsford, 1969; Young
and Peterson, 1969). Here  we shall make no distinction; the symbol Le (for "ear-weighted
sound  level") may be understood by the reader whichever way he chooses to stand for
loudness level, perceived noise level,  perceived level, D-weighted sound level or A-weighted
sound  level.  Since  the last two  are  the  easiest to measure and the last is by far the most
widely used, there is much to be said for identifying Le  with A-weighted sound level. At
least this will avoid false precision when we come to consider the less deterministic processes
that follow aural  transduction. It is a certain aspect of these processes with which we are
concerned, and  for want of a better term we shall refer to "general adverse reaction" (GAR)
to mean that  part of a person's total response to undesired community noise which cannot
be  attributed to a  specific disturbance such as speech  masking.  GAR may be roughly
equated with "annoyance" or  "dissatisfaction".
     The transduction process is largely independent of the conditions in which a sound is
heard. Contrariwise, GAR is very much a function of the  state of mind  and activity of the
hearer. Moreover, it is generally the  product of a continual succession of noises, rather than
single acoustic events. Whereas one  can speak legitimately of the loudness of an isolated
sound and evaluate it precisely in the phon  or sone scale, it is meaningless to speak of the
annoyance of a sound out of its context.  Exactly  which psychological entities  one can
equate with GAR is hard to say, but "arousal" is certainly a component and such concepts
are not susceptible of measurement with the same repeatability and simplicity as loudness.
The latter is simply a function of the intensity and spectrum of a sound; broadly speaking,

                                         777

-------
 duration is involved only when the whole event is brief compared with the auditory integra-
 tion time (a few hundred  milliseconds at most). Time enters into GAR in an altogether
 different way, however. With an ongoing noise, arousal will continue to increase for long
 periods, perhaps hours or days. Furthermore, .arousal is  augmented by perceived changes in
 the stimulus so that a noise which is varying evokes a greater reaction  than one which is
 constant Thus, both duration and temporal variation are involved,, and in order to represent
 GAR in terms of the physical parameters of the stimulus it is necessary to know the whole
 history of Lg as a function of time from start to finish of the exposure.
   ,  The development of loudness, theory began.with observations  at an experimental level
 and these were subsequently interpreted by hypotheses  about the functioning of the audi-
 tory mechanism. It culminated  in the elegant  procedure due to Zwicker in which each step
 can be identified with physiological mechanisms of the ear.  In  the case of GAR we are
 concerned with the functioning of the whole  auditory system plus its interactions with the
 other sensory modes  and the autonomic and behavioural states of the  auditory. It is no
 wonder that a close-knit theory of GAR, analogous to loudness theory, is elusive and may
 indeed be unattainable.
     In an idealized system one might envisage the environmental noise as a stressor defined
 by a number of physical  components and the induced GAR as a "strain" having many
 subjective components. The solution would then consist of determining the matrix of rela-
 tions between these components by some form of principal-factor analysis. Since the matrix
 would inevitably vary from person to person as well as from time to time, the usefulness of
 such an ambitious project would be in question  from the  outset. At a  lower level of
 sophistication, let us confine attention to an "average" or "typical"  response instead of that
 of the individual person and limit consideration to a single subjective dimension. The attrac-
 tion of this approach is that it will at least distinguish the importance of the various physical
 parameters.  Since  we  shall identify this single .subjective dimension with GAR we must
 attempt to. give it a more  specific definition.. In fact we shall equate it with the Guttman
 scales  of annoyance or unacceptability( that  have  been developed for use in  surveys, or
 alternatively with the scores obtained 9n rating scales and semantic differential question-
 naires used in laboratory studies. It is difficult to prove that these all measure the same thing
 but they seem to be  closely related to one another. Even with these Amplifications it is
 clearly  beyond the scope  of psychological theory  to  proceed from first principles to a
 formula relating the parameters of a noise environment  to GAR. We must advance, as with
 the history of loudness theory, from a starting point of observation and then see whether
 the results admit of interpretation according to accepted psychological theory. The observa-
 tions, naturally, are harder to come by tiian data from loudness experiments.
     The pragmatic method of studying  the relations between noise and adverse reaction
 which automatically takes  the context factors into account is by means of surveys, and a
retrospective study  ofi the  results of such investigations might be  expected to yield some
insight into the general workings of GAR formation. Unfortunately no, common method-
ology has been employed, and surveys have almost all been analysed individually. This has
fed to diverse  formulae.  Surveys are undeniably appropriate for determining absolute
reactions in  given  circumstances  and therefore,  by extension, to setting noise limits in
broadly similar situations. However, the experimental variables are not controllable and the

                                        778

-------
method does not lend itself to the testing of hypotheses or to the extraction of underlying
cause-effect relationships. In practice, the results of surveys are treated by multiple correla-
tion analysis to relate the obvious physical variables to the subjective scores by means of an
equation (usually linear) valid only  over the  range  of the variables encountered in  the
particular survey. Extrapolations from  such analyses are manifestly hazardous, and an
example will illustrate one of the limitations.  In the Heathrow Airport Survey of 1961
(McKennell,  1963), the only noise considered was that of the aircraft in flight. The resulting
index (NNI) has nevertheless  been  widely applied in planning studies for other airports
where it is almost certain that  the prevailing background noise is quite different from what
it was around Heathrow. One must view these applications with scepticism. In principle, the
background noise at  the sites surveyed around  Heathrow  could have been included as an
additional experimental observation  and introduced into the multiple regression analysis,
but in fact it was not. The fact that this would  have been  a difficult  thing to  do does  not
mean that it would have been irrelevant.
     Furthermore, surveys on different sources of noise have used different  methods for
expressing  the  physical parameters.  Consequently we  find formulae  cast in incompatible
moulds.  But  the fact  that all these studies concern the reactions of people in their normal
habitats  to  a common stimulus, namely a  noise stream, suggests the possibility that  the
results may  in fact be more alike than they have been made  to appear: the diversity of the
formulae may well conceal an underlying unity. In 1969 the writer attempted a synthesis of
some survey data on aircraft noise  and motor  traffic noise as  well  as certain laboratory
results on the trade-off between noise level and duration  for equal  perceived objectiona-
bleness.  It was found possible to do so in a very simple  way, in the formula for Noise
Pollution Level (Robinson,  1971). The motive  for this reexamination was not simply  the
academic pursuit  of  economical formulation although  such has often been the way of
scientific advance—for example, the progressive economy of the Ptolemaic, Copernican  and
Keplerian laws of motion of the celestial bodies. Nor was it an attempt to make a great leap
to some  formula  based inevitably on psychological theory. In fact the reason was the much
more prosaic one of handling prediction problems in situations not explicitly studied in  the
surveys,  namely  cases where  two or more  noises co-exist  or where noises of a new or
hypothetical kind are concerned. An example of the first category is  the case already cited
of aircraft noise around an airport superimposed on the prevailing ambient noise; typical of
the second kind is the consideration of proposed VTOL aircraft  operations in urban areas
already  subjected to a known environmental noise climate. It is impossible to handle these
mixed situations by means of the survey formulae as they stand.
     The concept of  Noise Pollution Level was arrived at inductively from certain experi-
mental results which clearly indicated  that the mean noise  level (more exactly, the equiva-
lent continuous noise  level Leq) fails to explain the differences of reaction to traffic noise at
different sites (Griffiths and Langdon, 1968). The less constant the noise level, the greater
the adverse  reaction.  On the other hand it is obvious that the mean  noise level must also
play a part,  and  the  simplest way  to embody these two principles was  to  form a new
measure  which increased with both factors.  Lgq being  the most straightforward measure of
mean level, standard deviation s being the basic  measure of variability, and a linear combi-
nation  of these two being the simplest algebraic entity, it required only to determine  the
value of one arbitrary  constant k to complete the Mark 1 formula LNP  = Leq + k-s.
                                         779

-------
     It happens that (Ljg - L9Q)/S K equal to 2.56 for a Gauss an distribution of noise levels
and since this value of k  fitted the data rather well it was selected for convenience. The
interesting consequences of the L^p formula when applied to aircraft noise are illustrated in
Fig.  1.  Here the growth  of LNP with the number of overflights is contrasted with the
empirical relation found from the Heathrow survey, namely an increase going as 15 log N.
The salient points are that  Lfjp implies a similar mean rate of increase (although the relation
is curvilinear) and that the  curve does not descend to an indeterminate value at N = O but to
a floor  determined by the ambient noise level.  The particular cases shown  are for back-
ground noise 20 and 25 dB below the average peak noise of the aircraft; note that the higher
background leads to a lower value of Lj^rp at high  traffic as found by Bottom (1971).
     Certain features of the  formula  can be verified in the laboratory, even though such
tests cannot be used to determine  absolute acceptability points on the scale. Three such
experiments are briefly described here. To induce a constant "set" in the subjects, a pencil-
and-paper task is given but the task scores are  not of primary interest. At the end of each
exposure lasting 30 minutes, reactions to the noise are  scored by a battery of rating scales
and a multi-item semantic differential test. The latter is found to be repeatable and fairly
sensitive to  changes in the noise environment (Anderson,  1971).  To establish a baseline
(Experiment I) the subjects received constant noise at 8 fixed levels from 60 to 95 dB(A)
with some repeats; a reasonably linear relation between test score  and noise level resulted
  1
 £»
 CD
 CO
 oc
              -HP
           Background
           high     low
                            10
20
50    1OO
2OO
500  N
               Figure 1. Dependence of rating value on the number off overflights N
                                       780

-------
(Anderson and Robinson, 1971). In Experiment II the noise was punctuated at irregular
intervals (averaging 3 per minute) by a louder noise peaking to 5, 10 or 15 dB  above the
steady level. The effect is greatly to increase the test scores. The peaks have, of course, some
effect  on Leq, but  a much greater effect on the term k-s  in  the formula  for  Ljyjp. The
subjective scores reflect this greater increase (Fuller and Robinson, 1973). In Experiment III
two steady noises differing  in level by 16 dB(A) were presented  alternately for a total of 15
minutes each. Four configurations were tested, the noise bursts being of 5-sec, 15-sec, 5-min
and 15-min duration respectively.  The test scores are slightly higher for the intermediate
burst durations  than for either the longest or shortest but much the greater effect is the
consistently high scores in  Experiment  III relative to the results for a steady noise of the
same  Leq. The results of the three experiments are plotted in Fig 2 against  Lgq  (left)  and
LNP (right). The "perfect" rating measure would bring the data points within a narrow band
whose width would  be determined solely by random error. The standard error of  the results
is typically 0.04 units on the ordinate scale. It is clear that L^q is considerably less effective
in explaining the results than Ljsjp. The finer points of the results show that LNP does not
     1-0
     0-8 -
  «
  k
  o
                                                                          100    110
                     Figure 2. Laboratory results on noise environment rating

