\
*
I
\
ss>
Of
NOISE EMISSION
MEASUREMENTS FOR
REGULATORY PURPOSES
NBS HANDBOOK 122
U.S. DEPARTMENT OF COM MERGE/National Bureau of Standards
in cooperation with
U.S. ENVIRONMENTAL PROTECTION AGENCY/Office of Noise Abatement
and Control
-------
NATIONAL BUREAU OF STANDARDS
The National Bureau of Standards1 was established by an act of Congress March 3, 1901. The Bureau's overall goal is to
strengthen and advance the Nation's science and technology and facilitate their effective application for public benefit. To this
end, the Bureau conducts research and provides: (1) a basis for the Nation's physical measurement system, (2) scientific and
technological services for industry and government, (3) a technical basis for equity in trade, and (4) technical services to pro-
mote public safety. The Bureau consists of the Institute for Basic Standards, the Institute for Materials Research, the Institute
for Applied Technology, the Institute for Computer Sciences and Technology, the Office for Information Programs, and the
Office of Experimental Technology Incentives Program,
THE INSTITUTE FOR BASIC STANDARDS provides the central basis within the United States of a complete and consist-
ent system of physical measurement; coordinates that system with measurement systems of other nations; and furnishes essen-
tial services leading to accurate and uniform physical measurements throughout the Nation's scientific community, industry,
and commerce. The Institute consists of the Office of Measurement Services, and the following center and divisions:
Applied Mathematics — Electricity — Mechanics — Heat — Optical Physics — Center for Radiation Research — Lab-
oratory Astrophysics2 — Cryogenics4 — Electromagnetics2 — Time and Frequency3.
THE INSTITUTE FOR MATERIALS RESEARCH conducts materials research leading to improved methods of measure-
ment, standards, and data on the properties of well-characterized materials needed by industry, commerce, educational insti-
tutions, and Government; provides advisory and research services to other Government agencies; and develops, produces, and
distributes standard reference materials. The Institute consists of the Office of Standard Reference Materials, the Office of Air
and Water Measurement, and the following divisions:
Analytical Chemistry — Polymers — Metallurgy — Inorganic Materials — Reactor Radiation — Physical Chemistry.
THE INSTITUTE FOR APPLIED TECHNOLOGY provides technical services developing and promoting the use of avail-
able technology; cooperates with public and private organizations in developing technological standards, codes, and test meth-
ods; and provides technical advice services, and information to Government agencies and the public. The Institute consists of
the following divisions and centers:
Standards Application and Analysis — Electronic Technology — Center for Consumer Product Technology: Product
Systems Analysis; Product Engineering — Center for Building Technology: Structures, Materials, and Safety; Building
Environment; Technical Evaluation and Application — Center for Fire Research: Fire Science; Fire Safety Engineering.
THE INSTITUTE FOR COMPUTER SCIENCES AND TECHNOLOGY conducts research and provides technical services
designed to aid Government agencies in improving cost effectiveness in the conduct of their programs through the selection,
acquisition, and effective utilization of automatic data processing equipment; and serves as the principal focus wthin the exec-
utive branch for the development of Federal standards for automatic data processing equipment, techniques, and computer
languages. The Institute consist of the following divisions:
Computer Services — Systems and Software — Computer Systems Engineering — Information Technology.
THE OFFICE OF EXPERIMENTAL TECHNOLOGY INCENTIVES PROGRAM seeks to affect public policy and process
to facilitate technological change in the private sector by examining and experimenting with Government policies and prac-
tices in order to identify and remove Government-related barriers and to correct inherent market imperfections that impede
the innovation process.
THE OFFICE FOR INFORMATION PROGRAMS promotes optimum dissemination and accessibility of scientific informa-
tion generated within NBS; promotes the development of the National Standard Reference Data System and a system of in-
formation analysis centers dealing with the broader aspects of the National Measurement System; provides appropriate services
to ensure that the NBS staff has optimum accessibility to the scientific information of the world. The Office consists of the
following organizational units:
Office of Standard Reference Data — Office of Information Activities — Office of Technical Publications — Library —
Office of International Standards — Office of International Relations.
1 Headquarters and Laboratories at Gaithersburg, Maryland, unless otherwise noted; mailing address Washington, D.C. 20234.
' Located at Boulder. Colorado ROV»
-------
Noise Emission Measurements for
Regulatory Purposes
D. R. Flynn
W. A. Leasure, Jr.
A. I. Rubin, and
M. A. Cadoff
National Bureau of Standards
Washington, D.C. 20234
In cooperation with the
U.S. Environmental Protection Agency
Office of Noise Abatement and Control
U.S. DEPARTMENT OF COMMERCE, Juanita M. Kreps, Secretary
Dr. Betsy Ancker-Johnson, Assistant Secretary for Science and Technology
NATIONAL BUREAU OF STANDARDS, Ernest Ambler, Acting Director
Issued March 1977
-------
Library of Congress Cataloging in Publication Data
United States. National Bureau of Standards.
Noise emission measurements for regulatory purposes.
(NBS Handbook; 122)
Supt. of Docs, no.: C13.11:122
1. Noise—Measurement. I. Flynn.D.R. II. United States. Environ-
mental Protection Agency. Office of Noise Abatement and Control.
HI. Title. IV. Series: United States. National Bureau of Standards.
Handbook; 122
QC100.U57 no. 468 [TD894] 620.2'3 76-608406
National Bureau of Standards Handbook 122
Nat. Bur. Stand, (U.S.), Handb, 122,193 pages (Mar. 1977)
CODEN: NBSHA
U.S. GOVERNMENT PRINTING OFFICE
WASHINGTON: 1977
For sale by the Superintendent of Documents, U.S. Government Printing Office
Washington, D.C. 20402 - Price $2.60
-------
Abstract
A review is given of the measurement needs attendant to regulation of
the noise generated and emitted by commercial products. The emphasis is
primarily on measurement procedures for use in conjunction with point-of-sale
regulations as opposed to regulations on the noise which a source actually
emits when in operation. The report is divided into three major parts. Part
I is a discussion of overall measurement requirements and the type of data
and information which are needed in order to promulgate regulations "based on
appropriate measurement techniques. Part II is designed as a checklist for
the evaluation of the suitability of a noise measurement standard for a
particular class of products or, in the absence of a suitable standard, as a
framework for development of one. The intent is to identify and discuss in
some detail those factors which can impact on the accuracy, precision, and
applicability of a noise measurement process. Part III consists of a series
of flow charts depicting the development of appropriate procedures for the
measurement of product noise emission.
Key Words: Acoustics; environmental pollution; machinery and equipment;
noise; noise abatement and control; noise emission; regulation;
sound.
iii
-------
Table of Contents
Page
Acknowledgements xiii
PART I: Overall Requirements 1
1. Introduction . 1
2. Measurement Methodology Development -— Overview h
2.1. Product Classification k
2.2. Normal Use Conditions k
2.3. Identification of Major "Effects" Parameters 4
2.U. Physical Measurements Correlated with-Major Effects. . h
3. Relationship of Measurement to Regulations 5
3.1. Role of Measurement 5
3.2. Measurement Requirements for Regulation 5
3.3. Measurement Uncertainty Considerations in Establishing
Regulatory Noise Limits 6
h. Informational Requirements for Setting Noise
Regulations: Product Classification Scheme 9
5. Informational Requirements for Setting Noise
Regulations: Operating Procedures and Environment 12
5,1. Typical Usage Data 12
5.2^ Effect on Noise Levels 12
a.. Operation 12
b.. Environment lU
c. Installation ill
6. Informational Requirements for Setting Noise
Regulations: Effects of Noise on People.. 16
6.1. Parties Affected 16
a. Operator 16
b. Passenger 16
c., Neighbor . . 16
d. Bystander 16
6.2. Nature of Effects 17
a. Hearing Loss 17
"b. Task Interference 17
c. Annoyance 17
7. Quantitative Description of Noise . 18
7..1. Identification of Major Parameters 18
a.. Relationship to Effects 18
b. Environment/Use Conditions 18
-------
Page
7.2. Measurement Parameters .-....,.,. v ^ ...» 19
a. Sound Level. ....„..' „.*..*„.' 19
t>. Spectral Quality .......... 19
c. Temporal Variation 19
8. Consensus Procedures for Measurement Standards and
Rating Schemes 21
8.1. Input from All Affected Parties ...... 21
8.2. International Trade ! ! ! ! . 21
9- Environmental Protection Agency Promulgation of
Regulations 22
9.1. Special Requirements !!!!!!!!! 22
a. Product Noise Emission Standards ......... 22
b. Labeling f 4 22
c. Low-Noise-Eraission Product Certification Procedures 23
9.2. Technical, Economic, and Administrative
Feasibility as It Relates to Measurement 23
9.3. Compatibility with Other Regulations and Needs .... 23
10. RecOTended Operational Procedure for Generating Measurement
Methodologies for Specific Product Classes 25
References for Part I 28
PART II: Contents of Measurement Standards . 29
11. Purpose and Applicability 30
11.1. Introduction 30
11.2. Scope and Purpose ! 30
11.3. Definitions and References . 31
11.4. Acoustical Quantities to be Measured ! 31
11.5. Applicability * * ' 35
a. Types of Noise .....!*.! 35
b. Nature and Size of Source.* ...... 3^
c. Measurement Uncertainty, !!!!!!'.!'*'! 38
12. Acoustic Environment 1^1
12.1. General Requirements 1^1
12.2. Criteria for Adequacy of the Test'Environment! ! ! . ^
a, Anechoic Environment U3
b. Hami-Anschoic Environment! . . . . . h&
c. Reverberant Environment. \ 53
d. In Situ ] 55
12.3. Criteria for Background*Noise! ! 68
vi
-------
Page
12. U. Criteria for Temperature, Barometric Pressure, Humidity,
and Wind TO
a. Temperature TO
b. Barometric Pressure Tl
c. Humidity T5
d. Wind TT
12.5. Criteria for Size of the Test Equipment TT
12.6. Criteria for Reflecting Surfaces TT
13. Instrumentation for Noise Measurements T8
13.1. General Requirements T8
13.2. Microphone and Ca"ble T9
13.3. Frequency Response of the Instrumentation System T9
13. U. Weighting Network and/or Frequency Analyzer 80
13.5. Signal Detection and Averaging 8l
13.6. Read-Out Device 83
13.T. Transient Response of Instrumentation System 85
13.8. Calibration and Maintenance of Instrumentation System . . 85
13.9. Precautions to be Taken When Selecting Instrumentation. . 86
Ik. Installation and Operation of Source 89
lU.l. Source Location 89
lU.2. Source Mounting and Installation 91
lU.3. Auxiliary Equipment 92
lU.U. Operation of Source During Test 9^
lU.5. Loading of Equipment During Test. .... 96
15. Measurement Procedures 98
15.1. General 98
15.2. Microphone Positions 112
a. Anechoic Space 112
b. Hemi-Anechoic Space 113
c. Reverberant Space 119
15.3. Source Positions 130
15.U. Period of Observation 131
15.5. Use of Diffusers 131
15.6. Background Noise Measurements 132
15.T. Characterization of Test Environment 133
a. Anechoic Chamber 133
b. Hemi-Anechoic Environment 133
c. Reverberant Environment 13^
15.8. Calibration 135
16. Calculation Procedures 136
16.1. Correction for Background Noise 136
16.2. Correction for Test Environment 13T
l6.3. Determination of Mean-Square Pressure 13T
16. k. Calculation of Sound Power 13T
16.5. Calculation of Noise Rating 1ST
16.6. Calculation of Measurement Uncertainty 13T
-------
Page
IT- Information to be Recorded .................. 139
17.1. Sound Source Under Test ................ 139
17.2. Sound Source Installation Details ........... 139
17.3. Sound Source Operating Procedures ...........
17.^. Acoustic Environment ..................
17.5. Instrumentation ....................
17-6. Special Measurement Procedures .............
17.7. Calibration History ........ .. .........
17.8. Acoustical Data and Related Information ........
17.9. Special Calculation Procedures .............
17.10. Measurement Uncertainty ................
18. Information to Be Reported .................. lU2
18.1. Identification of Source ................
18.2. Source Installation and Operating Procedure ......
18.3. Deviations from Standard Measurement Procedures . . . .
18. H. Acoustical Data and Related Information ........
l8.5. Measurement Uncertainty :. . . ............ .
References for Part II
PART III: Selection of Measurement Methodology Appropriate to a
Specific Product ....................... 153
Appendix A. List of Participants at Government/ Industry Meetings
on Noise Measurement Methodology for the Environmental
Protection Agency's Noise Emission Regulations ........
Appendix B. Pertinent Sections of the Noise Control Act of 1972 .....
Appendix C. Uncertainty of Measurement ................. l66
Appendix D. Possible GATT Code of Conduct for Preventing Technical
Barriers for Trade ....... .............. l69
References for Appendix D ...... . .......... 172
Appendix E. Methods of Labeling ................ .... 173
References for Appendix E ..... . ...... ... l80
viii
-------
List of Figures
Figure 1. Development of measurement methodology needed for
implementation of Public Law 92-57U: Sections 6, 8, 15,
IT and 18 ........................... 2
Figure 2. Recommended operational procedure for generating measure-
ment methodologies ...................... 26
Figure 3. Schematic representation of a sound source in an anechoic
chamber ............................ i+5
Figure h. Maximum range of uncertainty in the sound pressure level
in an anechoic chamber as a function of II and II , defined
in the text (for a single -wall, normal incidence; ....... ^7
Figure 5« Hypothetical variation of sound pressure level versus
distance as one wall of an anechoic chamber is approached. . . ^9
Figure 6. Schematic representation of a sound source in a free field
above a reflecting plane .......... » ........ 51
Figure 7. Limits for the effect of the image source on the observed
sound pressure level as a function of the absorption
coefficient of the reflecting plane with r/r' as a parameter. 52
Figure 8. Difference between the total sound level and that due to the
reverberant field alone, shown versus the effective distance
from the source ........................ 57
Figure 9. The effect of neglecting the 10 log (l-a ) term in eq. (l6) . . 59
Figure 10. Approximate number of normal modes in a 1/3-octave frequency
band for cubical rooms of the volume shown. ......... 6l
Figure 11. Frequency response of a reverberation room having equal-
strength modes at 95s 100, and 105 Hz for reverberation
times of 1 and 20 s, respectively ............... 63
Figure 12. Range of the variation (peak-to-valley in Figure ll) in
the frequency response of a reverberation room for fre-
quencies near 100 Hz as a function of reverberation time
for different modal spacings ................. 6k
Figure 13- Drawing illustrating the spacing, 6f, between normal modes,
the width, Af , of a normal mode, and the relationships among
Af, f /Q, T, and T ...................... 66
ix
-------
Figure lU. Pressure dependence of the acoustic resistance of a.ir ..... 73
Figure 15. Atmospheric absorption loss in air as a function of temper-
ature for relative humidities of 10 and TO percent at
frequencies of 1000 and ^000 Hz ................ 76
Figure 16. Schematic representation of instrumentation for sound level
measurements ......................... 78
Figure 17. Confidence intervals as functions of BT for measurements on
random noise ......................... 8U
Figure 18. Schematic representation of a rigid circular piston of
radius, a, contained in an infinite "baffle .......... 103
Figure 19. Variation of the axial intensity level due to a baffled
rigid piston of radius, a, vibrating at an angular fre-
quency, a). . . ........................ 105
Figure 20. Far-field radiation patterns showing the directionality for
a baffled rigid piston .................... 106
Figure 21. Axial intensity level due to an incoherent circular source. . 107
Figure 22. Schematic representation of sound radiation from an infinite
plate in vhich there is a plane bending vave propagating at
speed, CB ........................... 107
Figure 23. Far-field intensity level above the critical frequency for
an infinite plate in flexure ................. 110
Figure 2*t. Intensity level parallel to an infinite plate in flexure
at frequencies below the critical frequencies ......... Ill
Figure 25. A suggested set of continuous microphone traverses for
determination of sound power in a free field above a
reflecting plane ............ .. . . ........
Figure 26. Effect of bandwidth on directivity of sound field of a
point source at a distance, h, from a reflecting plane. _ _ ^
Figure 27. Error introduced by limited number of measuring points
when determining sound power output of a random point
source near a reflecting plane ..... . .......... 117
Figure 28. System for microphone traverses along meridional paths. . . . 118
Figure 29. Microphone array for a parallelepiped measurement surface. . . 120
-------
Figure lU. Pressure dependence of the acoustic resistance of Ri.r. . • -
Figure 15. Atmospheric absorption loss in air as a function of temper-
ature for relative humidities of 10 and 70 percent at
frequencies of 1000 and UOOO Hz
73
Figure 16. Schematic representation of instrumentation for sound level
measurements ......................... '
Figure 17. Confidence intervals as functions of BT for measurements on
random noise ......................... °
Figure 18. Schematic representation of a rigid circular piston of
radius, as contained in an infinite baffle .......... l0^
Figure 19. Variation of the axial intensity level, due to a baffled
rigid piston of radius, a, vibrating at an angular fre-
quency, u. . .
Figure 20. Far-field radiation patterns shoving the directionality for
a baffled rigid piston ........ . ........... I0b
Figure 21. Axial intensity level due to an incoherent circular source. • l0^
Figure 22. Schematic representation of sound radiation from an infinite
plate in which there is a plane bending wave propagating at
speed, c_ ................ ....... 107
B ......
Figure 23. Far-field intensity level above the critical frequency for
an infinite plate in flexure ................. ll0
Figure 2k. Intensity level parallel to an infinite plate in flexure
at frequencies below the critical frequencies
Figure 25- A suggested set of continuous microphone traverses for
determination of sound pover in a free field above a
reflecting plane ................
Figure 26. Effect of bandwidth on directivity of sound field of a
point source at a distance, h, from a reflecting plane. m . . 11 6
Figure 27. Error introduced by limited number of measuring points
when determining sound power output of a random point
source near a reflecting plane ..... t ...... » . . • H?
Figure 28. System for microphone traverses along meridional paths. . • •
Figure 29. Microphone array for a parallelepiped measurement surface. . •
-------
Figure 30. Microphone array for a "composed" measurement surface 121
Figure 31. Upper limit on the number of microphone positions in
rooms of the volumes shown if each position is at least
a half-wavelength from all other positions 122
Figure 32. Normalized variance of the mean-squared sound pressure in
a room having a reverberation time, T, and excited "by
random noise of "bandwidth, B 12U
Figure 33. Minimum number of microphone positions needed for 1/3-octave
hands of noise in order to have 95 percent confidence that
the spatial average of the mean-square sound pressure is
known within +1 dB 125
Figure 3^. Normalized variance of the mean-squared sound pressure in
a room, having a reverberation time, T, excited by M equal-
strength pure tones uniformly separated in frequency by 6f. . 128
Figure 35. Development of appropriate test procedure . . . 15U
Figure 36. Development of appropriate classification scheme 155
Figure 37. Identification of operational modes indicative of product
usage 156
Figure 38. Collection of noise effects data 157
Figure 39- Selection of appropriate measurement locations 158
Figure 1*0. Selection of appropriate test environment criteria 159
Figure Ul. Sample of the energy labels now found on many room air
conditioners
-------
Acknowledgements
The National Bureau of Standards gratefully acknowledges the assis-
tance of many people and organizations who contributed to the preparation
of this report. Their contributions have made it possible to produce
this document. The following groups and individuals deserve special recogni-
tion.
Between July 20, 1972 and December 20, 1972, a series of meetings was
conducted by NBS with representatives of governmental and industrial
organizations to obtain inputs prior to the development of this report.
During these meetings we discussed special problems and reviewed the existing
data base, including present noise levels, typical operational modes and
use statistics. Finally, we tried to identify the critically important
information necessary to evaluate the implications of noise measurement
methodologies. These include problems associated with the technical capa-
bility of an industry to use a particular measurement methodology and the
possible economic impact associated with alternative measurement require-
ments . We thank the following groups for their important contributions,
which in many instances included a written critical review of the initial
draft of this report.
o Acoustical and Insulating Materials Association
o Air Conditioning and Refrigeration Institute
o American Road Builders Association
o American Short Line Railroad Association
o American Trucking Associations
o Association of American Railroads
o Association of Home Appliance Manufacturers
o Construction Industry Manufacturers Association
o U. S. Department of Transportation (Federal Railroad
Administration and the Office of Noise Abatement)
o Engine Manufacturers Association
o General Services Administration (Public Building
Services)
o International Snowmobile Industry Association
o Motor Vehicle Manufacturers Association
o National Electrical Manufacturers Association
o SAE Motorized Snow Vehicle Subcommittee
o U. S. Postal Service
We would further like to recognize the significant contributions
of the following five acousticians for their critical review of the initial
draft of the report and their many useful and timely suggestions.
Dr. Erich K. Bender
Bolt Beranek and Newman, Inc.
-------
Mr. George M. Diehl
Ingersoll-Rand Company
Dr. Tony F. W. Embleton
National Research Council (Canada)
Mr. George W. Kamperman
Kamperman Associates, Inc.
Dr. William W. Lang
IBM Corporation
Joseph M. Cameron, Joan R. Rosenblatt, and Hsien H. Ku provided very
valuable input to the discussions of measurement uncertainty.
Particularly helpful guidance in the early stages of preparation of this
report was provided by Mr. John C. Schettino and Mrs. D. Elizabeth Cuadra of
the U. S. Environmental Protection Agency Office of Noise Abatement and
Control.
In addition we would like to acknowledge the editorial contributions of
Mr. Edward Case.
xiv
-------
PART I OVERALL REQUIREMENTS
1. Introduction
This report is the result of a project initiated, in May 1972, "by
the National Bureau of Standards at the request of the Office of Noise
Abatement and Control, U. S. Environmental Protection Agency, in
anticipation of the eventual passage of legislation authorizing the
Environmental Protection Agency to promulgate noise emission standards
and labeling requirements.
The intent of the project was to develop a working document to be
utilized by the Environmental Protection Agency, or its designated
agents, in the development of measurement methodologies for noise
emission or labeling standards. The questions that need to be addressed
during the development of such methodologies are presented here with
their appropriate technical backup. The specific directions of effort
and emphasis were developed in response to:
- Discussions with Environmental Protection Agency
personnel
- Meetings with trade associations
- Meetings with standards organizations
- Literature review
- Consultations with acoustical experts
- Legal requirements of the Noise Control Act of 1972.
A listing of the meetings with other organizations, including the individual
participants and the companies or agencies they represented at the meetings, is
given in Appendix A.
The general strategy of this report is depicted in Figure 1.
The Noise Control Act of 1972, Public Law 92-57^, vhich was signed into
law on October 27, 1972, requires or authorizes the Administrator of the
Environmental Protection Agency to control the emission from noise sources that
constitute a potential threat to the public health and welfare so as to provide
people with an environment which is free from noise that jeopardizes their
health and welfare. This policy will require actions for which appropriate
supporting measurement methodology must be developed. The actions are
summarized in Appendix B for the benefit of those readers who are not familiar
with the Act. The criteria documents and major noise source identification
reports developed in response to Sec. 5 of the Noise Control Act of 1972 are
the logical starting points for the development and assessment of the necessary
measurement methodology.
Figure 1 shows that one needs to collect data regarding product
classification, product usage, product noise production, and the resultant
effects of the noise on people. These data are to be used in determining the
-------
EPA Documents
- Criteria
- Levels
- Major
Sources
Review All Existing
Information to
Establish
Effects
- Who
- How
±
Operating
Characteristics
- Speed
- Load
- Duty Cycle
Select Most
Suitable
Measurement
Scale
-------
acoustical quantity to "be rated, the scheme to "be used for a noise rating
methodology which adequately correlates with human response, arid, finally, the
measurement standard to "be used in conjunction vith the emission standard or
other regulation. Note that the word "standard" is customarily used with two
very different meanings. An emission standard is a legal permissible limit,
for example, on noise emission. A measurement standard is a prescribed
procedure for conducting a measurement in such a way as to obtain reliable,
reproducible results with a specified level of accuracy.
Conceptually, there are three separate items which go into the noise
emission regulation, whether it be a noise emission standard, a labeling
requirement, or the procedure for certifying low-noise-emission products, (l)
There should be a measurement standard which prescribes the test environment,
operating procedures, and acoustical measurement methodology. (2) There should
be a scheme, or algorithm, by which the acoustical data can be used to obtain a
noise rating which relates to human response. (3) There should be a level, or
set of levels, which separate the noise rating into categories of performance
or acceptability. For example, a noise emission standard could set a single
upper limit on the noise rating — products exceeding that rating could not be
sold. Alternatively, it could set an upper limit which varied with some
operating parameter such as speed; the noise rating could not exceed that curve
at any value of the operating parameter. A labeling requirement might
establish several classes (e.g., A, B, C, and D] into one of which the noise
rating would fall.
Once a regulation has been promulgated, a reasonable and equitable
enforcement program is a necessity; for without enforcement, the regulation is
meaningless. The importance of highly repeatable, accurate measurements- should
be quite clear in the enforcement area. To stand up in court, a measurement
must be proved to be reliable, of accuracy sufficient for the purpose, and
appropriately related to the adverse impact of the noise on people. In
addition, a measurement assurance program is highly desirable to ensure that
the determination of a product's noise rating will be independent of the
organization performing the measurements — EPA, manufacturers, independent
test laboratories. Unless minimum requirements are placed on Instrumentation,
calibration, test facilities and personnel conducting tests, no assurance will
exist that a product actually conforms to the regulation.
Section 2 of this report provides an overview of the measurement
methodology development process vhile Section 3 is concerned with, the role of
measurement in relation to regulations. Ensuing sections develop in detail the
concepts depicted in Figure 1. It should be emphasized that the flow of action
is not as simple and unidirectional as indicated in Figure 1. Rarely, if ever,
would it be possible to progress directly- from the input data to the rating
scheme and measurement standard and then to the regulation without considerable
retracing. In the interest of simplicity, the many feedback loops. have been
omitted.
-------
2. Measurement Methodology Development ~ Overview
A sequence of activities must "be completed in assigning a meaningful
noise rating to a product. An overview of the steps involved in this process
is presented "below — a more detailed treatment appears in the succeeding
sections of Part I.
2.1. Product Classification
The vast number of available products which might be regulated precludes
handling each on an individual basis — this would constitute an endless task.
However, categories of products may be subdivided into product classes based on
critical parameters associated with noise. The scope and effectiveness of the
regulatory task will largely be determined by the degree of success achieved in
devising useful product classifications.
2.2. Normal Use Conditions
Discussions with organizations concerning product noise have invariably
dealt with the requirement to consider a device under realistic use conditions.
By ignoring the environmental and operating characteristics it might be easier
to develop standardized measurement techniques but it then becomes almost
impossible to relate the noise output to impact on people. If the noise impact
is not determined, the reason for making the measurement and the validity of
any rating are questionable.
2.3. Identification of Major "Effects" Parameters
The characteristics of noise associated with the major physiological,
psychological and sociological effects and the resultant impact form the basis
of the criterion measure to be ultimately employed by the Environmental
Protection Agency (along with considerations of technical and economic
feasibility). Among those features usually defined are "loudness" (or
noisiness) and annoyance. The disruption of tasks such as speech communication
is another major consequence of noise often noted.
2.4. Physical Measurements Correlated with Major Effects
When the major parameters associated with effects and population impact
have been identified and their physical correlates determined, a meaningful
test methodology can be developed. Measures of sound level, spectral
characteristics and temporal variation form the basis for these quantitative
descriptions.
-------
3. Relationship of Measurement to Regulations
3.1. Role of Measurement
Good measurements are essential to research and development activities.
Often solutions to problems cannot be implemented "because of the lack of
relevant data. Under these circumstances, an appropriate data "base must be
developed. This often calls for improved measures to develop control
techniques and/or to determine their effectiveness. In addition, generally
accepted and standardized test methods assume an important role so that data
acquired at different times or places can be compared and related meaningfully.
Noise measurements need to be conducted under both laboratory and field
conditions to meet the dual objectives of scientific accuracy and relevance to
real-use conditions. The primary goal is to make valid and reliable
measurements such that the actions taken on the basis of a measurement are only
negligibly affected by the errors in the measurement.
The primary questions are:
o What needs to be measured?
o What is now measured?
o How is it now measured?
o What are the current requirements concerning measurement
uncertainty? (Appendix C contains a discussion of
uncertainty of measurement).
o What are limitations of present instruments and
methodology?
If the answers to these questions result in poor or inadequate data,
another question must then be answered:
o What research and development need to be accomplished
to make the required measurements?
3.2. Measurement Requirements for Regulation
The assessment of noise problems and of alternative strategies
for noise abatement and control must ultimately rest on accurate,
reliable, and relevant measurement capability required to:
o ascertain the effects of a given noise exposure (what
exposure will produce how much hearing damage?)
o establish trends (is the average noise exposure of the
populace increasing and are more people being exposed?)
-------
o associate sources with environmental noise levels (how much of
urban noise comes from trucks? from planes? from air
conditioners?)
o promote equity in trade (are manufacturers "being treated
equally in voluntary or mandatory noise emission
standards programs?}
o provide adequate information to the consumer (which of
several products is really the quietest and are the
noise level differences among products significant?)
o permit selection of cost-effective solutions (what
degree of noise reduction should be attained and at
what cost?)
o monitor the effectiveness of control programs (is the
noise level actually responding to new control
procedures?)
o provide the factual (measurement) basis for legal action
should that prove necessary, (is a particular noise
source in violation? To stand up in court, a measure-
ment must be proved to have an uncertainty which is
sufficiently small for the purpose.)
3.3. Measurement Uncertainty Considerations in Establishing
Regulatory Noise Limits
How can effective noise regulation be carried out in such a way
as to minimize interference with such important values as individual
freedom of choice, dynamic operation of the marketplace, and progress
in the development and use of new science and technology?
There are three needs attendant to the problem of the abatement of
excess noise: first, an agency to execute the overall policy set by the
legislative process; second, a mechanism to determine allowable levels
of noise; and third, a means of enforcing the determined requirements.
Measurement is a key element in defining "allowable levels of noise" and
"enforcing the determined requirements."
More generally, measurements are essential to effective communica-
tions in transmitting the complex information required for noise abate-
ment and control. The results of measurements are data which are the
bases for judgements as to acceptability of products, violations of law,
etc.
The question arises as to whether close measurement tolerances —
attainable by careful control of environment, operating procedures, and
-------
measurement procedure — are justified in view of the uncertainties in the
application of the resultant noise rating to predict in-service noise levels
and the resultant effects. First, the total error in prediction involves the
sum of the errors in measurement and rating, in predicting in-service levels
and in predicting the resultant effects. Reducing the rating error thus
reduces the total error; however, it would be economically wasteful to insist
upon high levels of accuracy and precision for rating measurements if the
errors in predicting in-service levels and resultant effects are relatively
gross and difficult to control. Second, application errors and, to some
extent, effect-prediction errors tend to be averaged out from one application
to another; systematic errors are repetitive. Third, when noise ratings are
used to compare the performance of a product to the requirements of a
regulation or to the performance of a competitive product, if the ratings are
in error, difficulties in enforcement, inequities in trade, and excessive costs
may be incurred. Careful evaluation of measurement uncertainties is required
for assurance that they are neither too small nor too large in relation to the
purpose for which the measurements are made.
The setting of the enforcement level relative to the legal limit, for
test results to be based on some sample of the product, depends to a large
extent on the measurement uncertainty, the variability of the product, and the
desired levels of risks in accepting non-conforming units and rejecting
conforming units (e.g., see [l]). In general, for regulations to be effective
and reasonable, measurement uncertainties should be as small as practicable.
If, for example, the legal limit is 80 dB and the measurement uncertainty is _+5
dB, procedures will be needed for interpreting a noise level close to the legal
limit such as 83 dB or 77 dB. The existence of a violation can be sustained
more effectively when the noise level is above the legal limit by an amount
greater than the measurement uncertainty, and the public is better assured of
conforming products if the noise level is below the legal limit by a similarly
convincing amount. The difficulty is compounded by the need for uniformity of
enforcement, i.e., consistent measurements by different officials. The largest
uncertainty likely for measurements by the least competent official, using the
most inaccurate or imprecise equipment allowable, under the least favorable
test conditions, must be considered in the process of setting the enforcement
limit. If this measurement uncertainty is large, the enforcement agency's
problems are increased. If the enforcement level is set above the legal limit,
it appears that the agency is not enforcing the desired noise abatement. If
the enforcement level is set below the legal limit, manufacturers incur
increased compliance costs.
Measurement uncertainties should also be held as small as practical for
trade equity. For example, if products are graded A, B, C, D in terms of noise
emission, the attainable measurement errors should be considerably smaller than
the steps between the middle of adjacent grades. Otherwise, products could be
incorrectly labeled, resulting in unfair competitive situations. Similarly, of
two competing products with identical noise emission, one could be banned from
sale while the other would be allowed to be sold — both being tested against
the same emission standard.
-------
Cost of compliance with emission standards, or of meeting a certain
labeling grade, can be increased significantly due to measurement uncertain-
ties. As products are quieted more and more, the incremental cost of
quieting typically increases. Thus, for example, if due to measurement
uncertainties, it were necessary to quiet a product by 10 decibels to be
certain that it is at least 5 decibels quieter, the cost-to-quiet may be
several times greater than if the measurement uncertainty were, say, 1
decibel.
8
-------
4. Informational Requirements for Setting Noise Regulations
Product Classification Scheme
Consumer products are classified into categories which relate to
differences in the design, shape, style or quality of similar products.
Classification schemes are based on power, capacity, weight, etc.
Questions concerning product classification systems to be used for
noise ratings include:
• Does a classification system exist within a given
industry regardless of whether or not it is appro-
priate for noise measurement?
• Does the existing classification system lend itself to
acceptance as a categorization with noise as the
item of concern rather than weight, power, etc?
• Is the level of classification detail sufficient for the
measurement methodology to be universally applied?
In general, satisfactory classifications do exist for most products.
As an example, pumps can be classified:
• by capacity — 10 to a million gallons per minute.
• by application — industrial, commercial, agricultural,
municipal, domestic.
• by the materials that they are capable of handling —
water, sulphur, solids, slurry, liquid metal.
• by pressure range — high, medium, low, vacuum.
• by design — turbine, axial flow, centrifugal, rotary,
reciprocating.
• by salient features — submersible, self-priming,
proportioning, non-clogging, measuring.
• by the material used — stainless steel, plastic, bronze.
All of the above classification schemes are valid for specific applica-
tions; however, none of them has noise as its primary basis.
The fact that pumps can be subdivided into a myriad of classifications
should be considered with caution since some products are not so readily
classified. For instance, the construction industry has no classification
scheme for its products. As a first step in developing such a scheme, the
-------
Construction Industry Manufacturers Association asked its members to
define their concept of that piece of construction equipment known as a
roller/compactor. The answers identified approximately fifty different
devices ranging in weight from hundreds of pounds to several hundred-thousands
of pounds. The list represented self-propelled, towed, and even hand-controllS"
machines.
Motor vehicles offer another example. Since all motor vehicles do not
have the same noise generation characteristics, the category "motor vehicles'
must be further subdivided. According to functional characteristics, the next
level of detail would break into trucks, cars, buses and over-the-road
recreational vehicles. Utilizing gross vehicle weight as a further subdivision
at this level of classification, the Society of Automotive Engineers, Inc.
established its noise standards for motor vehicles (highway). One standard
exists for passenger cars and light trucks (vehicles of 6000 pounds or less)
and another applies to heavy trucks and buses. No specific standards now exist
for over-the-road recreational vehicles. Thus a passenger car could weigh more
than 6000 pounds and theoretically would be considered a truck from a noise
regulation standpoint. Licensing is not the final answer since the
distinctions between trucks and passenger cars for licensing purposes are
sometimes quite subtle. For instance, a four-wheel drive vehicle is licensed
in the state of Maryland as a truck unless it has a rear seat for carrying
passengers; then it is licensed as a car. Such an arbitrary system could
penalize a given vehicle due to an optional feature which has no bearing upon
the vehicle noise generation.
