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 -
) 80
UJ
CD
Ld
o
(/) 60
50
40
30
20
10
I st Session
PULSED
MASKED
I
1
I
0
10
20
TIME
IN
30 40
MINUTES
50
Figure 6. Mean learning curves for subjects who received monetary incentives.
organizing an attack upon the problems of learning to understand difficult speech. Indeed,
they hear very little that they recognize as being any kind of speech.
We cannot be certain that the apparent separation of subjects into exactly two groups
is correct, although it looks to be. And we have no testable explanation for why the original
differences in intelligibility exists. But it seems clear that some listeners do start off with
higher scores. Are they the sort who have learned to listen to dialects, perhaps? Whatever
the initial reason, after a listener has some success in pulling intelligence out of noise, he can
53
-------
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 >
^ «,
Walker (1969)
Cohen et al
(1966)
Cole« and Rice proposal! (1970)
J_
10
100 1000
Number of impulses
10000
100000
10 imp/sec
8 houri/day
Figure 2. Revised and extended corrections for numbers of impulses per exposure occasion. (After Coles
and Rice, 1970).
213
-------
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
///
Y//S.
/noise
///A
012345678
hours
Figure 4. Two hypothetical noise exposure patterns.
216
-------
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.
• \
A l.et.1
f 1 (^ |...tf
Figure 1: Warsaw.
724
-------
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
------- |