                                         781

-------
provide  a^ complete explanation either, but this is scarcely surprising in view of the rudi-
mentary derivation of the Mark 1 formula. There is no doubt that refinements would bring
the data into still closer alignment. Some indications of the direction which these develop-
ments might take are now given but at this point is is necessary to introduce a cautionary
note: can the added complication be justified in view of the relative success of the simple
formula?
     Shortcomings of the formula become apparent when the rate of noise level fluctuations
is either very slow or very fast. To take an extreme example of a noise which climbs 10 dB
during the morning and falls  10 dB during the afternoon at a very slow rate, it is intuitively
obvious that s, which is a gross measure of variability, cannot be correct; the listener might
well be quite unaware of the changes and consequently would suffer no arousal as a result.
An idea of how the  speed of the  fluctuations may  be accommodated  in a more  general
formula is  provided  by  Fig.  3 (Robinson,  1972). Here  the variations  of noise level are
represented in the frequency domain from about 5 //Hz (corresponding to a time of 8 hours)
to about 1.5 Hz (corresponding to a time constant of auditory temporal integration of 10
   ,.  10
   o
          -i
   o
   (0
10"
   TO
   I   io'3
       io'4-
 Shours  1hour  10min.  1min. 10sec. 1sec.  lOOmsec
	1	
                         T
                                  T
                                LNpMark1
                                       T
                                 Modified L
                                              NP
                                      I
                                                 I
  -6
10      10
                             -4
                           10       10"     10"     10
                          Fluctuation frequency (Hz)
                                                 -i
                                                                  10
                  Figure 3. Weighting of the noise level fluctuation spectrum

                                       782

-------
ms). This is a span of some 18 octaves and it would be amazing if the processes governing
GAR behaved uniformly  over such  a vast range. More likely the low and high  frequency
components of the noise level fluctuation spectrum are subjectively down weighted, some-
what as illustrated. The true shape of the appropriate weighting function could, in principle,
be determined by straightforward  though laborious experiments. The "cut-off frequencies"
are conjectural but they seem likely to be associated with time-constants of the  order of a
few  seconds  and  several minutes  respectively.  A computer realisation for a Mark 2 L^p
modified on these  lines is illustrated in  Fig. 4. The complication which  this refinement
entails is that the calculation no longer depends only on the statistic s obtainable from the
overall1 histogram of noise levels, but is essentially  an on-line process taking the level varia-
tions and their rates of change into account as they occur. An elementary explanation of
this scheme is that variations on the scale of a minute are fast enough for people  to be very
aware of them but not so fast they they easily habituate to them. For two different reasons,
slower or faster changes  are less  annoying. The modified  Lj^p formula converges to  the
simple one when  the periodicities  lie between the lower and upper limits mentioned, and to
Leq at the extremes of periodicity.
     Thus there are possibilities of refining the formula at  the expense of complicating the
calculation. The author's opinion is  that  such developments may be premature but that the
time is certainly opportune for general adoption of LNP in its simple Mark 1 form and we
shall conclude this  paper with an example  to illustrate  the powerful practical advantages
which it offers. This is the problem of determining the maximum acceptable noise level in a
residential  area arising from projected V/STOL aircraft operating in the vicinity, given an
I


t = 5 sec
• *
ITV-*- /"i" -»• a:2 -H» f* — >• 1/ -n no •»
|^_/— *• _/J * » * J 't LOG
"Ear- w
weighting
rT i«
fe i . fc */ - fc, I rf^O
J ' T LUU
\^ * I
T = 8 hour Leq| ^-^
DIGITAL i
Nf
COMPUTER l^^x

r
J


X




\
)J
>


T



2





k.s





^







)



^,
'r
= 5 rr

fT
J


•sac







lin






^




2
ac


•t/
IT

+
If

T =






^ '•.
^/^"l
fr\ J
\ '
N- rT
•^ j
^
: 8 hour


              I	;	I
                    Figure 4. Computer realization of modified L^p formula
                                         783

-------
engineering estimate of the overflight noise/time signature and of the traffic. The method
proceeds as follows. One first determines a typical background noise distribution for the
type of area affected and hence calculates the pre-existing L^p. By assigning a trial value of
the peak level of the aircraft noise, one then calculates the combined noise level distribution
and hence the new value of L^p- The result depends on the amount of traffic and to a lesser
extent on  the statistical variation of flight paths and source noise levels, these data being
part of the given engineering estimates. Using a series of trial values, peak aircraft noise level
is obtained as a function of the increment in LNP- It remains to fix a criterion of "salience"
(say 1 dB or 3 dB), meaning the maximum permissible increment in Ljsjp. The peak level can
then be read  off directly  and the problem is solved. The reader will  ask: How does one
decide on the criterion of salience? This step calls for a value judgement to be made and it is
possible to answer it in terms of "percentage of persons annoyed" from data in the litera-
ture. In conclusion it should be noted that the only  alternative to LNP which permits an
analysis on similar  lines is Lgq. But the disadvantage of this measure is its relatively  poor
correspondence  to GAR, as shown both in survey results and in the laboratory experiments
described here.

References

Anderson,  C.M.B. The measurement  of attitudes to noise and noises. National  Physical
     Laboratory, Acoustics Report Ac52 (1971).
Anderson,  C.M.B. and Robinson, D.W. The effect of interruption rate on the annoyance of
     an intermittent noise. National Physical Laboratory, Acoustics Report Ac53 (1971).
Botsford, J.H. Using sound levels  to gauge human response to noise. Sound and Vibration,
     3,16-28(1969).
Bottom, C.G. A social survey into annoyance caused by the interaction of aircraft noise and
     traffic noise. /. Sound Vib., 19,473-476 (1971).
Fuller,  Hilary C. and  Robinson  D.W. Subjective reactions to steady and varying  noise
     environments. National Physical Laboratory, Acoustics Report Ac62 (1973).
Griffiths, I.D. and Langdon, F.J. Subjective response to road traffic noise. J. Sound Vib.,  8,
     16-32(1968).
International Organization for Standardization. Method for calculating loudness level. ISO
     Recommendation R532 (1966).
McKennell, A.C. Aircraft noise annoyance around London (Heathrow) Airport. UK Central
     Office of Information Report SS337 (1963).
Robinson, D.W.  Towards a unified system of noise assessment. /. Sound Vib.,  14, 279  - 298
     (1971) (First issued as National Physical Laboratory Aero Report Ac38, 1969).
Robinson, D.W.  An essay in the comparison of environmental noise measures and prospects
     for a unified system. National Physical Laboratory, Acoustics Report Ac59 (1972).
Young,  R.W.  and Peterson, A. On estimating noisiness of aircraft sounds. /. acoust. Soc.
    Amer., 45, 834 - 838 (1969).
                                        784

-------
                 MOTOR VEHICLE NOISE: IDENTIFICATION AND
          ANALYSIS OF SITUATIONS CONTRIBUTING TO ANNOYANCE*

                                William J. Galloway
                                    Glenn Jones

                            Bolt Beranek and Newman Inc.
                               Canoga Park, California

    Noise is a common and frequent source  of annoyance in today's society.  In many
communities the prevalent sources of noise are motor vehicles. Laboratory psychoacoustic
evaluations can provide much insight on the comparative loudness of various vehicle sounds
and even provide some objective measure of  their acceptability under controlled situations.
Studies of the annoyance of people from noise exposure in  real living environments, how-
ever, show that attitudinal and situational  factors influence annoyance to the same, or even
greater, degree as the noise exposure  itself (1,2,3,4).
    This project  was  directed towards identifying a  number of motor-vehicle-noise-
generated "situations," quantifying the annoyance they create, and relating the physical
characteristics of noise to this annoyance.  The situations were identified by a social survey.
The physical evaluation involved both measurements of the statistics of the noise at the
interview sites and  the analysis  of the noise of discrete vehicle events. The  results of this
work are described in detail in a series of technical reports (5,6,7,8). The highlights of the
investigation are summarized in this paper.

ANNOYANCE FROM MOTOR VEHICLE NOISE

     In order to relate the physical noise of motor vehicles to the annoyance produced, it is
necessary first to have some notion of how these factors are identified and interact. A social
survey  was  designed and conducted to achieve this goal.  From the  onset, the program
sponsors and investigators assumed that annoyance is not a response that is solely  related to
a set  of sound stimuli  or noises.  Rather,  a complex  context was assumed  in which
annoyance is not only stimulated by the  physical and temporal characteristics of certain
sounds emitted by motor vehicles, but in which the annoyance is substantially conditioned
by the meanings that the  noise may have for people in terms of:
     vehicle maneuver, driver behavior, and driver responsibility; the activities in  which the
     auditors are engaged when they hear  the sounds and the effects of the noise on those
     activities; the barriers and distances between auditors and sources; the place where and
     times  when  the sounds are heard; the individuals'  susceptibility to noise; the indi-
     viduals' backgrounds and their station in life; the individuals' beliefs  and  attitudes,
     especially their attitudes towards vehicles and persons that make the noise.
     These  multidimensional contexts are called  "situations." While this study does not
assess  the contribution of each situational dimension to annoyance, it does  recognize  that
annoyance occurs in a context.
 *Research sponsored by the Automobile Manufacturers Association

                                         785

-------
     SURVEY RESEARCH DESIGN - The  research design considered  annoyance in two
 dimensions: extensity, the number(or proportion) of persons annoyed, and intensity, the
 degree to which people are annoyed. Previous studies tend to confound these dimensions by
 reporting the  number of persons above'a certain level of annoyance. One of the reasons
 results are difficult to compare from one'study to another is that these levels tend not to be
 uniform.
     An important objective has been to isolate and describe situations in which annoyance
 from motor vehicle noise occurs. Ideally,'these situations would constitute the independent
 variables. We  conceptualize the independent variables as all those forces external to the
 auditor that act in a situation to determine his annoyance. They include the sources of noise,
 the setting and-maneuvers of the noise-making vehicles, factors affecting the propagation of
 the sound, acoustical qualities  of the sounds that reach the auditor, and the auditor's
 location and activity.
     Even though they  are judged phenomenologically, the independent variables are con-
 ditions external to the  respondent.  The conditioning  variables,  on the other hand, are
 conditions that inhere in the respondent. Vft have deliberately limited these to perceptions
 about the source of qualities that would not change  the noise, but might influence an
 auditor's reaction to it, effects on activities, beliefs about the reactions of others, and status.