Even after the classification scheme is determined, all problems are
not solved. A tractor-trailer combination is considered as a truck by most
people. However, to the industry it is comprised of two independent parts
designed and fabricated by different manufacturers — the tractor and the
trailer. The eventual noise level produced by the combination depends on
a third party — the fleet or individual owner — who matches tractor to
trailer and determines which tires are mounted on each (tires usually
control noise emission at moderate-to-high speeds for well-maintained
vehicles equipped with adequate muffling systems). Since tractor-trailer
"trucks" are not sold as such in commerce, the Federal regulation of
these might have to be via product noise emission standards (Sec. 6 of
Public Law 92-5710 for the products (tractor, tires) which are sold separately
plus motor carrier noise emission standards (Sec. 18 of Public Law 92-571*)
for the overall vehicle.
Devices could also be classified by the noise they produce versus
such considerations as the service they perform (buses contrasted to auto-
mobiles), their location (urban versus remote or rural), and their effects
(how many people are affected and in what way?). For example, farm
machinery could be placed in a separate classification since the location
is usually rural and the only people impacted are the farmers themselves.
In this case the most economical noise control solution might be an
enclosed cab which would protect the operator of the machine.
10
-------
It should be evident that product classification is not the simple
matter that it may have seemed at first glance. Much thought must be
given to the relevance of existing systems and the development of new
schemes where classification does not now exist. With the high
costs associated with instrumentation, facilities, and the trained
manpower to make the measurements, one would hope that appropriate classi-
fication schemes can "be developed so that a separate and distinct
measurement methodology would not be necessary for each product.
11
-------
5. Informational Requirements for Setting Noise Regulations
Operating Procedures and Environment
"Scatter" typically is observed in test data. With the advent of
government noise regulations, it "becomes increasingly imperative to
determine the sources of variability in test results. Existing test
procedures should be revised and new procedures developed to be less
sensitive to device operation and test site conditions.
Test procedures should represent typical operating modes — that
is, relevant either to community or operator noise exposure to the
device under test.
5.1. Typical Usage Data
In general, products have a range of possible operational modes with
the noise emission characteristics dependent on how they operate. The
key to deciding upon the mode or modes of operation of a device that should
be specified for noise test purposes hinges on the knowledge of how the
device is "normally" operated. In order to determine "normal operation,"
a usage survey should be conducted.^. The results of one such survey[2]
in the outboard motor field are sfeown in Table 1.
Although the data are limited, it is interesting to note the very
low percentage of time that pleasure boats run at wide open throttle
(maximum noise position}. This example is not atypical. In the case of
railroad locomotives, for instance, over 30 percent of the time the
vehicle is at idle. These examples substantiate the fact that usually it
will not be sufficient to make measurements only during that operation
which produces maximum noise since most devices are not so operated for
long periods of time. In addition, people can be annoyed, or have their
conversations interfered with, at noise levels much lower than the
maximum.
5.2. Effect on Noise Levels
There are a great many noise sources and noise environments to
which people are exposed. The noise emission from a given machine in
a specific location is dependent not only on the sound radiating
characteristics of the machine itself but also upon the type of mounting,
the manner in which the machine is operated, and its environment. In
setting noise limits for such devices through regulations, the measure-
ment system should include specification of such items as operating
conditions and environment — both installation and weather.
a. Operation
Attention must be given to the operational procedure utilized
for a given test since the noise produced depends heavily on the way a
12
-------
Table 1. Outboard Motor Field Usage Study
Outboard
125 HP
100 HP
100 HP
55 HP (TWIN)
50 HP
50 HP (TWIN)
1*0 HP (TWIN)
9 1/2 HP
9 1/2 HP
9 1/2 HP
9 1/2 HP
Boat
17 Ft. Runabout
IT Ft. Runabout
18 Ft. Runabout
23 Ft. Cruiser
16 Ft. Runabout
20 Ft. Cruiser
16 Ft. Runabout
16 Ft. Fishing Boat
ll* Ft. Fishing Boat
ll* Ft. Fishing Boat
lit Ft. Fishing Boat
Hours
16.72
11.56
1*8.9!*
2l*.30
ll*.2l*
13.56
lit. 1*1*
21.56
lit. 28
10.2lt
10.16
PERCENTAGE TIME IN EACH SPEED RANGE
500
1000
RPM
3lt
32
2k
1*
13
5
3
6
12
i*
9
1000
1500
12
20
lit
19
19
12
13
13
11
11
11
1500
2000
3
7
U
15
5
11
12
10
7
12
9
2000
2500
5
3
2
5
10
5
7
6
9
19
U
2500
3000
6
i
i
2
5
5
3
6
3
11
1*
3000
3500
111
7
U
2
10
12
13
9
9
25
9
3500
ItOOO
12
13
11
8
7
2lt
13
8
5
15
22
i
1*000
1*500
10
11
33
37
10
22
19
19
25
3
19
U500
5000
3
1*
5
8
5
2
12
18
15
0
12
5000
5500
1
2
2
0
11
1
5
5
h
*
*
5500
6000
*
*
*
*
5
1
#
*
#
*
*
*0utboard motor was not set to run in this speed range at wide open throttle.
-------
piece of equipment is operated. Two devices might generate comparable
noise levels when operated in a certain mode yet might produce quite
different noise levels when operated in another mode.
b. Environment
The term environment, as used here, means the aggregate of all
external conditions and influences affecting the noise levels of a
given device, wherever the measurements are made.
Airborne sound from an outdoor source travels from the source to the
receiver through an atmosphere that is constantly in motion. Turbulence,
temperature and wind gradients, and reflections from the earth's surface all
affect the measured data. Essentially there are two distinct effects corres-
ponding to: (l) external factors (usually atmospheric) which affect the sound
pressure level at a particular point and (2) external factors which affect the
accuracy of its measurement.
One also must be concerned with the environment indoors. Room
volume, the sound absorption of the walls, floor and ceiling, and
the location of both the noise source and receiver must be carefully
evaluated. For instance, a sound level very near a noisy machine
may not be affected by sound-absorbing materials on the walls. How-
ever, such materials will affect the sound level measured farther from
the machine.
These are just a few of the factors that should be considered.
In many cases, these are factors over which the investigator may have
little or no control.
c. Installation
Another important factor affecting the noise level associated
with a device is the manner in which it is installed.
For example, the resultant noise level for a food-waste disposer
is dependent on (l) the effectiveness of the vibration isolation of
the disposer from the sink, (2) the damping characteristics of the
sink itself, (3) whether the connections between the disposer and
the drain pipe are flexible or rigid, (U) the effectiveness of the
closure of the mouth of the disposer and (5) whether the grinding
chamber and motor are enclosed.
Installation is likewise crucial when one considers any type
of component noise, for example, motors or engines. Returning to
the boating industry, every engine and drive unit, whether it is out-
board, stern drive, or inboard, will create a different noise
signature, depending on the boat to which it is coupled. Each boat will
respond in a different manner, and the overall boat noise can change
with different loads and operating conditions.
-------
When measurements are made to determine compliance with applicable
acoustic objectives, specifications or standards, the device should "be
installed in a manner similar to that in a typical customer's facility, if
that is practical.
-------
6. Informational Requirements for Setting Noise Regulations
Effects of Noise on People
Since noise regulations are designed to limit the exposure of noise on
people, it is necessary to determine who is affected, with what impact^ fey a
particular product and how these effects are manifested. These factors are
largely governed by the product and the way that it is normally used.
6.1. Parties Affected
The relationship between the operating environment and the determination
as to who is affected by a given noise becomes evident when products are
classified in terms of their mobility (or in terms of people's mobility
relative to the sources). For example, transportation vehicles have a far
different impact on a community than a home air-conditioning unit. An
exploration of these different types of noise sources will serve to illustrate
their general effects.
a. Operator
The operating procedures associated with a product will indicate how an
operator interacts with a product — what he does and where he is positioned.
These data provide the context for determining the effects of noise on a given
operator. The operator need not be the person exposed to the most severe noise
effects. In fact, in some instances, a product is designed to protect him from
the noise.
b. Passenger
When the many transportation vehicles in operation which are designed for
more than one person are considered, it becomes evident that being a passenger
is very common. The noise exposure experienced by a passenger can be quite
different from that of the operator and therefore merits independent
examination.
c. Neighbor
Another social role is that of a neighbor, who maintains a fixed and
rather permanent relationship to a noise source. (An air conditioning unit may
be positioned so as to produce minimum noise disturbance to owners, but may be
disturbing to neighbors.)
d. Bystander
For the three previous classes of people affected by noise (operator,
passenger, and neighbor), it may be anticipated that the duration of exposure
is usually considerable. However, in many situations the noise exposure is of
relatively short duration for an individual but a larger number of people are
impacted. The noise source may move through a community (a vehicle siren) or
many people may walk past a stationary source (a construction site).
16
-------
6.2. Nature of Effects
Traditionally, the effects of noise have been studied in terms of three
major approaches, each one identified with a method of measurement —
physiological changes (both temporary and permanent), psychological attributes
such as annoyance and, finally, task interference.
a. Hearing Loss
The most severe and damaging effect of noise exposure is permanent loss
of hearing. Exposure to noise of sufficient intensity for long periods of time
can produce temporary or permanent effects on the ability to hear. In some
instances, excessive noise exposure leads to the destruction of the primary
auditory receptors — the hair cells in the ear. Changes in hearing
experienced by the person suffering hearing impairment include distortions of
the clarity and quality of sounds as well as losses in the ability to detect
and understand sound. These changes can range from slight impairment to severe
deafness.
b. Task Interference
The effect of noise on the performance of desired activities has received
considerable research attention. The most unequivocal finding has been that
noise can seriously impair speech communication. The performance of other
complex tasks is also made difficult by noise, but it has been difficult to
adequately define the levels of noise acceptability (or unacceptability) or to
develop an adequate measure of task difficulty.
c. Annoyance
Many components associated with the disturbing characteristics of noise
have been subsumed under the term of annoyance. An International Organization
for Standardization study group defined annoyance as "that general quantity
that emerges from the various sociological surveys that concern themselves with
disturbance of various kinds around intense noise sources like aircraft."[ 3]
Annoyance can result from sleep disturbance, interference with radio and TV
listening, as well as many other subtle immediate and long term sociological,
Psychological and physiological reactions. Its definition and measurement have
posed a major problem for researchers for many years.
IT
-------
7. Quantitative Description of Noise
A viable method of assigning noise ratings for products depends
primarily on:
a. The identification of those aspects of noise which are of concern
in a particular situation.
b. A well-defined procedure for quantifying noise emission.
c. The ability to relate the results of measurement to human
response.
7.1. Identification of Major Parameters
The most important decision to be made in any measurement process is
the determination of what to measure. In noise abatement and control, there
is not general agreement as to what quantities should be measured —
primarily because of inadequate data relating the impact of noise to the
exposure which produced that impact.
a. Relationship to Effects
The attempt to associate particular characteristics of sounds with
effects on people have had mixed results. Some clear-cut findings have been
made — notably those dealing with intensity, speech interference and some
aspects of annoyance (high frequencies, pure tone components). However,
other effects often cited are less tangible although they might be
important. Interruption of activities such as sleep, work and recreation
are often caused by noises which are unpredictable, or of short duration an
need not be intense — yet might constitute major problems for the
individual from a psychological and physical health standpoint. For
example, a person who is awakened regularly from sleep nightly by distant
train whistles or aircraft overflights might be exposed to only moderate^
noise levels for a total of less than one minute during that time and still
have a major noise problem. Noise effects therefore cannot readily be
treated independently of the typical operating situation — including vho is
exposed, activities being performed and, finally, identifying the major
consequences of the noise.
b. Environment/Use Conditions
Other effects which cannot be readily dealt with in a rating scheme,
but nonetheless are often cited, concern environmental factors and use
conditions. For example, the low-frequency noise from trains penetrates
great distances under particular weather conditions, thereby affecting a
large number of people usually not bothered by this sound. Motorcycles an
other vehicles are often operated in a manner that increases noise output
rather than minimizes, it. These are annoying both because of the noise ^
intensity and the feeling that bystanders have that the noise is "louder
than is necessary.
18
-------
7.2. Measurement Parameters
If history is a valid predictor of the future, there vill 'be consider-
able disagreement as to particular rating schemes which should be used to
describe the noise emission of products in terms which vill appropriately
predict human response to noise. The important measurement parameters for
rating schemes are noise level, spectral quality, and temporal variation.
a. Sound Level
There is general agreement that, for a given spectrum shape and a given
temporal variation, annoyance, hearing loss, task interference, etc.,
increase monotonically with increasing sound level. The difficulties and
disagreements arise relative to the variation of noise level with frequency
(i.e., the spectrum shape) and with time.
b. Spectral Quality
Although there are many schemes for predicting loudness, noisiness,
annoyance, and so forth from the shape of a noise spectrum, it is perhaps
fortunate that only a limited number of these have been given official
status by regulatory agencies or standards bodies.
Relative to the common weighting networks used in sound level meters,
it is reasonably well established that the A-weighted sound level (LA)
predicts most human response much better than the B-weighted (L_) or C-
weighted (Lp) sound level. D-weighted sound level (L^J has been adopted
specifically for aircraft noise. Both Stevens' and dicker's methods for
computing loudness have been standardized, although significant
modifications have been proposed by Stevens that are not reflected in the
standardized version. Kryter's Perceived Noise Level (Lpjj) enjoys official
status through its use in calculating Effective Perceive^ Noise Level C.LgpN)
for Federal aircraft certification (and has also "been standardized by SAE
and ANSI). The Air-Conditioning and Refrigeration Institute has a sound
rating procedure somewhat similar to LpN- All of these quantities are used
to predict annoyance due to noise.
The Articulation Index (Al) has been standardized as a predictor of the
extent to which steady noise interferes- with speech communication (for male
speakers only). Owing to the complexity of the Articulation Index, several
versions of Speech Interference Level (SIL), based on the average of 3 or k
octave band levels, have been proposed.
c. Temporal Variation
For transient noise events, there is need for a means to assign an
overall noise rating to an event which includes appropriate consideration of
"the extent to which duration affects human response. The National Academy
°f Sciences-National Research Council Committee on Hearing, Bioacoustics,
19
-------
and Biomechanics (CHABA) proposed a hearing damage-risk criterion for
impulse noise (gunfire) which specifically includes measures of the duration
of the pressure wave. Both the L and the single event noise exposure
level (iwp,-.^) include corrections to account for duration. The
Environmental Protection Agency has recommended the A-weighted average sound
level (L ) and the day-night level (.L, ) as general descriptors of
environmental noise.
20
-------
8. Consensus Procedures for Measurement Standards
and Rating Schemes
8.1. Input from All Affected Parties
In writing measurement standards and establishing rating schemes, it is
important to consider the contributions to standards-writing of all the
affected groups. First, the voluntary standards bodies (for example, the
American National Standards Institute, American Society for Testing and
Materials) establish general standard test methods which reflect the most
knowledgeable engineering and scientific practice. The trade associations,
representing the manufacturers, must implement test methods, and therefore
in addition to technical input evaluate practicality, in terms of available
facilities, personnel, training, etc., in adopting or developing test
methods for their purposes. Trade associations, however, are limited in the
amount and kinds of information that can be channeled through them. Much of
the information the Environmental Protection Agency will need to ensure that
mandatory standards are reasonable, technologically feasible and
economically worthwhile will be supplied by industry and be of a
confidential and proprietary nature. Consequently, it will be necessary for
the Environmental Protection Agency to communicate with individual
manufacturers as well as with trade associations. Regulatory agencies,
acting on behalf of, and with input from, the consumer, set legal levels for
each class of product, levels which must be technically and economically
reasonable for the affected industry. Obviously, to consider only a single
element from this chain and ignore the contributions of the others, will
markedly decrease the effectiveness of the standard or rating scheme.
8.2. International Trade
Several years ago the concept of standardization in Western Europe
underwent changes that posed a serious threat to United States-European
"trade. These changes were the result of a projected "harmonization" program
"to eliminate intra-European technical non-tariff trade barriers caused by
differences in product standards. It was clear that upon completion of the
"harmonization" program, international trade within Western Europe would be
virtually barrier-free and trade into Western Europe almost impossible[U].
To maintain U. S. viability in European trade, the United States
delegation to the "Kennedy round" of trade negotiations has called for a new
scheme called the General Agreement for Tariffs and Trade (GATT). This
agreement has identified over 800 non-tariff trade barriers in existence
today. in February 1971, a new stage of the GATT non-tariff barrier work
was begun. Rather than face the overwhelming task of solving the whole
Program of non-tariff trade barriers, it was decided to concentrate on a few
— of which product standards is one area. Standards were chosen for
attention both because of their growing importance and because it appeared
that progress might be encouraged more readily in this area than in some of
the others[5]. A draft code on product standards was written and in May
1975, was accepted as a basis for negotiation[6]. A more thorough
Discussion of this code is given in Appendix E.
21
-------
9. Environmental Protection Agency Promulgation of Regulations
9.1. Special Requirements
This section highlights a few features to be considered in conjunction
with the regulations (see Section l) to "be issued "by the Environmental
Protection Agency. Pertinent sections of the Noise Control Act of 1972 are
reproduced in Appendix B.
a. Product Hoise Emission Standards
This report directly addresses product noise emission standards.
b. Labeling
A meaningful label on a noise-producing product (or noise control
product) is dependent upon the existence of an appropriate, reliable, and
repeatable measure of the.noise level that is reasonably well-tied to
subjective response to the noise. Whether the measure is a sound pover
rating, a weighted sound level at some distance from the device, or a set of
octave band sound pressure level data, the problem is the conversion of this
technical acoustic data into a labeling system which will be understood "by the
public and will provide the appropriate amount and type of information for &*
non-technical audience it is intended to reach.
Basically there are two types of audiences — the general public consume1"
and the business or industrial consumer. In general, the public will want a
label that is simple and easy to understand to permit a comparison to be made
between the noise expected from one device as opposed to another. The noise
label then serves as a criterion, along with performance characteristics,
aesthetic quality, safety features, cost, etc., which the consumer evaluates-
Would he be willing to pay, for example, $2.00 more for a blender that does
not annoy him or interfere with his listening to his hi-fi system during tnf
blender's operation? On the other hand, the business and industrial conununiw
will need additional information — either from the label itself or from-
application material developed to accompany the label. For instance, it isp
not enough to know that you have purchased a "quiet" machine, because if t*1?
machine is improperly placed you still could have a noise problem. An exairtp1
of this would be the building contractor who purchases the quietest air
conditioning units available and then installs them in a location under a
bedroom window or next to a patio. Even though the unit might be quiet,
complaints would be made due to loss of sleep or the deprived use of the Pa
due to annoying or speech-interfering noise. The label, or supplementary
material, should provide the purchaser with enough information so that he
sufficient data to determine whether or not his equipment .in situ will meet
the requirements of those noise ordinances which affect him, be they occupa"
tional safety and health regulations or sound level restrictions at property
lines.
22
-------
Product labels can take various forms. Appendix E contains a discussion
of some of the alternative methods which deserve consideration as possible
noise labels. It should be noted that it vas the general consensus of the
advisory and industry groups which reviewed the initial draft of this report
that noise labels should provide quantitative information as to the product's
noise emission.
c. Low-Noise-Emission Product Certification Procedures
To be eligible for certification as low-noise-emission products, the
measurement methodology and rating scheme should be established first. The
major task is to decide what further noise reduction is necessary for a
product to be certified and whether or not there could be several grades of
certification with associated procurement costs. If several grades are
established, the problems become very analogous to those discussed above
relative to labeling. However, in an industry composed of many small firms,
"with little research and development capability, the economic cost of
acoustical measurements may be a significant factor.
9.2. Technical, Economic, and Administrative Feasibility as It Relates to Measurement
Fair and effective noise regulations must be technically, economically,
and administratively feasible and reasonable. Generally, people consider
"technical feasibility only in terms of the technology available to quiet a
product adequately, the economical feasibility only in terms of the cost to do
so» and, if they think of it at all, administrative feasibility in terms of
enforcement of the regulations. However, the measurement methodology must
also be feasible and reasonable. The scientific and technical knowledge,
appropriate test facilities, and suitable instrumentation must be available to
carry out noise measurements of adequate accuracy and precision at a practical
cost. The measurement and rating methodology and the testing and quality
control procedures should be suitable for reasonable monitoring by an agent of
"the regulating body.
These factors should relate to the characteristics of the particular
industry, in selecting the measurement methodology for use in regulating a
Particular class of product. For example, the air conditioning industry has
voluntarily set up a noise rating system based on one-third octave band sound
Power levels measured in a reverberation chamber — measurements which require
expensive test facilities, complex instrumentation, and technically
sophisticated personnel.
9,3. Compatibility with Other Regulations and Needs
In promulgation of Federal noise emission regulations, care should be
taken to select measurement methodology, rating schemes, labeling practices,
etc-, which complement, to the greatest extent practicable, other regulations
and needs. For example:
23
-------
o Even if labeling requirements call for a single-number (or
letter, or color) rating scheme, it might be appropriate
for labels on industrial machinery to include octave-band
data so the purchaser can select and design to meet Occupa-
tional Safety and Health Administration (OSHA) regulations.
(While current OSHA regulations of occupational noise expo-
ure are expressed in terms of A-weighted levels, spectral
information is often needed for engineering design purposes.)
o Similarly, since many local zoning ordinances are written in^
terms of octave band sound pressure levels at the property line,
it might be appropriate for the noise labels on outdoor equip-
ment to provide sufficient data to permit prediction of
property-line octave-band sound levels.
o Federal product noise emission standards should, wherever
possible, utilize measurement methodology which is equivalent
to, or compatible with, state and local operating regulations
so a purchaser can determine the conditions under which he
will be able to operate a device in a given locality.
Many other examples could be cited. The major point is to carefully consider
all probable uses of a class of products and then structure the measurement
and rating methodology so as to provide as much information as practical in a
form to meet the purchaser's manifold requirements.
-------
10. Recommended Operational Procedure for Generating Measurement
Methodologies for Specific Product Classes
The recommended procedure for the efficient and timely generation of
measurement methodologies for specific prod-act classes is depicted in Fig-ore
2.
Once a specific product class has been identified as a major noise
source, the Environmental Protection Agency, or its designated contractor,
would form a small ad hoc task group to provide key input data for the needed
measurement methodology. This group would include, as a minimum,
representatives of the Environmental Protection Agency, independent acoustical
consultants, manufacturers and, where appropriate, users and representatives
of other government agencies at the Federal, state and local level. Careful
selection of task group members would ensure a "broad spectrum of expertise
including: detailed familiarity with the product category, its design and
operation; knowledge of quality control and product testing; expertise in
acoustics; and a familiarity with the regulatory process.
Following formulation of the task group, meetings would tie held to
determine the existence and relevance of data "bases and measurement procedures
appropriate to the specific product class in question. Three possibilities
exist: (l) a data base and/or measurement procedure exists and is relevant,
(2) a data base and/or measurement procedure exists but is not relevant, and
(3) a data base and/or measurement procedure does not exist.
If the task group determines that the existing measurement procedure and
attendant data base are relevant then the group vould recommend to EPA that a
Program be conducted to validate the data base as to its accuracy and
completeness.
On the other hand, if the task group determines that data bases and
measurement procedures either do not exist or are not relevant then the group
would recommend to EPA that an extensive investigative program be conducted to
evaluate alternative measurement procedures and to establish the needed data
^ase. Such an investigative program would include systematic variation of
source operating procedures, source loading, test site characteristics,
Aerophone locations and acoustic measurement procedures.
Following completion of the investigative and/or validation programs, the
task group would reconvene and evaluate the results. On the basis of these
results, they would prepare an outline for a draft measurement methodology.
EpA, or its designated contractor, would then prepare a draft measurement
methodology which would be circulated to all members of the task group and to
a number of affected parties for comment.
Once the draft methodology had been completed, the task group would
recommend to EPA a list of selected industry representatives who should be
to carry out measurements in accordance with the draft standard. Not
-------
Major Noise
Source
Identified
By EPA
Task Group
Established
YES
I
Does
Relevant
Data Base
Exist?
NO
Conduct
Program to
Validate
Data Base
«*•
Draft
Measurement
Methodology
Conduct
Investigative
Program to Establish
Data Base
Conduct
Experimental
Program to Validate
Methodology
Issue Final
Methodology
Figure 2. Recommended operational procedure for generating measurement
methodologies.
26
-------
only would this validate the measurement methodology, "but it would also direct
attention to key problem areas and obtain data in suppor-t of simplified
measurement procedures.
At the same time, representatives from EPA, or its designated contractor,
should visit a representative number of manufacturers or users of the major
noise source in question to: (l) observe the use of the draft measurement
methodology, (2) to discuss use conditions, operating procedures, and meas-
urement procedures, and (3) if necessary, to conduct supplemental
measurements.
Utilizing information gained via the above procedures, a final measure-
ment methodology, with supporting documentation, would be prepared for each of
the candidate major noise sources. To the greatest extent feasible, the final
measurement methodology should draw upon, and be consistent with, existing and
proposed voluntary measurement standards.
27
-------
References for Part I
[l] Griib"bs, F. E., and Coon, H. J., On setting limits relative to specifica-
tion limits, Industry Quality Control 10_(5) (195U).
[2] Private communication, R. H. Lincoln (Outboard Marine Corp., Milwaukee,
Wisconsin) to W. A. Leasure, Jr. (National Bureau of Standards,
Washington, D. C.), January 11, 1973.
[3] International Organization for Standardization Recommendation for
Acoustics, Assessment of Noise with Respect to Community Response,
R1996 (American National Standards Institute, New York, New York, 1971).
[U] Simpson, R. 0., The national impact of the GATT standards code, Materials
Research and Standards 12.C7) (July 1972).
[5] Krauthoff, L. C., II, World standards for a world market, ASTM
Standardization News l_(ll) (January 1973).
[6] Travaglini, V., Removing "barriers to trade through GATT, ASTM
Standardization News 3.(ll) (November 1975).
-------
PART II CONTENTS OF MEASUREMENT STANDARDS
Part II of this report encompasses the subject matter of measurement
standards for use in conjunction with noise emission standards to be
promulgated by the U. S. Environmental Protection Agency. Part II is
designed as a checklist for the evaluation of the suitability of a noise
measurement standard for a particular class of products or, in the absence of
a suitable standard, as a framework for development of one. The intent is to
identify and discuss those factors which can impact on the accuracy,
precision, and applicability of noise measurements.
Part II is structured similarly to a measurement standard. Each of
sections 12 through 18 contains information relevant to a number of different
types of measurement processes — e.g., measurements of sound power and sound
pressure; measurements made in anechoic, semi-anechoic, reverberant, and
semi-reverberant spaces; measurements made outdoors and indoors; measurements
made under laboratory and in situ conditions; measurements of overall
weighted sound level and of sound pressure levels in frequency bands; etc.
Obviously, much of this material will not be appropriate for any given
measurement standard.
At the beginning of each section, the major factors to be addressed are
outlined. The remainder of the section presents technical support material
needed to facilitate an appropriate choice between alternative options.
Information is given as to the type and size of errors which may result from
different assignable causes. These data are not intended to be comprehensive
or definitive, but merely illustrative.
Part II has been written with the assumption that it will be used by
People trained in the physical sciences. Accordingly, basic concepts and
terms_ are^ used without definition or discussion.
29
-------
11. Purpose and Applicability
• Every measurement standard should include an introduction
(1) the group that prepared the standard, (2) the review process used
for approval of the standard^ and (3) a clear statement of the
relationship of the standard with other standards.
m The scope of the measurement standard should be a concise abstract oj
its contents and should clearly delineate what is and what is not
covered by the standard.
• The purpose of the measurement standard should make clear the reasons
for establishing the standard.
• Existing standardized definitions should be utilized; those terms
to the measurement standard which have special technical meaning of
unique to a given industry require definition,
• All publications noted in the measurement standard should be listed ^
a separate "References" section.
• There should be a clear statement of the acoustical quantity to be
measured* e.g., sound pressure or sound power.
• The range of applicability should be clearly stated in the measuremen '
standard (i.e.; the types of noise sources (nature and size of
and types of noise (steady and non-steady; periodic, aperiodics ra
for which it is applicable) and guidance should be provided as to
accuracy and precision which oan be obtained utilizing the standard-
11.1. Introduction
Every measurement standard should include an introduction or forevord
identifying the group that prepared the standard, including organizations
officially represented on that group, and an indication, explicitly or W
reference, of the review process used for approval of the standard. If *
standard replaces, supplements, or complements one or more other standards'
that should be made clear.
11.2. Scope and Purpose
The scope of a standard should be a clear and concise abstract of its
contents. It should clearly delineate what is and vhat is not covered W
the standard — for example, classes of products to be included, types of
noise to be measured, range of acoustic environments, and so forth.
The purpose of the standard should be briefly stated so as to make
clear the main reasons for establishing the standard.
See also sections 11.U and 11.5.
30
-------
11.3. Definitions and References
Reference should "be made to existing internationally or nationally
recognized definitions vhere applicable. Unusual terms of measurement,
abbreviations, and symbols used in the standard should also be defined. It
is recommended that the International System of Units (SI) be used
throughout the standard and that equivalent values expressed in customary
U.S. units also be included when failure to do so would interfere with
effective communications. In the case of key quantities, e.g., distance
from the source to a microphone, it may be desirable to give the numerical
value in the text in both SI and U. S. customary units. For other
quantities, it may be preferable to refer to conversion factors in the
"Definitions" section. For further guidance, the reader is referred to
[1-3]*.
All publications referred to throughout the standard should be listed
in a separate "References" section. Each reference should be up-to-date and
should include all information required for clear and unambiguous
identification of the reference. In referencing other standards, the full
title of the document, its designation (and date of issue, if this is not
part of the designation) and the name and address of the issuing
organization should be given. If a standard has more than one official
designation (e.g., an American Society for Testing and Materials standard
which has been adopted as an American National Standard), all known
designations should be included in the reference. It should be made quite
clear whether the intention is to reference only a particular issue_of a
standard or whether, if a standard which is referenced is later revised, the
latest revision is intended. If reference is to part of the document only,
that part should be adequately identified so that in a later revision the
pertinent material can be identified.
Throughout the text, the following abbreviated reference forms may be
•used:
Standards — the name of the issuing organization and the desig-
nation of the document (e.g., American National Standard
SI.13-1971)
Laws, codes, and ordinances — title and number
Other publications —these may be simply referenced by number, by
author and year, footnoted, or the complete reference given
in the text.
11.4. Acoustical Quantities to be Measured
Under "Scope and Purpose," there should be a clear statement of the
acoustical quantity which is to be measured — sound pressures at one or
™>re particular locations, total sound power, or sound power plus direct-
ivity, if a frequency-weighted quantity (e.g., A-weighted sound pressure or
*Hiures in brackets indicate the literature references at the end of
Part II.
31
-------
power level) is to "be measured, this should "be so indicated. Usually this
will simply "be stated in very brief form without need for justification.
However, because of the prevailing arguments as to what quantity is most
appropriate, a discussion is presented below.
Noise ratings are derived from acoustical measurements which fall into
two groups :
o Effective (root-mean-squarel, or instantaneous, sound pressure level
o Sound power level, plus directivity if desired.
Some acousticians are quite adamant in their preference for expressing
ratings in terms of sound power rather than sound pressure — or vice versa--
Since sound power is almost always computed from measured sound pressures , 1
is difficult to argue that sound power is inherently superior to pressure.
However, the sound pressure resulting from a given source is strongly
dependent upon the point .at which the measurement is made and can depend up°
the environment in which both the source and the measuring microphone are
located. Sound power level is a measure of the total sound power radiated
the source in all directions and directivity is a measure of the spatial
intensity distribution. Sound power and directivity are basically
characteristics of the noise source and hence do .not need to be ^ i
.
terms of any particular measurement distance (other than that measurements
should be made in the far field). Sound power is less dependent on the
environment in which the source is located than is sound pressure. However*
both the radiated sound power and the directivity are influenced by nearby
reflecting surfaces, such as floors and walls. In enclosed spaces, the so
power usually is less affected by the environment than is the sound pressur
at distances far enough away from the source to be in the reverberant fiel
As was mentioned above, the sound power usually must be computed
sound pressure measurements. This is accomplished simply and accurately ^
two limiting cases: (l) in the region of a free field beyond the near fie
and (2) in a reverberant (diffuse) sound field. In such fields the sound
power can be calculated from the mean-square sound pressure, averaged over
appropriate surface enclosing the source (free field) or averaged over
volume of the room (reverberant field). Close approximations to free-
conditions can be achieved in anechoic chambers, hemi-anechoic chambers
(i.e., free field over a reflecting plane), or outdoors. Approximately
diffuse sound fields can be obtained in large, hard-walled reverberation
chambers. To determine the directivity of the source, an essentially
free-field environment must be used.
When acoustical data are expressed in terms of sound power, one shoul
be able to assume that the data correspond to the far field around the
source and that the sound power is based on measurements at a sufficient
32
-------
number of angular positions to sample adequately the field in all
directions or by adequately sampling a reverberant field. Thus one can
have a fair degree of confidence in using the sound power to predict sound
pressure at a particular location in a particular environment. If on the
other hand, acoustical data are expressed in terms of sound pressure, these
data should be accompanied vith sufficient information regarding the
measurement location and test environment to enable one to infer the extent
to vhich the data may be used to predict sound pressure at other locations
and in other environments. Compare the following statements:
The A-weighted total sound power level of the machine was found to be
93 dB re 1 pW.
The A-weighted sound pressure level of the machine, when resting on a
hard reflecting plane, was found to be 69 dB re 20 yPa at a location 7
m from the center of the machine, at an angle of ^5° relative to the
axis of the machine, at a height of 1.5m above the supporting plane.
No other reflecting surfaces were within 50m of the machine. Tests
established that the measurement location was in the far field and
that there were no significant variations in soured pressure near this
location due to phase cancellation between the signal reflected from
the supporting plane and the direct signal.
Note that the results of sound power determinations can be stated quite
simply since measurement location need not be specified. A statement
regarding local variations in sound pressure due to phase cancellation
frequently would not be made but can be quite important (see Sec- 15). In
the case of the statement of sound pressure level, there is no information
about sound pressure in other directions. Sound power determination
usually requires measurement of sound pressure levels at a number of
microphone locations. Directivity information may be available if desired
(provided, of course, the measurements were made under free-field
conditions). Some of the arguments in favor of expressing noise emission
in terms of sound power have been summarized by Lang and Flynn[^].
In spite of the above endorsement, sound power is frequently not the
appropriate quantity to use for noise ratings. As stated above, sound
power requires measurements in the far field of the source in a
well-defined acoustic environment. The extent of the far field of
Practical sources at any given frequency is not accurately known. For a
source whose radiating surface has a representative dimension, at the near
field is approximately circumscribed by the distances 2a /\ or 2a,
"whichever is larger. Here a, is the "dynamic" dimension of the source,
vhich may vary considerably with frequency and can be estimated
conservatively to be equal to the maximum source dimension, and X is the
33
-------
wavelength corresponding to the frequency of operation. Sometimes it is also
recommended that measurements be made at least one wavelength away from the
source. If measurements are made with the source located an effective
distance, h, above a reflecting plane, measurements should be made at a ?
distance away of at least Uh. The limitation to distances greater than 2a /A
tends to be the most serious one. At a frequency of 10 kHz, for instance,
the wavelength, X, is only . 03^m. Assuming a source dimension of a = 0.5m»
the measuring distance must be at least 15m. For real sources, the situation
would rarely be as bad as this. The 2a /A limitation is based on a source
for which all points radiate coherently. In many products, there are a
number of individual sources, a large number of vibrational modes are
excited, there may be aerodynamically-generated random noise, and the source
behaves more like an ensemble of small incoherent sources — in such a case,
the far field may begin very close to the source even at higher frequencies.
Recent investigations[5-7] have demonstrated that an accurate determination
of the sound power emitted may be derived from sound pressure measurements
which are made quite close, say, one meter, to the machine whose noise
emission is being measured. The validity of one-meter measurements has been
confirmed for all of the sources studied that emitted broad-band noise
without prominent discrete tones[5-7l« If such close-in measurements yield
sufficiently accurate results, smaller test facilities may be used and
ambient noise requirements may be relaxed.
Whether or not it is possible to make accurate sound power determina-
tions, there are instances where it is distinctly preferable to characterize
or rate machinery in terms of sound pressure. One of the most common of such
situations is when the noise rating relates to an operator location. If the
operator stays at a particular location, the sound pressure level at the
position of his head is the most useful quantity to measure. If the operator
moves around in some known area near to the machine, a spatial average of
sound levels over likely operator locations (e.g., a path around the machine
at ear level and arm's length from the machine) is appropriate. In some
cases where an operator works near one of a large assemblage of machines, it
may be necessary to know the close-in (operator positions) sound pressure for
his machine and, in addition, sound power of the other machines in order to
provide sufficient data to predict the total noise level at the operator
positions.
One may be primarily concerned with the noise radiated in a particular
direction so that total sound power information is not needed. Motor vehicle
noise ratings, for example, may not require total sound power information,
but only the sound pressure radiated to nearby communities. Of major concern
is the noise radiated to the side of the vehicle, while noise radiated to'the
front, back, and top is of much less concern and may be difficult to measure
for a moving source. Thus sound pressure at a far-field sideline position is
a logical measurement which relates directly to the noise levels in
communities.
Sound pressure is the preferred quantity for sources, such as firearms,
which produce impulsive noise. Here the instantaneous sound pressure and its
-------
time history are required in order to characterize adequately the sound and
assess the potential noise hazard.
In general, sound pressure is appropriate if the listener location is
well defined, the transmission to the listener does not differ much among
typical applications, and, if it is desired to extrapolate the data to other
locations, typical applications are such as to permit this to be done with
adequate confidence. Sound pover is usually more appropriate if information
is needed to predict sound pressure in a variety of environments which
significantly affect the resultant sound pressures. All other things being
equal, sound pover data are preferred for machines which are small enough to
"be tested under laboratory conditions; if directivity is a significant
factor, however, reverberation room measurements of sound power would not be
aPpropriate since no directivity information can be obtained.
11.5. Applicability
a. Types of Noise
The measurement standard should clearly state the types of noise for
which it is applicable. This is important since some measurement
Procedures are not at all appropriate for certain types of noise. The
folloving discussion of classification of noise by type is adopted from
American National Standard Methods for the Measurement of Sound Pressure
Levels, S1.13-19?l[8].
The noises usually encountered in practice are classified as steady or
nonsteady noise.
(l) Steady Noise. The level of a steady noise remains essentially
constant (that is, fluctuations are negligibly small) during the period of
observation. (See section 8.U of [8] for a discussion of observation
times). To the typical observer, a change in noise level of less than one
Decibel is not likely to be detectable while a six decibel change would be
considered significant. If the average noise level is relatively constant
but the spectral distribution of the sound changes during the period of
observation (as determined by listening), the noise shall be classified as
Unsteady.
(a) Steady Noise Without. Audible Discrete Tones. This type of
noise is frequently referred to as "broad-band" noise; prominent
discrete components and narrow-bands of noise are absent. The plot of
pressure spectrum level versus frequency is without pronounced
discontinuities.
(b) Steady Noise with Audible Discrete Tones. This type of
noise has components at one or more discrete frequencies which have
significantly greater amplitudes than those of the adjacent spectrum.
Clusters of such components or narrow-bands of noise may be observed.
35
-------
The spectrum obtained with a narrow-band analyzer has sharp peaks
(prominent single-frequency components) or steep gradients (narrow
bands of noise).
(2) Hongteady Noise. The level of a nonsteady noise shifts signifi-
cantly during the period of observation. This type of noise may or may not
contain audible discrete tones. The classification of nonsteady noises
depends upon the period of observation which must be defined for each meas-
urement .
(a) Fluctuating Noise. The sound pressure level varies over^a
range greater than six decibels with the "slow" meter characteristic
and does not equal the ambient level more than once during the period
of observation. Alternatively, the noise may fluctuate between two or
more steady levels six or more decibels apart when measured with the
"fast" meter characteristic of a sound level meter. Fluctuations
occur because of beats between two or more audible discrete tones
having nearly the same frequency.
(1°) Intermittent Hoise. The sound pressure level equals the
ambient level two or more times during the period of observation.
period of time during which the level of the noise remains at an
essentially constant value different from that of the ambient is of
the order of one second or more.
(°) Impulsive Noise (Bursts). Impulsive noise has brief
excursions of sound pressure (acoustic impulses) which significantly
exceed the ambient environmental sound pressure. The duration of a
single impulse is usually less than one second. Two subcategories of
impulsive noise are:
Isolated Bursts. One or more bursts occur during the peri°
of observation. The envelope of the burst waveform may be that
of a decaying transient or it may be of essentially constant
amplitude, for example, a tone burst. The burst spacing (time
interval between bursts) is such that each burst is individually
distinguishable with a sound-level meter.
Quasi-Steady Noise. A train of two or more bursts occur
during the period of observation. Individual bursts in the trai
may have equal or unequal amplitudes and the burst spacing (ti"16
interval between bursts) may be uniform or nonuniform. As the
burst repetition rate increases, the resolution of individual
bursts by a sound-level meter becomes difficult; the noise is
then classified as quasi-steady.
In specifying the measurement of impulsive noise, it is important to
be very clear as to the quantity desired: peak instantaneous sound
pressure or peak root-mean-square sound pressure averaged over a
short, and specified, averaging time.
36
-------
Examples of different types of noise sources are given in Table 2.
The acoustic environment has a strong influence on the types of noise
"which can be measured therein. Because a reverberant, or semi-reverberant,
room acts to average sound pressures over time, information cannot be obtained
concerning the short-term temporal variation of noise emission. Particular
caution is necessary in reverberant room measurements if the noise emitted by
the source contains significant discrete frequency and/or low frequency
components (see sections 12.2 and 15).
All types of noise can be measured in an anechoic space or a hemi-
anechoic space (i.e., free field conditions over a reflecting plane). In
conducting measurements over a reflecting plane, particular caution is required
if the noise emitted by the source contains significant discrete frequency
components or if a narrow-band frequency analysis is desired (see sections
12-2, 12.6 and 15).
Table 2
Examples of Sources of Different Types of Noise
Steady
Nonsteady
Without Audible
discrete tones
Distant city
Waterfall
With Audible
discrete tones
Circular saw
Transformer
Fluctuating
Heavy traffic (nearby)
Pounding surf
Intermittent
Aircraft fly-over
Automobile passing by
Train passing by
Impulsive Isolated Bursts
Drop forge hammer
Dog barking
Pistol shots
Door slamming
Electrical circuit breaker
Quasi-steady Noise
Riveting
Pneumatic hammer
Machine gun
37
-------
Different types of noise may require different types of instrumentation
(see section 13). Accordingly a measurement standard must call for specific
instrumentation performance characteristics; that standard will generally not
be suitable for all types of noise.
Departing now from the classification of noise by type as given in
American National Standard Sl,13-19Tlj it is also useful to classify noise as
being deterministic or nondeterministic. -Deterministic signals can be
described as an explicit function of time; nondeterministic signals cannot.
Deterministic signals are periodic or aperiodic. Periodic signals repeat
themselves in their entirety over some time interval. Aperiodic signals are
every other type of deterministic signal. Nondeterministic signals, also
referred to as random signals, may be described statistically. The probability
function for time-varying wave forms indicates the relative frequency of
occurrence of the various instantaneous values of the wave form.
The accuracy vith which an acoustical signal can be measured depends
upon, among other things, the rise time and the averaging time of the
measurement system. The signal must be "looked at" long enough for the
instrumentation to respond and, in the case of nondeterministic signals, long
enough to obtain a suitable average (see section 13).
b. Nature and Size of Source
A noise measurement standard should include an indication of the types of
noise sources encompassed. This should include such factors as stationary or
mobile, size of source, allowable operating characteristics, etc.
c. Measurement Uncertainty
A noise measurement standard should provide guidance as to the
measurement uncertainty which can be tolerated in each regulatory situation.
To have operational meaning, the uncertainty of a measured value must be given
relative to some actually achievable and nationally accepted reference. To do
this, a clear statement of what would constitute the nationally accepted answer
is needed. In some cases references to standards.as maintained by a national
laboratory may be sufficient; in others the reference may be a consensus of
many laboratories.
The current International Organization for Standardization standards and
draft standards for determination of sound power emitted by stationary noise
sources associate with each document[9-15] one of the following grades of
accuracy: precision, engineering, and survey. Within each document a
statement such as the following is made[l^]:
Measurements made in conformity with this International Standard tend to
result in standard deviations which are equal to or less than those given
in tables 2 and 3. The standard deviations of tables 2 and 3 reflect the
cumulative effects of all causes of measurement uncertainty, excluding
variations in the sound power of the source from test to test....
38
-------
Table 2. Uncertainty in determining sound pover levels of sound sources
in anechoic rooms
Octave Band
Center Frequencies
Hz
125 to 500
1000 to kOOO
8000
One-Third Octave Band
Center Frequencies
Hz
100 to 630
800 to 5000
6300 to 10000
Standard Deviation
of Mean Value
dB
1.0
0.5
1.0
Table 3. Uncertainty in determining sound pover levels of sound sources
in semi-anechoic rooms
Octave Band
Center Frequencies
Hz
125 to 500
1000 to HOOO
8000
One-Third Octave Band
Center Frequencies
Hz
100 to 630
800 to 5000
6300 to 10000
Standard Deviation
of Mean Value
dB
1.5
1.0
1.5
This particular example corresponds to precision measurements in anechoic and
hemi-anechoic environments. Analogous tables are contained in the other draft
standards. These documents also include discussions of the major causes of
uncertainty.
Whenever possible, the standard should specify the components of variance
which should enter into the uncertainty statement and the means for determining
a realistic limit to the offset of the process from the "accepted" process as
defined in the standard. In addition, the standard should specify the type of
evidence needed to establish that the measurements arise from a measurement
process which is in a state of control (i.e., has predictability).
If a noise regulation for a particular class of products requires all
noise measurements to be made in accordance with the same measurement standard,
and if the inter-laboratory reproducibility is excellent, that may be
acceptable even if a systematic bias is suspected since that bias would affect
all measurements similarly. On the other hand, a measurement process might be
capable of producing results that would average out, over a very large number
of measurements, to be quite accurate but the imprecision could be so large as
to be unacceptable. In general, and if sufficient knowledge exists, a
statement such as the following should be made:
Measurements made in conformanse with this standard tend to result in a
standard error of X dB and a systematic error of not more than ±Y dB.
Provided that separate statements of imprecision and systematic error are
included, it is acceptable to present a statement placing bounds on the
39
-------
inaccuracy, i.e., the overall uncertainty of measured values. In general, the
"bounds indicating the overall uncertainty should not "be numerically less than
the corresponding "bounds placed on the systematic error outwardly increased by.
at least three times the standard error. For example:
...result in an overall uncertainty of ±5 dB based on a standard error of
1.0 dB and an allowance of +2.0 dB for systematic error.
Considerable care must be taken in selecting values for the standard
error and systematic error which are appropriate for use in a measurement
standard. The standard error should correspond to the reproduciMlity among a
number of laboratories, rather than the repeatability in a given laboratory.
Similarly, the allowance for systematic error should correspond to that of the
measurement standard as realised in a given laboratory. Thus, the statement
just prior to this paragraph should be interpreted as indicating that, if a
very large number of laboratories vere to carry out measurements on a given
noise source, approximately 99% of the laboratories would obtain values within-
+_ 3 dB {corresponding to 3 standard errors) of the mean of all the values and
that mean value (assuming enough laboratories participated so that the mean of
the values obtained did not differ significantly from the mean of an infinite
population of laboratories) would not differ from the "true" value "by more than
+2 dB.
Neither the standard error nor the systematic error bound are easy to
obtain. Selection of appropriate values should be carefully considered "by the
group preparing the standard. A worthy, but difficult goal is to specify the
test environment, the source-operating procedures, and the measurement and
calibration procedures to maximize both accuracy and precision while minimi^illg
the time, difficulty, and cost of conducting the test.
All in all, it is the "quality" of measurements which is at issue. This
quality depends on a variety of factors including the operator, the environ-
ment t the item or quantity "being measured, the instrument, etc. Specification
on the instrumentation (e.g., traceability, re-calfbration schedule, etc.)
particularly when the results are done under the extremely restrictive con-
ditions of a laboratory may not give valid indications of the performance of
the instrument in routine use. In regulatory measurement, it is particularly
important that the focus be on the correctness of individual measurements,
because of the cross-examination" to which the measurement will be submitted
in legal proceedings. The decision as to the adequacy of measurement should
not he allowed to depend solely on pro form traceability requirements rel
to the scientific evidence which is needed to determine the uncertainty of
measurement (see Appendix C).
UO
-------
12. Acoustic Environment
• The measurement standard should specify and provide criteria for the
evaluation of the allowable test environment(s), e.g., aneohoia*
hemi-anechoio (free-field over a reflecting plane)3 reverberant and in_
situ.
• For those cases where the background noise is significantly3 but less
than 10 dB below the noise being measured* the measurement standard
should provide correction factors.
• The measurement standard should include restrictions on external
factors which can affect the sound pressure level at a particular
point or the accuracy of the measurement of that sound pressure level.
Allowable ranges of temperature, barometric pressure3 humidity and
wind over which measurements may be made as well as restrictions on
vertical and horizontal (ground plane) reflecting surfaces and
corrections for the effects should be included wherever possible.
In Section 5.2, a brief discussion was given of the effect of the
environment on the noise levels produced "by various sources. From this
discussion it should "be clear that meaningful measurements of noise
emission generally can be made only in a veil-characterized acoustic
environment. Whenever the noise source can "be moved readily into a lab-
oratory, or other controlled environment, it is generally preferable to
make acoustical measurements there rather than to be subject to the
vagaries of nature and the need to characterize a new acoustic environment
every time a test is carried out at a different location. Due to the size
or mobility of many noise sources, however, it is frequently necessary to
conduct measurements outdoors at a particular test site or, either outdoors
or indoors, in situ.
12.1. General Requirements
Four basic test environments frequently are used for acoustical
measurements:
Anechoic — An anechoic (meaning "without echo") environment is
usually obtained in an, indoor chamber, or room, in which the walls, floor,
and ceiling are specially treated to absorb virtually all of the incident
sound energy. An anechoic space can also be obtained outdoors, by
suspending the test specimen well away from the ground or other reflecting
objects. An anechoic space can sometimes be well approximated in a large
room, without special acoustic treatment of the room boundaries, if
measurements can be made close enough to the source that there is
negligible contribution from reflected sound energy.
In an anechoic environment, "free field" conditions prevail and the
noise emission characteristics of a source can be determined simply from
measurements of sound pressure in the far radiation field (see Sec. 15) of
the source. In general, any type of noise signal can be measured in an
Ul
-------
anechoic environment. A particular advantage of an anechoic environment is
that the directivity of the radiated sound field can be measured. Several
limitations to the use of anechoic environments are: (l) the high cost of
anechoic chambers, (2) difficulty in suspending large sources well away from
any reflecting surface, and (3) difficulty in providing for operating loads,
heat loads and exhaust emissions.*
Hemi-Anechoic — A "hemi-anechoic" environment is one in vhich free-field
conditions exist in the half-space on one side of a totally or partially
reflecting plane. Usually a hemi-anechoic space is the free space above a
hard, reflecting, horizontal surface but the concept can be extended to include
free-field measurements near any partially reflecting surface. A hemi-anechoic
space can be obtained in an indoor chamber with the walls and ceiling acousti-
cally treated as in an anechoic chamber but with the floor acoustically hard
and reflecting. A hemi-anechoic space can also be obtained outdoors, with the
test specimen resting on or above a paved surface. Occasionally outdoor
measurements are made oyer a partially reflecting surface such as grass.
In a hemi-anechoic space, as in an anechoic space, noise emission is
determined from sound pressure measurements in the far radiation field of the
source. Measurements are complicated, however, by interference effects due to
the combination of direct sound waves and waves, from the reflecting plane.
This results in local maxima and minima in the sound field and considerable
spatial sampling often is required to obtain meaningful results (see Sec. 15)-
In general, any type of noise signal can be measured in a hemi-anechoic
environment. Directivity inforiaation can be obtained under conditions typical
of use conditions for many types of equipment -- e.g., over a reflecting plane-
There is no particular difficulty in supporting or operating large or mobile
equipment. In conducting outdoor measurements considerable difficulties can
arise due to wind, rain, humidity and temperature.
An extension of the hemi-anechoic environment is a free field adjacent to
two or three intersecting reflecting surfaces. This could be useful when a
product is normally so installed.
*The time required to make measurements at a large number of microphone
locations, in order to obtain adequate spatial averaging in an anechoic
environment, has been indicated as a reason for preferring reverberant-
field measurements to free-field measurements. However, as the under-
standing of the requirements of reverberant room sound power determinations
has increased, it has become clear that many microphone locations, and
frequently several source locations, also are needed for adequate sampling
in reverberant fields. If only A-weighted sound power, rather than
spectral information, is needed, anechoic chamber measurements may be
easier than reverberation chamber measurements since, in reverberation room
measurements, it usually is necessary to determine sound power versus
frequency and compute A-weighted power rather than measuring A-weighted
levels directly as can be done in an anechoic chamber.
-------
Reverberant — Reverberant field measurements are carried out in a ro<
vith acoustically hard walls (i.e., low sound absorption) such that sc
waves undergo many reflections "before being absorbed. Reverberation ri ~.j
are typically equipped with stationary or moving diffusers to increase the
uniformity of the reverberant sound field.
In a reverberation room, one measures sound pressure level and
computes total sound power, the computation being based on the premise that
the mean-square sound pressure averaged in space and time is (l) directly
proportional to the sound power output of the source, (2) inversely
proportional to the total absorption in the room, and (3) otherwise dependent
only on the physical constants of air density and velocity of sound.
Reverberant room methods are particularly advantageous for sources which
produce steady noise and for which directivity information is not required.
Noise measurements in a reverberation room are dependent upon accurate
measurement of the sound energy density, in the frequency bands of
interest, in the reverberant field of the room with the noise source in
operation. In order to obtain such accurate measurements it is desirable
for the room volume to be sufficiently large (see Sec. 12.2) that many
normal modes are excited in the room. At low frequencies and/or when the
room is excited by pure tones or very narrow bands of noise, it is common
for only a few modes to be excited and accurate measurement of energy
density requires very extensive sampling of the sound field, preferably
with the use of moving diffusers.
In Situ — Certain types of equipment cannot, for practical reasons, be
operated outdoors or in a laboratory environment. When the noise emission
of a source must be measured in situ, it is highly desirable to utilize
procedures which approach, as nearly as possible, one of the three cases
just discussed. Two of the recent Draft International Standards for sound
power determination[l3,15] include provisions for correcting for a.
less-than-ideal acoustic environment. The need for such corrections would
almost always imply a degradation of measurement accuracy.
12.2. Criteria for Adequacy of the Test Environment
This section includes more detailed discussions of the criteria which
each of the above four types of acoustical environment should meet.
a. Anechoic Environment
The Draft International Standard for precision determination of sound
power in an anechoic chamber[lU] states (Annex G) that "the volume of the
test room shall be large enough so that the microphones can be placed in
the far radiation field of the sound source under test, without being too
close to the absorptive surfaces of the test room." It is instructive to
consider briefly the measurement error that might result due to reflections
from the wall of an anechoic chamber.
-------
Consider a situation such as shown in Figure 3, with a sound source in
an anechoic chamber and a measuring microphone in the far radiation field of
the source. It is assumed that the source emits a spherical sound wave at an
angular frequency w = 2-rrf. The instantaneous pressure of the emitted wave
can be expressed as
P
Pe(r,t) = - cos [u(r/c - t)], (1)
where P is the amplitude (Pa) of the pressure at a unit radial distance, r -
1 m from the source, t is time (s), and c is the velocity of sound (m/s) in
the medium. The wave reflected from the chamber wall may "be approximated by
a wave emitted from an image source a distance L to the right of the
effective location of the chamber wall.* The instantaneous pressure of the
reflected wave can be expressed as
RP
Pr(r',t) - p- cos[w(r'/c - t) + $], (2)
where r' is the distance from the image source to the measurement location,
is the magnitude of the pressure reflection coefficient of the chamber wall
normally-incident plane waves, and is a phase angle which depends on the
acoustical impedance of the chamber wall. For normal incidence, r1 = 2L - r.
The amplitude of the total pressure, p = p + p , will be between the following
limits:
the lower limit corresponding to destructive interference and the upper
corresponding to constructive interference.
*The analysis given in this section is fairly accurate for estimating the
effect of the reflecting plane in hemi-anechoic measurements or the effects
of the walls in reverberation measurements. It is less accurate for "soft"
walls such as those in an anechoic chamber. In addition, since the absorp-
tive lining of an anechoic chamber has considerable thickness, the "effective
location" of the wall is uncertain. Thus the analysis which follows is
intended to qualitatively show effects but should not be taken too seriously
in a quantitative sense. More exact analyses of the sound field due to a
point source near a partially reflecting boundary are given by[l6,17], which
also reference earlier work.
-------
EFFECTIVE LOCATION
OF CHAMBER WALL
SOURCE
MICROPHONE
T
IMAGE
SOURCE
Figure 3. Schematic representation of a sound source in an anechoic chamber.
The sound field produced by a source and its "image" only approxi-
mately represents the actual sound field.
-------
In terms of mean-square pressure ratios,
,2
Converting to sound pressure levels, the difference, AL, between the measured
level and the level in the absence of reflections is bracketed by:
20
log [l - ff] < AL < 20 log [l + |*J . (5)
where "log" designates the common logarithm (base iol and where rR/r'
represents the ratio of the magnitude of the reflected pressure, at any r, "to
the magnitude of the incident pressure, at the same r. Thus the measured
deviation of the pressure from that of a spherically spreading wave depends
on the product of two independent, non-dimensional parameters, R, and r/r'.^
R depends on the wall impedance only while r/r' accounts for the reduction i*1
the reflected pressure at any point r due to the spherical spreading of the
pressures from the real and image sources.
Defining
"a-*-
the range of the maximum sound pressure level deviations from a spherically
spreading wave becomes, from eq. (5),
20 log U-nR]
-------
IL
0.99
0,95
0.90 0.85
CD
•o
UJ
O
Z
k. Maximum range of uncertainty in the sound pressure level in an
anechoic chamber as a function of 1^ and E^t defined in the text
(for single wall, normal incidence).
-------
A standard for noise measurements in an anechoic chamber should provide
guidelines for constructing a suitable chamber and specify criteria for
ascertaining its adequacy (see Annexes A and G of
The usual procedure for determining the adequacy of an anechoic chamber
is to measure the pressure falloff as a microphone is moved a distance, r, away
from a source which has been selected so as to radiate essentially a spherical
wavefront so that the sound level should fall off as L = L -20 log r, where
L is the sound level at unit distance from the acoustic center of the source
(assuming that one is in the far field at a unit distance) — this corresponds
to a 6 dB decrease in sound pressure level for every doubling of the distance
from the source. Returning to eq. (6), it is illustrative to plot the sound
level versus measurement distance from the source. This is done in Figure 5
for R = 0.2 (a = 0.96); the values selected for other parameters are given in
the figure caption.
As indicated previously, the analysis above is only approximately correct
for real anechoic chambers. Particular difficulties could occur in some
instances, for example:
- microphones placed near room edges or corners could receive reflections
off two or three walls thus increasing the error
- highly directional sound sources can cause particular problems due to
sound emitted from the "louder" side of the source being reflected into
microphone locations on the "quiet" side of the source.
An additional, and quite useful method for checking for unwanted re-
flections in anechoic chambers is to use a pulsed sound source or a correlation
technique to determine not only the magnitude of the reflected energy but also
the surface (e.g., supporting structures) from which the 'sound is being
reflected.
The principal limitations on microphone locations in an anechoic chamber
are (l) to get the microphone far enough away from the source to be in the far
radiation field (see Sec. 15) and (2) to keep at least a quarter-wavelength, at
the lowest frequency of interest, away from any part of the absorptive lining
of the chamber to ensure that the sound pressure does not start to fall off
rapidly due to the proximity of the absorptive material.
Useful references concerning the design and qualification of anechoic
chambers include [ 18-26 ].
b. Hemi-Anechoic Environment
Qualification of a hemi- an echoic environment is essentially the same as
that of an anechoic environment , The test source generally should be placed as
close to the reflecting plane as possible to promote spherically divergent
waves and thus minimize the influence of reflections from the plane.
-------
CD
•o
: -5
UJ
UJ
DC
CO
a
cc
Q.
o
CO
UJ
<
cc
-10
-15
-25
I I r
\
t
J
•
1 1 1 1 1 1 1
10
DISTANCE, m
Figure 5. Hypothetical variation of sound pressure level versus distance as
one wall of an anechoic chamber is approached. The dashed line
indicates the fall-off (inverse square lav) that vould "be expected
in the absence of any reflections from the wall. The points were
computed from the sura of eqs. (l) and (2) assuming f = o)/2i7 -
100 Hz, $ - IT, and R = 0.2. The effective location of the
chamber wall was taken as 10 m from the source.
1*9
-------
In addition to measuring the falloff of sound pressure with distance to
determine the effectiveness of the five absorbing surfaces, it is necessary
to check the reflectivity, or absorption of the reflecting plane.* The
current Draft International Standards for "both engineering[ 13] and
precision[lU] determinations of sound power in a free field over a reflecting
plane call for the normal-incidence energy absorption coefficient, a, to be
less than 0.06 (R > .97). This criterion normally would be met for dense
concrete but the absorption might be too large in the case of unsealed
asphalt pavings or certain types of floor coverings.
An indication of the reason for setting what may seem to be a rather
strict requirement on the absorption of the reflecting plane follows from an
extension of the analysis given in the previous section. Consider Figure 6,
which shows a spherical source located above a reflecting plane. The
microphone is no longer required to be directly between the source and the
image source. With r and r1 defined as shown in Figure 6, eqs. (l) through
(9) remain valid, for a locally-reacting plane. The limits of eq. (.5) are
shown in Figure 7, plotted against both R and a with r/rr as a parameter.
The upper limit on the measured sound pressure level is seen to be rather
insensitive to the value of a since when the reflected signal is in phase
with the direct signal and r/r'—*•!, the combined signal level can be at most
6 dB above the direct signal. However when the reflected signal is out of
phase with the direct signal, the combined sound pressure-*-0 as r/rr—*•! an
cfr—*-0. In order to be in the far radiation field Csee Sec. 15), it is
generally necessary that r/r?—+-1 so the sound level in regions of phase
cancellation depends strongly on the value of a. For a source which radiates
uniformly in all directions, requiring a to be less than about 0.06 ensures
that the sound level in such regions will be sufficiently below the level in
regions of phase reinforcement that no significant error, relative to the
levels corresponding to a perfectly-reflecting plane, in the measured noise
emission will occur (provided, of course, measurements are made at enough
locations to ensure adequate spatial averaging). For highly directional
sources, still stricter requirements on a might be needed.
The normal-incidence energy absorption of the material constituting the
reflecting plane should be measured, for example, in accordance with[27]-
Measurement of such low absorption coefficients is difficult by an impedance
tube technique and care is required to obtain accurate data. A direct
measurement of the pressure reflection coefficient might be preferable, for
example by pulse or correlation techniques, but no current standard method of
measurement exists.
*In the case of outdoor measurements, of course, there will be no walls and
ceiling to be concerned with. However, site qualification is very useful to
determine the influence, if any, of nearby reflecting or absorbing surfaces.
-------
MICROPHONE
SOURCE
REFLECTING
PLANE
IMAGE
SOURCE
Figure 6. Schematic representation of a sound source in a free field above
a reflecting plane.
-------
0.99
0.98
0.96 0.94 0.90
0.8
CD
•o
o
o
OJ
I I I I I I I I
r/r'
-40
0.01 0.02
Figure 7. Limits for the effect of the image source (see Figure 6) on the
observed sound pressure level as a function of the absorption
coefficient of the reflecting plane with r/rr as a parameter
solid lines correspond to destructive interference while the
line corresponds to constructive interference.
5.2
-------
If the reflecting surface is not an integral part of the hemi-anechoic
environment care must be taken to ensure that noise sources cannot excite the
reflecting surface into vibration of such magnitude as to radiate significant
sound energy. If there is any question that this might occur, acoustical
and/or vibration measurements should be made to estimate the energy radiated by
(as opposed to reflected from) the reflecting plane.
Corrections for the influence of the acoustic environment, other than
the major reflecting plane, are discussed"in Section l6.2.
c. Reverberant Environment
Particularly in the case of noise measurements in reverberant envir-
onments, it is difficult to separate the adequacy of the test environment
from the adequacy of the entire measurement process. In Section 15,
attention will be directed to such factors as the variation in sound pover
output due to source location, spatial variation of sound pressure, and the
use of reflecting elements to increase the "diffusehess" of the sound
field. In the present section the properties which the room should have
will be considered.
Early workers developed the basic reverberation room theory using
certain simplifying assumptions: (l) uniform, diffuse distribution of
sound energy throughout the room at any instant; (2) equal probability of
propagation of sound in all directions; (3) continuous absorption of sound
by the room boundaries. The assumptions were those of geometrical
acoustics, in which sound energy is considered to travel in rays and all
wave phenomena are neglected.
The sound energy in a reverberant room can be considered as two com-
ponents, that in the direct field of the source, which has not yet suffered a
reflection from a room boundary, and that in the "reverberant field", which
can be somewhat arbitrarily defined as that sound which has undergone one or
more reflections[28, p.311]. Of the total power, W, emitted from the source
a fraction, a W, will be absorbed at the first reflection, leaving a total
power of (l -°a )W in the reverberant field. Here the absorption
coefficient, a ? corresponds to an appropriate average over the surfaces of
the room, with°account being taken of variations in the absorption
coefficient as a function of location and angle of incidence. Note that a
does not correspond to a diffuse field absorption coefficient since there °
certainly is not an equal probability of the energy, directly from the
source, arriving from all possible directions.
Conservation of energy requires that:
Rate of increase
of reverberant
energy in room
Rate of emission of
energy from source
into reverberant field
Rate of absorption"
of reverberant
_energy
53
-------
In the spirit of the assumption of a uniform diffuse reverberant sound
the rate of absorption of reverberant energy at any given instant of t
should be proportional to the sound energy density in the room at that
instant. The total energy in the reverberant sound fiel V * ™R ' ^ber ant
the volume (m3) of the room and D is the energy density in the reverberan
sound field (Jm-3). Thus the differential equation which follows from the
above statement of energy conservation is
(10)
where T is a characteristic time which defines the rate of sound absorption
in the room, whether by the walls, the air in the room, or by other
surfaces such as diffusers. Under steady-state conditions (dDR/dt = 0}»
the energy density is given by
T(! - a )W
R 2V '
The energy density in a diffuse sound field can be shown to be equal to
p /pc , where p is the mean-square sound pressure (Pa ) and p is the
density of air (kg m ) in the room; thus
_ pc2T(l - a )W
- 2 = _ - . - °__ (12)
PR 2V
is the mean-square sound pressure in the reverberant field only.
If the reverberant sound energy density is assumed to have a constant
value, and, at time t = 0 the sound power input is suddenly turned off*
energy density will decay as
_
DR dt " T '
so that T also is seen to be the time constant which defines the rate of
decay of the reverberant sound field; alternatively
^ 2
dt T
where "An" designates the natural logarithm (to the base e - 2.71828. ...)>
indicating that T can be directly obtained by measuring the rate of decay
of the sound pressure level when the source of sound power is turned off*
-------
2 2
Converting to sound pressure level, L = 10 log (pR /p ), where p
20 micropascals (yPa) is the reference pressure, and sound power level, L.,
= 10 log (W/Wn), where W~ = 1 picowatt (pW) is the reference power, eq.
(12) becomes
+ 10
+ 10 l°g dB. (15)
The time constant, T, is related to the commonly used reverberation time, T
— the time required for the sound field to decay by 60 dB — by the relation
T = 6.91-r, the non-integer number arising from the conversion between natural
and common logarithms. The last term in eq. Cl5) can be evaluated by
recalling that for an ideal gas, pc = yB» where y is the ratio of specific
heats (= l.U for air) and B is barometric pressure (Pa). Utilizing these
quantities, eq. (15) can also be written as
a )]
j +
Lp =1^ + 10 log - y - + 10 logB - 36.0 dB. (16)
Equation (l6) is analogous to the corresponding expressions in the current
international [10] and national [29] standards for determination of sound
power in reverberation rooms, with the following differences:
_ the constant term differs because the barometric pressure in the
standards is expressed in millibars (l bar = 10 Pa);
_ there is an additional term in the standards arising from interference
effects near the room boundaries; the above derivation neglects such
effects ;
-the standards do not include the term 10 log (l - aQ); this will be
discussed shortly.
In order to ensure that the sound field "sees" the same effective absorp-
tion in both the steady-state power determination and the transient decay
rate for reverberation time measurement, the sound source used to excite
the room for reverberation time measurements preferably should be in the
same location and have the same directivity as the source whose sound power
is being determined.
An alternative method of measuring the effective room absorption is by
use of a reference sound source of known sound power output. This known
source should be at the same location as the unknown source and preferably
have a similar directivity. Looking at eq. (l6) for tests on two sources
under otherwise identical conditions,
55
-------
so that no explicit knowledge of room properties is required.
These expressions are based upon measurements of the mean-squared sound
pressure of the reyerbjerant field. Thus it is necessary to examine the
relative strengths of the direct and reverberant fields to determine where
microphones may be located. The mean-square sound pressure due to the direct
field may be expressed as
'D
2 .
pc
4-irr2
(18)
where the directivity factor Q is defined as the ratio of (l) the mean-
square sound pressure measured at angle 9 and distance r from an actual
source radiating a power W to (2) the mean-square sound pressure measured at
the same distance from a nondirective (spherical) source radiating the same
total acoustic power W.
The total mean-square pressure due to "both the direct and the rever-
berant field follows from eqs. (12) and (18):
= pc'W
4-rrr2
2V
(19)
The difference between the total sound pressure level and the reverberant
field sound pressure level is
10
1 +
2V
47rr2
(20)
Substituting T = 6.91T and c = 3^3.U m/s (air at 20° C),
1 + 0.0032
V
(21)
The difference, L - L , between the total sound level and the reverberant
T R 1/2
field level is plotted vs r/Q , the equivalent distance from the acoustic
center of the source, in Figure 8 with V/[T(.l-a } ] as a parameter. It can W
seen that as V/T becomes large, it is necessary°to be much farther from the
source in order for the direct field contribution to the sound field to
become negligible. While it is rather obvious that, for a chamber of a given
volume, the direct field contribution decreases as the reverberation time
increases, it is perhaps less evident that for a given reverberation time, it
is necessary to be farther from the source in a large chamber than it is in a
small chamber.
-------
CD
CK
Q.
I 1 1—I I
Figure 8. Difference between the total sound level and that due to the
reverberant field alone, shown versus the effective distance from
the source.