            Variable Type              Measure

            Perceptions of source        — if automobile: type, sport or
                                          family
                                       — if automobile: manufacture,
                                          foreign or domestic
                                       — if automobile or motorcycle:
                                          driver, young or old
                                       - legality of vehicle operation
                                       — driver control of noise
            Effect on activity            - effect: none, makes harder,
                                          interrupts, stops
            Sharedness of annoyance     — annoyed: respondent alone,
                                          personal acquaintances, public
            Status                     - sex
                                       — age
                                       — occupation of head of household
                                       — education of respondent
                                       — income of household

Three classes of variables thus constitute the elements in the research design: independent
variables are dimensions of situations, intervening variables are attributes of persons, the
dependent variable is annoyance. The object of the research is to determine the levels of
annoyance induced by situations that are varied combinations of sources and settings and to
parcel out the  individual attributes that influence the independent-dependent variable rela-
tionship.

                                        786

-------
     SAMPLE DESIGN - The sampling process was consistently guided by the fact that the
theoretical unit  of analysis in this study is the situation in which people are annoyed by
noise arising from  motor vehicles. The parameters of the population of these situations are
not known, therefore our sampling of them has been indirect. Theoretically, the population
to be sampled consisted of the situations in which people experienced annoyance from noise
made by  motor vehicles. A direct sampling  of  this population is not feasible since its
parameters cannot be estimated. However,  we can assume  that these situations are distrib-
uted like the  population of persons exposed to motor vehicle noise, hence when we sample
these people,  and elicit from  them  the motor vehicle noise situations that cause them
annoyance, we assume that we have sampled the situations themselves.
     Our  resources were insufficient  to sample the entire  population exposed to motor
vehicle noise, hence  we confined ourselves to a systematic sampling in two metropolitan
areas—Los Angeles and Boston. This choice was based on two considerations:  first they were
accessible and second they offer very  different environmental conditions. A smaller sample
was  also taken in Detroit in order to provide sites that could more conveniently be explored
in depth by research personnel from the automobile industry.
     We wanted  to be sure that three  kinds of noise environments were represented in the
neighborhoods that were selected: (1) neighborhoods near enough to limited access high-
ways that the sounds of individual vehicles could frequently be distinguished, (2) neighbor-
hoods that were more distant from  freeways, but where noise from them was still audible,
and  (3) neighborhoods that were dominated by stop-and-go  traffic. A stratified sample was
planned in order  to increase the certainty that neighborhoods with these characteristics
would be  included. This entailed the systematic drawing of a large first-phase sample, strati-
fying the neighborhoods, then drawing from each stratum. The final sample thus remained
unweighted.
     Dwellings within neighborhoods  were to be characterized by uniform exterior noise
environments. A fixed  quota of households was chosen for each neighborhood. Within each
household, a systematic selection of the individual to be interviewed was planned by rotating
sets  of instructions to the interviewers that designated the person to be interviewed by the
size  of households. Persons were identified  by ranking them according to sex and age. The
research design  required that the levels and  sources of annoyance in each neighborhood
could be specified and, at the same time, that the neighborhoods constitute a representative
sample  of each of the two major metropolitan areas. This created a trade-off between the
number of respondents in each neighborhood and the  number of neighborhoods in each
area. In pre-testing the questionnaire, it was determined that  descriptively  and intuitively
one  could characterize  a neighborhood's noise environment after about twenty completed
interviews. This  allowed 25 neighborhood sites each for Boston and Los Angeles (and  10 for
Detroit) to be sampled.
     SAMPLE SELECTION - Once neighborhoods had been selected, a set  of dwellings in
the neighborhood  that would provide suitable targets for interviews had to  be chosen. The
crucial problem in making these choices was to select dwellings whose noise environment
was  uniform.  Because the solution to this problem rested primarily on judgments  about
noise, the task of  choosing the eligible dwellings for each neighborhood site was performed
by acousticians. The acoustician, following a prescribed set of site appraisal procedures, was
asked to take the  address of a single dwelling and build a researchable neighborhood around

                                         787

-------
it; he was asked also  to provide an impressionistic description of the neighborhood he
defined.
     He recorded in detail the traffic  flow, physical characteristics, and noise patterns of
each of the roadways from which traffic noise was audible from the sidewalk in front of the
target address. He also made a general record of non-traffic noises and their sources. His
descriptions were impressionistic and were limited to a single point in time.
     Having thus evaluated  the  target address,  his next  task was  to select a group of
dwellings that had very similar external noise exposures. After the acoustician had selected a
quota of  eligible dwellings in a neighborhood, he was asked to make an  impressionistic
characterization  of it in non-acoustical terms. He was asked to draw a map showing noise
sources  and eligible dwellings, to describe land use including prevailing building types and
distributions, to  rate the level of neighborhood maintenance, and to typify  the population
in terms of racial mixture, age distribution, occupational patterns, and income level. He also
photographed the target address and the street on which the target address fronted.
     INTERVIEWS - The 1200 interviews were conducted by telephone, selecting telephone
numbers from a  reverse order telephone directory on the basis of the addresses identified in
the site  evaluation. The method of interview, selection of individual within  the household,
and  the interview protocol  itself are  discussed in detail in Reference 5 and will  not be
treated further here.
     Certain results  of the interview  will help in understanding the composition of the
respondents. Although an even balance was desired, 62% of the respondents were women,
38% men. No one under  18 years old was interviewed, and relatively balanced percentages
of 5 year increment groups  were generally  obtained from "20" to "over  65" years old.
Education  level,  income level,  occupational status (thus allowing computation  of the
Duncan  Index of socio-economic status*) were obtained from all respondents, whether or
not they described a vehicle noise situation.  Detailed further data was obtained from only
the 549 respondents who identified any annoying vehicle noise situation.
     ANNOYANCE  FROM ALL  MOTOR VEHICLES - In analysis,  motor vehicle  noise
(regardless of the type  of vehicle or the part  of the vehicle from which the noise arose) may
first  be  considered  as a single entity.  Annoyance from different motor vehicle sources is
considered later in the  section. One of the first categorizations determined was the respond-
ent's rating of the noisiness of his home and working environments. This was evaluated on a
seven point scale ranging from "not noisy at all" (rated 1) to "unbearably noisy" (rated 7).
On this basis 72% rated  their neighborhoods "noisy" (i.e., from 2 to 7) with an  average
score of 3.2. In  terms  of their work environments, 66% rated their environment "noisy"
with an average score of 3.9.
     Next, the respondents were asked to identify the proportional contribution of various
noise sources to the total noisiness of their neighborhood. While  the percentages by source
differed slightly from city to city, the overall average is quite representative of the overall
survey.  By general  classification  the following  percentage contributions were stated  for
various noise sources:
*This scale assesses occupations in terms of prestige, education and income. The larger the
 index number, the higher is the status associated with the occupation.

                                        788

-------
                        Aircraft                         15%
                        Construction                      1%
                        Industrial                        <1%
                        Motor Vehicles                   55%
                        Lawn Mowers, Snowblowers        2%
                        Radio and TV                     2%
                        Voices                          12%
                        Other Noises                      6%
                        Not Ascertained                   8%

The  remaining bulk of  the  analysis is concerned with scores  of annoyance rather than
noisiness. When asked to identify whether they were annoyed by vehicle noise 54% were not
annoyed, while 46% were, with an average intensity of annoyance of 4.2 on a scale where 3
stood for "quite annoying," 4 for "definitely annoying," and  5 for "strongly annoying."
The  range  of annoyance intensity is quite large, having a standard deviation for individual
judgements of 1.6.

     With respect to other variables, the following points were observed:
     a)   Men and women exhibit about the same degree of annoyance to vehicle noise.
     b)   The age group  from 20-29 is 1.8 times as likely to be annoyed as the average, and
          those over 60 are only 0.6 times as likely to be annoyed. The other age groups are
         essentially equal to the overall average.
     c)   Education level  enters strongly into the  percentage  of annoyed versus not
         annoyed:

                     Highest School        % of       % Annoyed/
                     Level Attended       Total       % Not Annoyed

                     Elementary            9             0.4
                     High School            46             0.8
                     College                42             1.6
                     Not Ascertained        3

     d)    On an income basis, those with annual income closest to $5,000 were 0.7 as likely
          to  express  annoyance  as the sample whole; those with $25,000 were  twice as
          likely to express annoyance.

     Similarly, some general results on situational characteristics were identified:
     a)    Eighty percent of all annoying vehicles are moving.
     b)   Where no individual vehicle is identified, leaving  a residual of "traffic noise," the
          annoyance  is less intense, and can  be rated in terms of gross flow parameters.
          Heavy traffic,  57% of the total, had an annoyance score of 4.0 on the seven point
          scale, slightly lower than the 4.2 observed for all categories. Intermediate traffic,
          27% of the total, and light traffic, 8% of the total, with both having scores  of 3.2.

                                         789

-------
    c)   Although half the moving vehicle responses were stated to be under "accelerating"
         conditions,  the differences  in  annoyance as a  function of speed  state were
         not statistically significant
    d)   Over half the annoying noise situations from identifiable vehicles originate on city
         streets, 30% on main streets, 27% on side streets. Twice as many originate on
         main roads-boulevards, thoroughfares, highways (18%)-as  originate on limited
         access roadways—freeways, expressways, turnpikes (9%). Heavily travelled (main)
         roads and streets seem to be the source of more  annoyance than side roads and
         streets.
    e)   Situations are  ranked in  intensity according to distance; 90% of all annoying
         situations occur within a block of the noise source, 70% within  an estimated 100
         feet
    The respondents were queried about various acoustical factors associated with the vehicle
noise situations. Among their judgments the following results were obtained:
    a)   When asked to judge the loudness of the situations  they described  on a seven
         point scale, the respondents rated the following distribution:

                     Category            %        Average
                                                Intensity Score

                     Deafening  (7)        9         6.2
                     Extremely  (6)       13         5.1
                     Very      (5)      25         4.6
                     Oust plain) (4)       19         4.1
                     Quite      (3)       18         3.3
                     Mildly     (2)       13         3.1
                     Not at all  (1)        2         2.5
                     Not relevant           3	

                         Average                    4.2
    b)   Loudness  accounted for one-third  of the variance  in  overall  annoyance; the
         average loudness score was equal to the average annoyance score.
    c)   Judgments of pitch (e.g. frequency) showed little variation in intensity of annoy-
         ance; low and mixed frequency noises predominated.
    d)   With respect to duration, sounds longer than five seconds but short of continuous
         seemed to predominate. No particular dependence of annoyance intensity with
         duration was observed.
    e)   Between 90% and 95% of the annoying noises recur at least weekly, most of them
         daily.  However, except for  a very few cases, frequency of onset does not affect
         intensity of annoyance.