-------
Equation (2l) and Figure 8 provide information as to the criteria which
a reverberation chamber should meet in order to have a negligible
contribution from the direct field at reasonable distances from the source.
The key points are
T/V should be as large as possible; this is equivalent to the total
absorption in the room being as small as possible;
highly directive sources require that measurements in the directions J
of maximum noise emission be taken at distances from the source which
are larger than when Q-l.
There is still another reason to require the sound absorption in the
room to be very small. Returning to eq. (l6), the term 10 log (l - ctQ) is
normally neglected in computing sound power from mean-square sound pressure
levels in the reverberant field. As discussed at the beginning of this
section, the particular value of a may depend on the location and
directivity of the noise source ana, in general, cannot be measured directly
in any simple manner. Figure 9 shows -10 log (l-ct ) plotted vs a ; it is
seen to become a significant correction when a exceeds a value or about 0.1-
An approximate value for a can be computed from the reverberation time and
room geometry but the percentage uncertainty in the value so obtained could
be fairly large.
The above discussion does not address wave effects. "When consideration
is also given to wave effects, further guidance on room size, shape, and
absorption can be obtained. The frequency of a normal mode of vibration
(so-called "resonance frequency") in a rectangular room with hard boundaries
is given by
2TT
(22)
where nx, n , and n are integers and S, , I , and i are, respectively, the
length, width, and Height of the room. It ?an be sSown that as f increases
the number of normal modes in a frequency band of width Af centered on f
approaches the value
AN =
Af , (23)
where V is the volume of the room, S is the total surface area of the walls»
and L is the sum of the lengths of all edges of the room. For frequency
analysis in 1/3-octave or 1/1-octave bands, eq. (23) may be written as
(24a)
-------
P .6 —
o
o
o
I
.20
Figure 9. The effect of neglecting the 10 log (l-a ) term in eq. (l6)
59
-------
and
._ 8.89V f 3 1.11S
3 C o
c3 c*-
(24b)
where f is the "band center frequency. For a given room volume, these
equations predict that a cubical room vill have the lowest modal density of
any "basically rectilinear room shape. Figure 10 shows the approximate number
of modes in 1/3-octave frequency "bands, as a function of frequency, for
cubical rooms U, 6, 8 and 10 m on a side. There would be approximately three
times as many modes in an octave "band having the same center frequency.
A knowledge of the characteristic frequencies of a reverberation room is
important in terms of understanding its properties as a measurement tool. I
excited at a location where there is not a node in the pressure standing wave
pattern, a room will act as a resonator and may respond strongly to impresse
sound energy at frequencies near to the characteristic normal mode
frequencies. The extent to which the room responds is dependent upon the
reverberation time. In electrical circuit theory it is customary to talk
about the "Q" of a circuit element and in microwave theory, the "Q" of a
cavity. In the same sense it is useful to consider the Q of a reverberation
room. Q is defined as
Q • -V ' 05 • (25)
where f is a natural frequency of the room and T = 6.91t is the rever-
beration time. The frequency response of the mean-square sound pressure,
large Q, is given approximately by
_ f2/f 4
p2 (f) « ° . (26)
(f2/fo2 - 1)2 + i/Q2
Substituting eq. (25) into eq. (26),
f2/f "
P2 (f) « —— s- = Ll!o (27)
-------
31.5 63 125 250
BAND CENTER FREQUENCY, H*
500
Figure 10. Approximate number of normal modes in a 1/3-octave frequency
band for cubical rooms of the volume shown.
61
-------
100 Hz, there are only three normal modes, at frequencies of 95> 100, and 105
Hz, respectively. It is further assumed that these modes all have the same
reverberation time (see discussion below). Figure 11 shows the frequency
response of the room, when all three modes are of equal "strength", for
reverberation times of 1 and 20 seconds. It is seen that when T is large the
average sound pressure level in the room varies rapidly with frequency.
Figure 12 illustrates how the "range" of this variation, defined as shown in
Figure 11, varies with the reverberation time of the room for different
spacings of the normal mode frequencies. In Figure 12, it has been assumed
that there are a number of modes of equal "strength" spaced uniformly, at a
separation of 6f = 1, 2, 3, ht or 5 Hz, within a frequency band centered at
100 Hz. (Alternatively, the modal spacing can be thought of as being 1, 2, 3,
h, or 5 percent of any band center frequency.) Mote that both T and 6f depend
upon the room volume.
If a noise source were to emit only a pure tone at some frequency in the
band pass region of the filter, the resultant mean-square sound pressure in
the room would be dependent upon how close the driving frequency is to one or
more resonant frequencies. In order for the room to be relatively insensitive
to such an effect, it would be desirable to have a large number of modes in
the measurement bandwidth, to have those modes as uniformly spaced as
possible, and to have the resonance peaks broad enough to "fill in the gaps"
between adjacent peaks. The implications of these observations are:
-the room should be large to increase the number of modes in a given
frequency interval,
-the geometry of the room should be selected to maximize the number of
modes and promote uniformity of spacing,
-the reverberation time should be low enough to broaden the resonance
peaks by an amount comparable to the modal spacing.
Summarizing, the number of modes is controlled by the size of the room, while
the modal spacing is controlled by the shape of the room. If the room is very
hard (i.e., long reverberation time corresponding to low absorption} the
resonance peaks will be sharp — at lower frequencies where the modal density
is low and the modal spacing may not be uniform, the room response, as a
function of frequency, may not be as uniform as desired. Thus it is desirable
to add low frequency absorption to enhance modal overlap. However, additional
absorption will decrease the spatial volume in which the reverberant sound
field is well above the direct field from the source. A compromise should be
sought between these two effects.
Inspection of eq. (22) shows that a phenomenon, known as degeneracy, in
which the same resonance frequencies occur more than once, will be exhibited
any time the dimensions of the room are in the ratio of small integers. Among
62
-------
95
T=20s
100
FREQUENCY, H*
105
Figure 11. Frequency response of a reverberation room having equal-strength
modes at 95, 100, and 105 Hz for reverberation times of 1 and
20 s, respectively.
-------
i i—i—rn
REVERBERATION TIME.s
Figure 12. Range of the variation (peak-to-valley in Figure 111 in the fre-
quency response of a reverberation room for frequencies near^l°°
as a function of reverberation time for different modal spacing3•
-------
rectangular rooms, a cubical shape exhibits the largest number of degenerate
modes and hence clearly should be avoided. Current international[10] and
national[29] standards list the following room proportions as having been
found to give satisfactory mode spacing for rooms of about 200 m volume (the
"preferred minimum room volume" for measurements down to the 1/3-octave band
centered at 100 Hz):
S. /£
y' x
0.83
0.83
0.79
0.68
0.70
i, /a
Z X
o.ki
0.65
0.63
0.1*2
0.59
The width of a resonance line shape can be characterized by the width
of the peak (see Figure 13) between the frequencies at which the energy
density is one-half of what it is at the exact resonance frequency. This
width, given by
Af = fo/Q
I/TTT = 2.20/T
(28)
defines the frequency region in which the mean-square pressure is within 3 dB
of that at the resonance frequency. It would seem reasonable to select the
room's reverberation time such that the width of a resonance line is
significantly greater than the average spacing between normal modes for the
lowest frequency of interest (i.e., modal overlap, M = Af/6f, significantly
greater than unity); thus, from eq. (23)> one would desire
T
InrV
3
f2,
(29)
where M is the desired amount of modal overlap and only the most significant
term in eq. (23) has been retained. With c = 3^3. ~
expressed in cubic meters, this becomes
m s~ (air at 20°C) and V
(30)
If one selects M = 3, the lowest frequency at which a reverberation room could
be used and still meet the selected criterion for modal overlap would be
/m\l/2
f >_ 2100 l±] . (31)
All of the analyses in this sub-section are rather approximate. For more
rigorous treatments of room acoustics, see[30-36].
-------
UJ
LU
UJ
cr
CO
co
UJ
en
CL
Q
3
o
CO
Af =
1°.
Q
TTT
FREQUENCY-
Figure 13. Drawing illustrating the spacing, 5f, "between normal modes, the
width, Af, of a normal mode, and the relationships among Af»
f /Q, T, and T.
66
-------
As discussed earlier, it is desirable to make the reverberation time as
large as possible so that the uncertainty due to inadequate knowledge of a is
acceptable and, over most of the room volume, the contribution from the direct
field is negligible compared to the reverberant field. At lower frequencies,
these desiderata are in conflict with the just-discussed criterion calling for
a reverberation time small enough to spread the resonance peaks in the room
response. For any given room it is desirable to analyze all of these factors
and select an adequate compromise for the situation at hand.
The current American standard[29] requires that the room volume be "at
least 180 m and preferably 200 m for measurements including the 125 Hz
octave band, and 70 m for measurements covering the 250 Hz and higher octave
bands, but excluding the 125 Hz band." It further states that
"the floor of the test room shall be reflective with an absorption
coefficient below 0.06. Apart from the floor, none of the surface should
have absorptive properties significantly deviating from those of the
other surfaces. For each one-third octave band within the frequency
range of interest the mean value of the absorption coefficient of each
wall and of the ceiling should thus be within 0.5 and 1.5 times the mean
value of the absorption coefficient of all walls and ceiling."
The following guidelines are given with regard to the absorption of the room:
"The sound absorption coefficients of the surfaces of the reverberant
room must be small enough to insure an adequate reverberant field. The
coefficient must be large enough to minimize the effect of source
position on the sound power produced by the source (refer to the
qualification procedure of Section ll). The average sound absorption
coefficient of all surfaces of the reverberation room should not exceed
0.06 over the frequency range of interest, except that additional
absorption below a frequency given by
- _ 2000
V1/3
is usually desirable in order to increase the bandwidth of the resonance
curves of the normal modes of the room. The highest value of the average
sound absorption coefficient, at any frequency, should not exceed O.l6."
The international standard[lO] includes very similar requirements.
Both the American and international standards[10, 11, 29] include quite
detailed room qualification procedures — separately for the measurement of
broad band sound and for the measurement of discrete frequency components.
These procedures involve not only the reverberation chamber but source loca-
tions, microphone locations, diffusing elements, and, in effect, the entire
measurement procedure. Anyone planning to conduct reverberation room
measurements should study these standards carefully.
67
-------
In the present section, the relations among room absorption, source
location, and pover output have not been discussed. These will be covered i*1
Section 15-3.
d. In Situ
It is not possible even to attempt to cover all of the different types
of test environments that might be encountered in carrying out In situ tests.
In general, measurements should "be made to determine the extent to which the
test environment approaches an anechoic, semi-anechoic or reverberant
environment. The types of measurements which should be made should be fairly
obvious from the preceding discussions.
Whenever the sound source can be removed from the test environment, the
environmental influence should be checked by placing a reference sound source
at selected points which define the boundaries where the actual source will
be and then measuring the sound pressure at various distances from each
reference source location. Procedures for correcting sound power
measurements for the influence of the test environment are discussed in
Section 16.2.
12.3. Criteria for Background Noise
In some locations, particularly for in situ measurements, it may not be
possible to maintain background noise sufficiently below the noise emitted
from the source for the background noise to be neglected. In that case, the
measured values at each microphone location may be adjusted using the
corrections given in Table 3.
These corrections are predicated upon the background noise not being
coherent with the noise emitted from the source. As an example of a case
where this might not be true, the noise from a ventilating fan which emits
discrete frequency components could have a definite phase relation to the
noise from the test source when both are connected to the 60 Hz line power.
Then the above corrections for background noise would be inappropriate.
In some circumstances, it may be difficult to separate the noise of a
particular piece of equipment from the noise of its supporting equipment.
For example, in measuring the noise of a pneumatic tool, care has to be taken
to ensure that the noise from the air compressor is not influencing the
results. Background noise measurements made with the tool inactive and the
compressor operating might not be appropriate since the noise from the
compressor may increase when it is called upon to supply air for the tool.
Another problem area is time-varying background noise. Consider an
outdoor test site fairly near a highway. Background noise measurements have
little meaning unless one can be confident that the average background noise
was the same during the time interval when the noise emitted from the source
was measured.
68
-------
Table 3 Corrections for Background Sound Pressure Levels
Difference (in decibels)
between sound pressure
level measured with sound
source operating and
background sound pressure
level alone
6
1
8
9
10
11
12
13
1U
15
Correction (in decibels)
to be subtracted from sound
pressure level measured with.
sound source operating to
obtain sound pressure level
due to sound source alone
1.3
1.0
0.8
0.6
O.U
0.3
0.3
0.2
0.2
0.1
69
-------
12.4. Criteria for Temperature, Barometric Pressure, Humidity, and Wind
a. Temperature
A measurement standard should include restrictions on the allowable
temperature range over which measurements may "be conducted and, if practical*
provide corrections for the effects of temperature. Temperature can affect
the measurement results in the following -ways:
- the performance of the test source may "be a function of temperature with
the result that the noise emitted varies with temperature even though
the operating load appears to "be constant;
- the properties of the air surrounding the test source may vary suf-
ficiently with temperature to affect the sound pressure at the
measurement locations;
- temperature gradients or inhomogeneities in the atmosphere cause
refraction or scattering of the sound;
- the accuracy of the measuring instrumentation may be affected by
temperature.
The acoustic impedance of air is affected by both temperature and by
barometric pressure and will be discussed in the section on barometric
pressure. Absorption of sound by the air is influenced by temperature,
barometric pressure, and humidity; it will be discussed in the section on
humidity.
In making measurements outdoors over a reflecting plane, errors can
occur due to refraction of sound arising from the variation of the speed of
sound due to a steep temperature gradient above the reflecting plane. Such^
an effect would be expected to be particularly serious for sources which emit
pure tones and to be worse at higher frequencies where the wave length of "the .
sound is much shorter than the height of the source and/or the measuring
microphone above the reflecting plane. To a certain extent this could be
compensated for by multiple microphone locations or by traversing the
microphone over an appropriate path (see Sec. 15.2).
It is difficult to generalize on the effects of temperature on measuring
instrumentation. The American standard for sound level meters[37] includes
the following statement:
"•7
7.1 Temperature. The temperature range over which the sensitivity of
the sound level meter varies less than 0.5 decibel at any frequency
shall be stated by the manufacturer. If this range does not include the
extremes of -10° to 50° C, the manufacturer shall supply temperature
correction values over that range. If provision for internal
70
-------
calibration is made in the sound level meter, the manufacturer shall
state the effect, if any, of temperature upon the calibration system,
and thence upon the self-calibrated sound level meter over the
temperature range of -10° to 50° C. The manufacturer shall state the
temperature limits beyond which permanent damage to the sound level
meter may occur,"
If it is intended to make measurements over a broad temperature range,
the instrumentation system should be calibrated over that range — care
should be taken that the calibration procedure itself does not have
unaccounted-for temperature effects.
b. Barometric Pressure
In Section 12.2, it was indicated (see eq. (12)) that in a reverberation
chamber, the mean-square sound pressure is proportional to pc W. A similar
analysis of free-field radiation would show that the mean-square sound
pressure is proportional to pcW. Thus the computation of sound power
requires inclusion of a term of the form 10 log (pc ) for reverberation room
sound power level* and a term of the form 10 log (pc) for free field sound
power level. These terms do not correct the data to correspond to standard
conditions of temperature and barometric pressure, but are simply part of the
procedure to calculate sound power level from sound pressure level data (or
vice versa).
The effect of a variation in temperature and pressure on reverberation
room measurements is more readily seen by rewriting eq. (15) as
- 10 log
2V
- 10 log
400
- 10 log
(32)
thus if the calculation were made using a standard value, p c , the
o o
2 2
correction term to be subtracted is 10 log (pc /PQC )•
For free field measurements over an imaginary sphere at radius r from a
point source of power W, the mean-square sound pressure is
2 pcW
-
(33)
from which the equation for power level is seen to be
p c PC
Lp + 10 log[47rr*] - 10 log
- 10 log
•]
J
(34)
*If reverberation room calculations are made using total room absorption, A,
rather than reverberation time, T, to account for the power loss at the
boundaries and in the air, the term T/V is replaced by cAA, the c cancels
one of the c's in pc and the correction term becomes 10 log (pc) — the
same as for free-field radiation.
10
-------
The dependence of 10 log (pc/p CQ) on temperature and barometric
pressure is shown in Figure lU. The curve corresponding to 0°C also
represents the dependence of 10 log (pc /p c ) on pressure since pc is
independent of temperature. These dependencies stem from the following
equations, "based on the ideal gas law:
where p c is the value of the acoustic resistance at barometric pressure BQ
and absolute temperature T , and
A
As stated previously, the above discussion only concerns the influence °
temperature and barometric pressure on the calculation of sound power level
from the appropriate sound pressure level. It does not correct the results o*
a calculation to correspond to the noise emission of a source under some
standard conditions other than those which existed at the time of measurement.
Temperature and barometric pressure directly influence both the radiated soun
power and the sound pressure at a particular location, but in a manner that i
different for different types of sources. Sources with very high internal
acoustic impedance (constant velocity sources) will be affected differently
from sources having very low internal impedance (constant near-field
pressure).* In addition, sources having different directionality
characteristics are affected differently by temperature and barometric
pressure. Expressions for monopole, dipole, and quadrupole point sources,
radiating into a free field (air) have been derived from equations in Morse
and Ingard[32, pp 306-318] and are listed in Table U for sound sources having
either constant velocity or constant near-field pressure. It can be seen tha
the dependence on temperature and the dependence on barometric pressure can "be
quite different for different types of sources. Thus corrections to standard
conditions cannot be made reliably without some knowledge of the
characteristics of the noise source.
A related phenomenon that also can affect the accuracy of acoustical
measurements is the dependence of sound pressure level produced by piston-
phones, and other types of microphone calibrators, on barometric pressure.
*See Section 15.1 for discussion of particle velocity, volume velocity,
near field, and far field.
72
-------
20
I I
m
o
3.
CD
3
PRESSURE, in. Hg
25
30
PRESSURE, mm Hg
500 600 700
800
0.6 0.7 0.8 0.9 J.O
PRESSURE, otm
Figure lU. Pressure dependence of the acoustic resistance of air,
73
-------
Table h. Dependence of far-field sound pressure and radiated sound power on barometric pressure
and temperature for various types of sound sources. The numbers in parentheses repre-
sent the rates of change for the source in air at T^O mm Hg (1.013 x 10 Pa) and 20°C
(293.15 K).
Type of Source
Monopole
Dipole
Quadrupole
Constant Volume Velocity
mean- square pressure
ftfftf
(.011 dB/mm Hg)
(-.030 dB/°C)
W'ftf
(.011 dB/mm Hg)
(-.OUU dB/°C)
/B \2/T \~h
W (To)
(.011 dB/mm Hg)
(-.059 dB/°C)
sound power
B /T\-3/2
Bc IV
(.006 dB/mm Hg)
(-.022 dB/°C)
B /T \~5/2
B ( T 1
o\ o/
(.006 dB/mm Hg)
(-.037 dB/°C)
B /T \~T/2
B(T)
°\ °/
(.006 dB/mm Hg)
(-.052 dB/°C)
Constant Near-Field Pressure
mean-square pressure
1
(0 dB/mm Hg)
(0 dB/°C)
(iT
(0 dB/mm Hg)
(-.015 aB/°C)
/T\-2
VToJ
(0 dB/mm Hg)
(-.030 dB/°C)
sound power
/if/Tf2
(BOJ (TO;
(-.006 dB/mm Hg)
(.007 dB/°C)
ftTftT
(-.006 dB/mm Hg)
(-.007 dB/°C)
(I)" (if"
(-.006 dB/mm Hg)
(-.022 dB/°C)
-------
c. Humidity
The two major effects of humidity on noise measurements are (l) the
effect on absorption on sound "by the air and (2) effects on measurement
systems.
Sound energy is attenuated "by molecular absorption, "by viscosity, and "by
heat conduction. For a plane wave, the mean-square sound pressure will
attenuate with distance as
p2(x) = p2(0)exp(-mx), (37)
p
where p (0) is the mean-square pressure at x = 0. Converting to sound
pressure level,
L = L(0) - MX , (38)
where M = I*.3lv3m is the attenuation in decibels/meter. The attenuation coef-
ficient varies in a complicated manner with temperature, humidity, and baro-
metric pressure. It is larger at high frequencies and, at normal
temperatures, peaks at rather low relative humidities. Representative curves
are given in Figure 15. For additional information see[38, 39] and the
references therein.
The effect of air absorption is quite important, especially at higher
frequencies, in reverberation room measurements since, with hard walls, the
reverberation time is severely limited by absorption of sound by the air.
This requires that care be taken to measure reverberation times -under the
same atmospheric conditions as when steady-state sound energy density meas-
urements are made. Current national[29J and international]10] standards
require that the temperature t (°C) and the relative humidity rh (percent) be
controlled such that the product rh(t+5) does not differ by more than +. 10
percent from the value of this product which prevailed during the
measurements of the reverberation time (for the direct method) or reference
sound source (for the comparison method).
Even when carefully controlled, high air absorption at high frequencies,
coupled with the fact that sound sources tend to be more directive in their
radiation pattern at high frequencies, can make it difficult to achieve a
diffuse reverberant field that is sufficiently larger than the direct field
from the source.
A major effect of high humidity on acoustical measurement systems is
condensation of moisture in microphones, resulting, in the case of condenser
microphones, in electrical leakage. Many condenser microphones are heated to
avoid this problem. High humidity can also lead to electrical leakage in
75
-------
m
•o
CO
CO
o
a.
oc
o
o
a:
UJ
x
a.
CO
o
10%
RELATIVE
HUMIDITY
0 10 20
TEMPERATURE, °C
10%
RELATIVE
HUMIDITY
0 10 20
TEMPERATURE, «C
Figure 15- Atmospheric absorption loss in air as a function of temperature
hUmidities of 10 and TO percent at frequencies of
-------
cables, connectors, or instruments.
d. Wind
A serious effect of wind on outdoor measurements is the extraneous noise
generated by the wind blowing across the microphones. In general, a wind
screen should always be used and a measurement standard should set limits,
analogous to those for background noise, on the allowable wind noise[35l»
Wind can also influence the propagation of noise from the source. A
shadow zone, into which no direct sound can penetrate, may be encountered
upwind from a source because the typical positive wind gradient (i.e., wind
speed increasing with height above the ground) bends the sound rays
upward[33,35]. If measurements must be made a considerable distance from the
source, tighter limits (than for close-in measurements) should be placed on
allowable wind speeds.
12.5. Criteria for Size of the Test Equipment
A given test facility should be sufficiently larger than the equipment
being tested to enable measurements to be made outside of the near field of
the noise source. The International Organization for Standardization draft
on laboratory measurements in anechoic and hemi-anechoic chambers[lU]
recommends that the volume of the source be less than 0.5 percent of the
volume of the test room (e.g., aim source may be tested in a 200 m
chamber). For large equipment, such a restriction would be too severe for
practicality. However, the measurement standard should adequately address
this point.
12.6. Criteria for Reflecting Surfaces
A discussion was given in Section 12.2, in conjunction with the material
on measurements in a semi-anechoic space, of the importance of the reflecting
plane having a very low sound absorption. This can be particularly difficult
in outdoor measurements on large, or mobile, sources where it may be
difficult to find large paved areas with sufficiently low sound absorption.
In general, measurements over grass or soil are not recommended — the
surface is neither hard enough to approximate a perfect reflector nor soft
enough to approximate a perfect absorber.
Even when the source of the noise is very close to the ground, so that
there are essentially no reflections, difficulties may arise. A sound wave
traveling parallel to an absorptive surface can lose energy into that surface
if the "sound rays" pass close, in terms of wavelength, to the surface.
In addition to criteria for the reflecting plane, measurement standards
should include restrictions on the presence of nearby objects which could
reflect sound energy in such a way as to influence the test results.
77
-------
13. Instrumentation for Noise Measurements
• The measurement standard should require that each instrument meet
specified requirements of existing national or international standards.
For instruments for which standards do not exist or where eosisting
standards are not sufficient, the measurement standard should include
specific criteria for evaluating the performance of such devices* e'3"
for sources which produce transient signals the standard might include
allowable tolerances for system response to one or more well-defined
transient events.
• The measurement standard should clearly state the allowable tolerances
for frequency response, environmental effects* harmonic distortion, e^
which the instruments are required to meet. These specifications s
be applied not only to specific components of the system but to the
overall system as well.
• The measurement standard should require overall system calibration at
stipulated intervals. The fact that each component of the system
satisfactory does not ensure that the system performance will be
acceptable.
• The measurement standard should require that the overall system
measurement error not be degraded below that allowed for direct
measurementsj regardless of the instrumentation configuration.
13.1. General Requirements
Instrumentation for noise measurements consists, generally, of the
components shown in Figure 16. The sound pressure is converted into an
electrical signal by a microphone. This signal is amplified and passed
through a filter which weights the various frequency components of the
The filtered a-c signal is then detected (usually converted to a d-c value
equivalent to the root-mean-square value of the a-c signal) and averaged °ve
an appropriate time interval. This detected signal is then displayed via d
read-out device. The signal may be recorded on a magnetic tape recorder
brought back to the laboratory for analysis. If this is necessary, the
recording and playback operation should not degrade the overall measurement
error below that allowed for direct measurements.
MICRO-
PHONE
AMPLI-
FIER
FILTER
DETECTOR
READ-OUT
Figure l6. Schematic representation of instrumentation for sound level
measurements.
78
-------
In acoustical measurements where a frequency analysis is required, it
has been traditional to use a bank of band pass filters, and switch through
them sequentially to obtain sound pressure levels in each frequency band.
More recently, real-time analyzers have become available which have parallel
filters, with detection and read-out for each frequency band. The output of
these can be read visually or they can be interfaced to a computer.
Alternatively, the instantaneous voltage from the microphone can be passed
through an analog-to-digital converter which is interfaced to a computer.
Digital data are taken at a rate at least twice the highest frequency of
interest. Filtering and detection are done digitally within the computer.
Any of these techniques is acceptable if done properly. In the dis-
cussion below it will generally be assumed that the various elements in Figure
16 are analog devices which can be treated separately. However, the
measurement standard should apply to the overall measurement process and not
assume any particular configuration of components.
13.2. Microphone and Cable
There are a number of different types of microphones in use. For
acoustical measurements, condenser, electret, and piezoelectric microphones
are most common although dynamic microphones are still used occasionally.
The important features of microphones are frequency response, sensitiv-
ity, and freedom from adverse environmental effects. The American
standard[29] for reverberation room determinations of sound power requires
that:
"The microphone shall have a flat frequency response for randomly
incident sound over the frequency range of interest. The microphone
shall meet the requirements of ANSI SI.12-196?. The microphone and its
associated cable shall be chosen so that their sensitivity does not
change by more than 0.5 dB in the temperature range encountered in the
measurement. If a moving microphone is used, care shall be exercised to
avoid introducing acoustical or electrical noise (e.g., from gears,
flexing cables, or sliding contacts) that could interfere with the
measurements."
An American standardise] describes types of laboratory microphones that are
suitable for calibration by an absolute method such as the reciprocity
technique described in the American standard for the calibration of
microphones[kl].
13.3, Frequency Response of the Instrumentation System
The American[37l and international[U2,U3] standards for sound level
meters prescribe tolerances on frequency response, omni-directional response,
and the effects of environmental conditions.
The American standard states that:
79.
-------
"the frequency response of the instrumentation calibrated for randomly
incident sound shall "be determined according to the procedures of ANSI
Sl.10-1966 and the random incidence response shall be uniform within the
tolerances given "below:
Frequency Tolerance Limits
Ha <3JB
50 to 80 +1.5
100 to iiOOO + i"
5000 to 8000 +1.5
10000 72* "
13.4. Weighting Network and/or Frequency Analyzer
If direct measurements are to be made of the A-, B-» or C-weighted sound
level, the measurement standard should clearly reference the allowable
tolerances on the frequency weighting. The international standard for
precision sound level meters[^3] gives tolerances for all three weighting
networks. The American standard[3T] defines three types, or classes, of sound
level meter with different tolerances. The American Type 1 tolerances 1371 are
essentially identical to the international tolerances[U3] except below 10° Kz
where the American tolerances are tighter.
It is anticipated that frequency analyses required for regulatory actions
will not require measurements in frequency bands narrower than 1/1- or
1/3-octaves. The international standard for band-pass filters[WO defines the
center frequencies and sets limits on terminating impedances, effective
bandwidth, attenuation in the pass-band, attenuation outside the pass-band,
overall tolerances, harmonic distortion, and the effects due to environmental
conditions. The American standard[^5] is a rather more detailed document
which establishes three classes of band filters, Classes I and II for octave
band filters and Classes II and III for half-octave and third-octave band
filters.
The choice of a filter for a given measurement is based upon the accuracy
required. The bandwidth error of a filter depends upon its transmission l°sS
at the band edges, the slope of the transmission loss characteristic outside
the band, and the input noise spectrum slope. Appendix B of [U5l discusses
the subject and gives data and references allowing selection of filter
characteristics which will yield measurements falling within specified error
limits at various noise spectrum slopes.
The international tolerances[UU] are less restrictive than the American
Class III but somewhat more restrictive than Class II. The American standard
for determination of sound power in reverberation rooms[29] requires that an
octave band or one-third octave band filter set meeting at least the
requirements for Class II filters of AUSI SI.11-1966, or latest revision,
80
-------
shall be used." International standards[10,11] and draft standards[l2-lH] for
sound pover determination specify a band filter set meeting the requirements
of [U].
13.5 Signal Detection and Averaging
The electric signal from a microphone is typically comprised of the sum
of components at many frequencies. The measuring instrumentation should
"detect" the root-mean-square value of this signal, defined as
1/2
Pi fL. . . ,r> . I
E
rms
i /[eft)]2 dtl (39)
where e(t) is the time varying voltage and, T, the integration time, should be
long enough to provide adequate averaging.
Many voltmeters measure the average absolute value of the signal, defined
as
m
(*0)
The value of E is uniquely related to E for a sinusoidal signal so that
avg -"^
rectified average detectors are quite satisfactory for that purpose. However,
they can result in very large errors for more complex signals. Table 5
indicates the error of a rectified average meter for several types of signals.
Since typical noise signals differ considerably from pure sine waves, it is
important to specify a true rms detector. The International standard on
precision sound level meters [43] requires that instrumentation complying with
that standard be able to measure the combination of signals of two
non-harmonic frequencies 'to within ±0.1 dB of the true rms value.
After the signal is "detected", it is necessary to carry out the
integration indicated in eq. 39- Preferably, this should be done using a
"true integrator", either analog or digital. Many acoustical instruments
utilize RC-averaging in which the squared signal is the input to a low-pass
RC-filter . If the rms value of the input voltage to an RC-integrator is
changed from one value to another (e.g., the system is switched to a different
microphone or filter), the output of the integrator will vary with time as
the initial value E decays with a time constant RG and the new value ^ is
approached asymptotically with the same time constant. In terms of the
voltage change,
" (42)
E2-E1
81
-------
Table 5
Wave Form
Sine Wave
Sine Wave Plus 100$
Third Harmonic in Phase
Sine Wave Plus 100$
Third Harmonic out of Phase
Square Wave
Gaussian Noise
Pulse Train: 10$
on and 90$ off
Pulse Train: 1$
on and 99$ off
Error in reading of rectified
average meter
dB
0
-0.50
-6.52
+0.91
-l.OU
-19.17
Table 6
t/RC
1
2
3
*
5
6
7
i-.-*/*
0.632
0.865
0.950
0.982
0.993
0.998
0.999
20 log10(l-e"t/RC)
-3.98 dB
-1.26
-0.1*1*
-0.16
-0.06
-0.02
-0.01
82
-------
Table 6 shows this function for different times after the integrator input is
switched from E to E .
Thus if there originally was no input to the integrator, one should wait 5
time constants for the output of the integrator to be within less than 0.1 dB
of its final value.* Similarly, if one wishes to take successive independent
readings of the same signal, an interval of several time constants should be
allowed.
For nondeterministic, or random, signals (see Section ll) the averaging
time must be sufficiently long for the statistical error to become small
enough to be acceptable. For bandwidth limited Gaussian white noise, the
standard error (normalized) in the root-mean-square pressure (or voltage) is
given approximately by
where T is the averaging time and B is the effective bandwidth of the filter.
The 90, 95, and 99 percent confidence limits for bands of random noise are
shown in Figure 17 as functions of bandwidth and averaging time. For
additional details see
In the case of a deterministic process, the averaging time depends only
on the response of the filter and detector. That is, one must wait long
enough for the filter to respond to the signal and for the detector to respond
to the signal. If one defines the rise time of a filter as the time it takes
for response to rise from 10$ to 90$ of its final signal, then the rise time,
T is T = 0.88/B where B is the effective bandwidth of the filter. If an
RC-integrator is used in tie detector and a 0.1 dB error is accepted, compared
to the response to a unit step change in voltage of a true integrator, one
wait 5 RC time constants . The total observation time should be
-^+ 5RC
Be
*•»
In most cases of concern, the rise time will be much less than 5 RC time
constants.
13.6. Read-Out Device
If the averaging has been accomplished by a true or RC integrator, the
readout can be in analog form via a meter or a strip chart recorder, or
digitally through a digital voltmeter or a computer. If the averaging was
done digitally, the read-out would usually be via a computer.
*The squared voltage from the detector will contain a-c components
superimposed on the mean-square value. Each frequency component will be
attenuated in squared-voltage amplitude by (l/uRC) . For small values of RC,
this can result in noticeable "ripple voltage" passing through the integrator
if there is considerable low frequency voltage present.
-------
EQUIVALENT DEGREES OF FREEDOM
5 10 20 50 100 200 500 1000 2000
90% 95% 99%
—i—'I'M 1 1—•—•———,f^f\r\
2 5 10 20 50 100 200 500 l°°u
(EFFECTIVE BAND WI DTH)x( AVER AGING TIME)
Figure 17. Confidence intervals as functions- of BT for measurements on ra-n
noise.
-------
Graphic level recorders are frequently used to provide a time history of
sound level. When used for such purposes, the paper speed and, especially,
the writing speed should be selected carefully and should be reported with the
results of the measurement . When graphic level recorders are used in
conjunction with & filter to provide read-out of frequency analyses, the
"writing speed should be carefully selected with regard to the frequency sweep
speed [Ii8, 1
It is quite common to use a slow writing speed on a graphic level
recorder in an attempt to obtain a long averaging time such as might be
indicated in reverberation room measurements. Typically, there is an RC
integrator, following the square-law detector^ which has a low effective
averaging time T , The writing speed of the recorder in effect acts as
another integrator with an averaging time Tg. Thus the recorded level is
A = [e^J2 dt dt • (43)
1 1/2
dt
J
*f T > T , the system is unstable. However, if T^ is chosen to be much
greater than T and the frequencies of the fluctuations in the detector input
level are of tie order of 1/T , the recorded signal level will be the "mean
detected value" (i.e., "mean rms"). This quantity is of dubious physical
meaning and from a theoretical point of view it would be better to increase
eliminate the second integration!1^].
13.7. Transient Response of Instrumentation System
Although noise measurement instrumentation with different principles of
operation may be calibrated to yield the same results on steady-state signals ,
such may not be the case for transient effects such as motor vehicle pass-bys.
For one example of the differences among various measurements of time-varying
&oises see [50]. It is recommended that measurement standards for sources
which produce transient sounds include specific criteria for system response
to one or more well-defined transient events (e.g., a pure tone that is
amplitude-modulated in a specified manner).
13.8. Calibration and Maintenance of Instrumentation System
An acoustical instrumentation system can be quite complex. Accordingly,
it ig necessary to calibrate the overall system frequently and not to rely on
the system performance being acceptable simply because each component of the
system appears satisfactory.
The American standard for the measurement of sound pressure levels [8]
includes the following statement with regard to calibration and maintenance of
instrumentation :
"The instruments used for the acoustical measurements shall be serviced
at least once every twelve months in accordance with the manufacturer's
-------
instructions. This shall include checking the performance of all
mechanical components and electrical circuits and replacing substandard
items. The date of most recent servicing shall "be written on tags
attached to the instruments. To ensure that the calibration of the
equipment has not changed during a series of measurements, the
instrumentation system shall "be calibrated acoustically according to the
manufacturer's instructions. A comparative calibration provided "by a
sound-level calibrator or pistonphone of known sound pressure level is
usually satisfactory for this purpose. The frequency response of the
complete instrumentation system shall be checked periodically to insure
that the requirements of 5.^-2 are satisfied. For the laboratory method,
microphones shall be calibrated by comparison with reference standard
microphones which are calibrated according to American National Standard
Method for the Calibration of Microphones, SI.10-1966 (see Section 12).
13.9. Precautions to be Taken When Selecting Instrumentation
The following is also taken from the American standard for measurement of
sound pressure levels[8]:
"5.T.1 Precautions (Field and Laboratory Methods)
5.7.1.1 Wind (Field Method Only). To perform sound pressure level
measurements in a moving air stream, a suitably designed windscreen or
nose cone shall be utilized to minimize the influence of the air stream
on the output of the microphone. No such precaution is necessary if the
wind noise is 10 or more decibels below the signal being measured in. each
frequency band of interest. Corrections for changes in microphone
sensitivity for the windscreen or nose cone used during the measurements
shall be applied to the observed sound pressure levels.
5.7.1.2 Humidity and Temperature. High humidity or temperature
will change the sensitivity or damage many types of microphones. The
microphone manufacturer's instructions shall be carefully followed to
avoid such effects.
5.7.1.3 High Sound Pressure Levels. Many piezoelectric,
moving-coil, and capacitor microphones may be used for the measurement of
sound pressure levels up to approximately lUO dB re 20 uN/m . At higher
levels, specially designed microphones with stiff diaphragms shall be
used: these shall be calibrated at the levels to be measured and, if
possible, over the entire frequency range of interest. At high sound ^
levels, special precautions shall be taken to ensure that "microphonics
are not generated by the transmission of mechanical vibration to the
microphone or instrumentation. These include:
(l) Installing the microphone and instrumentation on a soft
mounting.
(2) Removing the instrumentation from the high sound levels and
utilizing long cables: precautions are necessary to minimize cable
86
-------
noise, that is, the noise produced when the cable itself is subject to
vibration or flexing.
(3) Installing the instrumentation behind suitable barriers or
enclosures: a mechanically soft mounting shall be used for the
lov-sensitivity microphones that are utilized for the measurements of
high sound levels.
Determining electrical noise and possible "microphonics"
problems by replacing the microphone with a highly insensitive (dummy)
microphone .
5.7.1.1^ Lov Sound Pressure Levels. A microphone used to measure
low sound pressure levels must have high sensitivity and low internal
noise. When connected to suitable low-noise amplifiers, many
piezoelectric, moving-coil, and capacitor microphones are suitable for
measurements of sound pressure levels below 20 dB re 20 yN/m .
5.7.1.5 Low-Frequency Noise. Piezoelectric and some capacitor
microphones are suitable for measuring sound pressures at frequencies
down to fractions of a hertz. Special amplifiers are required for
measurements of low- frequency noise. The low frequency sensitivity of
a microphone may vary considerably from the mid-frequency sensitivity
due to the presence of a pressure-equalizing leak. Calibration shall
be performed over the frequency range of interest.
5.7.1.6 High-Frequency Noise. For measurements above 20000 Hz.
miniature capacitor or piezoelectric microphones usually give the most
satisfactory results.
5.7.1.7 Hum Pickup. When sound pressure levels are to be measured
near electrical equipment, a moving-coil microphone shall not be used.
The instrumentation shall be checked to make certain there is no hum
pickup in the instruments themselves. Hum can be reduced by moving the
instruments away from the source of the magnetic field or by selecting a
proper orientation of the instruments with respect to the magnetic field.
5.7.1.8 Cables. When a cable is used between the microphone and
the acoustical instrumentation, the system shall be calibrated according
to the manufacturer's instructions with the cable in use.
5.7.2 Precautions (Survey Method). Sound-level meters with integral
microphones are generally not suitable for a measurement program that
requires the observance of the special precautions' of 5-7.1.
5-7.3 Additional Effects on Measured Data
5.7.3.1 Effect of Observer and Meter Case on Measured Data
5.7.3.1.1 Survey Method. The sound-level meter shall be held in
front of'the'observer. The observer shall be oriented with' respect to
87
-------
the principal sound source so that the sound energy arrives at the
microphone from the side unless some other orientation is specified by
the instrument manufacturer.
5.7.3.2 Field and Laboratory Methods. In order to minimize the
obstacle effect caused by the insertion into the sound field of the
sound-level meter and the experimenter holding it, the microphone shall
be connected to the sound analysis equipment by means of an appropriate
cable or extension connector and mounted on a tripod or other suspension
system. The observer and all acoustical instrumentation except
microphone(s), associated preamplifiers and cables should be located
outside the test area.
5.T.3.2 Microphone Response and Orientation
5.7.3.2.1 General. The microphone calibration applied to compute
sound pressure level shall conform to the way the microphone is used in
the measurement; for example, free-field calibration at the appropriate
angle of incidence. It should be recognized that microphone calibrations
are often furnished in terms of the pressure response, vhich may differ
from the free-field response at high frequencies by as much as 9.5 dB for
one-inch diameter microphones.
5-7.3.2.2 Survey Method. See 5.7.3.2.1
5.7.3.2.3 Field and Laboratory Methods. The microphone shall be
oriented with respect to the source so that sound strikes the diaphragm
at the angle for which the microphone was calibrated to have the flattest
frequency response characteristic. The variation of the response with
frequency shall be taken into account in each frequency band for maximum
accuracy. It should be noted that microphones are usually most sensitive.
for sound propagating perpendicular to the microphone diaphragm.
However, the angle required to obtain the flattest response vs frequency
will be a function of the microphone design. It is imperative that
reliable calibration data be used to determine the angle of operation for
the flattest response. It should be noted that a microphone may be
extremely sensitive at -high frequencies to small changes in orientation
for sound waves arriving parallel to the diaphragm. Therefore, during a
measurement of sound which contains significant high-frequency
components, it is advisable to maintain the microphone orientation to
within +_ 5 degrees for the survey and field methods and to within ± 2
degrees for the laboratory method."
88
-------
14. Installation and Operation of Source
• The measurement standard should specify that the device under test be
located in its use configuration or alternatively the location should be
governed by the test environment and the quantity to be measured3 e.g.,
source located near the center of the room for anechoia measurements of
sound power.
• The measurement standard should specify that the device under test be
mounted under conditions similar to those recommended for normal
installation. Care should be taken to ensure that (1) adequate
isolation is provided to minimize extraneous airborne noise due to
vibration excitation and (2) the process noise does not exceed the sound
of the device itself.
• The measurement standard should require the input mass and energy -and
the output energy to be brought to and removed from the source under
test without influencing the quantity being measured.
• The measurement standard should specify the number of operational modes
under which tests are to be carried out.
• The measurement standard should specify the extent of the loading and
the manner of application of the load to the source under test so these
are similar to actual use conditions wherever possible.
The noise level produced by a specific device is not only dependent upon
the sound radiating characteristics of the machine itself but also on the way
the machine is operated and/or installed and the specific environment in
which it is used. In setting noise limits for such devices through noise
emission or labeling standards, test procedures and measurement methodology
should include such items as loading, operating speeds, installation
requirements, and the location and specification of needed auxiliary
equipment.
14.1. Source Location
Sound pressure level measurements for a given device are obtained by
measuring at a specified distance from the source in essentially free-field
conditions. For sound power determinations, in an acoustically controlled
environment, the source is usually located near the center of the room for
anechoic measurements, near the center of the floor for herai-anechoic
measurements, while for reverberant measurements the source could be located
at various locations. For devices normally mounted on or against a wall,
they should be tested in their "use configuration."
As an example of a typical source location specification, consider
American Society for Heating, Refrigerating, and Air-Conditioning Engineers
89
-------
Standard Methods for Testing for Sound Rating Heating, Refrigerating, and
Air-Conditioning Equipment[51]:
"5-1 EQUIPMENT LOCATION
The equipment to be tested shall be placed in the sound test room in a
position representative of normal usage (see Fig. 1.)
5.1.1 Equipment Used Against a ."Wall: Equipment normally used
against a wall shall be positioned against a wall, at least 5 ft from a
corner of the room, and not on a center line of the wall.
5.1.2 Equipment Used Away From a Wall: Equipment normally mounted
on the floor or ceiling away from a wall shall be located no closer than
5 ft to any wall, and away from any position of room symmetry.
5-1.3 Equipment Mounted Through Window, Wall or Ceiling: Equip-
ment normally mounted through a window, wall or ceiling shall be mounted
through the wall or ceiling of the test room and shall be located at
least 5 ft from any corner and away from any position of room symmetry*
except that equipment normally mounted near a corner shall be located at
the normal distance from such corner."
A - Equipment location, Par. 5.1.2
B = Equipment location, Par. 5.1.1
C = Equipment location, Par. 5.1.3
D = Diameter of circular microphone traverse
E «Length of microphone traverse (arc or
linear)
F = Location of reference sound source, Par. 5.2
X = 5 ft minimum
Fig. 1 Location of Equipment in the Test Room
90
-------
Except for devices normally mounted on or against walls, the most criti-
cal requirement is that sources not "be located nearer to a wall, edge, or x
corner than a distance of approximately A/H, where X is the wavelength of the
lovest frequency of interest. When sources are located near reflecting
surfaces the sound power output is changed.
14.2. Source Mounting and Installation
The acoustic radiation of a device can depend on its installation. In
general, the device to he tested should "be mounted under conditions similar
to those recommended by the manufacturer for normal installation. If a given
machine is to be securely "bolted to a heavy concrete floor in use, it should
be tested that way. Many times it is impractical to simulate mounting con-
ditions such as exist in the actual machine installation. In this case, the
usual alternative is to use a very resilient mounting. Since devices can
produce forces which may excite vibration in the base, floor, or surrounding
structure, and since these may in turn produce airborne noise, precautions
should be taken to ensure that adequate vibration isolation is supplied so
that extraneous airborne noises of this type are minimized.
The mounting specifications should be well-defined — either the final
use mounting configuration or a resilient mounting should be utilized. Tests
should not be conducted using other mountings. As an example of the level of
detail needed to adequately specify the mounting/installation of the device
during test, American Society for Heating, Refrigerating, and
Air-Conditioning Engineers Standard Methods of Testing for Sound Rating
Heating, Refrigerating, and Air-Conditioning Equipment[5l] is-again cited.
"5.^.3 To minimize wall vibration effects, the mounting wall shall be
of heavy masonry or equivalent, or an auxiliary mounting platform
similar to that shown in Fig. 2 shall be provided.
Wall
Seal
(Rubber or Non-
hardening Caulk-
ing Compound)
Test Room Side.
Test Unit
Vibration
Isolation
Pad
Concrete
Block
wall
Test Unit
Vibration
Isolation Pad
Supports
Fig. 2 Auxiliary mountings
91
-------
5.^.H The equipment shall be mounted according to the manufacturer's
instructions. Openings "between the equipment casing and vail shall be
sealed with a gasketed sound isolation plug similar to that shown in
Fig. 3."
Test Unit
Gasket (Foamed
Rubber or Equiv.)
wail < ;
X 1 /
r ;
r: s Y
_1 J A
Plug Filled
with Glass
Fiber -~.
Intulation or
Equivalent
i 1
1
I*1
1
1
^
\
t
1
., — • Gasket ~ — -^.
(Foamed Rubber) *
or Equivalent)
11
1
1
1
1
1
1
I
1
L
Fig. 3 Typical sound isolation plug
Mounting problems during test are not limited to stationary equipment.
For instance, the Compressed Air and Gas Institute - Test Code for the
Measurement of Sound from Pneumatic Equipment[52] specifies that in cases
where the process noise far exceeds the sound output of the machine itself,
e.g., riveters, and it is necessary to consider the machine on its own, "it
should be run with the working tool embedded in a shock absorbing body whose
secondary sound level is at least ten (lO) dB below the machines own output
in each octave band of interest. (For example, the tool is to be embedded in
or running on rubber, sand, etc.).11
For in.-situ measurements, either indoors or out-of-doors, made on non-
stationary sources, e.g., motor vehicles, no special provisions are usually
necessary concerning mounting or installation.
14.3. Auxiliary Equipment
In general, machines are governed.by the conservation of energy prin-
ciple relating the balance of input/output energy. For example, the fuel
(gasoline) that is used in an automobile is converted into power (including
noise and vibration) and heat during the combustion process. It should be
obvious that this energy flow must be accounted for during noise measurement
tests, especially those'conducted in enclosed spaces such as anechoic or
reverberation rooms. If a car is to be tested in an anechoic room, the room
must be cooled to dissipate the heat produced by the combustion process, the
exhaust emissions must be removed from the room, and the power to the wheels
must be dissipated through a dynamometer.
92
-------
In the case of an automobile the output energy is the main problem. For
devices which operate on electricity, the input energy is also a problem. A
good power supply is necessary to supply any piece of electrical machinery
during test. Machine noise may be influenced by irregularities in the power
supply output. Noise in alternating current motors may be influenced by
voltage unbalance and/or harmonic content; noise in direct-current motors may
be affected by ripples in the power supply.
Auxiliary equipment may also be a problem out-of-doors. The diesel
train locomotive, for instance, incorporates a diesel engine which drives an
electrical generator which in turn provides power to traction motors on each
axle of the locomotive. The diesel engine is water cooled and utilizes
water-to-air heat exchange radiators and associated cooling fans. Dynamic
braking is used on many locomotives to slow the locomotive and train at
higher speeds or on steep grades. This is accomplished by disconnecting the
traction motors from the main generator and using them as generators. The
high electrical currents that result are dissipated as heat through heavy
duty resistor grids with the use of separate cooling fans located in the roof
of the locomotive. When such a locomotive is evaluated for noise emission
utilizing a stationary test, the generator load output must be dissipated
into a resistor grid load box. This load box facility must be isolated from
the locomotive under test so that the grid cooling fan noise will not affect
the measurements.
The basic requirement is that the input mass and energy — gasoline,
air, electricity, etc. — and output energy — power, heat, air emissions,
etc. — be brought to and removed from the device under test without influ-
encing the quantity being measured.
Such requirements should be specified in the measurement standards.
American Society for Heating, Refrigerating, and Air-Conditioning Engineers
Standard Methods for Testing for Sound Rating Heating, Refrigerating, and
Air-Conditioning Equipment[5l] stipulates the following requirements for
auxiliary equipment:
"5.5 AUXILIARY FACILITIES FOR SOUND TESTING AIR CONTROL & TERMINALS
(ACT) DEVICES
5.5.1 General
5.5.1.1 A quiet air system shall be provided and arranged to
absorb sound generated by the fan or duct system so that it does not
affect measurements of sound power generated by the ACT device.
Correction to sound measurements for background noise from the fan or
duct system shall not be permitted.
5.5,1.2 Background sound entering the test room through paths not
involving the ACT device are corrected per Table III, provided that the
background sound level in the test room is measured with the air duct
into the test room blanked and the exterior noise shall, not exceed that
93
-------
to prevail during measurement of the sound generation of the ACT device.
5.5-l«3 Air flow control accessories (damper, deflectors,
straighteners, equalizers, etc.) normally used in conjunction with the
ACT device under test, whether an integral part of the device or not,
shall be included in the test setup. They shall "be located and set in
the same manner recommended for the application of the ACT device.
5.5-l.U ACT devices shall be tested with an outlet duct of
recommended size terminating in the test room. In addition, ACT devices
which are recommended to be used in combination with lined ducts, ells
or silencers shall "be tested together with these with an outlet duct of
recommended size terminating in the test room.
5.5.1.5 When required, return air shall be vented from the test
room through a sound trap to prevent pressure build-up within the room.
All sound measurements of test equipment, reference sound measurements
of test equipment, reference sound source and background noise shall be
made with the return sound trap in place in a consistent manner, per
Par. 8.1.1."
14.4. Operation of Source During Test
The range of noise levels generated by a device is dependent on the
range of operational modes. A noise emission or labeling standard should
specify tests at a sufficient number of operational modes to fully character-
ize the device. For example, a truck is characterized by two operational
modes although the vehicle can operate over a wide range of speeds under the
various loads. Low speed operation is measured by a maximum acceleration
test (engine/exhaust noise) while high speed operation is measured by a
coastby test (tire noise). The Society of Automotive Engineers Recommended
Practice for the Exterior Sound Level for Heavy Trucks and Buses[53] defines
in great detail the operational procedure for measuring maximum noise:
"U. Procedure
U.I Vehicle Operation - Full throttle acceleration and closed
throttle deceleration tests are to be used. A beginning engine speed
and proper gear ratio must be determined for use during measurements.
U.I.I Select the highest rear axle and/or transmission gear
( highest gear is used in the usual sense: it is synonymous to the
lowest numerical ratio) and an initial vehicle speed such that at
wide-open throttle the vehicle will accelerate from the acceleration
point:
(a) Starting at no more than two-thirds (66%) of maximum rated or
of governed engine speed.
(b) Reaching maximum rated or governed engine speed within the end
zone.
-------
(c) Without exceeding 35 mph (56 km/h) before reaching the end point.
U.I.1.1 Should maximum rated or governed rpm not "be attained
before reaching the end zone, decrease the approach rpm in 100 rpm
increments until maximum rated or governed rpm is attained within the
end zone.
U.I.1.2 Should maximum rated or governed rpm not be attained until
beyond the end zone, select the next lower gear until maximum rated or
governed rpm is attained within the end zone.
U.I.1.3 Should the lowest gear still result in reaching maximum
rated or governed rpm beyond the permissible end zone, unload the
vehicle and/or increase the approach rpm in 100 rpm increments until the
maximum rated or governed rpm is reached within the end zone.
U.I.2 For the acceleration test, approach the acceleration point
using the engine speed and gear ratio selected in paragraph U.I.I and at
the acceleration point rapidly establish wide-open throttle. The
vehicle reference shall be as indicated in paragraph 3.7. Acceleration
shall continue until maximum rated or governed engine speed is reached.
U.I.3 Wheel slip which affects maximum sound level must be
avoided.
U.l.U For the deceleration test, approach the microphone point at
maximum rated or governed engine speed in the gear selected for the
acceleration test. At the microphone point, close the throttle and
allow the vehicle to decelerate to one-half of maximum rated or of
governed engine speed. The vehicle reference shall be as indicated in
paragraph 3.7. If the vehicle is equipped with an exhaust brake, this
deceleration test is to be repeated with the brake full on immediately
following closing of the throttle.
U.2 Measurements
U.2.1 The meter shall be set for "fast" response and the A-
weighted network.
U.2.2 The meter shall be observed during the period while the
vehicle is accelerating or decelerating. The applicable reading shall
be the highest sound level obtained for the run, ignoring unrelated
peaks due to extraneous ambient noises. Readings shall be taken on both
sides of the vehicle.
U.2.3 The sound level for each side of the vehicle shall be the
average of the two highest readings which are within 2 dB of each other.
Report the sound level for the side of the vehicle with the highest
readings."
95
-------
An operational procedure for measuring tire noise would include the vehicle
speed, load per tire, and the pavement surface on which the truck could run.
In this case the pavement surface not only has an affect on the propagation
of sound but also on the sound generation process.
Devices such as dishwashers and clothes washers operate according to a
prescribed cycle and the noise levels generated depend on the particular
operational characteristics of each individual portion of the total cycle. A
measurement during a single operational mode for such a device would be
meaningless. To fully characterize such devices, a measurement would have to
be made during the rinse operation, the water filling operation, the
spin-drying operation, etc. This is somewhat analogous to the air emission
tests conducted on automobiles in which emission measurements are made while
the vehicle runs through a prescribed series of operations known as the
driving cycle.
The Compressed Air and Gas Institute - European Committee of Manufact-
urers of Compressed Air Equipment Code for the Measurement of Sound from
Pneumatic Equipment[52] specifies for percussive machines the working
pressure, the material to be penetrated, the depth of penetration and the
feeding force. For other pneumatic tools tests are to be run at no-load
(running free), at a rated load and speed, at governed speed under load, at
idle, at maximum performance, etc.
To accurately characterize sources having noise levels dependent on the
manner in which they are operated, operational constraints, in conjunction
with precise measurement and calibration procedures, should be incorporated
into the standard measurement procedures.
14.5. Loading of Equipment During Test
A garbage disposal obviously will produce a different noise level when
grinding bones than when grinding regular food. Likewise, a dryer with a
load of regular wash tumbling sounds very different than if several pairs of
tennis shoes are drying. For such devices there is a need for a "standard
load so that comparisons can be made among the noise levels produced by such
equipment.
^Stationary tests on moving equipment require a different type of loading
specification. In such cases a dynamometer or brake may be utilized to apply
a specified load to the device under test, thus simulating the road load
characteristic of normal vehicle operation.
The static load carried by a vehicle also has an influence on the noise
generated by the vehicle. For instance the loaded vehicle weight influences
the noise generated by tires. Depending on the tread design, the influence
can be significant.
The operational procedure can be such that the loading is not important.
Tests run according to the operational procedure specified in SAE J366b[53]
96
-------
(see Sec. ik.k for details) are intended to yield the same noise level
whether a tractor is tested by itself or the tractor is pulling a load of
70,000 pounds. The operational procedure specified hopefully ensures that
the engine is loaded properly and thus the load pulled is not important.
The extent of the loading and the manner of application of the load to a
device under test are important considerations which should "be addressed.
Loading should "be similar to actual use conditions wherever possible and
operational procedures should include loading requirements along with those
for speed, gear selection, and other operational parameters.
97
-------
15. Measurement Procedures
• The measurement standard should specify the location of all microphone
positions.
• The position of the source with respect to the test environment should
be defined. If (for example* in reverberation room measurements)
multiple source positions may be indicated* criteria should be given for
ascertaining how many source positions are required in order to attain
the desired level of precision and accuracy.
• The number of observations3 and the averaging time for each, necessary
for each sound level measurement should be specified in the measurement
standard.
• Criteria should be given to enable -determining whether the use of
diffusers is indicated in reverberation room measurements.
• Procedures for determining background noise should be specified.
• Techniques and procedures for characterizing> or "qualifying" the test
environment should be clearly laid out in the measurement standard.
• Instrumentation and facility calibration requirements and procedures
should be given.
Previous sections of Part II have included discussions of the acoustic
quantities of concern in noise measurements, the types of acoustic
environments and how they relate to measurements of sound pressure and
sound power, instrumentation used for acoustic measurements> and the
influence of the way the source is installed and operated. *In the present
section, these factors are incorporated into a discussion of overall
procedures for conducting acoustic measurements.
15.1. General
It is essential to keep in mind that the sound power radiated "by a given
source, and the sound pressure at any given location relative to that source,
will depend upon the acoustical properties of the environment in which the
source is located, the transmission path(s) between the source and the
microphone, and the properties of the environment in which the microphone is
located — all in addition to the properties of the source itself.
For a more complete description of sound propagation in a medium, it is
useful to introduce the particle velocity, u, in addition to the sound
pressure, p, and the density, p. If the acoustic pressure and particle
velocity can be expressed as harmonic functions, then these quantities* are
related by:
*An underlined symbol denotes a complex quantity having, in general, both real
and imaginary parts. An arrow over a symbol denotes a vector quantity.
98
-------
where u = 2irf is the circular frequency, i = -1, and V designates the
gradient operator.
In general, p_ is of the form Pe1'e*""Ud , where P is the amplitude and <|>
is the phase angle at t = 0. In acoustics it is customary to suppress the
time dependence and simply write p_ = Pe . With this convention, the
time-average power per unit area, known as the intensity, is
I = {pu)t = |fie (£*u) = ^Re (pi*) (>5)
"rms cos «• (UT)
whereby designates a .time average, $ is the phase angle "between the
acoustic pressure and the particle velocity, and "rms" indicates the
root-mean-square values of pressure and velocity.
The total time-average power flow across any closed surface surrounding
the source is
(W)
where n is an outward directed unit vector perpendicular to the elemental
surface area, dS. The integration is over the entire closed surface. In
terms of the root-mean-square pressure and velocity,
r * *
W - J P u «n cos 6 dS. (Ijo)
^ ^rms rms ^^y'
D
Since W represents the total radiated power, it is independent of the size or
shape of the surface of integration, providing the medium is non-absorptive.
99
-------
Ideally, the total sound power emitted by a source would be determined by
direct application of eq. (^9). Although there have been "intensity meters"
constructed (e.g., see[55-6l] and there is current interest[62] in this method
of sound power determination, intensity meters are not commercially available
and there are certain difficulties in applying eq. (1*9) to real-world
situations[63].
It also is of interest to consider the energy density (or total energy
unit volume), wt in the medium through which the sound is propagating:
w = - -^ + y Pu2 ; (50)
2 pc2
where p2 = |p|2 and u2 E |u|2.
At sufficiently large distances from the source,
(51a)
(51b)
pc
pcu2 <51c>
(51d)
lul
I ' pc
w
pc2
w - pu2 <51e>
The region where all of these relations hold true is known as the far
field. The particle velocity is in phase with the acoustic pressure (as is
the case for a simple plane wave) and there is no reactive component of
energy density. Thus all of the energy density is radiant energy. In the
near-field region, close to the source, there is a large component of the
particle velocity which is out of phase with the acoustic pressure,
resulting in reactive energy which does not radiate outward!32].
Since instruments for direct measurement of acoustic intensity are not
generally available and since, over the frequency range of interest for many
noise sources, acoustic pressure can be measured more readily than particle
velocity, it is customary to carry out determinations of the noise emission
of sources by measuring the mean-square pressure and assuming that |l]=p /pc-
Note that this assumption is often implicitly made whether or not one is
interested in computing the sound power. If one measures the sound level due
to, say, a motor vehicle passby, one wishes to be able to predict the sound
level at other distances from £hat measured at a particular distance this
implicitly requires that I « p .
If it is desired to determine the total sound power from measurements of
mean-square pressure, eqs. (1*8) and (^9) are replaced by
1QQ
-------
W~— f 2
* PC yg P dS ; (52)
this equation involves two distinct assumptions:
2
- the acoustic intensity can be accurately estimated by p /pc.
- the surface of integration has been selected such that the flow of radiant
power is normal to that surface at all points on the surface.
Deviations from either of these conditions will result in errors in the deter-
mination of sound power. In addition, of course, errors may result from inade-
quate sampling of the sound field over the surface of integration.
Note that it is necessary for all five conditions given in eqs. (.51) to be
true in order to be in the far field. However, it is only necessary that eq. (51b)
be true in order for eq. (52) to be valid. Thus, in some cases, accurate deter-
minations of total sound power may be based on measurements of mean-square pressure
gven when such measurements are not made in the true far field. An example of
this is the periodic simple source which is considered below.
Conceptually, one of the simplest sound sources is a "pulsating sphere"
from which acoustic energy is radiated uniformly in all directions into free
space. The instantaneous sound pressure due to such a source is, in complex
form,
P ±u(r/c - t) (53)
£ = - e
where P is the amplitude at unit distance from the center of the source; the
real part of eq. (53) is seen to be as given in eq. (l). Using eq. (hk), the
radial particle velocity is
_ _i s£. _ / _ic\ P eioj(r/c - t) _ (54)
This can also be expressed as
_ p i[w(r/c - t) + 4>] ^ (55)
— ~ per cos<|>
where
c (56)
* E arctan —
is the phase angle between p and u.
101
-------
Thus the intensity is given, from eq. (H?), by
(5T)
at any radius. The intensity is, of course, radially directed. The total
sound power, from either eq. C^9l or eq.. C52] is
'h& (58)
. *I . Zi (59)
pc 2
which, as it must "be, is independent of r.
The energy density at any radius r is given, from eq.. (50), "by
-.
2
Thus, the energy density appro ache Sgits far-field value only when (cur/c)
»1, or, equivalently, when (X/2irr) «1, X "being the wavelength of sound at
the frequency of interest. Yet, even in the very near field, as ru/c-*0,
the intensity and sound power are given exactly "by the usual "far-field
formulae", eqs. (58) and (59).
For a dipole point source (i.e., two point sources of equal strength Taut
opposite phase located close together) the radial intensity- at radius r is
related to the mean-square pressure at the same location "by132, p. 312]
(61)
Thus, for a dipole source, estimates of sound power based on measurements of
mean-square sound pressure will give somewhat high results, the approximation
getting better as the distance from the source increases.
For more complicated sources, analytical expressions giving a general
description of the relations among intensity, acoustic pressure, and particle
velocity are, if attainable, frequently quite complex. Accordingly, the
following discussion relates only to the intensity in a particular direction
102
-------
(for which a simple solution exists) rather than to the total radiated sound
power.
Consider a rigid, circular piston of radius, a, located in an infinite
rigid baffle, as shown in Figure 18. If the piston oscillates in the
z-direction with a velocity Ue"~ , the sound pressure on the z-axis is (see,
e.g.»[61»,65])
£ = pcu
Utilizing eq. (kk), the particle velocity on the z-axis and in the z-direction
is
1/2
u = UJe
ito(z/c - t)
z/a
/ (1 + z2/a2)1/2
Using eq. (1*5), the axial intensity is
lu[(1 + z2/aV a/c - t]
pcU2
1 +
2172
sin2
+ z2/a2)
The intensity is related to the mean-square pressure by
I L , _ z/a \
21,,.
pc
:2/a2)
1/2
(63)
(64)
(65)
At z = 0, p /pc overestimates the axial intensity by a factor of two, or 3 dB.
As z/a—»•», I —»-p /pc.
2*
w
L «. J
f tG *l
* 11
M i
Figure 18.
Schematic representation of a rigid circular piston of radius, a,
contained in an infinite baffle.
103
-------
Figure 19 indicates the variation with distance of the axial intensity
level for three values of wa/c. It is seen that for large wa/c both the
intensity and pressure level oscillate rapidly for small z/a and do not
decrease monotonically until
z_ \_ t Ola _ ji
a > 2-rr* c ~" X
where X is the wavelength of sound at a frequency w/2ir. This phenomenon
occurs because sound radiated from different regions on the piston results in
interference phenomena. The curves in Figure 19 are normalized to the
far-field on-axis levels, extrapolated "back, "being set equal to 0 dB at
z/a =1. The differences, in dB, "between the solid curve and the dotted line
of slope -6 dB/double-distance represent the errors that would result from
predicting far-field on-axis levels from near-field on-axis levels at any
particular z/a.
The dashed line in Figure 19 indicates the intensity that would "be
predicted from far-field measurements of the total radiated sound power rather
than that radiated in the z-direction. It is seen that the far-field
intensity in the axial direction is greater than the average intensity,
particularly for wa/c » 1. This is better seen in the directivity plots
shown in Figure 20, where the far-field intensity level is shown, as a
function of angle, relative to the average intensity level (i.e.,
corresponding to the total sound power radiated into the half-space
corresponding to positive values of z) [65,66]. For large values of wa/c,
most of the sound power is radiated into a narrow beam along the z-axis.
The rapid oscillations in intensity and sound pressure levels and the
directive radiation patterns result from the fact that.the piston is a
coherent source — the velocities at all po?'nts on the piston have a definite
phase relationship with one another (for a rigid piston they are exactly in
phase). It is instructive to contrast this with the case where all points on
the piston radiate independently with no phase relations. For an ensemble of
incoherent point sources, each radiating uniformly in all directions, the
intensities can simply be summed and the on-axis intensity due to an
incoherent circular source is[67-68]
(66)
pc
where W is the total sound power radiated into the half-space (z positive).
This is plotted in Figure 21. It is seen that the intensity decreases
monotonically with distance away from the piston and approaches its far-field
dependence on distance very closely for all values of z/a greater than unity.
The far-field radiation from such an incoherently radiating circular source is
independent of both angle and frequency.
In order to introduce another phenomenon that is characteristic of sound
radiation, consider an infinite plate that is vibrating in flexure as
indicated in Figure 22. If the velocity of the plate normal to its surface is
described by
u(0,y) - Ueia)(y/cB ~ c> (67)
-------
o
-------
10 0 -10 -20 -10
CJQ
10 0 -10 -20 -10 0
OJQ
10
aja
100
30 20 10 0 -10 -20 -10 0 10 20 30
RELATIVE SOUND PRESSURE LEVEL, dB
Figure 20. Far-field radiation patterns showing the directionality for a
"baffled rigid piston.
106
-------
CD
^^
•*
_J
Ul
LU
K
CO
UJ
J-
Figure 21. Axial intensity level due to an incoherent circular source.
B
Figure 22. Schematic representation of sound radiation from an infinite plate
in which there is a plane "bending wave propagating at speed, c.
107
-------
where c is the velocity of sound for flexural (bending) waves in the plate,
the sound pressure can "be shown to "be [33, (&> 70j:
P(z,y) = , HUU eiUJU""B ' 'V- - - '-B r ~ t] , CD >c (68a)
- c2/v
B
- t]
B
When CB >c, the pressure at z = 0 is in phase with the velocity; when CB < c,
the pressure and velocity at z = 0 are out of phase.
The velocity, CB, of flexural waves in an infinite plate is not constant
but is related to frequency by
c - a\4 , (69)
B
where a is a constant which is dependent upon the plate thickness and elastic
properties. Introducing a "critical frequency", w = c /a , and using eq.
(hh) , the particle velocities in the z- and y-directions are
pc
(70a)
(70b)
„ ,n
£. , all u .
Using eq. (^5), the corresponding intensities are
T _ PcU2 ._.. (72a)
(72b)
—z 2\/l - w /o> , o
T 0
(73a)
2 w - »« , w
-------
III
pcU2 1
2 1 - co/0)
(75a)
-2 w , the power is radiated as a plane wave in the
direction 0, where °
6 =
so that
Sln
c A
/-ir\
(76)
(77)
where, as shown in Figure 22, X is the wavelength of the radiated sound and X
is the wavelength of the flexural wave in the plate. The angle, 6, at which
the sound is radiated must be such that the radiated wave has a "trace", or
projection, onto the plate that is equal to the wavelength of the plate wave.
In the case where
-------
i i—i—i—i—i i i i
i 1 1—i—i i i i
100
!00
Figure 23. Far-field intensity level (lower figure) above the critical
frequency for an infinite plate in flexure. The upper figure shows
the direction (relative to the plane of the plate) in which sound
is radiated.
110
-------
-30
.01
Figure 2k.. Intensity level parallel to an infinite plate in flexure at
frequencies "below the critical frequencies. The different
curves correspond to different distances from the plate. The
reference level (corresponding to 0 dE) is the same as in
Figure 23.
Ill
-------
- coherently radiating sources are frequently highly directional and in
some cases the intensity in a given direction oscillates widely with
distance from the source
- vibrating plates have associated with them a "critical frequency" below
which very little sound energy is radiated into the far field and above
which energy is radiated very efficiently in a specific direction
When sources radiate prominent discrete tones, particular care should be taken
to ensure that microphone locations are far enough away from the source
(unless, of course, primary interest is in near-field measurements, e.g.* at
an operator location). For many practical sources, however, reliable sound
power determinations can be based on measurements made rather close to the
source [5-7].
15.2. Microphone Positions
a. Anechoic Space
If sound pressure measurements are made over a hypothetical sphere, of
radius r, surrounding a sound source, the total sound power is given by
«• ^ (78)
where\p /is "the spatial average of the time-averaged squared sound pressure
over the surface of the sphere, provided the conditions for the validity of
eq. (52) are met. The total radiated sound power level can be computed from
r = L~ + 20 log r + 10 log 4n - 10 log ^ 3 (79)
in metric units, where L is the spatial average of the sound pressure level*
the average being taken ,Pon a mean-square pressure basis, over the surface of
the sphere.