                                       790

-------
     Questions related to the setting-iit.whiqh the respondent thought about the vehicle
noise situations revealed:
     a)    Seventy-seven percent 'experienced the noises in: their home; 12% while in transit;
          only 5% while at work.
     b)    Roughly two-thirds of the annoying motor vehicle situations can be characterized
          by the time of day in which they occur. Over twice as many situations are found
          at evening or at night as in the daytime. However, these factors make no signifi-
          cant difference in levels of annoyance. In nearly a third of the situations, time of
          day is immaterial.
     c)    Annoyance occurs most frequently when  people are sleeping, followed in order
          by listening to TV, radio or recordings; mental activity such as reading, writing or
          just thinking; driving; conversing; resting; and walking.
     d)    Motor vehicle noise stops only a few activities, however in nearly two-thirds of
          the  cases it either interrupts them or it makes them more difficult. In 16% of the
          cases it does not affect tf\e activities.

     Investigation of the conditioning variables, those which may predispose annoyance but
not intrinsically influence the sound,  revealed:
     a)    The question of whether family cars or  sports cars were more annoying was
          irrelevant in  half the cases; in the other half,  sports cars  were slightly more
          annoying than family cars,  and were specified twice as often as family cars.
     b)    Whether an automobile was foreign or "domestic was also irrelevant half the time;
          however, when specified, domestic cars were cited three times as often as foreign
          cars.
     c)    Half the time drivers under 25 years old were cited as being responsible for
          annoying operation of vehicles; one-third of the time age was irrelevant.
     d)    In over half the cases, drivers were thought to be operating their automobiles and
          motorcycles  legally; in  nearly  30% they were thought to be operating illegally.
          This perception raises the annoyance one-half a scale point relative to  the legal
          operation.
     e)    In almost half the cases, it was thought that the noise produced by the vehicle was
          easily within the control of the operator; in 15% more, it was thought  he could
          control the noise if he tried. In only slightly less than one-third of the instances
          was it thought that the driver could not control noise through his operation of the
          vehicle.
     f)    The annoyed individuals believe that in over 60% of the situations the annoyance
          is shared by the public at large, in 20% the  annoyance  is shared by individual
          acquaintances, and in only 5% is it the respondent alone who is annoyed.
     ANALYSIS OF  VEHICLE NOISE SOURCE PATTERNS - In  each interview an
 attempt was made to have the respondents identify, wherever possible, not only vehicle
 type/situation, but what part of the vehicle seemed to be causing the annoyance. In the
 resultant analysis, five gross categories could be defined as follows:

                                         791

-------
                Vehicle Type

                Busses
                Motorcycles
                Trucks
                Automobiles
                Traffic
Average Intensity
   of Annoyance

       5.1
       4.5
       4.3
       4.2
       3.7
 Percent of
Total Cases

     3
    23
    20
    36
    17
     In order to retain an adequate sample size, busses and gasoline powered trucks could
not be subdivided, diesel trucks could be divided into two categories, automobiles into six,
while no attempt was made to subdivide traffic in general (the residual when no specific
vehicle is identifiable) or motorcycles.  For the resulting twelve categories, the annoyance
analysis of  Table  I results. This approach  to classification precludes a straightforward
interpretation of extensity since all categories are not the result of a uniform procedure.
One can manipulate extensity of annoyance by the fineness of his divisions. Hence, one can
logically compare  bus  and motorcycle  noise but not motorcycle  noise  and automobile
exhaust noise. Intensity, on the other hand, is logically comparable between any groups
since the acts of classification does not directly alter it Because of the mass of data and the
results of the general analysis, the number of dimensions examined  for the twelve specific
sources was reduced to the following 19:
                     Variable

              Independent

                Source


                Propagation



                Vehicle setting



                Acoustical properties
                Auditor's
                  orientation
                  Dimension
         (If traffic) volume
         Part of vehicle

         Distance
         Obstructions
         Season

         Movement
         (If moving) speed state
         (If moving) type of roadway

         Loudness
         Pitch
         Duration

         Location of respondent
         Activity
         Effect on Activity
         Time of day
                                        792

-------
              Conditioning Variables

                Auditor's                   Driver control of noise
                   perceptions               Legality of operation

                Public nature                Public reactions
                   of annoyance

     A detailed evaluation of each source pattern is provided in Reference 5, including a
textual summary of the pattern it provides. An overview of each source's relationship to
annoyance is provided in Table I.
INTERVIEW SITE NOISE MEASUREMENTS

     On-site noise measurements were gathered at 20 sites out of the total of 60 sites at
which interviews were taken. The purposes for these measurements were three-fold:
     1.   To provide a quantitative description of the noise environments at the different
         sites,
     2.   To provide noise data for correlation with interview judgments of sites noisiness
         and vehicle annoyance,
     3.   To provide data for comparing the several different noise measures which  are
         often used in describing traffic and community noise.
     SITE NOISE MEASUREMENT TECHNIQUES - At each of the sites, a noise measure-
ment position was located at or near  the residence of one or more of the survey respond-
ents. The microphone was placed 35 feet from the roadway wherever possible and at least
10 feet from any large object or wall.  At each position, a 10-min noise sample was recorded
approximately once each hour over  the course of an entire 24-hour day. Although  the
measurements were not necessarily  taken  in successive  hours, all measurements were
obtained on week days. All noise measurements were obtained by recording the signals on
magnetic tape with later analysis of the tapes in the laboratory. During the field recording, a
log was kept of the various discrete noise intrusions that occurred.
     In the laboratory, the recorded  tapes were analyzed to obtain the A-weighted sound
level, with  more complete spectral analyses in  one-third and full octave bands reserved  for
selected discrete vehicle noise events occurring during measurement periods.
     For these measurements, the A-level was felt to be a good choice for weighting  the
frequency spectra because of previous  high correlations of A-level with subjective judgments
of individual vehicles in terms of noisiness or loudness. However, one of the major problems
in community noise analysis is that of describing the time-varying character of the noise
levels. To permit comparison of several of the community noise measures that have recently
been advocated, the recorded 10-min noise samples were analyzed with a statistical distribu-
tion analyzer. The distribution analyzer determined the proportion of time that the noise
signal fell within specified noise level ranges. From this distribution data, the levels exceeded
90% of time  (1,90), those exceeded 50% of the time (LSQ) and those exceeded
and  1% of the time (Lj) were obtained.

                                         793

-------
                                      Table 1.
Pattern
                   ALL VEHICLES: ANNOYANCE BY SOURCE PATTERN

                                                    Annoyance
Busses
Diesel trucks:
  Audi tor 'In transit
Automobiles:  Tire  screech
Automobiles:  Exhaust  noise
Motorcycles
Diesel trucks:
  Auditor at hone or at work
  (or location Is not  relevant)

Automobiles :  Noise origin
  irrelevant, brake squeal
Automobiles:  Siren noise
Automobiles:  -Engine noise
Gasoline trucks
Traffic flovr
Automobiles:  Horns and
  other noises

Unclassified

Total

Recapitulation :

  Busses
  Motorcycles
  Trucks
  Automobiles
  Traffic flow
  Hot relevant
Extensity

Rank
12
6
8
3
1
4
8
11
5
10
2
8
-
_

Number
20
30
27
97
160
88
27
21
44
26
115
27
11
693

Percent
3
4
4
14
23
12
4
3
6
4
17
4
2
100

Rank
1
2
3
4
5
6
7
8
9
10
11
12
-
—
Intensity

Mean
5.1
1.9
4.7
1.5
1.5
1.2
1.1
1.0
3.8
3.8
3.7
3.3
5.6
1.2
Standard
Deviation
1.8
1.2
1.7
1.7
1.5
1-6
1.5
1.8
1.4
1.5
1.5
1.7
0.6
1.6
5
2
3
1
1
 20
160
HI
218
115
  6
 3
23
20
36
17
 1
1
2
3
1
5
5-1
1.5
1.3
1.2
3.7
6.3
                                                                           1-8
                                                                           1.5
                                                                           1.5
                                                                           1.7
                                                                           1.5
                                                                           0.8
     From the statistical distribution, the energy mean value was also computed, as well as
 two noise rating scales that have recently been developed. These are the traffic noise index
 (TNI) and the noise pollution level (NPL). Both of these measures are attempts to provide a
 meaningful noise measure which reflects both the variability of the noise environment and
 the magnitude of the average or background noise level. In  differing degrees, each measure
 accounts for the belief that one's annoyance may be influenced by the time varying charac-
 ter of the noise signal as wefl as the absolute level of the noise. These measures are defined
 as (9,10):
                                NPL = Lgq + 2.56 s


                                       794
            (1)

-------
where Leq is the energy mean noise level and s is the standard deviation in level

                            TNI = L50 + 4(L10 - L90) - 30                         (2)

     SELECTION OF SITES - The 20 sites selected (9 in Los Angeles, 8 in Boston and 3 in
Detroit) include those which ranked highest in motor vehicle annoyance as interpreted in
terms of the product of intensity and extensity of annoyance scores for the following motor
vehicle categories:
     automobiles  (engine  and exhaust noise), diesel  trucks (observed  at home or work),
     motorcycles, gasoline trucks, busses  and traffic flow.
Selection was further conditioned by review of  traffic situations  at or near the site plus
screening of sites  to eliminate extremes in sociological or economic influences. Several sites
were added to obtain a sampling of sites where very low annoyance scores were observed.
     In terms of their location with respect to traffic-i.e., near light traffic, heavy traffic or
limited access highways—three of the 20 sites were located near or adjacent to limited-access
highways (freeways), 11 were located near heavy street traffic and the remaining 6 positions
were exposed to light traffic. (See Table III for traffic definitions.)
     NOISE LEVEL VARIATION AMONG SITES - There was considerable variation in the
noise environment among the different  sites. In addition to the changes in magnitude of
noise levels  between sites, there are also distinct  differences in the patterns of noise levels
during the day, with the  Los Angeles site showing a greater change  between daytime and
nighttime levels.
     To simplify  comparisons,  the hourly noise  patterns have been summarized to obtain
average noise measures for day, evening and nighttime periods and further discussion will be
confined to noise measures  averaged over these three  daily periods. In terms of the daytime
period (arbitrarily taken as 7 a.m. to 7 p.m.) and the nighttime period (from 10 p.m. to 7
a.m.), Table II lists the minimum, the maximum and the median noise levels for the 20 sites.
Also shown is the range between minimum and maximum noise levels among the  20 sites. In
terms of the statistical noise measures,  the  noise level variation among sites ranged from
about  20  dB during the day to the order of 30 dB during the night. The NPL and TNI
measures generally show a much wider dynamic spread.
     Also shown in Table II is the variation in TNI values when computed from an average
distribution of levels over a 24-hour period, in accordance with the way  in which the TNI
measure was originally defined. On a 24-hour basis, the range of TNI values is markedly
reduced.
     NOISE CHARACTERIZATION OF SITES - Comparison of noise levels for sites in the
same classification with regard to traffic shows that  the noise levels for  the three freeway
sites clustered quite  closely together but that the range of noise levels  for sites classified as
light  traffic  or heavy traffic  spread  over a considerable  overlapping range. There is an
increase of approximately 6 to 10 dB between  the  LSQ levels from light to heavy traffic
situations and  another increase  of 2 to 6 dB  for  freeway sites. Sites  near freeway traffic
show a much smaller change in noise levels from day to night then do light traffic sites.
     A detailed breakdown of the identity of specific noise intrusions is  given in Table III.
The upper portion  of the table lists the average number of noise intrusions per 10-min

                                        795

-------
                                       Table 2.