In carrying out sound power determinations in an anechoic chamber, it is
frequently desired to measure the directivity pattern of the noise source.
It is customary to define a directivity factor, Q , defined as the ratio of
the intensity measured at angle 9 and distance r from the source to the
average intensity at distance r . Thus ,
% - V1 - Pft = . (80)
The directional gain, at angle 9, can be defined as
DIe = 10 log Qfi = L^ - IT . (81)
The sound pressure level at angle 8 and distance r is related to the total
sound power and the directivity index by
DIe - 20 1Q8 r - 10 ^g 4u + 10 log
112
-------
The current American standard[7l] and the draft international
standard[ll|] for determination of sound power in an anechoic chamber recommend
that the sound pressure be measured at locations corresponding to the 20
surfaces of a regular icosahedron. Other possible microphone arrays are given
in [72-73]. In lieu of a stationary array of microphones, measurements can be
made along a number of continuous paths, either by moving the microphone(s) or
by rotating the source (see b., below).
The major considerations in selecting microphone locations for measure-
ments in an anechoic environment are (l) to be far enough away from the source
to be assured of being in the far field yet close enough that reflected
signals and background noise are negligible compared with the direct signal
from the source and (2) that sufficient microphone locations are used to
obtain adequate spatial averaging, especially for directional sources.
Whenever there is any question concerning these points, it is advisable to
measurements at additional angular positions and/or at two or more radii.
b. Hemi-Anechoic Space
In carrying out measurements in a free field over a reflecting plane, it
be desired to obtain the sound pressure level at one or more specific
microphone locations (e.g., a microphone at a distance of 50 feet from the
center line of a passing vehicle) or to obtain the sound power level (with or'
without directivity information). In the case of sound pressure level
measurements, the microphone location(s) is(are) selected to correspond to
"typical listener locations or to provide information that can be reliably
extrapolated to other locations. In the case of sound power level
determination, measurements of sound pressure level are typically made over a
hypothetical surface surrounding the source. If a hemispherical surface is
used, eqs. (79) and (82) apply, but with fcir replaced by 2tr and the directivity
index computed from
DIa = L - TT + 3 dB, <83)
e PQ P
where L~" is now the average over the test hemisphere. The 3 dB arises from
the facl that averaging is not carried over the space below the plane.
A number of microphone arrays corresponding to hemispherical measurement
surfaces are suggested in the literature. The current American standard[7l]
endorses a 12-point array which is one of several arrays suggested in [72].
Alternative 10-point arrays are given in [72], in [33,13], and in [IV]. One
recent investigation^] utilized a 73-point array. In addition to these fixed
microphone arrays, it is common to use continuous microphone traverses (either
ty moving the microphone around the source or by holding the microphone
stationary and rotating the source) along circular paths parallel to the
reflecting plane. One such configuration is shown in Figure 25.
113
-------
AXIS OF ROTATION OF MICROPHONE
TRAVERSING MECHANISM
ELEVATION OF
MICROPHONE TRAVERSES
HEIGHT OF CORRESPONDING
AREAS OF HEMISPHERE
7777
Figure 25. A suggested set of continuous micropfione traverses for determinate
of sound power In a free field above a reflecting plane[l^l«
llU
-------
Selection of microphone locations for measurements over a reflecting plane
is complicated by variations in sound pressure with angle due to constructive
and destructive interference between sound waves radiated directly to the
microphone and reflected waves. The type of difficulties that may be
encountered can be illustrated by considering a point source located above a
perfectly reflecting plane. Assuming the measurement locations are
sufficiently far away to be in the far field of the source and its image, the
variation of sound pressure level with angle is as shown in Figure 26, where
the three curves shown correspond to a pure tone, a 1/3-octave band of noise,
and a 1/1-octave band of noise. Baade[?3], from whom this figure was taken,
draws the following conclusions:
"(a) At low frequencies, sound reflection does not cause any significant
directivity as long as the wave length is more than 10 times the distance
between the source and the reflecting plane, (b) High frequency random
sound is radiated fairly uniformly in all directions except those almost
parallel to the reflecting plane. Microphone readings taken near the
reflecting plane therefore tend to have low accuracy. This low accuracy
zone shrinks with increasing frequency and band width, (c) At medium
frequencies, the directivity pattern is very pronounced, even for random
sound of one octave effective band width, (d) The "valleys" in the curves
[of Figure 26] occur in regular intervals. At any given frequency and
source location, low readings will be obtained at several microphone
positions spaced in the ratio of 1:3:5=7 from the reflecting plane. Odd
multiple spacings should therefore be avoided."
Figure 27 illustrates the errors in determining the total sound power from
a random point source a distance h above a reflecting plane, when measurements
are made in l/3_octave band widths, for the microphone arrays indicated. The
difference between the mean for each array (broken curves) and the true mean
(solid curve) represents the error associated with the particular array.
One way to obtain essentially perfect vertical averaging is to traverse
the microphone along a meridian as shown in Figure 28.
The curves shown in Figures 26 and 27 were calculated for the case of a
point source above a perfectly reflecting plane. In general, increasing the
source size will result in less variation of the sound pressure level with
angle.
Other references relevant to the effect of the reflecting plane on the
selection of appropriate microphone positions include [74-77J.
Several recent investigations (e.g., [75, 5-Tl) have found that rather
accurate determinations of the total sound power emitted by real machines can
be made using measurements of sound pressure level made quite close (0.3 to 1
*) to the source. Accordingly, a draft international standard[13] suggests two
arrays of microphone positions to be used for such close-in determinations of
sound power.
115
-------
Figure 26.
Effect of "bandwidth on directivity of sound field of a point source
at a distance, h, from a reflecting plane U = vave length at
center frequency of
pure tone
1/3 octave '
full octave
^effective "band width
116
-------
,TRUE MEAN LEVEL
OBTAINED BY INTEGRATION
V)
_]
UJ
go
o
UJ
o
-3
-6
-12
-15
-v*
MEAN OF (4) POINT
HEMISPHERE
MEAN OF (6) POINT
HEMISPHERE
0.2 0.3 * O6 1.0 2O 10
DISTANCE FROM SOURCE TO REFLECTING PLANE
WAVE LENGTH
0.1
5.0
£
X
Figure 27. Error introduced "by limited number of measuring points when deter-
mining sound power output of a random point source near a reflecting
plane[73-7^]• The error is the difference "between the curve corres-
ponding to a given microphone array and the solid curve which corres-
ponds to the true mean level. All curves correspond to a 1/3-octave
"band width.
117
-------
TRAVELING MICROPHONE
COSINE FUNCTION
POTENTIOMETER FOR
AREA WEIGHTING
NOISE SOURCE ON
REVOLVING PLATFORM
Figure 28. System for microphone traverses along meridional patns[lU].
118
-------
In the first of these, a "reference parallelepiped", as shown in Figure
29, is imagined to just enclose the source. A measurement surface is
hypothesized to have its faces parallel to, and spaced a constant distance
(typically 1 m) from, the reference parallelepiped. Nine key microphone
locations are established, corresponding to the (approximate) centers of the
five faces, plus the four upper corners, of the measurement surface.
Procedures are given in [13] for adding additional microphone positions for
large sources.
Figure 30 shows an example of the "composed measurement surface" given in
[13]. It consists of a parallelepiped with rounded edges and corners so as to
"be everywhere equidistant from the reference parallelepiped which just
encloses the source. There are eight key microphone positions, four on the
side faces and four which usually lie on the upper curved edges of the
measurement surface.
HolmertT] has carried out extensive comparisons of data (on 17 portable
air compressors) obtained using the surfaces of Figures 29 and 30 with data
obtained using a 73-point hemispherical array of 7 m radius.
c. Reverberant Space
As stated at the beginning of Section 12. 2. c, elementary reverberation
room theory is based on geometrical acoustics, in which wave phenomena are
neglected. If that assumption were true, a single microphone placed in the
reverberant field would suffice. However, wave phenomena result in local
variations in the sound pressure level, particularly for pure-tones. The
question of how best to sample the sound field in a reverberant room has been
the subject of active research over the past decade [ 79-103 ].
The sound power emitted from a source is related, as discussed in Section
12, to the sound pressure in the reverberant field averaged in space and time
on a mean-square basis. In practice, spatial averaging over a finite path
length (using a traversing microphone) or over a fixed number of microphone
positions leads only to an estimate of the true mean-square sound pressure.
Theory shows that in order to have essentially independent samples of the
sound field, microphones must be located at least (approximately) one-half
wavelength apart at each frequency of interest. In addition, interference
phenomena occur near reflecting surfaces so that microphones typically are not
located within one-half wave length of any room boundary or diffuser. For a
close-packed (hexagonal close-packed or face-centered cubic) array of
microphones located one-half wavelength from each other and at least one-half
wavelength from any wall, the number of microphones that can be accommodated
is less than
where V is the volume of the room and X is the wavelength of sound of
frequency f traveling at speed c. This upper limit is shown in Figure 31 as a
function of frequency for rooms of different volumes. In practice the number
of independent microphone positions which could be located in a reverberation
room would usually be significantly less than this upper limit.
119
-------
MICROPHONE
POSITIONS
MEASUREMENT
SURFACE
L.
REFERENCE
PARALLELEPIPED
Figure 29. Microphone array for a parallelepiped measurement surface[13]
120
-------
REFERENCE
PARALLELEPIPED-
////A//////////////////////
MEASUREMENT SURFACE
REFERENCE
PARALLELEPIPED
Figure 30. Microphone array for a "composed" measurement surface[l3].
121
-------
1000
500
co
O
CO
O
Q.
UJ
z
O
I
Q.
O
o:
o
U.
O
o:
ui
CD
X
<
200 -
100
63
125
FREQUENCY,
250
500
Figure 31. Upper limit on the number of microphone positions in rooms of the
volumes shown if each position is at least a half-wavelength from
all other positions.
122
-------
When a reverberation room is excited by a random noise source, the
normalized variance of the mean-square sound pressure in the room is given
approximately by the following expression[8U, 93* 100] provided the modal
overlap is sufficiently great (e.g., see eq. (31)) and there are no moving
diffusers in the room:
V2 = !• arctan z -^-ln (1+ z2) , (85)
*
where z = BT/2.2, T being the reverberation time and B the bandwidth of the
filter. For small values of z, eq. (85) becomes
2 z2 z4 z6
V - 1 - 27J + JTs - 4^7 + '''' ' (86)
Q
so that V —*-l as BT—^0. For large values of z, eq. (85) can be written as
v2 = £ " \ C1 + ln z) ^—zr + fi ~ (87a>
z z2 1-3-z 2-5-zb
5[f + °'8] " [erfl*0-8] ' Z>>1* (87b)
p P
so that V —*-TT/Z » 6.91/BT as z>>l. The normalized variance, V , computed
from eq. (85), is shown in Figure 32. The spatial variance of the mean-square
sound pressure decreases with an increase in the bandwidth of the noise signal
since the number of modes excited in the room is approximately proportional to
the bandwidth. As the reverberation time increases the amount of modal
overlap decreases so that the effective number of independent samples in a
given bandwidth increases. It should be emphasized that eq. (83) is only
valid for modal overlap greater than about 3 so that if the reverberation time
is too large, the behavior shown in Figure 32 can no longer be expected to be
observed.
If N independent samples of the sound field are taken, the mean value of
the (normalized) squared sound pressure is given approximately by s = V /N.
The 95 percent confidence interval is given by +e = +1.965. Thus the number
of independent microphone positions required in order to have 95 percent con-
fidence that the fractional error in the mean value of the squared pressure is
less than +_e is
N > A.96\ „
_ ^~-l V . (88)
As an example of the use of eq. (88), let e=.259» corresponding to 95 percent
confidence limits of +1 dB. Let B = 23 Hz (l/3-octave band centered at 100
Hz) and T = 2 s. Thus z = BT/2.2 = 20.9. From eq. (83) or Figure 32, V =
0.132. Thus a value of N greater than (l.96/0.259)2V2 = 57.3V2 = 7.6 is
required.
Figure 33, which was generated using eq. (85) and (88), shows the minimum
number of microphone positions needed in order to be 95 percent confident that
for l/3-octave bands of random noise, the spatial average of the mean-square
sound pressure is known to within +1 dB. This number is shown as a function of
123
-------
I I I I I I 1 I
100 200 500 1000
BT
Figure 32. normalized variance, ((Q?2)2) -
-------
100
CO
2 50
CO
o
O_
LJ
CL
o
01
o
or
LU
m
20
10
= 2
T I I I I I I I i l
PURE TONE
31.5 63 125 250 500 1000 2000 4000
BAND CENTER FREQUENCY, H*
Figure 33. Minimum number of microphone positions needed for 1/3-octave "bands
of noise in order to have 95 percent confidence that the spatial
average of the mean-square sound pressure is fcnovn within +1 dB.
125
-------
frequency (utilizing the fact that the "bandwidth of a standard 1/3-octave
filter is equal to 0.232 times the center frequency, so that z = fT/9.5) with
reverberation time as a parameter. This figure may "be used for other
reverberation times simply "by entering the abcissa at a numerical value equal
to fT and reading the minimum number of microphone positions from the curve
corresponding to T = Is.
Suppose additional absorption had "been added to a reverberation room,
lowering the reverberation time from hs to Is, in order to obtain a more
uniform frequency response in the room. Assuming modal overlap were high
enough in both cases for eq. (83) to be valid, it would be necessary to
increase the number of microphone positions from k to Ik in order to still have
95 percent confidence that the average sound pressure level of a 1/3-octave
band of noise was known within +_ 1 dB. This example illustrates that once
enough absorption has been added to achieve adequate modal overlap at the
lowest frequency of interest, it is generally inadvisable to further reduce the
reverberation time.
When a reverberation-room is excited by a number of pure tones, the
normalized variance of the mean-square sound pressure in the room, provided the
modal overlap is sufficiently great and there are no moving diffusers in the
room, is given approximately by [8U, 93, 100]
(89)
>
where A. (or A. or A. ) is the mean-square sound pressure of the tone at
frequency f, (or f, ) and T is the reverberation time. In its complete form
this expression is perhaps too complex to see easily the effects of the
reverberation time and the spacing between tones. However, several special
cases do yield considerable insight.
If the frequency separation, f. - f,, between tones is large compared to
the modal bandwidth, 2.2/T, only the terms for j=k contribute significantly to
the summations and eq. (89) reduces to
V
,2
M
m=l
M
A 2
2 ' (90)
(A)
If, further, all the tones are of equal strength,
V =M ' (91)
Thus for well-separated tones of equal strength, the normalized variance of the
mean-square sound pressure is simply equal to the reciprocal of the number of
tones.
If the tones are of equal strength but are no longer well-separated, an
126
-------
interesting special case is that vhere the M tones are uniformly separated by
6f Hz. With this simplification, eq. (89) reduces to
M
„ M-l M-n
V - c
(nT6f/2.2)
2
(92)
Figure 3^ illustrates the variance, computed from eq.. ( £2) as a function of the
number of tones, with T6f, the product of the reverberation time and the tone
spacing, as a parameter. It is seen that when TSf is very small, so that the
room responses to the several tones are highly correlated, the variance remains
near unity, the value for a single tones, until M increases enough so that some
of the tones are far enough apart in frequency to significantly reduce this
correlation. When T5f "becomes large, the variance approaches 1/M, the value
for well-separated tones (eq. (91))- If the tones are not of equal strength,
the variance will be greater than that indicated by eq. (92) and "by Figure 3^
tut the effect of reverberation time will be approximately the same provided
more than one tone contribute significantly to the overall mean-square sound
pressure. It is again seen that the reverberation time should be as large as
possible provided only that it is small enough to provide sufficient modal
overlap in order that the reverberation room have a fairly uniform frequency
response.
If the number, strength, and spacing of tones is known, the variance com-
puted from one of eqs. (89) - (92) can be used in conjunction with eq. (88) to
estimate the minimum number of microphones needed in order to determine the
mean-square sound pressure within the desired confidence limits.
Equation (Qk) and Figure 31 indicated upper limits on the number of
independent"stationary microphone positions which can be accommodated in a
reverberation room of a given volume. It is frequently more convenient to use
a single microphone which is moved slowly over a particular path so as to
sample the sound field at a number of positions. Since the sound field at one
location can be highly correlated with the sound field at a nearby position, it
is useful to consider the equivalent number of microphone positions
corresponding to a given path. If a continuous linear microphone traverse over
a path length L is used the equivalent number of independent microphone
positions is [85-87, 91, 93]:
2L-f - Ne < 2L-| + 1 , (93)
For a circular path of circumference L [93,96],
N £
eq
1 , 2L'£ < 1
2L-- , 2L-- >. 1
r c
(94)
At high frequencies, where adequate spatial sampling can readily be achieved,
it may be easier to use a continuous traverse. At lower frequencies, a fixed
array will usually enable more independent samples than can simply be obtained
with a continuous traverse.
127
-------
CM
>
M
Figure 3^. Normalized variance of the mean-squared sound pressure in a room,
having a reverberation time, T, excited by M equal-strength pure
tones uniformly separated in frequency by 6f. The curves
correspond to eq.. (92) which is valid for modal overlaps greater
than about 3.
128
-------
The current national[29] and international[lO] standards for determination
of sound power in a reverberation room state that for broad-band sound sources,
space averaging of the sound field shall be accomplished by one of the
following two procedures:
(l) Traversing a microphone at constant speed over a path at least 3 m in
length while the signal if being averaged on a mean-square basis.
The path may be a line, an arc as obtained by swinging the
microphone, a circle, or some other geometric figure.
(2) Using an array of at least three fixed microphones (or microphone
positions) spaced at least X/2 from each other, where A is the
wavelength of sound corresponding to the lowest frequency in the
frequency range of interest. The outputs of the microphones shall be
either scanned automatically and averaged on a mean-square basis by
the indicating device, or the average shall be computed from the
mean-square outputs of each individual microphone position.
A path length of 3 m for the traverse and three positions for the array are the
Minimum requirements. It may be necessary to use a more extensive microphone
traverse or array, or use moving or stationary sound diffusers, or both, in
order to meet the requirements of the standard.
For pure tones sources the national[29] and international[11] standards
require the minimum number of independent microphone positions shown in the
following table:
OCTAVE BAWD
CONTAINING
DISCRETE
FREQUENCY
COMPONENT
125 Hz
250 Hz
500 Hz
1000 Hz
Above 1000 Hz
THIRD-OCTAVE
BAND CONTAINING
DISCRETE
FREQUENCY
COMPONENT
100 to 160 Hz
200 to 315 Hz
UOO to 630 Hz
800 to 1250 Hz
Above 1250 Hz
MINIMUM
NUMBER OF
INDEPENDENT
MICROPHONE
POSITIONS
6
12
2l»
30
30
The corresponding approximate minimum path lengths, if a traversing microphone
is used, follow from eqs. (9l) or (92). These standards[10,29] also require
that:
129
-------
The microphone traverse or array shall be within that part of the test
room where the reverberant sound field dominates and where the
contribution of the direct field to the measured mean-square pressure is
negligible. To ensure that the chosen microphone traverse or array is
within the reverberant field, the following criteria shall be met:
(l) The minimum distance between the sound source and the nearest
microphone position shall not be less than
d . = 0.08
nun
where
V = volume of test room in cubic meters
T - reverberation time in seconds
(2) ....
This criterion corresponds (see eq.. (21)) to the total (direct and
reverberant) sound field not being more than 1.8 dB above the reverberant
sound field (i.e., the direct field is 3 dB below the reverberant field),
provided the source is essentially omni-directional. If the source is
directional, one should place the nearest microphone(s) still further from the
source (see eq.. (21)) to assure that the direct field is at least 3 dB below
the reverberant field. Since typically only one or two of the several
microphone positions would be close enough to the source to be affected by the
direct field, the bias introduced is rather small (e.g., if only one of six
microphones is biased upwards by 1.8 dB, the estimate of the average sound
level will be biased upward by 0.3 dB; if more microphones are used, the bias
would typically be less).
15.3. Source Positions
The importance of selecting an appropriate source location„ typical of
normal operation, was mentioned in Section lU.l. In carrying put measurements
under conditions of a free field, or a free field over a reflecting plane,
there are no additional measurement procedures specifically concerned with
source position.
The current national[29] and international[ll] standards for
determination of sound power in a reverberation chamber give an empirical
formula for calculating the recommended number of source locations when the
source produces pure tones (see also [97]). This number depends on the
reverberation time and volume of the room and on the frequency of the tone(s).
The use of multiple source positions reduces the error due to low modal
density because the extent to which a given mode is excited depends on the
source positional*]. "The error due to incomplete space averaging is reduced
because the total number of samples of the sound field is the product of the
number of microphone positions used for each source position times the number
of the source positions"[29].
130
-------
15.4. Period of Observation
If the sound level at any given microphone position is steady, the
required sampling time follows directly from considerations such as those
discussed in Sec. 13. In carrying out measurements of the noise emitted from
a particular source, several factors can cause the sound level at a given
location to vary:
1* The noise emission from the source may be inherently variable as a
function of changes in some operating parameter (e.g., speed, load,
normal operating cycle).
2. There may be changes in the environment through which the sound
propagates. These could include wind, air turbulence, temperature
gradients, etc.
3. Some aspect of the measurement process may vary the sound level.
Typical examples would be motion of the microphone relative to the
source or modulation of the sound field in a reverberation room by a
moving diffuser.
In case 1, it is necessary to decide whether one is interested in the
noise emission during one or more specific portions of an operating cycle or
whether one is interested in an appropriate average over a complete operating
cycle. Once this decision has been made, the appropriate period of
observation follows rather easily.
In case 2, one is probably in trouble if the situation arises and hence
one should seek a more favorable environment. However, if that is not
possible, sufficient independent data points should be taken to permit
averaging out statistical variations.
In case 3, two situations merit specific mention. If a source is being
rotated in an anechoic (or hemi-anechoic) environment (or if a microphone is
being moved around the source) it is important that the motion be slow enough
to permit valid sampling. This is particularly important for highly
directional sources and for sources which produce random noise. If sound power
determinations are being made in a reverberation chamber it is very common to
use a moving microphone and/or a moving diffuser. The national[29] and
international[10, 11] standards provide guidance on the appropriate period of
observation.
15.5. Use of Diffuse™
It has been customary for many years to use fixed and/or moving
"diffusers" to affect the accuracy and precision of sound measurements carried
out in reverberation rooms. Dodd and Doak[105l have pointed out the reasons
why stationary diffusers are probably rather ineffective in improving
reverberation room determinations of sound power:
131
-------
"Also, it is perfectly clear that, whatever else they may do to
affect sound pressure level distributions in reverberant rooms* fixed
diffusers in no way reduce, or otherwise affect, the statistical spatial
and frequency fluctuations in sound pressure level described above. Thus
use of fixed diffusers will not, in general, reduce the amount of
frequency and space averaging required to give unambiguous measurements.
Fixed diffusers could, in theory, reduce eigenfrequency degeneracy and
also eigenfunction degeneracy (i.e., the tendency of eigenfunctions in
geometrically simple rooms to fall into classes having maxima or minima
at particular points). As, however, reverberant rooms are seldom
constructed in the shapes of perfect spheres, cylinders, or cubes, and
also because when used in practical measurements they usually contain at
least one scattering object, it would appear that use of fixed diffusers
contributes much more to optical aesthetics than it does to acoustics.
Of course, there is one notable exception to this remark. When
relatively large areas of absorbing material are placed in reverberation
rooms for absorption coefficient measurements — on the floor, for
example — the reverberant field may no longer be directionally
isotropic. In this case, it is possible that suitable fixed diffusers
can restore a measure.of directional isotropy to the field in the
neighbourhood of the area of the absorbent material."
On the other hand, it has been well established that moving diffusers can
very significantly improve the determination of the average steady-state sound
pressure level in a reverberation room. Consider first the situation where a
diffuser is incrementally moved through a range of stationary orientations.
Each change in orientation can result in a perturbation of the standing wave
pattern in the room, with both the eigenfrequencies and the eigenfunctions
being somewhat changed. In addition, as shown by Ebbing[lo6], the radiation
impedance seen by the source will vary with diffuser orientation so that the
actual sound power radiated by the source also will vary. Thus, even when a
diffuser is moved very slowly, the various orientations of the diffuser lead
to an improved average for the mean-square sound pressure in the room.
When the diffuser is moved more rapidly, there is both amplitude and fre-
quency modulation of the sound field[107-110, 93, 97-100] with the result that
the sound energy from a pure tone is converted into a number of tones —.as
seen in Section 15.2, the spatial variance due to a multitone can be much less
than that due to a single tone. In addition, the rotation of the diffuser
does not permit modes to build up to the full strength they could have if the
diffuser were stationary.
While the design of moving diffusers remains somewhat of an art, it is
recognized[llO] that the major design parameters are size, shape, percent open
area, surface density, speed, number of panels, and panel' damping. A few
general guidelines for the design of rotating diffusers are given in current
standards[29,11].
15.6. Background Noise Measurements
A general discussion of criteria for background noise was given in
Section 12.3. It is important to assure that the background noise' is the same
132
-------
vhen it is measured as it is when the source is operated. If corrections for
"background noise are to "be made, additional measurements may "be necessary to
ascertain whether or not there is coherence between the source noise and the
"background noise.
Peterson[lll] has recently examined the uncertainties which may occur in
the background noise correction when the device noise and the background noise
are both random in character.
15.7. Characterization of Test Environment
Section 12.2 includes discussions of the adequacy of various types of
test environment. In the present section, a brief summary is given of the
means for characterizing the test environment which are spelled out in current
standards.
a. Anechoic Chamber
The current draft international standard[lU] for precision determinations
of sound power in anechoic rooms requires measuring the change of sound level
with distance along at least eight straight paths away from the center of an
omnidirectional sound source which is radiating a pure tone. The range of
distances and frequencies for which the measured sound levels agree, within
specified tolerances, with levels predicted by the inverse square law define
the usable volume and frequency range for the anechoic chamber.
b. Hemi-Anechoic Environment
The current draft international standard[ik] for precision determinations
of sound power in hemi-anechoic rooms utilizes the qualification procedure
just described but allows larger deviations than in the case of anechoic
chambers. No corrections are permitted for the influence of the test
environment.
The current draft international standard[l3] for engineering methods of
determining sound power under free-field conditions over a reflecting plane
permits making a correction of up to 2 dB for the influence of the test
environment. Three alternative methods are given for qualifying the acoustic
environment and determining the "environmental correction":
1. Replace the device whose sound power level is being determined with
a reference sound source of known sound power output. The environ-
mental correction is then taken as the difference between the known
power level and the power level computed using the procedures of the
standard.
2. Replace the device whose sound power level is being determined with
a broad band test source (of unknown power output). Determinations
133.
-------
of the sound power of this test source are made using three measure-
ment surfaces of similar shape "but different size. The
environmental correction is obtained from the differences among the
three sound power levels obtained in this manner.
3. Compute the environmental correction from measurements of the rever-
beration time of the test room.
The current draft international survey method[l5] for determination of sound
power levels permits environmental corrections of up to J dB based on the
absorption (estimated or computed from the reverberation time) of the test
room.
The above-described procedures are based mainly on the work of
Hubner[5-6], Diehl[112-113] has investigated determination of environmental
corrections by the "two-surface method" in which two measurement surfaces of
different area are utilized vith the actual device under investigation (rather
than a special test source).
Hubner[6] has pointed out that none of the above procedures for
determining environmental corrections can account for the influence of the
acoustic impedance of the reflecting plane. The draft international
standards[13-1^] for sound power determination simply require that "the
absorption coefficient of the plane should be less than 0.06 over the
frequency range of interest." In the case of outdoor measurements, whether of
sound power level or of sound pressure level at a specified location, it may
be difficult to ensure that the absorption is as low as desired. At test
sites for measuring motor vehicle noise emission, very large differences have
been observed between the acoustical absorption of sealed and unsealed
asphalt[llU], The flatness of reflecting surfaces can also lead to
problems[115]. Statistically significant differences have been observed among
results obtained at various test sites[ll6]. While there has been progress in
developing means to correct for the effects of the reflecting plane( e.g.* see
[T6-T8]), much further work is needed. At present it appears that dense
concrete, sealed asphalt, or a material at least equally dense and free from
porosities should be specified for the reflecting plane.
c. Reverberant Environment
The current national and international standards for reverberation-room
determinations of both broad-band[10, 29] and pure-tone[11, 29] sources give
specific room qualification procedures.
In the case of broad-band sound, determinations of the apparent sound
power level of a reference sound source are carried out for at least eight
different source locations. In order for the room to be qualified, the
standard deviations of this set of band power levels must not exceed the
limits tabulated (as a function of frequency).
In the case of sources containing discrete-frequency components, two
alternative qualification procedures are given. In the first of these, an
13U
-------
array of six microphone positions is used to estimate the spatial variance of
the mean-square pressure in the room while the device under test is emitting
noise. This information is used to enter a table which gives the minimum
number of required microphone procedures. Procedures are also given for
determining the minimum number of required source positions.
The second qualification procedure for the measurement of discrete
frequency components involves measuring the frequency response of the
reverberation room. This is done by measuring the space/time averaged sound
pressure level at each of a specified series of frequencies. A loudspeaker,
excited by an oscillator, is used and adjustments are made for the frequency
response of the loudspeaker and measuring instrumentation. The apparent
standard deviation of the frequency response over each frequency band must not
exceed a tabulated limit.
15.8. Calibration
A measurement standard should specify what calibration procedures are
required in conjunction with normal testing procedures and also what
calibration procedures are required (e.g., annually] in order to ensure
proper functioning of all instrumentation.
135
-------
16. Calculation Procedures
• The measurement standard should clearly and unambiguously specify all
calculation procedures that are required in order to carry out
measurements in accordance with the standard.
16.1. Correction for Background Noise
Corrections for background noise were discussed in Sections 12.3 and
15.6. A measurement standard should clearly indicate how much background
noise is permissible and whether or not corrections for background noise are
to be made. If corrections are to be made, the standard should clearly spell
out the correction procedure to be used.
Consider the following differences in the approach to background noise
corrections as taken in the current (draft and approved) international
standards for sound power determination:
Precision methods for anechoic and semi-anechoic rooms [lU]
The background noise must be at least 6 dB below the measured sound
pressure levels. For background noise between 6 and 15 dB down, the cor-
rections given in Table 3 (Section 12.3) are to be applied. The
corrections are rounded to the nearest 0.1 dB.
Precision methods in reverberation rooms[10,11]
Same as above except no corrections are applied when the background
noise is more than 10 dB down. (This reflects the greater uncertainty in
reverberation room measurements.)
Engineering methods for free-field conditions over a reflecting plane[13]
For background noise that is 6 to 8 dB down, a 1.0 dB correction is
applied. For background noise that is 9 to 10 dB down, a 0.5 dB
correction is applied.
Engineering methods for special reverberation test rooms[12]
For background noise that is U to 5 dB down, a 2 dB correction is
applied. For background noise that is 6 to 9 dB down, a 1 dB correction
is applied.
Survey method (free-field conditions over a reflecting plane)[15]
For background noise which is 3 dB down, a 3 dB correction is
applied. Otherwise, same as previous standard[l2].
136
-------
Many standards require the "background noise to be at least 10 dB down and
Permit no correction.
16.2. Correction for Test Environment
4 measurement standard should clearly delineate what corrections, if any,
are to be made for the influence of such factors as:
— test room
reflecting planes(s)
temperature and barometric pressure
wind
16.3. Determination of Mean-Square Pressure
A measurement standard should specify how individual determinations of
sound pressure level are to be combined in order to obtain the appropriate
average value for the mean-square pressure or the sound pressure level. This
involves both time and spatial averaging.
Some standards permit averaging of levels, rather than of mean-square
pressure, when the range of levels is not too large. If such is the case, the
allowable range should be indicated.
16.4. Calculation of Sound Power
A measurement standard should present explicit equations for calculating
sound power level for the measured data.
16.5. Calculation of Noise Rating
If the final data (sound pressure level or sound power level as a
function of frequency) are to be used to compute a single-figure rating that
is intended to correlate with subjective response, the computation procedure
should be clearly and unambiguously given. For example, a procedure which
involves application of a "pure tone penalty" should specify quantitatively
how the presence and magnitude of the pure tone is to be determined.
If other documents are to be referenced, the particular issues and
relevant portions of those documents should be indicated.
16.6. Calculation of Measurement Uncertainty
"Examination of noise literature reveals a general absence of estimates
of uncertainties associated with measurement or predictive procedures or with
actual measured data. Even when such estimates are given, they frequently are
ambiguous or inadequate.
"In noise control engineering, there has been little opportunity for
137
-------
direct comparison of data obtained by different investigators on nominally
identical specimens. Thus, there have been few direct indications of
experimental error. Furthermore, in view of the vagaries of human response,
it has been probably justifiable to take the attitude that a "few decibels"
are not of much consequence. Perhaps for these reasons, the use of
uncertainty estimates has not evolved in noise control engineering.
"The emergence of many new and pending noise regulations has changed this
situation drastically. Enforcement of these regulations requires
manufacturers, independent laboratories, and regulatory agencies to conduct
measurement on similar specimens. The requirement that all of them obtain
essentially the same answer creates a strong need for realistic, reliable
estimates of measurement uncertainty throughout the field of noise control
engineering. Measurement uncertainties must be known not only to enforce
equitably noise regulations but also to enable rational selection of
economical noise control solutions and to enable reliable monitoring of the
noise environment."[117]
Although little work has been done in the past toward assessment of the
uncertainty (see Appendix C for a discussion) of noise measurements, efforts
should be made in future measurement standards to incorporate specific calcu-
lation procedures for estimating the uncertainty of the final number emerging
from a noise measurement procedure.
138
-------
17. Information to Be Recorded
The measurement standard should require the following essential
items to be recorded:
• The size, dimensions, design characteristics and noise
performance claims for the source under test.
• The location, mounting and/or installation details of the source,
• The operational and loading characteristics of the source during
the test.
• A description of the acoustic environment including test
facility* background noise levels and environmental conditions,
• Identification of instrumentation utilized.
• Documentation of unavoidable deviations from the prescribed test
procedures.
• A maintenance and calibration record to indicate the current
calibration status of all instrumentation. Calibration methods
and periodicity, accuracy and traceability of the calibration
devices need to be detailed.
• All significant data collected during the test.
• Documentation of calculation procedures utilized in transforming
the raij) data into its final form.
• An indication of the accuracy and precision of the data.
17.1. Sound Source Under Test
A complete description of the test specimen should "he recorded. This
should include:
(l) Size and dimensions of specimen
(2) Detailed design and construction characteristics
(3) Expected performance requirements (manufacturer's claims)!.
17.2. Sound Source Installation Details
A detailed description or a photograph of the noise source as it is
normally installed for use should be included. Following this, a
description should he given of:
139
-------
(l) Location of source
(2) How source is mounted
(3) Installation of source
Insofar as possible, the test installation should "be representative of normal
use conditions.
17.3. Sound Source Operating Procedures
The information to "be included under sound source operating procedures
is:
(l) Auxiliary equipment used to power the source (if any)
{2} How the source was operated during the test
(3) What loading, if any, was applied during the test.
A copy of the operating instructions for the noise source should be made
part of the record.
17.4. Acoustic Environment
Information on the acoustic environment should include a detailed des-
cription of the test facility or site, the nature and levels of any background
noise, and temperature, humidity, barometric pressure and wind conditions (as
pertinent).
17.5. Instrumentation
A complete list of the instrumentation utilized when conducting the tests
should be recorded, including the following:
(l) Name of instrument
(2) Manufacturer
(3) Model Number
(U) Serial Number
In addition, for each sophisticated test device there should b.e a schematic,
parts list, technical description of operation, a complete accuracy statement
(absolute error, repeatability, effect of environmental and operational
factors, etc.), and & maintenance and calibration schedule (note that, for
purchased equipment, most or all of this information may be supplied by the
manufacturer or vendor). The information noted above should be readily
available on request.
1UO
-------
17.6. Special Measurement Procedures
Any deviations from the standard method of test, or any additional tests
which are conducted should lie carefully recorded.
17.7. Calibration History
For each instrument a fully implemented calibration procedure should
prescribe the method and periodicity of calibration and the accuracy and
traceability of calibration devices used. All instruments should be assigned
an exclusive identification number and complete records maintained to indicate
current calibration status at all times. On key instrumentation, labels
should be affixed to indicate status and next required calibration. In
addition, the calibration procedures utilized before and after each test
should be detailed, and the results of test environment qualification checks
should be referenced.
17.8. Acoustical Data and Related Information
All significant data collected during the tests should be recorded.
Also, variable settings on the equipment which have an influence on the data
should be noted (for example, gain setting on a measuring amplifier, voltage
and speed settings on recorders, etc.).
17.9. Special Calculation Procedures
If any special procedures are used to convert the data to some other
measure (for example, converting sound pressure level readings to sound power,
voltage readings to sound pressure level readings), these should be
documented.
17.10 Measurement Uncertainty
An indication of the measurement uncertainty should be calculated
from the data (see Appendix C).
-------
18. Information to Be Reported
The measurement standard should require that the following essential
items ~be reported:
• A complete description of the product tested.
• A detailed description of (1) the acoustic environment in which the
tests are carried out and (2) how the device is operated under test.
Detailed diagrams should be utilized where appropriate.
• When unavoidable deviations from the prescribed measurement pro-
cedures are necessary> a description of the substituted procedures
and a justification for the modification.
• A tabulation of the acoustical data in its final form plus the
notation of any factor which is thought to have influenced the data.
• A statement of measurement uncertainty including (1) the degree of
confidence placed on the measurement results and (2) an indication
of the representativeness of the sample tested.
Not every test report would include exactly the same information. To
make it a requirement that each test report contain information on X number of
attributes would not only "be wasteful and unnecessary, "but also detrimental to
the whole testing process. In general, however, there are five areas where
information should always appear in a test report — these are briefly
described "below.
18.1. Identification of Source
The test report should contain a complete description of the product
which was tested. Information to "be included would "be the manufacturer's name
and address, model and serial numbers of the product, a description (.and if
appropriate, a diagram) of the appearance of the device, its conventional
operating characteristics, and the performance claims of the manufacturer
which relate to noise.
18.2. Source Installation and Operating Procedure
A detailed description, witn diagrams where appropriate, of the acoustic
environment in which the tests are carried out should "be included in the-
report. As part of this description, the location of the sound source in the
test environment, its mounting configuration, and its spatial relationship to
the measuring system should be discussed. A description of how the device is
operated under test is an essential part of the report. If the device is
operated in any manner other than its conventional operating mode, this
-------
operating procedure should be completely described.
18.3. Deviations from Standard Measurement Procedures
When there are no deviations from the standard measurement procedure, it
is sufficient in the test report to merely cite the measurement procedure
employed. If there are deviations from the measurement procedure, the
following information should be included in the test report:
(l) The section(s) of the standard test method vhich has (have)
been deviated from.
(2) A complete description of the method and procedures which have
been substituted.
(3) A convincing Justification of why the standard measurement
method was not followed.
(M An estimate of the uncertainty due to deviations from standard
measurement procedures.
18.4. Acoustical Data and Related Information
A tabulation of the acoustical data in its final form (that is, if any
conversion factors have been applied to the original data) should appear in
the test report. If there are any factors, such as relative humidity,
barometric pressure, wind speed and temperature, which the investigator
Believes may have influenced the data, these should be accounted for in the
test report.
18.5. Measurement Uncertainty
The section on measurement uncertainty should include two parts. First,
there should be some indicator of the degree of confidence which can be
Placed in the measurement results. This could be expressed in terms of
standard error, confidence limits, or some other appropriate statistical
factor. Secondly, there should be some statement that indicates how
confident the testing laboratory is that the sample it has chosen and tested
!s representative of the product line. Again, this should be stated
statistically. If a number of supposedly identical products were tested, the
standard error and the range should be reported and an attempt made to
estimate how much of the variance is due to sample differences and how much
to measurement uncertainties.
-------
References for Part II
[l] The International System of Units (Si), NBS Special Publication 330,
19TU Edition.
[2] Some References on Metric Information, NBS Special Publication 389
(1973).
[3] Mechtly, E. A., The International System of Units — Physical Constants
and Conversion Factors (Revised), NASA SP-7.012 (National Aeronautics and
Space Administration, Washington, D. C., 19705.
[k] Lang, W. W., and Flynn, D. R. , A Noise Emission Number for product
designation, Noise Control Engineering^, 108-113 (1975); errata in NCE
i, 3 (1975).
[5] Hubner, G., Analysis of errors in measuring machine noise under
free-field conditions, J. Acoust. Soc. Amer. 5U, 967-977 (1973).
[6] Hubner, G., and Meurers, H., Investigations of sound propagation over
non-ideal reflecting planes and the influence of the measurement
distance on the accuracy of sound power determination of sound sources
operating outdoors, Proceedings of Le Bruit des Machines et
LrEnvironment, Premier Congres Europeen d'Acoustique, F.A.S.E. 75» PP"
559-567 (Paris, 1975).
[7] Holmer, C. I., Procedures for Estimating Sound Power from Measurements
of Sound Pressure: An Experimental Investigation with Application to
Noise from Portable Air Compressors, NBSIR 75-652, EPA-550/8-76-001
(National Bureau of Standards, July, 1975). Available from the National
Technical Information Service, Springfield, 7a. 22151, as Accession No.
COM-75-11399.
[8] American National Standard Methods for the Measurement of Sound Pressure
Levels, SI.13-1971 (American Rational Standards Institute, New York,
1971).
[9] Acoustics — Determination of Sound Power Levels of Noise Sources —
Guidelines for the Use of Basic International Standards and for the
Preparation of Noise Test Codes, Draft International Standard ISO/DIS
37^0 (international Organization for Standardization, Geneva, 1975)*
[10] Acoustics — Determination of Sound Power Levels of Noise Sources —
Precision Methods for Broad-Band Sources in Reverberation Rooms,
International Standard ISO/37^1 (international Organization for
Standardization, Geneva, 1975).
[11] Acoustics — Determination of Sound Power Levels of Noise Sources —
Precision Methods for Discrete-Frequency and Narrow-Band Sources in
Reverberation Rooms, International Standard 130/37^2 (international
Organization for Standardization, Geneva, 1975).
-------
[12] Acoustics — Determination of Sound Power Levels of Noise Sources —
Engineering Methods for Reverberation Rooms, Draft International
Standard ISO/DIS 3743 (international Organization for Standardization,
Geneva, 1975).
[13] Acoustics — Determination of Sound Power Levels of Noise Sources ~
Engineering Methods for Free-Field Conditions over a Reflecting Plane,
Draft International Standard ISO/DIS 37^ ( International Organization
for Standardization, Geneva, 1975).
ISO 3745, Acoustics ~ Determination of Sound Power Levels of Noise
Sources — Precision Methods for Anechoic and Semi-Anechoic Rooms, Draft
International Standard ISO/DIS 3745 (international Organisation for
Standardization, Geneva, 1975).
[15] Acoustics ~ Determination of Sound Power Levels of Noise Sources —
Survey Method, Draft International Standard ISO/DIS 3746 (international
Organization for Standardization, Geneva, 1975).
[16] Delany, M. E., and Bazley, E. N. , Monopole radiation in the presence of
an absorbing plane, J. Sound Vib. 13, 269-279 (.1970).
[17] Van Moorhem, W. K., Reflection of a spherical wave from a plane surface,
J. Sound Vib. 42, 201-208 (1975).
[18] Beranek, L. L., and Sleeper, H. P., The design and construction of
anechoic sound chambers, J. Acoust. Soc. Amer. 18, 140-150 U94bJ.
[19] Meyer, E., Kurtze, G., Severin, E, , and Tanm, K. , Ein neuer grosser
reflexionsfreier Raum fur Schallwellen und kurze elektromagnetische
Wellen, Acustica 3, 410-420 (1953).
[20] Ingerslev, F., Pedersen, 0. J. , Moller, P.K., and Kristensen, J., New
rooms for acoustic measurements at the Danish Technical University,
Acustica !£, 185-199 (1967-68}.
[21] Diestel, H.G., Messung des mittleren Reflexionsfaktors der
Wandauskleidung in einem reflexionsarmen Raum, Acustica 20, y
(1968).
fe2] Delany, M. E., and Bazley, E. N., A note on the sound field due to a
point source inside an absorbent-lined enclosure, J. Sound Vib. 14,
151-157 (1971).
[23] Parkin, P. H., and Stacy, E. F. , The anechoic and reverberant rooms at
the Building Research Station, J. Sound Vib. 1£, 277-286 (1971).
Davern, W. A., A design criterion for sound-absorbent lining of an
anechoic chamber, Appl. Acoust. J., 69-71 (1972).
145
-------
[25] Koidan, W., Hruska, G. R., and Pickett, M. A., Wedge design for National
Bureau of Standards anechoic chamber, J. Acoust. Soc. Amer. 52,
10T1-10T6 (1972).
[26] Bell, E. C., Hulley, L. N., and Mazunder, N. C., The steady-state
evaluation of small anechoic chambers, Appl. Acoust. 6_, 91-109 (1973)-
[27] Standard Method of Test for Impedance and Absorption of Acoustical
Materials "by the Tube Method, C381;-58 (1972), pp 115-126, Part l8, 1975
Annual Book of ASTM Standards (American Society for Testing and
Materials, Philadelphia, 1975).
[28] Beranek, L. L., Acoustics (McGraw-Hill Book Company, New York, 1951*) •
[29] American National Standard Methods for the Determination of Sound Power
Levels of Small Sources in Reverberation Rooms, SI.21-1972 (American
National Standards Institute, New York, 1972).
[30] Morse, P. M., and Bolt, R. H., Sound waves in rooms. Rev. Mod. Phys.
16(2), 69-150 (19UU).
[31] Morse, P. M., Vibration and Sound, Second Edition (McGraw-Hill Book
Company, New York, 19U8).
[32] Morse, P.M., and Ingard, K. U., Theoretical Acoustics (McGraw-Hill Book
Company, New York, 1968).
[33] Beranek, L. L., Noise and Vibration Control (McGraw-Hill Book Company,
New York, 1971).
[3U] Kuttruff, H., Room Acoustics (John Wiley & Sons, New York, 1973).
[35] Crocker, M. J., and Price, A. J., Noise and Noise Control, Vol. I CCBC
Press, Cleveland, 1975).
[36] Magrab, E. B., Environmental Noise Control (John Wiley & Sons, New York,
1915/«
[37] American National Standard Specification for Sound Level Meters,
Sl.U-1971 (American National Standards Institute, New York, 1971).
[38] Standard Values of Atmospheric Absorption as a Function of Temperature
and Humidity, SAE Aerospace Recommended Practice ARP-866A (Society of
Automotive Engineers, Warrendale, Pa., 1975).
[39] Sutherland, L. C., Review of Experimental Data in Support of a Proposed
New Method for Computing Atmospheric Absorption Losses, DQMSM5-87
IU.S. Department of Transportation, Office of Noise Abatement,
Washington, D.C. 20590, May, 1975). «"««« »
1U6
-------
American National Standard Specifications for Laboratory Standard
Microphones, SI. 12-1967 (American National Standards Institute, New
York, 1967}.
American National Standard Method for the Calibration of Microphones,
Sl.10-1966 (American National Standards Institute, New York, 1966).
Recommendations for Sound Level Meters, IEC Publication 123 (Inter-
national Electrotechnical Commission, Geneva, 1961).
Precision Sound Level Meters, IEC Publication 179 (international
Electrotechnical Commission, Geneva, 1965) -
Octave, Half -Octave, and Third-Octave Band Filters Intended for the
Analysis of Sounds and Vibrations, IEC Publication 225 (International
Electrotechnical Commission, Geneva, 1966).
American National Standard Specification for Octave, Half -Octave, and
Third-Octave Band Filter Sets, SI. 11-1966 (R1971) (American National
Standards Institute, New York, 1966).
Magrab, E. B., and Blomquist, D. S., The Measurement of Time-Varying
Phenomena (Wiley-Interscience, New York, 1971) .
Bendat, J. S., and Pier sol, A. G., Random Data: Analysis and
Measurement Procedures ( Wiley-Interscience, New York, 1971).
Donovan, J., and Ketcham, T., Transportation noise — Its measurement
and evaluation, Sound and Vibration £( 10 ), h-8 (October 1973).
Olesen, H. P., and Zaveri, K. , Measurements of averaging times of Level
Recorders Types 2305 and 2307, Bruel and KJaer Technical Review No. 1 -
[50] Leasure, W. A., Jr., and Mathews, D. E. , Pecos Truck Tire Noise Study: A
Summary of Results, NBSIR Jh-kh6 (National Bureau of Standards,
Washington, D. C., 1971*) •
[51] Standard Methods for Testing for Sound Rating Heating, Refrigerating,
and Air-Conditioning Equipment, ASHRAE Std. 36-72 (American Society for
Heating, Refrigerating and Air-Conditioning, New York, 1972).
[52] CAGI-PNEUROP Test Code for the Measurement of Sound from Pneumatic
Equipment (also American National Standard S5.l) (Compressed Air and Gas
Institute, New York, 1969).
t53] Exterior Sound Level for .Heavy Trucks and Buses, SAE J366b (Society of
Automotive Engineers, Warrendale, Pa., 1973).
-------
Skudrzyk, E. , The Foundations of Acoustics: Basic Mathematics and
Basic Acoustics (Springer-Verlag, New York, 1971).
[55] Olson, H. F., System Responsive to the Energy Flov in Sound Waves,
U.S. Patent 1,892,6^ (1932).
[56] Clapp, C. W. , and Firestone, F. A., An acoustic vattmeter, an
instrument for measuring sound energy flow, J. Acoust. Soc . Amer. 13.*
[57] Baker, S., An acoustic intensity meter, J. Acoust. Soc. Amer. 2J_,
269-2T3 (1955).
[58] Schultz, T. J., Acoustic wattmeter, J. Acoust. Soc. Amer. 28, 693-699
(1956). —
[59] Kurze, U. , Zur Entwicklung eines Gerates fur komplexe
Schallfeldmessungen, Acustica 20_, 308-310 (1968).
[60] Jhaveri, A. G., Design and development of a portable acoustic
impedance meter, Annals New York Acad. Sci. 172, 7Vf-756 Cl97l).
[6l] Olson, H. F., Field-type acoustic wattmeter, J. Audio Eng. Soc. 22,
321-328 (197*0.
[62] van Zyl, B. G. , and Anderson, F., Evaluation of the intensity method
of sound power determination, J. Acoust. Soc. Amer. 57, 682-686
(1975). JLL
[63] Schultz, T. J. , Smith, P. W., Jr., and Malme, C. I., Measurement of
acoustic intensity in reactive sound field, J. Acoust. Soc. Amer. Jl»
1263-1268 (1975).
[6U] Rogers, P. H., and Williams, A. 0., Jr., Acoustic field of circular
plane piston in limits of short wavelength or large radius, J. Acoust.
Soc. Amer. £2, 865-870 (1972).
[65] Stenzel, H., and Brosze, 0., Leitfaden zur Berechnung von
Schallvorgangen ( Springer-Verlag, Berlin, 1958). An English
translation of the. 1939 edition of this took is available as Naval
Research Laboratory Translation No. 130 (available from
Photoduplication Service, Library of Congress, as PB 1061571.
[66] Kinsler, L. E., and Frey, A. R., Fundamentals of Acoustics, Second
Edition (John Wiley & Sons, New York, 1962).
[67] Maekawa, Z., Noise reduction by distance from sources of various
shapes, Appl. Acoust. 3, 225-238 (1970).
[68] Tatge, R. B., Noise radiation by plane arrays of incoherent sources,
J. Acoust. Soc. Amer. ^2., 732-736 (1972).
148
-------
[69] Junger, M. C., and Feit, D. , Sound, Structures, and Their Interaction
(MIT Press, Cambridge, Mass., 1972).
t70] Cremer, L. , Heckl, M., and Ungar, E. E. , Structure-Borne Sound:
Structural Vibrations and Sound Radiation at Audio Frequencies
(Springer-Verlag, New York, 19J3) •
[71] American National Standard Method for the Physical Measurement of Sound,
SI. 2-ig62 (American National Standards Institute, New York, 1962).
[72] Peterson, A. P. G. , and Gross, E. E.s Handbook of Boise Measurement, 7th
Edition (General Radio Company, Concord, Mass., 1972).
[73] Baade, P. K. , Sound radiation of air-conditioning equipment; measurement
in the free field above a reflecting plane, Trans. Amer. Soc. Heating,
Refrig., and Air-Condi tioning Engrs. JO, 217-227
Baade, P. K. , Standardization of machinery sound measurement, ASME Paper
Ho. 69-WA/FE-30 (Amer. Soc. Mech. Engrs., Mew York, 1969).
tT5] Ellison, A. J.t Moore, C. J. , and Yang, S. J.t Methods of measurement of
acoustic noise radiated by an electric machine, Proc. Inst. Elec. Engrs.
116, ll*19-ll»30 (1969).
[76] Moore, C/ J., A solution to the problem of measuring the ^ sound field of
a source in the presence of a ground surface, J. Sound Vib. 16, 269-282
(1971).
[77] Acoustic Effects Produced by a Reflecting Plane, Proposed Aerospace
Information Report 1327 (Society of Automotive Engineers, Warrendale,
Pa.). Not yet approved as of this writing.
[76] Miles, J. E., Analysis of Ground Reflection of Jet Noise Obtained with
Various Microphone Arrays over an Asphalt Surface, NASA TM X-71696 -
(National Aeronautics and Space Administration, Levis Research Center,
Cleveland, Ohio, 1975).
Cook, R. K., Waterhouse, R. V., Berendt, R. D., Edel^an, S., and
Thompson, M. C., Jr., Measurement of correlation coefficients in rever-
berant sowid fields, J. Acoust. Soc. Amer. 2£, 1072-1077 U955J-
Lyon, R. H., Direct and reverberant field amplitude distributions in
nearly hard rooms, J. Acoust. Soc. Amer. 33, 1699-170^ U961).
Diestel, H. G. , Probability distribution of sinusoidal sound pressure in
a room, J. Acoust. Soc. Atner. 3i, 2019-2022 (1963).
Andres, H. G., A law describing the random spatial variation of the
sound field in rooms and its application to sound power measurements ,
Acustica 16, 278-29^ (1965).
149
-------
[83] Waterhouse, R. V., Statistical properties of reverberant sound fields,
J. Acoust. Soc. Amer. ^3_» 1U36-1UUU (1968).
[8U] Lubman, D. , Fluctuations of sound with position in a reverberant room,
J. Acoust. Soc. Amer. liU. 1^91-1502 (1968).
[85] Schroeder, M. R., The effect of frequency and space averaging on the
transmission responses of multimode media, J. Acoust. Soc. Amer. jt6,
2T7-283 (1969).
[86] Lubman, D., Spatial averaging in a diffuse sound field, J. Acoust. Soc
Amer. U6., 532-53^ (1969).
[87] Schroeder, M. R., Spatial averaging in a diffuse sound field and the
equivalent number of independent measurements, J. Acoust. Soc. Amer.
M» 53U (1969).
[88] Lyon, R. H., Statistical analysis of power injection and response in
structures and rooms, J. Acoust. Soc. Amer. 1+5, 51*5-565 (.1969) .
[89] Tichy, J., The sound field in a reverberation chamber at low
frequencies, J. Acoust. Soc. Amer. Hj_t 117(A) Cl970) .
[90] Waterhouse, R., Sampling statistics for an acoustic mode, J. Acoust.
Soc. Amer. Hj, 961-967 (1970).
[9ll Waterhouse, R, V., and Lubman, D. , Discrete versus continuous space
averaging in a reverberant sound field, J. Acoust. Soc. Amer. US, 1-5
(1970). —
[92] Morrow, C. T., Point-to-point correlation of sound pressures in
reverberation chambers, J. Sound Vib. 16, 29-U2 (1971).
[93] Lubman, D., Spatial averaging in sound power measurements, J. Sound
16, U3-58 (1971).
in a
[95] Tichy, J The measurement of sound power in a reverberation chamber at
kP7C^TrT!nfeSs PrOC' Inter-Koise 72 (M. J. Crocker, Ed.), pp
1972) (lnStltute °f Noise Contro1 Engineering, Poughkeepsie, i.Y.,
[96] Lubman, D. , Waterhouse, R. V., and Chien, C., Effectiveness of
[97] Haling, G. C., Jr., Guidelines for determination of the average sound
power radiated by discrete-frequency sources in a reverberation room, J
Acoust. Soc. Amer. 53 106U-1
-cy sour
Acoust. Soc. Amer. .53_, 106U-1069 (1973).
150
-------
•ion
L98] Ebbing, C. E., and Maling, G. C., Jr., Reverberation-room qual if i cat ~«
for determination of sound pover of sources of discrete-frequency sound,
J. Acoust. Soc. Amer. 2±, 935-9^9 (1973).
[99] Tichy, J., and Baade, P. K. , Effect of rotating diffusers and sampling
techniques on sound-pressure averaging in reverberation rooms, J.
Acoust. Soc. Amer. ^6, 137-1^3 (197*0.
[100] Lubman, D. , Precision of reverberant sound power measurements, J.
Acoust. Soc. Amer. £&, 523-533 (197*0.
[101] Lubman, D. , and Associates, Review of Reverberant Sound Power Measurement
Standard and Recommendations for Further Research, NBS Tech. Note 8^1
(National Bureau of Standards, Washington, D. C., 197*0-
[102] Dodd, G., Steady state room transmission response parameters — their
distributions and relationships, and implications of these
relationships, J. Sound Vib . 36 , bj3-k83 (197*0.
[103] Chien, C., Waterhouse, R., and Lubman, D., Spherical averaging in a
diffuse sound field, J. Acoust. Soc. Amer. 57 » 972-975 (1975).
[10*0 Waterhouse, R. V., Power output of a point source exciting a single
Cartesian mode, J. Acoust. Soc. Amer. ]f£, 9-l6 (1971).
[105] Dodd, S. D., and Doak, P.E., Some aspects of the theory of diffusion and
diffusers, J. Sound Vib. 16, 89-98 (1971 ).
[106] Ebbing, C. E., Experimental evaluation of moving sound diffusers for
reverberant rooms, J. Sound Vib. 16, 99-118 C19.71).
[107] Lubman, D. , Distribution of reverberant sound in large rooms, J. Acoust.
Soc. Amer. 3£, 1266 (A) (1966).
[108] Tichy, J., The effect of rotating vanes on the sound field in rever-
beration chambers, J. Acoust. Soc. Amer. _*£, 82U) U-97D.
Tichy, J., The effect of boundary conditions on the statistical
properties of the sound field in enclosures, J. Acoust. Soc. Amer. £0,
98(A) (1971).
Lubman, D., Guidelines for the design of rotating diffusers, J. Acoust.
Soc. Amer. £2, U1*7-U*8(A) (197*0.
Peterson, A. P. G. , Device noise and background noise from a statistical
point of view, Noise Control Engineering J£, 76-83 (1975 J-
[H2] Diehl, G. M., Machinery Acoustics (John Wiley & Sons, New York, 1973).
[H3] Diehl, G. M., Sound power measurements on large machinery installed
indoors, Compressed Air Magazine I£, 8-12 (January 1974).
151
-------
Leasure, W. A., Jr., and Quindry, T. L., Methodology and Supporting
Documentation for the Measurement of Noise from Medium and Heavy Trucks,
NBSIR 7^-517 (National Bureau of Standards, Washington, D. C., 197M •
[115] Bettis, R. A., and Sexton, M. Z., The effect of test site topography in
vehicle noise measurement, J. Acoust. Soc. Amer. 5U, 332(A) (1973).
[116] Sheth, N. J., and Gegesky, P. S., The influence of test site on exterior
vehicle noise measurements, SAE Paper Wo. 7^0967, presented at
Automobile Engineering Meeting, Toronto, Canada (Society of Automotive
Engineers, Warrendale, Pa., 197*0-
[117] Flynn, D. R., Accuracy and precision (an editorial), Noise Control
Engineering 3., 2 (197*0.
152
-------
PART III SELECTION OF MEASUREMENT METHODOLOGY
APPROPRIATE TO A SPECIFIC PRODUCT
This part of the report consists of a series of flow charts which
depict the development of appropriate procedures for measuring the
noise emission of particular classes of products. These charts are
intended to serve as reminders and check lists of the factors which
are discussed in detail in Parts I and II.
Figure 35 shows the overall development of appropriate test
procedures. The 'five boxes shown with heavy borders are expanded
in Figures 36-UO.
153
-------
APPROPRIATE
CLASSIFICATION
SCHEME
OPERATIONAL MODES
INDICATIVE OF
PRODUCT USAGE
I
QUANTITY TO BE
MEASURED
APPROPRIATE
INSTALLATION
AND OPERATION
OF SOURCE
APPROPRIATE
MEASUREMENT
LOCATIONS
CONSENSUS
STANDARDS
AVAILABILITY
OF
EQUIPMENT
AVAILABILITY
OF
FACILITIES
PHYSICAL
MEASUREMENT
CONSIDERATIONS
CONSTRAINTS
c
NOISE
EFFECTS
DATA
APPROPRIATE
TEST
ENVIRONMENT
CRITERIA
APPROPRIATE
INSTRUMENTATION
SPECIFICATIONS
ACCURACY AND
PRECISION
REQUIREMENTS
AVAILABILITY
OF TRAINED
MANPOWER
TOTAL
ECONOMIC
BURDEN
APPROPRIATE TEST PROCEDURE
Figure 35- Development of appropriate test procedure.
-------
DOES A CLASSIFICATION
SYSTEM EXIST
FOR THE PRODUCT
IN QUESTION?
NO
I
YES
DOES EXISTING SYSTEM LEND
ITSELF TO ACCEPTANCE AS A
CATEGORIZATION WITH NOISE
AS THE ITEM OF CONCERN?
NO
DEVELOP A CLASSIFICATION
SCHEME APPROPRIATE
FOR PRODUCT NOISE EMISSION
I
YES
DEFINE THE LEVEL OF DETAIL
THAT MUST BE REACHED BEFORE
A MEASUREMENT PROCEDURE
CAN BE UNIVERSALLY APPLIED
i
APPROPRIATE
CLASSIFICATION
SCHEME
Figure 36. Development of appropriate classification scheme.
155
-------
APPROPRIATE
CLASSIFICATION
SCHEME
CONTROL OF
PRODUCT
OPERATION
(E.G., OPERATOR
OR AUTOMATIC)
1
i
SOURCE
MOBILITY
(STATIONARY
OR MOBILE)
NOISE
EFFECTS
DATA
1
DO USAGE DATA EXIST WHICH ESTABLISH THE
"NORMAL OPERATION" OF THE PRODUCT?
I
DEVELOP THE NECESSARY
DATA BASE THROUGH
THE CONDUCT OF A
USAGE SURVEY
YES
OBTAIN A DATA BASE THAT CORRELATES THE PRODUCT NOISE
OVER THE RANGE OF OPERATIONAL MODES, WITH THE EXPOSURE
OPERATIONAL MODES
INDICATIVE OF
PRODUCT USAGE
Figure 37. Identification of operational modes indicative of product
usage.
156
-------
WHO IS AFFECTED?
OPERATOR
PASSENGER
BYSTANDER
NEIGHBOR
WHAT ARE THE EFFECTS?
HEARING DAMAGE
COMMUNICATION
INTERFERENCE
ANNOYANCE
COMPLAINTS
EXTENT OF
NOISE EXPOSURE
NOISE
EFFECTS
DATA
'Figure 38. Collection of noise effects data.
157
-------
c
QUANTITY TO BE MEASURED
WHO IS
AFFECTED?
*
FREQUENCY
RANGE OF
INTEREST
*
t
SIZE OF
SOURCE
*
MOBILITY
OF SOURCE
+
IS TOTAL NOISE EMISSION DESIRED?
i
YES
IS SOURCE
DIRECTIONAL?
i
NO
NOISE IN PARTICULAR
DIRECTION OR LOCATION
APPROPRIATE TEST
ENVIRONMENT
CRITERIA
i
NUMBER AND DEPLOYMENT OF
MICROPHONE LOCATIONS
SOURCE-TO-MICROPHONE DISTANCES(S)
MICROPHONE HEIGHT(S)
i
APPROPRIATE
MEASUREMENT
LOCATIONS
Figure 39. Selection of appropriate measurement locations,
158
-------
OPERATIONAL MODES
INDICATIVE OF
PRODUCT USAGE
C
i
QUANTITY TO
BE MEASURED
PRODUCT
CHARACTERISTICS
(PHYSICAL SIZE,
MASS AND ENERGY
INPUT/OUTPUT)
,
TYPICAL USE
LOCATION
(OUT-OF-DOORS,
INDOORS)
1
i
WOULD DATA TAKEN INDOORS
ALLOW FOR PREDICTION OF
NOISE LEVELS IN ACTUAL USE?
i
YES
CAN TESTS BE PERFORMED INDOORS?
YES
i
NO
NO
LABORATORY
IN-SITU
OUTDOOR SITE
FACILITY
REQUIRE-
MENTS
EFFECT OF
INSTALLA-
TION
SITE
REQUIRE-
MENTS
EFFECT OF
WEATHER
i
APPROPRIATE TEST ENVIRONMENT CRITERIA
Figure
Selection of appropriate test environment criteria.
159
-------
Appendix A.
List of Participants at Government/Industry Meetings on Noise
Measurement Methodology for the Environmental Protection
Agency's Noise Emission Regulations
Meet''ng
Representatives of
the Assn. of Home
Appliance Manufacturers
(AHAM)
Representatives of
the American Road
Builders Assn. (AREA),
the Construction
Industry Manufacturers
Assn. (CIMA), and
contractors
Date
July 7, 1972
Representative of
the General Services
Administration (GSA),
Public Building
Services
Representative of
the Acoustical and
Insulating Materials
Assn. (AIMA)
Representatives of
the U, S. Postal
Service (P, 0.)
July 20, 1972
November 13, 1972
November 29, 1972
December U, 1972
P ar t icop ant s
D. Flynn (NBS)
W. Leasure, Jr. (KBS)
R. Musa (Westinghouse)
H. Phillips (AHAM)
J. Wfdzeorick (AHAM)
J. Benson (CIMA)
J. Codell III (Codell
Construction Co.)
R. Crowe (ARBA)
G. Diehl (ingersoll-
Rand)
D. Flynn (NBS)
W. Land (Contractor)
H. Larmore (CIMA)
W. Leasure, Jr. ( NBS)
B. Miller (AREA)
J. Oman (Oman
Construction Co.)
D. Powlson (Tennessee
Road Builders)
W. Leasure, Jr. (NBS)
R, Rice (GSA)
A. Rubin (NBS)
D. Flynn (NBS)
R. LaCosse ( AIMA)
W. Leasure, Jr. (NBS)
Cornog (P. 0.)
Flohr (P. 0.)
Hull (P. 0.)
Leasure, Jr. (NBS)
D.
R.
¥.
¥.
A. Rubin (NBS)
160
-------
Representatives of
the Dept. of Trans-
portation (DOT),
locomotive manufac-
turers, the Assn. of
American Railroads
(AAR), and the
American Short Line
Railroad Assn. (ASLRA)
December 5, 1972
Representatives of
the Air Conditioning
and Refrigeration
Institute (ARI)
Representatives of
the National Elec-
trical Manufacturers
Assn. (NEMA)
Representatives of
the International
Snowmobile Industry
Assn. (ISIA) and the
Society of Automotive
Engineers Motorized
Snow Vehicle Sub-
committee
December 6, 1972
December 7, 1972
December 7, 1972
P. Baker (General
Electric)
D. Bray (DOT/FRA)
W. Close (DOT/ONA)
J. Coxey (AAR)
H. Croft (ASLRA)
D. Flynn (NBS)
K. Hawthorne (AAR)
W. Leasure, Jr. (NBS)
R. Lucas (DOT/FRA)
R. Pribramsky (General
Motors)
A. Rubin (NBS)
J. Wesler (DOT/ONA)
W. Bayless (Borg-
Warner)
R. Kelto (Air Temp.)
W. Leasure, Jr. (MBS)
A. Meling (ARI)
A. Rubin (NBS)
J. Schreiner (Carrier)
W. Leasure, Jr. (NBS)
H. Michener (NEMA)
R. Nims (NEMA)
R. O'Brien (NEMA)
A. Rubin (NBS)
J. Werner (NEMA)
J. Arbuckle (ISIA)
R. Croteau (Bombadier
Ltd.)
J. Giesen (Deere &
Co.)
G. Gowing (Bombadier
Ltd.)
L. Haas (Scorpion,
Inc.)
W. Leasure, Jr. (NBS)
J. Nesbitt (ISIA)
A. Rubin (NBS)
J. Spechko (Outboard
Marine Corporation)
161
-------
Representatives of
the Engine Manu-
facturers Assn.
(EMA) and the Society
of Automotive Engineers
(SAE)
December 8, 1972
Representatives of
the American Trucking
Assns. ( ATA)
December 12, 1972
Representatives of
the Motor Vehicle
Manufacturers Assn.
(MVMA)
December 20, 1972
R. Canfield (American
Motors)
J. Crowley (Case)
W. Hamilton (Allis
Chalmers)
T. Hutton (EMA)
J. Jensen (John Deere)
J. Johnson (Caterpillar)
D. Kabele (SAE)
R. law (Detroit
Diesel Allison)
W. Leasure, Jr. (NBS)
C. Leber (Caterpillar)
R. Lincoln (Outboard
Marine)
J. McWally (Caterpillar)
J. Nadolny (Teledyne
Wisconsin)
A. Rubin (NBS)
C. Salter (EMA)
R. Staadt (inter-
national Harvester)
D. Stephenson (Out-
board Marine)
T. Wu (international
Harvester)
T. Young (EMA)
W. Gibson (ATA)
L. Kibbee (ATA)
W. Leasure, Jr. (UBS)
A. Rubin (UBS)
R. Tilley (Safeway
Stores)
L. Bridenstine (MVMA)
J. Damian (Ford)
T. Dolan (General
Motors)
F. Kishline (American
Motors)
W. Leasure, Jr. (NBS)
R. Ratering (General
Motors)
A. .Rubin (NBS)
R. Staadt (inter-
national Harvester)
R. Vievig (MVMA)
R. Wasko (MVMA)
162
-------
Appendix B.
Pertinent Sections of the Noise Control Act of 1972
For the "benefit of those readers not familiar with the Noise Control
Act of 1972, Public Lav 92-5?U, the following summary of pertinent sections
of the Act is presented.
NOISE EMISSION STANDARDS FOR PRODUCTS DISTRIBUTED IN COMMERCE
Sec. 6. EPA is given authority to prescribe and amend standards limit-
ing noise generation characteristics for any product or class
of products which has been identified as a major source of
noise and which falls in the following categories: construc-
tion equipment, transportation equipment (including recreational
vehicles), any motor or engine, and electrical or electronic
equipment. EPA may issue regulations for products in other
categories if it is necessary to protect the public health or
welfare. The standards must be " based on criteria pub-
lished under Section 5»" and "requisite to protect the public
health and welfare, taking into account the magnitude and con-
ditions of use of such product (alone or in combination with
other noise sources), the degree of noise reduction achievable
through application of the best available technology, and
the cost of compliance.11
The manufacturer of regulated products must warrant that its
product is designed and built so as to conform at the time of
sale with such regulation. The cost of this warranty cannot
be passed on by the manufacturer. States and political sub-
divisions are prohibited from setting noise emission levels
different from those promulgated by EPA, but remain able to
regulate use, operation or movement of products.