              VARIATION IN NOISE LEVELS AMONG 20 NEIGHBORHOOD SITES
Noise
Measure

L90
L50
L10
Ll
NPL
TXI

LSQ
L50
*•!(>
Ll
K?L
TKI

L50
WI

Minimum
A. Day
36
41.5
53
63.5
65-5.
54.5
B. Night
26.5
30
35-5
45.5
46
45
C. 24-Kour
36.5
45-5
Noise Level
Median

49-5
56
64
70
75
72

38.5
43.5
54
63
67
73
Average
51
59
in d3
Maximum

63
65.5
70.5
86
88.5
103

54
60.5
67
75-5
91
134

64.5
87

Range

27
24
17.5
22.5
23
48.5

27-5
30.5
31.5
30.0
45
89

28
41.5
sample for the three site classifications for day and night periods. Succeeding rows in the
table show the relative frequency of noise intrusion sources. Of particular note is the sharp
decrease in frequency of diesel truck intrusions in other than freeway situations, and the
sizable increase in proportion of human and animal sounds at the light traffic sites. These
changes reflect changes in motor vehicle compositions with site classifications and changes
in background noise levels, with  possibilities  of  many human and animal sounds being
masked by the higher noise levels typical of freeway sites.
                                        796

-------
                                     Table3.
  AVERAGE NUMBER OF NOISE INTRUSIONS AND RELATIVE FREQUENCY OF OCCURRENCE
                    FOR DIFFERENT SOURCES OF NOISE INTRUSION
Average Number  of
Noise Intrusions  Per
10 Minute Sample

Relative Frequency
of Noise Intrusions
  Automobiles
  Diesel Trucks
  Gasoline Trucks
  Motorcycles
  Buses
  Brake or Tire Squeal
  Horn, Siren
Total Motor Vehicle
 .Aircraft
  Human and Aninal
Total Non-Motor Vehicle
  Freeway .
Day     Night


11.8    li.o
                                              Heavy Traffic*
                                              Day     Night
12.5
9.7
                   Light  Traffic**
                   Day      Night
17.8
8.2
31.9?
31.1
10.5
2.1
2.2
0.3
3-2
81.6
9.3
9.1
18.1
39. 9*
12.8
5.7
0.3
2.1
3.2
2.5
96.8
2.8
0.3
3.1
67.2Jf
1.7
5-2
1.1
<0.01
0.1
3.3
78.6
10.2
11.0
21.2
81. 5*
1.7
2.7
0.1
0
2.6
2.6
91.6
1.3
3.8
8.3
58.6$
1.5
7.1
0
0.6
0
1.1
69-3
7.1
23.6
30.7
55.6?
1.8
1.1
0.1
. 0.8
1.3
1.3
62.6
1.7
32.7
37.1
                                              * Less than  8  vehicles/minute
                                             ** Between 1  and  8  vehicles/minute
     Also to be noted in Table III is the extremely small proportion of motorcycle noise
intrustions recorded at the 20 sites. These proportions can be compared with the interview
results where motorcycles accounted for 23% of the total cases of motor vehicle annoyance!
     CORRELATION  BETWEEN VARIOUS NOISE LEVEL MEASURES  - The correla-
tions that may exist between various noise level measures are of interest since there are
sizable differences in  the complexities determining the various measures. Thus if a simple
measure is very highly  correlated with a more complex  one, the simple measure may often
suffice.  To shed  some light on this measurement  concern, Table IV lists the correlation
coefficient and standard errors  of estimate  for several  noise level comparisons. There is a
high correlation between the  L\Q and L$Q values, with  somewhat lower correlations of the
other measures with the LfQ values.
     CORRELATION OF NOISE MEASUREMENTS WITH INTERVIEW RESPONSES -
As previously described, one of the first questions asked respondents was their rating of the
noisiness of their home and their working environment. Noisiness was evaluated on a 7-point
scale that varied  from "not noisy at all," (rated 1) to  "unbearably noisy"  (rated 7). The
percentage of those responding who classified their neighborhoods as noisy (a rating of 2 or
                                       797

-------
                                       Table 4

                   CORRELATIONS OF SEVERAL SITE NOISE MEASUR ES
                          (DAY. EVENING AND NIGHT VALUES)
Noise Measures
L9Q versus I*5Q
T " T
L10 L50
T n T
Ll "50
NPL « L5Q
TNI » L5Q
Ll " L10
TOI " NPL.
T M T
Leq L50
Correlation
Coefficient,
r
0.97
0.93
0.83
0.65
0.28
0.91
0.82
0.91
Standard Error
of Estimate, d3
2.0
3.0
it. 2
6.9
15-9
2.5
9.6
3.1
more on the 7-point scale) ranged from 67% in Detroit and 68% in Los Angeles, to 77% in
Boston. For individual sites, the noisiness scores ranged from 2.2 to 4.8.
     Comparison of the  site noise levels with the average site noisiness scores shows quite
good correlation between several of the  site noise measures and noisiness scores. Table V
lists the Pearson product moment regression coefficients for several noise measures. Regres-
sion coefficients of the order of 0.6 or slightly better are observed for several of the noise
measures, including L|,  LIQ,  LSQ and NPL. Slightly lower correlation is observed for the
TNI averaged over a 24-hour period, with distinctly lower correlation observed for the TNI
computed separately for day and evening periods.
     Correlation of the same site noise data with the scores for vehicle annoyance (averaged
over the individual sites) show much more scatter and significantly  lower correlation coeffi-
cient values. For example, the correlation coefficient for comparison of site noise levels with
the extensity of vehicle annoyance drops to 0.3  or less. When correlated with the intensity
of the annoyance scores, the correlation coefficients are similarly small, and, in fact, the
slope of the regression line becomes negative! Thus, while the respondent judgments of site
noisiness are in quite satisfactory agreement with site noise measurements, vehicle annoy-
ance scores (considering only site-related motor vehicle noise) are poorly related to the site
noise levels.

CORRELATION OF ANNOYANCE WITH LOUDNESS

     In the analysis of the survey data on judged annoyance, two  factors stand  out. First,
the largest component of variance was intensity of Ibudness, accounting for about a third of
                                        798

-------
                                       Tables

                   CORRELATION OF VARIOUS NOISE MEASURES WITH
                   INTENSITY OF NEIGHBORHOOD NOISINESS RATINGS
Noise
Measure
L50

L10

Ll
NPL

TNI


Tine
Peribd
day
night
21-hr
day
night
21-hr
day
night
day
night
day
night
21-hr
Standard Error
of Estimate, dB
S*k
5.3 dB
6. '9
6.1
4.2
7.1
5-3
1.6
6.7
6.1
9-5
12.7
21.0
10.1
Correlation
Coefficient,
f
0.61
0.57
0.60
0.66
0.61
0.66
0.61
0.66
0.18
0.59
0.17
0.13
0.55
the variance  in  annoyance  (or  about 0.6  of the standard deviation) in almost all cases.
Secondly, the average intensity of annoyance, by situation, however, was sometimes greater
and sometimes less than the grand averages. Closer inspection of these results shows that, by
noise situation, three separate correlations can be made of judged loudness with annoyance.
These data are  shown in Fig. 1. This figure indicates a classification of the situations into
three categories  (although  we cannot be sure that we have correctly identified the factors
that distinguish them):
     1.   Standard - situations where normally anticipated  operations occur,  presumably
         reasonable arid legal use in involved, and  the driver is more or less doing as ex-
         pected. The loudness of the vehicle noise source correlates directly to annoyance.
         In this category are the following:
              Automobiles - exhaust, engine, transmission
              Diesel trucks - heard at home or work
              Motorcycles
              Traffic in general
     2.   Extra-Annoying - situations which are unusually close to the observer when he is
         in  transit, - busses  along curbside,  diesel  trucks while passing another vehicle,
                                        799

-------
                       MAJOR VEHICLE SITUATIONS

                    1  Busses
                    2  Diesel Trucks - Transit
                    3  Auto Tire Squeal
                    4  Auto Exhaust
                    5  Motorcycles
                    6  Diesel Trucks - Home & Work
                    7  Auto - Brake Squeal, Misc.
                    8  Sirens
                    9  Auto - Engine, Transmission
                   10  Gas Trades
                   11  Traffic
                   12  Auto Horn & Other
 u

 I5
<
t*.
 o
 g 4
  2
    2345
                           Loudness Intensity
                            Figure 1.
                               800

-------
         situations  where unusual  abuse of automobiles is involved, presumedly  con-
         trollable and possible illegal - e.g. "peeling rubber."
     3.   Sub-Annoying - situations which are loud, but because of their presumed utility,
         legality, infrequency of occurrence, and location of the auditor, are less annoying
         than their loudness would indicate - sirens, delivery trucks, horns.
     These three  categorizations provide correlation coefficients of 0.98 better for each of
the three categories. Without the categorization, i.e., without assigning a semantic content
to the situation, the gross correlation  of loudness with annoyance drops to 0.73.
     One of the interesting points of this categorization is the similarity in slopes of annoy-
ance versus loudness for the three cases. This permits several assumptions:
     1.   The relative annoyance of  standard category  noises can be intercompared strictly
         on  the basis of noise level and noise exposure.
     2.   Within  a category, a  similar statement can be made for the other cases.
     3.   Alteration of the non-acoustical parameters in the non-standard  cases can be
         translated into an equivalent physical noise reduction or increase.