LABELING
Sec. 8. For any product which (a) emits noise capable of adversely
affecting the public health or welfare, or (b) is sold wholly
or in part on the basis of its effectiveness in reducing
noise, the EPA must require the manufacturer of such product
to give notice of the noise level or its effectiveness in re-
ducing noise to the consumer. EPA's regulations must indicate
the form of such notice and the method and unit of measurement
must be prescribed.
RAILROAD NOISE EMISSION STANDARDS
Sec. 17 After consultation with the Department of Transportation,
EPA is required to promulgate regulations for surface car-
riers engaged in interstate commerce, including regulations
163
-------
governing noise emission from the operation of equipment
and facilities of such carriers. The effective date for such
regulations must permit the development and application of
the requisite technology. The Secretary of Transportation
is charged with the responsibility of assuring compliance
with EPA's regulations. State and local governments are pro-
hibited from establishing operational noise emission limits
different from applicable federal standards, "but the Adminis-
trator may allow a different standard if he determines in con-
sultation with the Secretary of Transportation that local con-
ditions necessitate such different regulations.
MOTOR CARRIER NOISE EMISSION STANDARDS
Sec. 18 The provisions of this section are nearly identical to Sec. 16
except that they apply to "a common carrier by motor vehicle,
a contract carrier by motor vehicle, and a private carrier of
property by motor vehicle as those terms are defined in the
Interstate Commerce Act (H9 U.S.C. 303(a))."
DEVELOPMENT OF LOW-NOISE EMISSION PRODUCTS
Sec. 15 Provides for Federal procurement of and public notice about
products certified as "low-noise-emission products" (defined
as: any product which emits noise in amounts significantly
below the levels specified in noise emission standards under
regulations applicable under Sec. 6 at the time of procure-
ment to that type of product). The Administrator is allowed
to establish a Low-Noise-Emission Product Advisory Committee
to assist him in determining which products qualify. Once an
application for certification is received and the product is
determined to be a low-noise-emission product, the Administrator
must certify the product as such if he determines that the pro-
duct is a suitable substitute for a type of product at that
time in use by agencies of the Federal government. Various
instructions as to when the Federal government is required to
purchase such products and when the EPA is required to pub-
lish information about its determinations are given."
In setting noise emission standards for products distributed in commerce,
the Environmental Protection Agency is required (see Sec. 6 of Public Law 92-
51 L A ff standards on criteria published under Sec. 5. Section 5
of the Act is summarized below:
IDENTIFICATION OF MAJOR NOISE SOURCES, NOISE CRITERIA
AND CONTROL TECHNOLOGY
Sec. 5 (1) requires EPA to publish criteria which reflect the kind
or w!ff^ a^ldeflfiatle eff€CtS °n the P*>lic hea"*
or welfare resulting from differing quantities and qualities
164
-------
of noise (within 9 months); (2) requires EPA to publish infor-
mation on levels of environmental noise which in defined areas
under various conditions are requisite to protect the public
health and welfare with an adequate margin of safety (within
12 months); (3) requires EPA to publish a report identifying
major sources of noise, and giving information on techniques
for control of noise (within 18 months).
165
-------
Appendix C.
Uncertainty of Measurement
If quantitative noise regulations are to be effective and equitable, it
is essential to have a good understanding of the uncertainties extant in the
associated measurements. In legal proceedings this uncertainty represents the
"shadow of doubt" associated with a measurement. A clear statement of the
factors entering into the uncertainty computation and the data on vhich it is
based should be available for "cross-examination."
In a legal proceeding \.and in regulatory operations generally) a
statement as to what measurement process would be accepted as correct is a
necessity — otherwise regulation would not be possible. The uncertainty of a
measurement should therefore be stated in terms of the results which would have
been obtained by the "accepted" process. In the case of some standards,
reference to the unit as maintained by the National Bureau of Standards may be
appropriate. In other cases, results by a selected agency or by an average of
the measurement processes of several organizations may serve as a reference.
Two characteristics of measurements must be accounted for. First, that
successive measurements of the same quantity will.disagree, and second, that
the long-run average by two difference realizations of the same method of
measurement will differ. One's model of a measurement process must therefore
be enlarged to include both the variability and possible offset of the process
from that which would be accepted as correct.
It is in the concept of a repetition of a measurement that the
uncertainty of measurement can be given operational meaning. Measurements can
be regarded as arising from a process whose properties can be determined from
an appropriate sequence of such repetitions. It is only when one can attribute
the properties of the process to the isolated single measurement that a
defendable statement of its uncertainty can be made. To be able to do this the
process must have predictability, i.e., be in a state of statistical control at
the time of the measurement so that the use of the current values of the
process parameters is valid.
Limits for the Effects of Random Error
The crucial step in assessing the effects of random error is that of
defining the set of repetitions over which the measurement is to apply. At the
very minimum it would involve repetitions with the same instrument-operator-
procedure configuration. It would include sampling variability when that is
appropriate and include a number of components of variance such as those
associated with day-to-day differences, operators, instruments, etc.
All of these factors which enter into the random error calculation could,
in principle, be varied in repeated measurement so that their effects could be
mitigated by averaging over the set of repetitions. Those which cannot be so
averaged out are regarded as systematic errors which account for the offset of
166
-------
one process relative to another.
To assess the possible extent of random error of the quantity, y, where y
is a specified function of random variables x,, x . , . x and constants c.,
Cp . . . so that y = f (x^ Xg . . . x , c.^, Cg . . .), let us assume that a
standard deviation, s , is available,, This standard deviation will, in
general, involve many^components of variance, the constants, c.., and will
depend on the functional form f(). To go from this standard deviation to a
bound for the random error (e.g., three standard deviations) involves some
arbitrariness both in the assumed nature of the probability distribution
involved and the desired degree of coverage.
Offset of Measurement Process
The offset of one measurement process relative to another tor of a
process from that accepted nationally) may arise from deficiencies in the
mathematical model or in realizations of the specified process. To this one
must add systematic error in the prescribed standards, and the fact that the
corrections for environmental and other effects may not account for all the
effects of such variables on the particular measurement process.
The procedures for arriving at bounds to the possible offset of the
process will involve direct measurement by introducing changes into the process
and observing the effect, the use of values from critical experiments run on
similar processes, and other similar techniques which have in common the fact
that they are based on observation (not judgement).
In some cases it is possible to determine the offset of a measurement
process by measuring a "control" in the form of some reference meter or
standard signal. It is necessary that the reference items be similar in all
important respects to the items being measured, and that it be measured by the
same procedures used in the regular workload. When the bias correction is
made, the process would be regarded as being free from systematic errors from
the source represented by the item. The random error in'the applied correction
becomes part of the random error of the process, of course.
Control of Measurement
At some point in time one will have values for the bound, R, to the
effect of random error and a bound, E, for the possible offset of the process
relative to nationally accepted reference standards or measurement processes.
These values are used to characterize the process and these properties can be
applied to individual measurements from the process if the process remains in a
state of control.
Some evidence is therefore needed to establish that the process is "in
control." Measurements in a reference item made periodically throughout the
year are an example of the type of redundancy needed to provide the assurance
of consistency. When coupled with an independent outside check on the offset
of the process, one has evidence of the validity of the uncertainty statement.
167
-------
An out of control condition signifies that predictability has been lost
and one should therefore redetennine the process parameters to arrive at a nev
uncertainty statement.
Uncertainty
The uncertainty of a reported value could, in principle be reduced to +E,
the offset or systematic error of the process, by increasing the number of
measurements to be made. Any such increase in the number of measurements would
not change the process average, of course. If one or only a few measurements
are made, then the uncertainty is increased by the random errors so that the
uncertainty* of a measurement from a process in control is
uncertainty = +.(E + R)
*The modification for asymmetrical limits are obvious.
168
-------
Appendix D.
Possible GATT Code of Conduct for Preventing Technical
Barriers for Trade
The United States and other major trading nations through the General
Agreement on Tariffs and Trade5 have been investigating the international
trade problems arising from the development of different regional, national,
and local product standards and technical regulations and the various pro-
cedures for assuring conformity with them[l]. Most of these standards
barriers are side effects of efforts to protect the public welfare rather than
deliberate discrimination, and the GATT negotiators are trying to find a
solution that will not hamper the achievement of these objectives[2J.
The following discussion on the GATT code of conduct is reprinted from
the ASTM Standardisation Hevs with permission of the American Society for
Testing and Materials[l].
"It was the recognition of the problems presented by standards and
the inadequacy of the GATT provisions for dealing with them that led the
contracting parties to begin the development of a standards code.
Furthermore, the United States and other countries were becoming
increasingly concerned about European plans to conclude regional
standards arrangements on an exclusive basis. In 1971 Working Group III
of the Committee on Trade in Industrial Products began the drafting of
such a code. The working group agreed on certain hypotheses of which
two should be noted. One was that the solution developed should take
the form of a binding code subject to a reservation that the final
product may be changed to a voluntary code or set of principles.
Another was that the GATT should not become involved in writing
standards or certifying that products conform to standards.
"The draft GATT standards code which was developed deals separately
with the preparation, adoption, and use of mandatory standards by
central and local government bodies; the preparation of voluntary
standards by central and local government bodies; and the preparation of
standards by voluntary standards bodies. In each case standards are to
be formulated and applied so as not to afford protection to domestic
production, and they are to be based upon 'appropriate1 international
standards. There shall be active participation in international
standards organizations. Wherever appropriate, standards are to be
specified in terms of performance rather than design. Proposed
standards not based on international standards must be published,
consideration given to comments received, and a reasonable time allowed
for foreign suppliers to adapt to the standards (except when urgent
problems of health, safety, environmental protection, or national
security exist). All mandatory standards must also be published. In
addition, adherents of regional standards organizations shall use their
"best efforts1' to ensure that the standards will expedite progress
toward the preparation of international standards and that the regional
169
-------
organizations comply with the requirements Just described.
"Compliance by local government "bodies is subject to 'reasonable
means' being used to ensure compliance with these requirements.
Similarly, standards of voluntary bodies (such as ANSI) are conditioned
"by a 'best efforts1 stipulation.
"In determining conformity with mandatory standards, central
governments must likewise employ methods that do not afford protection
to domestic production and test methods must be harmonised, as far as
practicable » on an international basis and be published. Procedures
should permit tests to be carried out in the exporting country and
should recognize other equivalent test methods,
"Khere a positive assurance is required that imported products
conform to a mandatory standard, whenever possible a declaration by the
supplier or by a quality assurance body in another member country should
be accepted. If tests are carried out in the importing country they
should be on the same basis as tests of domestic products.
"Where quality assurance systems are operated or relied on by the
&SSUre conf «"»«*• 1th its mandatory standards,
r Safeguards as to ^irneSS a*d nondiscrimi nation. Here
major difference of viev exists, particularly when the quality
SC " internati°nal « regional in nature. ^The u! S. view
ass^ance ^sterns should be open to
to mi a in the initial stages of the system.
™ r
to the code, especially the developing countries.
' ^Xpressed reservations about the retroactive
-oorunry
afford urotectloT, t^ J««, ™-LU^lTary — or quality assurance systems
industrles» then signatories shall bring
s wouid have different
changes required. Tor
vould
regulation. niM tM
other sections of the code. O" WlU sralt «»Pl=tion of
to Trade'
n
code, to investigate ^i2 to C°MDlt °D ^^^n^^ °f
««ipj.aints, and to recommend actions after
170
-------
completion of an investigation. It is also proposed that the committee
be empowered to authorize suspension of code obligations for violators
and refer serious violations to the GATT contracting parties for
appropriate action. This question of sanctions is controversial and
some countries favor other, less severe enforcement provisions.
"Some of the drafters also have reservations about the binding
nature of the code, particularly over the difference in how it would
affect the obligation of a federal as opposed to a unitary form of
government. In the former many mandatory standards are developed by
state and local bodies which the national government in some cases may
not effectively control . ' In the latter all mandatory standards are
developed by the national government. A related problem is that some
countries maintain only mandatory standards while in others there are
many voluntary and so-called 'quasi-voluntary1 ones. (Quasi -voluntary
standards are those that are developed in some European countries by a
cooperative effort between industry and government, but are ^ voluntary
with the proviso that they can be made mandatory at the option of the
government.) Thus, those countries that rely mainly on mandatory
standards would assume greater obligations under the code than those
that rely on voluntary or quasi-voluntary standards. The draft code
will address these problems by placing different levels of obligations
on the signatories.
"Obviously the effects a binding international code of standards
would have cannot be fully measured until final agreement is reached o
specific provisions; however, a few conclusions can be ventured with a
fair amount of confidence. In the first place the role of internation
standardization and the organizations which perform it will become muc,
more important. From that fact will flow greatly increased
responsibilities on the American National Standards Instxtute (MSI) a,
its activities, since ANSI represents the U. S. interest in
international standardization work by the nontreaty organizations such
as the ISO and IEC (the International Electrotechnical Commission),
which in fact perform much of the international standards work. ANSI
will therefore need the full support of the whole spectrum of American
industry if it is to have the strength and resources to fulfill its
task.
"American industry should benefit from the code. The code will
encourage open certification systems in which our industry can
participate if it is willing to assume the resf>nsf ^^ °*Ln
membership. Thus the certification arrangements of CM (European
Standards Coordinating Committee) and CMELEC (European Committee for
the Coordination of Electrotechnical Standards) would have to be open
In < *
instead of restricted to Europeans as they are now.
code would place even greater pressure on the E^°*ea^°™t^
use international norms in developing their regional standards and
certification systems.
"Pinallv the code may necessitate changes in current U. S.
legislated ^inXtraUve practices of regulatory bodies. This
aspect is complex and is being carefully explored.
171
-------
References for Appendix D
[l] Travaglini, V., Removing "barriers to trade through GATT, ASTM
Standardization News J3(ll) (November 1975).
[2] Travaglini reports on GATT code negotiations, ANSI Reporter (July T>
1975).
172
-------
Appendix E.
Methods of Labeling
In the development of appropriate procedures for labeling products as to
their noise emission, it is useful to distinguish among the following four
concepts:
o Determination of the noise emission of machinery or equipment
involves the actual measurements "which are carried out according to
a specified test procedure.
o Designation of the noise emission is in terms of a particular
quantity (such as A-weighted sound power level or perceived noise
level) expressed in a particular unit (such as decibel).
Designation frequently involves combining the measured data in a
prescribed manner in order to obtain some single-figure rating,
usually one that can be expected to correlate with human response.
o Classification involves systematic arrangement of designated noise
emission values into groups or categories according to established
criteria. Classification typically implies subdivision of the range
of noise emission values into a series of intervals which, are large
enough that the difference between adjacent classes is
"significant", in some sense. These intervals may be of fixed,
equal size (e.g., 5 dE) or maybe derived in a statistical sense
(e.g., quartiles). Classification frequently, but not always,
implies some form of coding in which a system of words, numbers,
letters, or other symbols is used to represent the several classes.
The coding system can either be "absolute", so that a given noise
emission value would always be assigned the same coding symbol, or
"relative", so that the coding symbol conveys information concerning
the relationship between the noise emission value of a particular
machine or model and the distribution of noise emission values for
the family of similar equipment. The coding may also convey
information about the particular family to which the machinery
belongs, the year to which the classification corresponds, or other
information necessary to minimize any possible confusion.
o Labeling means to furnish or affix written or printed matter in
order to furnish the purchaser or user with, information concerning
the noise emission of the product. This information may include the
designation of the actual noise emission and/or the classification
of the noise emission.
Voluntary standards organizations have been and are addressing the
questions of how to designate and/or classify the noise emission of products.
A recent national standard[l] designates the noise emission of small
stationary sources in terms of the Product Noise Rating (PNR), which is
obtained from the A-weighted sound power level, adjusted by a constant such
173
-------
that the rating is approximately equal to the space average of the A-weighted
sound pressure level at the specified rating distance (usually 1 meter).
Radiation from a sound source on a reflecting surface is assumed.
At the international level there currently is interest in expressing
noise emission directly in terms of the A-weighted sound power level, with the
unit being the "bel rather than the decibel so as to reduce confusion "between
sound pressure level and sound power level[2]. Such a document has just been
approved at the national level[3].
A group under the jurisdiction of ISO/TC&3 (Acoustics)/SCI (.Noise) is
currently considering a possible international standard on classification and
labeling of equipment and machinery as to noise emission. In the current
draft of this proposal[U], noise emission would be designated in terms of the
A-weighted sound power level. An absolute classification scheme is proposed
in which the sound power level is rounded up to the nearest 5 dB. The
proposed relative classification scheme also uses a 5 dB class interval: Class
1 would include machinery which is "representative of the guaranteed noise
emission number to be expected from well designed machines of good workmanship
belonging to the family* in question;1' Class 2 would include machines for
which the guaranteed noise emission is 5 dB less than in Class 1; Class 3
machines would be at least 10 dB quieter than those in Class 1, etc. It has
been suggested that the Class number be identified by showing the appropriate
number of stars.
In considering classification or labeling schemes for noise emission it
is important to consider who is the intended user or audience. A label, or
other documentation, which conveys the noise emission designation and/or
classification should enable fair comparisons among competing products and
should assist purchasers or users in assessing the noise impact of the
product, The user, however, may be, for example, (l) an engineer who needs
to determine the effect of the product on compliance with hearing conservation
regulations, (2) a construction superintendent who has to comply with zoning
regulations, or (3) a private citizen who simply wants a "quiet" appliance for
his home. Both the information needs and the technical sophistication of
these users will vary widely and the information conveyed must be adequate for
the more sophisticated user while at the same time must not be confusing to
persons with less technical backgrounds.
The acoustical consulting firm of Bolt Beranek and Newman has just
completed, for the U. S. Environmental Protection Agency, a study of problems
and considerations involved in labeling products as to their noise emission or
control characteristics. Thus in this brief appendix there is no need to
delve very deeply into the many questions involved with establishing a viable
noise labeling program. However, there are a few key points that can
beneficially be mentioned.
*Mote: A family of machines or equipment is a group... of a similar design or
type or meeting the same performance requirements for which it is reasonable
to establish a noise class."
-------
Labeling data could "be provided in two ways. The label itself might be
in the form of detailed acoustic data, e.g., octave-band sound power levels
vhich the user could utilize in placement of the machines within the
workplace. Or there could be a "cookbook" application guide which would be
used in conjunction with the simple general public label which would allow the
purchaser to predict what sound levels could be expected at some distance from
the noise source, e.g., at the property line. There are several possible
labeling schemes which deserve consideration. They include single number
ratings, single letter ratings, color codes, actual sound level values or sets
of data.
A single number rating has, as its main value, simplicity. However,
public education would be necessary to assure acceptance and thus success of a
single number labeling scheme. Assume that products can be labeled on a
single number scale ranging from 1-10. The initial problem is that the
difference between adjacent numbers (e.g., difference between a 3 and a k)
should be acoustically significant. In other words, the public would have to
be able to discern the product improvement from one number to another without
the assistance of sophisticated instrumentation. In acoustics, it is
generally accepted that approximately a 3 dB difference must exist before the
human ear consistently can detect any difference. This question of the
acoustic level change between adjacent numbers, letters, colors, etc., is an
inherent problem with any labeling system.
A second problem is determining what meaning the number system already
has with the public. One might naively expect that a numbering system would
work perfectly with 1 the quietest and 10 the noisiest. However, one could
think of 10 as being larger than 1; therefore, a product rated 10 would
provide more of something than a product rated 1. The more in this case could
either be more noise or more noise abatement. In addition, in the educational
system utilizing a numerical grading system, 100 is a perfect score.
The single letter rating, like the single number, has the advantage of
being simple. Moreover, the problem of public education might be somewhat
easier, because most people in the United States have been exposed in varying
degrees to the formal education process that traditionally utilizes an A, B, C
grading system. In all cases A is the best. Also people have been further
familiarized with this grading system since it is now applied to the grading
of meat and dairy products. Because of this, it seems that if the best
product — in this case the quietest — were labeled A and Z the noisiest,
consumers would not find this rating scheme to be inconsistent with other
grading systems with which they are familiar.
Use of a color code is another easy method of labeling but it too
involves educational problems. There are only three colors — red (stop),
yellow (caution), and green Cgo) — whose meanings most people would recognize
immediately. Three colors are probably not enough to establish an adequate
labeling scheme. The addition of other colors to the t.hree already mentioned
could result in a. massive educational problem. Furthermore, a significant
percentage of the population is partially color blind.
175
-------
Another alternative would be to list the noise emission (.sound pressure
or sound power level) produced by the product directly on the package itself.
This method has the advantage of giving the consumer an absolute value vhich
he can directly compare with the sound values of other products. It has "been
stated (e.g. [5-6]) that consumers desire more specific information and are
capable of understanding more than they are usually given credit for. For
example, women's hosiery is graded according to its "gauge". The fact that
gauge indicates how many needles are used in 1-1/2 inches of the loom is not
important to know. It is enough that the consumer knows that the higher the
gauge, generally the more durable the hosiery. The same may "be true about
sound level values. It may not be necessary fpr a person to understand the
technical aspects of the decibel. They merely need to be told enough so that
they can interpret in an adequate way how A-weighted sound level is a measure
of a product's "noise quality". This could probably be accomplished by
inserting a well-written article in some of the more popular magazines and
newspapers discussing in layman's terms the meaning of the terminology and
units.
A final labeling measure which could be used is a set of data which
describes the noise emission spectrum of the product. Such a label might not
really be needed or useful for the general public; however, this type of
information might be the only label thus far mentioned that would satisfy some
of the needs of the business and industrial community. The typical plant
owner needs information that will tell him whether the equipment he is
purchasing will violate any noise regulations under his use conditions. This
type of information is necessary so that plant layouts can be designed to
achieve minimum noise levels, utilizing natural shielding of machinery and
equipment, avoiding location of high-intensity noise sources near walls,
reflective surfaces, work stations, and areas frequented by employees. A
color code or single number or letter rating by itself will not provide the
user the kind and amount of information he needs. Any of the simple labels
coupled with an application guide in cookbook form could serve his needs if
the guide were properly designed. The decisions made by the business com-
munity can involve considerable expenditures and the purchaser could avoid a
mistake if it is possible to provide him with the necessary information.
Thsre is one voluntary labeling systems which already exists and is well
accepted in the area of acoustics. The Air-Conditioning and Refrigeration
Institute utilizes 1/3-octave band sound power determinations in a reverberant
room as the basis for its sound labeling system. The sound power data are
converted into a single number rating (numerical classification scheme ranging
from 1U-21 with increasing noise level corresponding to increasing numbers —
a change of 1 corresponds to about 3.3 dB) and an application guide that
provides the_steps necessary to convert a single number into a predicted sound
level at a given distance from the source. The labeling system is primarily
aimed at the distributors and installation contractors; however, education
booklets describing the meaning of the ratings have been developed for the
consumer.
176.
-------
Since this labeling scheme does presently exist and has some history of
acceptance, it will be briefly discussed to provide some additional insight
into the problems associated with the development of a labeling scheme that
will benefit and be understood by the audience for which the label is
intended.
The Air-Conditioning and Refrigeration Institute began "sound rating"
outdoor -units in 1971 with publication of its first Directory of Sound- Rated
Outdoor Unitary Equipment.
Under this voluntary program, all participating manufacturers (the models
listed in the most recent directory represent more than 90 percent of the
total U. S. output of these types of equipment) are required to rate the noise
emission of their outdoor units in accordance with specified procedures. Each
manufacturer certifies his own equipment and he must certify his entire line,
not Just selected models, to be listed in the directory. Air-Conditioning and
Refrigeration Institute member laboratories have the specialized facilities
and trained manpower to make sound power measurements in a reverberant field.
Based on an American Society for Heating, Refrigerating and Air-
Conditioning Engineers test method, the Air-Conditioning and Refrigeration
Institute standards and the Certification Program have been developed. One
standard is for rating and certification through independent laboratory tests
of the sound-generating characteristics of air conditioning equipment and
provides a uniform method for assigning a single rating number to this
equipment. Most units rate between ih and 2k on the Air-Conditioning and
Refrigeration Institute rating scale. The rating has built into it a penalty
for whines, screeches, and whistles —p the kind of noise that is disturbing
but may not be adequately indicated by meter readings.
At random, over a 3 year period, each model of each manufacturer is
retested by an independent laboratory to ensure the accuracy of the noise
rating. Enforcement procedures are strict: if a unit is tested and found to
"be inaccurately rated, a manufacturer must change the. sound rating, improve
the unit to meet the original rating, or withdraw from the directory and thus
lose the right to display the Air-Conditioning and/Refrigeration- Institute
Sound Rating Seal of Certification on his units.
The Air-Conditioning and Refrigeration Institute has also developed an
application standard which is basically a "cookbook" approach to converting
the sound rating number into an expected A-weighted sound level that will be
produced at given points of evaluation, e.g., the property line.
Another program that is of interest because of analogous problems to
noise emission labeling is the voluntary labeling program which applies to
energy-consuming home appliances[7-c1]*. Figure 35 shows an example of the
*This program will be modified in the near future since the Energy Policy and
Conservation Act (P.L. 9^-l63), which was passed on 22 December 1975, makes it
mandate'v for >nany products to be labeled as to their energy consumption.
This program will be carried out" by the Federal Energy Administration with the
design of appropriate labels being the responsibility of the Federal Trade
Commission.
177
-------
energy guide
ASDF Corp. Model 5508A10
8,000 Btu per hour
(cooling capacity)
115 volts 860 watts 7.5 amperes
E E R=9.3
Energy Efficiency Ratio expressed in Btu per watt-hour
IMPORTANT... for units with the same cooling
capacity, higher EER means:
Lower energy consumption
Lower cost to use!
For available 7,500 to 8,500 Btu per hour 115 volt
window models the EER range is
[ EER 5.4 to EER 9.9 |
For information on cost of operation and selection
of correct cooling capacity, ask your dealer for NBS
Publication LC 1053 or write to National Bureau of
Standards, 411.00, Washington, D.C. 20234
Data on this label Tested in accordance with
for this unit certified by »\t*"" 01
-v<* .^
*""«>«
Figure In. Sample of the energy labels now found on many room air con-
ditioners.
-------
label to be affixed to a room air conditioner. The cooling capacity of a unit
is determined, in accordance with an appropriate test procedure, in units of
Btu per hour. The electrical power requirement is determined in units of
watts. The efficiency of a unit is designated in terms of the Energy
Efficiency Ratio (EER) which is the ratio of the cooling capacity to the
electrical power requirement. The class, interval is 0.1 Btu per watt-hour,
obtained simply by rounding off the calculated ratio. Thus this is an
absolute classification scheme in that the number on the label directly
expresses a measure of the efficiency of the unit. The performance of a
particular unit can be judged from the range of EER values shown on the label
for available units of comparable cooling capacity. Thus this voluntary
labeling program retains the benefits of an absolute classification scheme
while at the same time giving some relative information.
Returning to noise emission, most acousticians who have been contacted by
"the authors believe that as a minimum, a label, or other documentation, should
convey information as to the absolute noise emission. This can either be in
terms of the designated noise emission, to whatever precision is reasonable,
or in terms of a simple classification scheme such as the A-weighted sound
power level rounded up to the nearest 0.5 B'or 5 dB. (Note that such a
rudimentary classification system involves division of the range of
designations into intervals of "significant" size, but does not involve
arbitrary coding). There seems to be very little reason to favor coding of an
absolute classification scheme in any vay (such a & Class- 1, 2, 3, .... or Class
A» B, C, ...) that obscures the relationship to the actual noise emission.
There have been some proposals for relative classification schemes based
on statistical distributions of noise emissions for a family of equipment.
Difficulties with this type of system include the following:
a particular model cannot be classified until essentially all other
products in the same family have been measured. This makes one
manufacturer too dependent upon the time schedules of other manu-
facturers .
the classification of a product is sensitive to errors in the noise
emission designations for other products in the same family.
the classification scheme can be "manipulated" by catalog listings
of particularly quiet or particularly noisy products, even though
such products are not normally sold.
the classification system changes with time so that one cannot
readily compare products sold in different years.
It is recommended that serious consideration be given to a labeling
system based on an absolute classification scheme, plus information on the
range of noise emissions for competing products. This would be analogous to
"the energy guide label shown in Figure 35-
179
-------
References for Appendix E
[l] American National Standard Method for Rating the Sound Power Spectra of
Small Stationary Noise Sources, S3. 17-1975 5 also Acoust. Soc . Amer. Std.
14-1975 (American National Standards Institute, New York, 1975)-
[2] Lang, W. W., and Flynn, D. R. , A Noise Emission Number for product
designation, Noise Control Engineering ^_, 108-113 (1975); errata in NCE
I, 3 (1975).
[3] American National Standard Method for the Designation of Sound Power
Emitted by Machinery and Equipment, SI. 23-1976; also Acoust. Society
Amer. Std. 5-1976 (American National Standards Institute, New York,
1976).
[U] First Draft Proposal ISO/DP U871 for Acoustics — Noise Classification
and Labelling of Equipment and Machinery, Document No. ISO/TC U3/SC 1
(Secretariat - 203) 27^ (January 1976").
[5] Coles, J. V., Standards and Labels for Consumer Goods (Ronald Press Co.»
New York,
[6] Dickerson, F. R., Product Safety in Household Goods (Bobbs-Merrill Co.,
New York, 1968).
[7] Air conditioner labels show how to keep coolest for least $, Commerce
Today, Vol. IV, No. 18, pp 10-11 (June 10,. 197*0.
[8] Conserving Energy Through Appliance Labeling, Dimensions/NBS 58, 22^-225
& 238 (197*0 .
180
-------
NBS.1I4A (REV. 7.73)
U.S. DEPT. OF COMM.
BIBLIOGRAPH4C DATA
SHEET
1. PUBLICATION OR REPORT NO.
NBS HB-122 and
EPA 5SQ/
-------
NBS TECHNICAL PUBLICATIONS
PERIODICALS
JOURNAL OF RESEARCH reports National Bureau
of Standards research and development in physics,
mathematics, and chemistry. It is published in two
sections, available separately:
• Physics and Chemistry (Section A)
Papers of interest primarily to scientists working in
these fields. This section covers a broad range of physi-
cal and chemical research, with major emphasis on
standards of physical measurement, fundamental con-
stants, and properties of matter. Issued six times a year.
Annual subscription: Domestic, $17.00; Foreign, $21.25.
• Mathematical Sciences (Section B)
Studies and compilations designed mainly for the math-
ematician and theoretical physicist. Topics in mathemat-
ical statistics, theory of experiment design, numerical
analysis, theoretical physics and chemistry, logical de-
sign and programming of computers and computer sys-
tems. Short numerical tables. Issued quarterly. Annual
subscription: Domestic, $9.00; Foreign, $11.25.
DIMENSIONS/NBS (formerly Technical News Bulle-
*V This monthly magazine is published to inform
scientists, engineers, businessmen, industry, teachers,
students, and consumers of the latest advances in
science and technology, with primary emphasis on the
work at NBS. The magazine highlights and reviews
such issues as energy research, fire protection, building
technology, metric conversion, pollution abatement,
health and safety, and consumer product performance.
fn addition, it reports the results of Bureau programs
m measurement standards and techniques, properties of
matter and materials, engineering standards and serv-
ices, instrumentation, and automatic data processing.
Annual subscription: Domestic, $12.50; Foreign, $15.65.
NONPERIODICALS
Monographs—Major contributions to the technical liter-
ature on various subjects related to the Bureau's scien-
tific and technical activities.
Handbooks—Recommended codes of engineering and
industrial practice (including safety codes) developed
m cooperation with interested industries, professional
organizations, and regulatory bodies.
Special Publications—Include proceedings of conferences
sponsored by NBS, NBS annual reports, and other
special publications appropriate to this grouping such
as wall charts, pocket cards, and bibliographies.
Applied Mathematics Series—Mathematical tables, man-
uals, and studies of special interest to physicists, engi-
neers, chemists, biologists, mathematicians, com-
puter programmers, and others engaged in scientific
and technical work.
National Standard Reference Data Series—Provides
quantitative data on the physical and chemical proper-
ties of materials, compiled from the world's literature
and critically evaluated. Developed under a world-wide
Program coordinated by NBS. Program under authority
« National Standard Data Act (Public Law 90-396).
NOTE: At present the principal publication outlet for
these data is the Journal of Physical and Chemical
Reference Data (JPCRD) published quarterly for NBS
by the American Chemical Society (ACS) and the Amer-
ican Institute of Physics (AIP). Subscriptions, reprints,
and supplements available from ACS, 1155 Sixteenth
St. N.W., Wash. D. C. 20056.
Building Science Series—Disseminates technical infor-
mation developed at the Bureau on building materials,
components, systems, and whole structures. The series
presents research results, test methods, and perform-
ance criteria related to the structural and environmental
functions and the durability and safety characteristics
of building elements and systems.
Technical Notes—Studies or reports which are complete
in themselves but restrictive in their treatment of a
subject. Analogous to monographs but not so compre-
hensive in scope or definitive in treatment of the sub-
ject area. Often serve as a vehicle for final reports of
work performed at NBS under the sponsorship of other
government agencies.
Voluntary Product Standards—Developed under proce-
dures published by the Department of Commerce in Part
10, Title 15, of the Code of Federal Regulations. The
purpose of the standards is to establish nationally rec-
ognized requirements for products, and to provide all
concerned interests with a basis for common under-
standing of the characteristics of the products. NBS
administers this program as a supplement to the activi-
ties of the private sector standardizing organizations.
Consumer Information Series—Practical information,
based on NBS research and experience, covering areas
of interest to the consumer. Easily understandable lang-
uage and illustrations provide useful background knowl-
edge for shopping in today's technological marketplace.
Order above NBS publications from: Superintendent
of Documents, Government Printing Office, Washington,
B.C. 20^02.
Order following NBS publications—NBSIR's and FIPS
from the National Technical Information Services,
Springfield, Va. 22161.
Federal Information Processing Standards Publications
(FIPS PUBS)—Publications in this series collectively
constitute the Federal Information Processing Stand-
ards Register. Register serves as the official source of
information in the Federal Government regarding stand-
ards issued by NBS pursuant to the Federal Property
and Administrative Services Act of 1949 as amended,
Public Law 89-306 (79 Stat. 1127), and as implemented
bv Executive Order 11717 (38 FR 12315, dated May 11,
1973) and Part 6 of Title 15 CFR (Code of Federal
Regulations).
NBS Interagency Reports (NBSIR)—A special series of
interim or final reports on work performed by NBS for
outside sponsors (both government and non-govern-
ment) In general, initial distribution is handled by the
sponsor; public distribution is by the National Techni-
cal Information Services (Springfield, Va. 22161) in
paper copy or microfiche form.
BIBLIOGRAPHIC SUBSCRIPTION SERVICES
The following current-awareness and literature-survey Superconducting Devices and MateriaIs A literature
bibliographies are issued periodically by the Bureau: survey issued quarterly. Annual subscription. $*>.UU.
Cryogenic Data Center Current Awareness Service. A Send subscription orders and remittances for the pre-
jiterature survey issued biweekly. Annual subscrip- cedin£, bibliographic services to National Bureau of
— - • Standards, Cryogenic Data Center (275.02) Boulder,
ttion: Domestic, $20.00; Foreign, $25.00. C0. f ,a Cr^
Liquified Natural Gas. A literature survey issued quar- Standards Cry
te*ly. Annual subscription: $20.00. Colorado 80302.
------- |