SELECTIVE  NOISE REDUCTION

     The consequences of these observations and insight  on the situational and attitudinal
aspects  of  noise  provided by  the  survey results suggest  a way of quantifying the relative
effects on  annoyance in mixed traffic situations. Clearly, mandating an equal noise reduc-
tion on all vehicles of, say  5 dB, is an  inefficient way to  improve the environment.
     In  all  the situations described above, within a category, the louder the noise, the more
annoying. Therefore, noise control of the loudest noise sources should be of first  priority.
Since the loudest noise sources are  generally  the smallest in  number in the total vehicle
population, incremental  noise reduction  of these sources might  materially improve the
environment.
     We can  examine several cases to evaluate this effect quantitatively, utilizing the Noise
Pollution Level concept which is specifically designed  to examine  the effect of composite
noise sources on  annoyance. Robinson (9) has suggested a numerical  value of 72  for NPL
expressed in  units of A-weighted sound level  as an  upper limit of acceptability; in his
analysis, this  constitutes an acceptable environment for  about two-thirds of a population.
     Consider the effect of reducing diesel truck noise levels by 5 dB in a freeway situation
of 2000 vehicles per hour moving at an average speed of 50 mph (80 km/h) where the trucks
are 5%  of the mix.  At 100 ft. (30 m) a typical NPL for current automobiles alone is 72.
With a  5% mix of contemporary  diesels  the value of  NPL rises to 83. Reducing the peak
noise of trucks by 5 dB reduces the composite NPL to 72.3, or a level essentially the same as
that for automobiles alone.
     As another example, consider the impingement of diesel truck noise at night on an area
having a relatively low background noise level. At a maximum A-weighted sound level of 80
dB, superposed on a background level of 40 dBA, 30 trucks per hour  at a speed of 60 mph
(96 km/h) generate an NPL value of 94 at  100 ft. One can calculate  that the acceptable NPL
value of 72  would be reached if the maximum level of the truck noise  were  reduced by a
little more than  13 dB. It is worth noting that 30 automobiles per hour would generate an
NPL of about 63.

                                         801

-------
     These examples show that moderate decreases in the noise level of noisy vehicles can
achieve a significant reduction in the overajl noise environment. It is not necessary to reduce
all vehicle noise to levels comparable with automobile noise levels in order to make major
improvements.
     What steps can be taken to improve the acceptability of the "extra-annoying" situa-
tions? Clearly, those noise events caused by misuse or abuse of vehicles are usually beyond
the control of the manufacturer. The annoyance due to noise from busses and diesel trucks,
however, is within his control. Reduction in composite  noise environments  caused by
improved noise control of these events can be computed in a fashion similar to the examples
given above, since the slopes of annoyance and loudness are similar for the "extra-annoying"
and "standard" categories.

                                   SUMMARY

     A social survey has been used to describe twelve motor vehicle generated noise "situa-
tions^ which cause annoyance. Comprehensive noise measurements at the interview sites
and  for selected individual vehicle events indicate that physical description of noise at the
sites has approximately a 0.6 correlation  with annoyance. Segregation of the individual
"situations" into three categories which are dependent on attitudinal aspects of the "situa-
tion** provide correlations between judged loudness and annoyance of 0.97 and higher.

                             ACKNOWLEDGEMENT

     The authors wish to express their thanks to the Vehicle Noise Subcommittee of the
Automobile Manufacturers Association for  their advice and criticism during, the project. A
number of our staff members contributed to the project. Major contributions were made by
Ricarda Bennett and Barbara Freeman in the conduct of the social survey, by Douglas
Dodds and Austin Henderson in its analysis, by Salvatore Pecora  and Chris Paulhus in site
evaluation, and Dwight Bishop, Myles Simpson, and Richard Horonjeff  in planning and
performing the noise measurements.

                                  REFERENCES

 1.  A. Wilson (chmn.), "Noise," Final Report to Parliament, Cmnd. 2056,  HMSO,  1963.
 2.  A. C McKennell and E. A. Hunt, "Noise Annoyance in Central London." U.K. Govern-
     ment Social Survey Report SS. 332,1966.
 3.  J. a Kerrick, D. C. Nagel, and  R. L. Bennett, "Multiple Ratings of Sound Stimuli." J.
     Acoust Soc. Am., Vol. 45, No. 4, 1014-1017, April 1969.
 4.  Tracer, Inc., "Community Reaction to Airport Noise, Vol. 1." NASA CR-1761, July
     1971.
 5.  G. Jones, "A Study of Annoyance from Motor Vehicle Noise." BBN Report No.  2112
     to Automobile Manufacturers Association, June 1971.
 6.  D. Bishop and M. Simpson, "Community Noise Measurements in Los Angeles, Detroit
     and Boston." BBN Report No. 2078 to Automobile Manufacturers Association, June
     1971.

                                       802

-------
 7.  R. Horonjeff and D. Findley, "Noise Measurements of Motorcycles and Trucks." BBN
    Report No. 2079 to Automobile Manufacturers Association, June 1971.
 8.  W. Galloway, "Motor Vehicle Noise: Identification  and Analysis of Situations Con-
    tributing  to  Annoyance."  BBN Report  No. 2082 to Automobile Manufacturers
    Association, June 1971.
 9.  D. W. Robinson, "The Concept of Noise Pollution Level." NPL Aero Report AC 38,
    National Physical Laboratory, March  1969.
10.  I.D. Griffiths and F. J. Langdon, "Subjective Response to Traffic Noise." J. Sound and
    Vibr. Vol. 8, No. 1, pp 16-32, 1968.
                                        803

-------
         SESSION 9

SUMMARY AND INTEGRATION

Chairmen: G. Zarkovic, Yugoslavia
       W.D. Ward, USA
            805

-------
                                     SUMMARY


                                       Ira Hirsh
                                 Central Institute for
                                      The Deaf
                                  St. Louis, Missouri

  Mr. Chairman, thank you for that kind introduction.
  Ladies and gentlemen, let me begin with a preface to my remarks. This is by way of
excuses. First, I received only one of the papers before coming to Dubrovnik, and I received
four papers since coming to Dubrovnik; but then I knew that I would not have time to read
them anyway because I had come here to listen. Thus, this summary will reflect only what
we have all heard in the spoken lectures, and sometimes through translations. Second, there
is a strong  temptation to be a critic or discussant as opposed to  summarizer; I have tried to
resist that  temptation but  I  have not succeeded completely. Third, since no one can be a
specialist in all of the  fields discussed,  the summary may fail to represent an adequate
understanding on my part of some of the lectures. For this, I can only express regret that
the task was not given to someone as broad as Leibnitz.
  I propose to organize the summary in the  following way. We will take the title of the
conference and  turn  it  around, and  ask:  "Is  Noise Exposure a  Health  Problem?"
Consideration  of that question leads to a preceding question, namely, "Are there effects
from noise exposure?", and then,  "Do such  effects constitute a health  problem?" That
question leads  to still another; "Can we give criteria, in the sense that Mr. Suess used the
term, for good health against which such effects can be evaluated?" Then finally,  "Is it
possible to specify levels of noise exposure such that these criteria for good health can be
maintained?"

Auditory effects — laboratory studies

  We consider  first the  auditory effects - the effects of noise on  the auditory system. By
now it is quite clear from laboratory studies and from  field studies that  hearing function
deteriorates during noise exposure (masking) and after noise exposure. We know something
of the mechanism and quite a lot about the physical aspects of noise that will predict the
amount of permanent hearing loss. So there are after-effects. They represent a departure
from good  health, and the main controversies concern how much hearing loss represents an
adequate criterion of good  health, how we calculate exposure over time, and what are the
relations among temporary, asymptotic, and permanent threshold shifts.
  From the laboratory  we have heard many reports of biological studies on the auditory
system. The general paper by Dr. Eldredge reported on both  temporary and permanent
shifts,  observed  changes  of electrical potentials  from the  cochlea and  also electrical
potentials from the  auditory nerve.  He reminded us that the older literature had already
established that high-intensity sounds do  damage hearing functions and damage tissues of
the auditory system. These were demonstrated in other papers - by Dr. Jankowski, and also,
in response to impulses, by Dr. Dieroff. But Dr. Eldredge was interested in bringing to our
attention laboratory studies with animals, where the noise levels were more representative of
those to which man is exposed, lower noise levels and longer exposure durations. The other
novel contribution of his paper was the careful correlation between anatomical, physio-
logical and behavioral studies on the same subjects with respect to temporary threshhold

                                         807

-------
 shift and  asymptotic  threshold shift. Of particular interest  here was  the  asymptotic
 threshold  shift  indicating a  kind  of equilibrium  that obtains for levels of  acoustic
 stimulation, with  a very regular set of rules that  relate this "plateau" or asymptotic
 threshold shift to the stimulus intensity — the same rule, by the way, for the octave-band
 centered at 500 Hz and  the  octave-band centered at 4000 Hz, if only  you correct for
 differences in sensitivity. He is speaking of how damage, as well  as  this equilibrated
 threshold shift, occurs in response to levels on the order  of 65 dB. The anatomical studies
 point to damage,  particularly of outer hair cells,  the thickening of  their walls, and Dr.
 Haider reminded us that the cell damage that is often brought by certain drugs is enhanced
 in the presence also of noise.

     I have included, under biological studies of auditory effects, some of the observations
 of Dr. Haider, particularly the combination of noise  and other physical agents, especially
 vibration. He  appeared  to be rather discouraging  in his report,  since I believe that he
 reported that sometimes  there were additive effects, sometimes synergistic,  sometimes
 antagonistic, sometimes none. In the same context, I should remind you of special kinds of
 acoustic stimulation, the low frequencies about which Dr. Nixon was concerned, and for
 which  we  seem to  tolerate  very  high  levels and very long  durations  without serious
temporary  threshold shift  Similar  conclusions  appear, according  to   Dr. Acton, for
ultrasonic sounds.
   Now before we come to translate these kinds of observations into functional changes in
 hearing, we must be reminded that  changes in hearing - hearing loss or deafness - also
 develop for other reasons. Dr. Spoor told us that the differentiation between noise-induced
 hearing loss and presbyacusis is not always easy, especially since, for the population at large,
 there is considerable overlap .between the distributions of hearing loss that characterize the
 different decades of age. In fact, the upper and lower quartiles of successive decades of age
 almost touch each other, and sometimes overlap. His discussion of presbyacusis, and those
 relations emphasized by Dr. Bochenek, reminded us that there is a close  relation between
 disturbances of the circulatory system and the auditory system, and the two of course
 become very seriously entangled.
   Now we can speak of auditory effects studied psychophysically, that is, in the laboratory
 - audiologically, if you prefer. We didn't hear very much about temporary threshold shift
 studies. These, as you know, are studies  for which our  chairman, Dr. Ward, has become
 famous and his own studies, and  those of others, became the basis  for the  CHABA
 damage-risk curves. The relationships appear to be reasonably secure with some important
 details about frequency and cumulative exposure still  to be worked out. And here we must
 thank Dr. Kraak for his contribution, for his relations between recovery and duration of
 exposure and the mathematical representation that I think moves us forward toward closing
 a gap between temporary and permanent threshold shift.
   I  will not dwell here on issues raised by Dr. Ward's discussion of susceptibility, mostly
 because he did not seem very enthusiastic about the  possibilities. The agreement between
 temporary  and permanent shift is apparently not perfect. Dr. Kylin told  us, for example,
 that young adults studied in the laboratory yielded, from a specified noise exposure, less
 temporary  threshold shift after two minutes than the  permanent threshold shift that would
 have been predicted by the present rules under the ISO Recommendation  1999. Temporary
 threshold shift is also evident in recordings of evoked potentials from the brain, according to
 Dr. Gruberova. But in  addition, she contends that there are additional, more integrative
 functions that heed to be accounted for in those potential changes.


                                        808

-------
Auditory effects — field studies

   Along with the psychophysical studies that emphasize temporary threshold shift, we have
had many reports in this conference about genuine field studies of changes in the hearing of
workers exposed  to  more  or less specified noise levels. Hearing loss  is not merely  of
laboratory interest. The comparability among different studies  in different countries is
probably now improved by  the international agreements on ways of representing the noise
through the ISO Recommendations.
   The diagnostic characterization of workers exposed habitually to noise, as alluded to by
Dr. Sulkowski, reminds us again of some of the difficulties in  separating this group from
other kinds of sensorineural losses. Some of us who thought that Recommendation 1999
was predicting too conservatively were rather surprised by Dr. Raber's findings. He reported
much  lower  incidences  of  hearing  loss  than  would  have  been predicted  by the
Recommendation.  We  are  told  that there  was  considerable discussion about  that
recommendation, and in particular the way in which noise level is cumulated over time. We
won't go back over that in detail, but I would remind you that there are at least two ways of
making that cumulation over time; one that has been accepted internationally, and one that
is proposed, at  least in the United States,  where the trading of intensity  and time is
represented by a ratio of 5 dB per doubling in time as opposed to 3 dB per doubling hi tune.
I was interested to hear in Mrs. Passchier-Vermeer's review that the only justification for the
5-dB trade comes from relatively high-intensity studies, and involves predictions made for
what we now call the speech frequencies 500, 1000, 2000 Hz.
   For a long time acoustic impulses were considered rather special and  outside  of the
general rules that would be given for continuous noise and intermittent noise. But Dr. Coles
seems to tell us that maybe they're less special than we thought, and in fact the changes in
the auditory functions that follow impulse stimulation may be brought also under the
equal-energy rule. In  this  connection, Dr.  Dieroff told of  physiological or biological
reactions to impulses that are quite similar to those of noise  except for some temporal
disparity between the cochlear potentials and the nerve potentials. Similarly, Dr. Kuzniarz
took this impulse work into the field and reported to us on the hearing  losses that attend
the operation of drop forges.

Auditory effects — speech perception

   Now I will return in a  few  minutes to the problem  of hearing loss  following noise
exposure, and in particular to the question of what is "good health" with respect to hearing.
But before I do that, I want to talk about another land of auditory effect: not the hearing
threshold change, but rather the simultaneous reduction or interference with speech percep-
tion, another auditory function.  Here, Dr.  Webster reviewed  for us mostly laboratory
studies. Keep  in mind,  however, that almost all  of the social surveys that were discussed
subsequently quite often mention interference with speech communication or television as
one of the important components of annoyance. I think the only exception was Dr. Gallo-
wayj and it may be that in California people do not listen to speech very much.
   The criterion with respect to speech interference has been too much neglected. More than
any other effect of noise, it can be specifically related to spectrum. Webster told us that the
frequencies in the noise that are most important for predicting speech interference depend
upon the level of speech performance that will be required. That's an important point. If the
criterion is understanding sentences, then octave bands centered  at 500, 1000, 2000 Hz will

                                        809

-------
predict  the failure of intelligibility very well. If there is a higher level of performance -
single-word intelligibility, for example, at perhaps 75 percent - then you would do better to
average  the frequencies at  1000 and 2000 Hz. And if it is a very high level performance,
perhaps 90 or 95 percent intelligibility of one-syllable words, then you should use all four
frequencies, as now recommended  in an international draft recommendation on speech
interference: 500, 1000, 2000 and 4000 Hz.
  We must bear in mind that these considerations are for rather ordinary  speech-hearing
conditions, and are based on acoustic factors alone. Surely, Dr. Tobias demonstrated some
other factors, strategies that listeners can learn to use in difficult communications situations,
and these have not been part of our audiometric calculations.
   There is a clear effect of noise on speech and the question is, "Is it a health problem?"
Certainly, satisfactory speech communication could be brought under the concept of human
well being or welfare. And even if not, even if we must use a health criterion, then I would
remind  you of some of the discussion offered by Dr. Hausman on the psychological effect
of deafness. I do not believe it  appropriate to have brought that discussion to bear upon
persons who lose their hearing due to exposure to  noise, because the authorities whom he
quoted  were talking about the psychological accompaniments of substantial hearing losses,
on the  order of 50 to 60 dB  in the speech frequencies. These are people who, in the
presence of a live talker, do not hear his sounds. These sounds of a live talker are not audible
to such people, much less understandable. Such extensive hearing loss would be unusual in
the individual who is exposed to occupational noise, where we have first high-frequency
losses and, only after a very long time, substantial losses in the speech frequencies. However,
I think that it is proper to use  these psychological considerations with respect to speech
interference in the presence of noise. If it is true, as Dr. Hausman suggested, and as I think
Dr. Herridge implied later  on, that the sense of isolation that results  from inability to
communicate with one's fellow is a serious matter of mental health, then intrusive noise that
makes speech perception impossible becomes a serious health problem. Personally, I believe
that if you look at Webster's figure, when the Conference is published, and think about a
perfectly reasonable criterion for good health  with respect to speech interference, I should
think it would be the ability  to hear one's fellow beings when they are speaking at a distance
of something like two meters.

Hearing impairment

   Let me use this discussion of auditory effects with respect to speech perception to return
to what constitutes hearing impairment,  when hearing is lost as a result of noise exposure.
The present rules, referred to in our discussions by the expression "AAOO" - the American
Academy  of Ophthalmology and Otolaryngology, in the United States - have two items,
both questioned in the discussions presented by Dr. Glorig and Dr. Kryter. One of the two
items is the so-called "low fence", that is to say, that hearing impairment does not exist - is
not significant, if you like, or does not represent bad health - until it exceeds 25 dB on the
average at the frequencies 500, 1000, 2000 Hz. Now I am coming back to these audiometric
frequencies by analogy  to  the  octave-band  systems discussed by  Dr. Webster in his
discussion of speech interference: 25 dB low fence; average hearing loss at 500, 1000, 2000
Hz. Now in point  of fact, and here I am just departing for a moment from my role as
summarizer, those items that were agreed to come from some relations that were not spelled
out completely in our discussion the other day. If you take a large group of hard-of-hearing
patients and  correlate the  hearing losses at  individual pure  tone frequencies  with the

                                        810

-------
 threshold  for.speech intelligibility (that means the levels at which the listeners obtain 50
 percent correct intelligibility), the highest correlation is between that speech threshold and
 the frequencies 500, 1000, and 2000 Hz. That threshold  in absolute terms is about 20 dB
 Or .L.
   Speech  is spoken at a level of about 70 dB at a distance of one meter, although as Dr.
 Kryter pointed out, people are often farther away than 1 meter and not everybody talks at
 70 dB, so  maybe the more appropriate figure is 60. At any rate, the typical level is 40 dB
 above the  normal threshold for speech reception, a fairly good safety factor. You still have
 15 dB left, so  to speak, if you discount  25  dB  hearing loss for speech and call that the
 threshold  of hearing  impairment.  Similarly,  with respect   to  frequency, those three
 frequencies that I mentioned are sufficient to predict the threshold for speech.
   Now the point that's made in that discussion is exactly the same point that Webster made
 in discussing speech interference. Fifty percent of a list of words represents a rather low
 level of intelligibility, but adequate for a high degree of understanding of sentences, which is
 what the committee in the U.S. thought they were predicting. If you wish to provide your
 listener  with  better speech perception than  that, then  you  may have to  include higher
 frequencies. You may have to move toward the conditions that Dr. Webster described as an
 articulation index of .35 or .5 as opposed  to .2. In this case the criterion for health becomes
 a specification of articulation indices. What are you shooting for? What is the goal? Then we
 can  tell you what frequencies to use. I do not believe that 4000 Hz should be included by
 reasoning about speech perception,  when in fact the people who are suggesting that it be
 included are doing so because they want to predict damage to the auditory system, rather
 than damage to speech perception. If you want to predict damage to the auditory system,
 then you  might as  well  use the audiometry proposed by Dr. Fletcher  in the very  high
 frequencies above 8000 Hz and that will tell you about damage to the auditory system even
 sooner. But neither Dr. Fletcher  nor  Dr. Flottbrp nor Dr. Dieroff has told us about the
 relation between those high frequencies and speech perception. (By the way, if we can agree
 that a reasonable criterion for speech intelligibility is being able to hear another talker at a
 distance of 2 meters, then this coincides pretty well with a hearing loss in the quiet of about
 25 decibels. It's the same criterion.)


 General biological effects

  With respect  to more general biological effects,  we thank Dr. Jansen for his review. The
 laboratory  studies that he spoke about involved activation responses that include inhibition
 of gastric juices, lower skin  resistance, modified  pulse  rate, increased metabolism - in
 general,  modifications  in  what  have  been called the vegetative functions, some with
 orienting components. Furthermore, most of these effects are quite clear for noise levels of
 90-95 dBA and above, although some  of them, involving blood volume changes, occur in
 response to noises as low as 60 or 70 dBA. Even here he  notes, as did several others, that
 there is psychological interaction with these vegetative responses to noise, in the sense that
you get changes in the response depending on such things as the meaning of the sound. Dr.
 Kryter's report emphasized  more the change, than the actual level of sound - change either
 from noise to quiet, or change from quiet to noise. In fact his report made me think of an
older psychological  theory (attributed, in our  country at least, to Harry Helson) called
adaptation  level. Any  change represents a departure from what is the customary level of
sensory levelof stimulation.


                                         811

-------
  Dr. Griefahn found it useful to use the ovarian cycle as a controlled variation in general
activity and could conclude on the basis of her study that the associated autonomic nervous
system responses are smaller as the general autonomic level is higher. We had also a report
from Dr. Markiewicz on chemical effects in the laboratory, and here he and we were reminded
that the rat may be more active in this regard than man. It's quite clear that his interest in
that kind of study and others like it is not so much in providing a direct assessment of these
effects in man but rather studying the mechanism whereby these changes take place.
  As we move  from these general biological  effects to  the  question  of whether they
represent a health problem, we find considerably more by way of field studies than we did
some time ago. There may be correlates of these general biological effects  that become
health problems. Dr. Bochenek, for example, reminded us about the incidence  of heart
disease in persons exposed to noise. We have  Dr. Cohen's report on the variety of medical
symptoms as well as the  increase in accident rate for those exposed to noise, and, again, the
difference for those exposed to levels above 90  or 95 dB and those exposed below. On the
other hand, Dr. Carlestam found  some trouble showing great  effects, either in fatigue or
adrenalin in comparable exposures. The problem is again the relation between these effects
and any long-term cumulative ones. When I say  this 1 wish I had before me the proceedings
of the first of these Conferences, held in the  United States in 1968, where  I think we had
the same sentence, that what we need are cumulative long-term studies.


Sleep

  We were surprised by some of the material hi the session on sleep interference as another
effect. I thought that things were quite stable in that speciality, but now I find that such
concepts as levels of sleep, as represented by  stages hi EEC records, are being questioned.  I
was particularly interested to listen, as perhaps some of you were, to Dr. Williams' emphasis
on  individuals' variation,  relative to arousal by noise, habituation, quality and meaning of
the sounds, and such things as the fact that infrequent sounds will arouse at lower levels
than frequent sounds. He warned  us not to assume that, since sounds must be well above
threshold to awaken  persons,  the  sensory system itself was asleep, but  rather that more
central processes were involved. We heard about sleep responses after exposures to aircraft
noise from  Dr.  Lukas,  to sonic  booms from Dr. Collins, and during  many  days of
stimulation by brief pure  tones called "pings" by Dr. Johnson. The EEC indices appear to
be  somewhat ambiguous  with respect to dimensions of noise exposure.  But what  was
interesting in the point of agreement, I thought, between Dr. Johnson and Dr. Muzet was
that despite variability in  the physiological indicators, like the EEC records, the one thing
that stood out  was the  subjective report, "I find difficulty  in getting to sleep." What
becomes interesting about that,  at least for legal or governmental authorities, is that that
becomes the basis of annoyance  much more clearly than any physiological response of
which  the  complainant is unaware. Dr.  Friedman's interesting comparison  of noise and
non-noise couples gives a quite clear indication that airport noise involves less time in deep
sleep and more time in light sleep and more awakening. "Is this a health problem?" again we
ask. Surely,  there is an effect. People can be awakened, and this can be studied with sleep
indicators similar to those mentioned already.
  Dr.  Herbert,  in reporting his  experiments on  psychological  performance  after  sleep
deprivation, mentioned interesting but small interactions, but no  major decrements in
performance. Here again we must note that he studied that behavior only over a period of a


                                        812

-------
few days, and did not study subjects who had been awakened over a period of weeks or
months.

Human performance

  We now  come  to  the psychological  or performance aspects  of noise,  whether a
performance undertaken during noise exposure, or following the noise, is affected. We have
to thank Dr. Gulian for a very good review of that experimental work. I personally thank
her also for including annoyance under these psychological aspects. As you know, it has
been difficult over the years to show clear decrements in performance due to noise. And so
both the public and the engineers and the government authorities who are asked to consider
these aspects of noise say, "The psychologists don't know how to do it or they don't know
what they're talking about, because this one gets a positive result, that one a negative result,
this one gets zero results." I think that the field is now becoming clear, due to workers like
Dr. Gulian and others who can tell us what are the important variables.
  The  tasks themselves  turn out  to be very important: whether one  measures speed or
accuracy, whether the  task involves high levels of thought or cognition, whether the task
itself is represented by a high level of complexity with respect to, for example, the number
of stimuli among  which a subject must choose. Then the noise itself turns out to be
important, not only with respect  to its acoustical or physical aspects, but such things as
whether it is continuous or intermittent, whether it has predominantly high frequencies or
low (and a very unfortunate interaction between those first  two terms), and finally, the
meaning of  the noise. Then there are still other features, as pointed quite clearly by Dr.
Broadbent, like the level of arousal and its role in the task being undertaken. There are some
tasks where arousal represents something more  like  distraction, where the performance
shows a decrement, but there are others where performance will benefit from increasing the
state of arousal. There is certainly a lack of unanimity among these results, and I interpret
the absence of unanimity as meaning that this kind of study must continue. Certainly the
reports by  Dr. Hartley, Dr. Harris and Dr. Broadbent serve as examples of the  interaction
among these kinds of variables.
  Within that same group, I think all of us were impressed with the way in which Dr. Glass's
laboratory study could point out the role of controllability of the one's environment or the
feeling that  "I can do something about it." That was under laboratory control and appeared
to reflect very well the importance of this item, which is called "misfeasance" by some of
the survey people  when  they  talk about  the  important  non-acoustical  variables in the
responses of subjects.  This seems  quite  close  also to the conception that  Dr.  Herridge
introduced in his studies of psychiatric symptoms of the residents of the Heathrow ghetto.
  Study of these  three groups of effects, the general biological, sleep and psychological
performances  will undoubtedly continue. They do not seem ready for incorporation in
health criteria; they probably represent some of the bases of what are annoyance reactions
in surveys and  thus they remain important, practically.
  Now, it is with some timidity that  I approach the subject of annoyance and what Dr.
Robinson called  "general  auditory reaction" in the last discussion. But  Dr. Ward in  his
introduction had more trepidation than I do; in fact he seems to worry about things like a
person's feelings. Yet that's what annoyance is all about  - people's feelings - and this is
manifested in questionnaires and interviews that comprise the basis of surveys.
  There  are three important problems,  I  believe. One has  to do with  consistency and
meaning of annoyance  responses;  the second has to do with the dependence  of these

                                         813

-------
responses on non-acoustical variables like  fear; and the third  has  to  do with the best
predictive combination or calculation of level, duration and number of occurrences.
   For the first two items, Dr. Jonsson was quite discouraging, at least about the present. He
would like to extend the methodology, and perhaps I infer too much, but he seems to say
that that is not quite ready yet for utilization by authorities.
   Dr. Alexandre  was  somewhat  more  sanguine. He finds, for example,  considerable
agreement among European surveys that relate annoyance scores to exposure. But we were
all reminded of the importance of such psychological  factors as fear, the feelings about who
is responsible for noise and so on, by a variety of our survey reports: Dr. McKennell, Dr.
Leonard, Dr. Grandjean and his colleagues and  Dr. Rohrrnann as well.
   You have to be careful, I think, about the conclusions that you might make about some
of these reports. If you find that there is  a low correlation between annoyance response
(however defined) and exposure  level,  that  does not  mean that  exposure  level  is
unimportant or that we can forget about noise control. The alternative meaning is that if
you want to  predict  annoyance reactions precisely with  correlations of 0.9,  then the
psychosocial factors must  be added  to the noise levels for those predictions. For  noise
control.and for planning we can work with low correlations that average across great ranges
of these psychological  factors in large groups of individuals. We can work with the kind of
average relations displayed hi Dr.  Alexandra's slides.  Furthermore, you know, we can
legislate about the noise level, but we cannot legislate the incidence of fear.
   Now, as for the best representation of noise exposure level or index, there are several
problems  that are quite similar to those involved when we attempt to specify damage risk
with respect to hearing loss.  Dr.  McKennell noted  in his  summary that the number of
operations per day predicts pretty well the annoyance of response, without regard to level.
Contrarily, Dr. Sorensen told  us that number is  most useful when levels are high, and
Rylander  finds that level alone is sufficient after the number reaches a high value. I suppose
we have to remind ourselves,  as we consider  each  report, that the correlation between
annoyance and any one variable, "X", is always low whenever the range of variation in "X"
is small.
   Now with respect to the weighting coefficient that should be attached to  a number of
operations,  whether it  is 6.6 or 10 or 15 times the logarithm, that seems a little  more
difficult  to  select.  Authorities have selected for us in a sense. You  see the difference
between this coefficient and that one, and such things as which weighting is best, can only
be  addressed by  the  relatively large  and  admittedly  clumsy surveys done  on  large
populations. We have been  told that the variability associated with such  techniques is quite
high. Surveys represent, therefore, rather insensitive measures to changes in such things as
weighting, whether perceived  noise level or  A-weighting. Laboratory methods can  show
differences reliably. But in  the case of the logarithm  of a number  of flights and its
coefficient, you cannot take a laboratory judgment  of something that must extend over
months. The effects of such extended numbers over time can be told only in the sort of
attitude that gets built up over time, and is revealed in  the survey techniques.
   Dr.  Robinson's emphasis this morning, on  the  necessity  that fluctuation must also be
introduced into this overall measure of noise exposure, is agreed to by many. For example,
Dr. Eldred, who gave us a  review of some of these calculation schemes, reminded us that
among some  of the first ones,  background level was considered very important. It's not
quite exactly the same  as a fluctuation measured in terms of the standard deviation, but at
least the two take care  of the same omission from more  frequently-used schemes. The
evaluation of an improvement in prediction by Noise Pollution Level and its inclusion of

                                       814

-------
fluctuation should continue. It should take place under the study of scientists of various
sorts, but I am not sure that other procedures should be abandoned  in the presence of its
candidacy, because we can see now some influence of standardization and use of standard
rules by the  authorities, and I don't think that should be undone,  even if some of these
procedures might be changed for the better in the future.
  Now it may be silly to summarize a summary but let me remind you of several general
observations. There appear to be two groups of levels of concern to us. In the region of 50
to 60 dBA  there ar,e clear changes demonstrated in the  laboratory as after-effects of
stimulation. In addition, that region represents a marginal one for speech interference. Then
we  have, several tens of decibels  above  that, another region - 90-95  dB  - where the
biological effects become clear and where the auditory system can be permanently damaged.
In fact, that  90 dB, which has been used as kind of a floor in the United States with respect
to occupational hearing loss, is probably  too high, in view of some of the reports we've
heard at this conference about permanent loss following such noise exposure. It was the
alpha and the  omega of the  conference that  attracted most of our attention  -  that is,
concern  with hearing loss and speech interference in the first day, and with annoyance in
the last  days. I don't know whether this is because those two represent fertile fields for
argumentation about  how to define noise exposure, or whether those two are most clearly
subject to interpretation as health problems. My guess is that it is the latter: the hearing loss
is clear, the annoyance and  its accompanying feelings - particularly that were described to
us by our psychiatric colleagues yesterday - are serious.
  The  work on psychological  performance, on  biological  effects, on  changes in sleep
records,  and  so on -  that will continue. It is of great interest to the persons  who work in
those fields,  who understand each other  well, and appreciate this or  that detail being
brought  under control. They do not appear yet to provide us with the kind of sure evidence
that will be a matter of concern and control by authorities. It is important that we stand on
past work, as it has been applied. We must not kick it aside as we take further steps forward.
                                         815
                                                    4U.S. GOVERNMENT PRINTING OFFICE: 1974 S46-314/201 1-3

-------