INTERNATIONAL PROGRAMME ON CHEMICAL SAFETY
ENVIRONMENTAL HEALTH CRITERIA 12
NOISE
This report contains the collective views of an
international group of experts and does not
necessarily represent the decisions or the stated
policy of either the World Health Organization
or the United Nations Environment Programme
Published under the joint sponsorship of
the United Nations Environment Programme
and the World Health Organization
World Health Organization Geneva, 1980
ISBN 92 4 154072 9
(c) World Health Organization 1980
Publications of the World Health Organization enjoy copyright
protection in accordance with the provisions of Protocol 2 of the
Universal Copyright Convention. For rights of reproduction or
translation of WHO publications, in part or in toto, application
should be made to the Office of Publications, World Health
Organization, Geneva, Switzerland. The World Health Organization
welcomes such applications.
The designations employed and the presentation of the material in
this publication do not imply the expression of any opinion whatsoever
on the part of the Secretariat of the World Health Organization
concerning the legal status of any country, territory, city or area or
of its authorities, or concerning the delimitation of its frontiers or
boundaries.
The mention of specific companies or of certain manufacturers'
products does not imply that they are endorsed or recommended by the
World Health Organization in preference to others of a similar nature
that are not mentioned. Errors and omissions excepted, the names of
proprietary products are distinguished by initial capital letters.
CONTENTS
ENVIRONMENTAL HEALTH CRITERIA FOR NOISE
1. SUMMARY AND RECOMMENDATIONS FOR FURTHER STUDIES
1.1. Summary
1.1.1. Introduction
1.1.2. Noise measurement
1.1.3. Effects of noise
1.1.3.1 Interference with communication
1.1.3.2 Hearing loss
1.1.3.3 Disturbance of sleep
1.1.3.4 Stress
1.1.3.5 Annoyance
1.1.3.6 Effects on performance
1.1.3.7 Miscellaneous effects
1.1.4. Summary of recommended noise exposure limits
1.2. Recommendations for further studies
2. PROPERTIES AND MEASUREMENT OF NOISE
2.1. Physical properties and measurements
2.2. Sound perception and its measurement
2.2.1. Loudness and loudness level
2.2.2. Calculation and measurement of loudness level
2.2.3. Sound level and noise level
2.2.4. The time factor
2.2.5. Noise exposure scales
2.2.6. Equivalent continuous sound pressure level
2.2.7. Level distribution
2.3. Sources of noise
2.3.1. Industry
2.3.2. Road traffic
2.3.3. Rail traffic
2.3.4. Air traffic
2.3.5. Sonic booms
2.3.6. Construction and public works
2.3.7. Indoor sources
2.3.8. Miscellaneous sources
3. EFFECTS OF NOISE
3.1. Noise-induced hearing loss
3.1.1. Hearing impairment
3.1.1.1 Hearing level, noise-induced threshold
shift, and hearing impairment
3.1.1.2 Noise-induced temporary threshold shift
3.1.1.3 Noise-induced permanent threshold shift
3.1.1.4 Incidence of noise-induced permanent
hearing loss
3.1.2. Relation between noise exposure and hearing loss
3.1.2.1 Laboratory studies
3.1.2.2 Occupational hearing loss
3.1.2.3 Factors that may influence the incidence
of noise-induced permanent threshold
shift
3.1.2.4 Combined effects of intensity and
duration of noise exposure
3.1.2.5 Estimation of hearing impairment risk
3.1.2.6 The importance of high-frequency hearing
3.1.3. Effects of impulsive noise
3.1.4. Infrasound and ultrasound
3.2. Interference with communication
3.2.1. Masking and intelligibility
3.2.2. Speech interference indices
3.2.2.1 Articulation index
3.2.2.2 Speech Interference Level
3.2.2.3 A-weighted sound pressure level
3.2.3. Perception of speech out-of-doors
3.2.4. Indoor speech communication
3.3. Pain
3.4. Sleep
3.4.1. Nature of sleep disturbance
3.4.2. Influence of noise characteristics
3.4.3. Influence of age and sex
3.4.4. Influence of previous sleep deprivation,
adaptation, and motivation
3.4.5. Long-term effects of sleep disturbance by noise
3.5. Nonspecific effects
3.5.1. The stress response
3.5.2. Circulatory system responses
3.5.3. The startle reflex and orienting response
3.5.4. Effects on equilibrium
3.5.5. Fatigue
3.6. Clinical health effects
3.6.1. Background
3.6.2. General health
3.6.3. Mental health
3.7. Annoyance
3.7.1. Definition and measurement
3.7.2. Instantaneous noise dose
3.7.3. Long-term noise dose
3.7.3.1 Aircraft noise
3.7.3.2 Road traffic noise
3.7.3.3 General environmental noise
3.7.4. Correlation between noise exposure and annoyance
3.7.5. Overt reaction
3.8. Effects on task performance
3.8.1. Noise as a distracting stimulus
3.8.2. Effects on tasks involving motor or monotonous
activities
3.8.3. Effects on tasks involving mental activities
4. EVALUATION OF HEALTH RISKS TO MAN FROM EXPOSURE TO NOISE
4.1. Environmental noise
4.2. Populations affected
4.3. Specific health criteria
4.3.1. Physical injury
4.3.2. Hearing loss
4.3.3. Nonspecific health effects
4.3.4. Interference effects
4.4. General health, welfare, and annoyance criteria
5. NOISE CONTROL AND HEALTH PROTECTION
5.1. Noise control at source
5.2. Control of sound transmission
5.3. Reduction in length of exposure
5.4. Education of workers
5.5. Ear protection
5.6. Audiometry
REFERENCES
NOTE TO READERS OF THE CRITERIA DOCUMENTS
While every effort has been made to present information in the
criteria documents as accurately as possible without unduly delaying
their publication, mistakes might have occurred and are likely to
occur in the future. In the interest of all users of the environmental
health criteria documents, readers are kindly requested to communicate
any errors found to the Division of Environmental Health, World Health
Organization, Geneva, Switzerland, in order that they may be included
in corrigenda which will appear in subsequent volumes.
In addition, experts in any particular field dealt with in the
criteria documents are kindly requested to make available to the WHO
Secretariat any important published information that may have
inadvertently been omitted and which may change the evaluation of
health risks from exposure to the environmental agent under
examination, so that the information may be considered in the event of
updating and re-evaluation of the conclusions contained in the
criteria documents.
WHO TASK GROUP ON ENVIRONMENTAL HEALTH CRITERIA FOR NOISE
Members
Dr H. E. von Gierke, Department of the Air Force, Aerospace Medical
Research Laboratory, Wright Patterson Air Force Base, OH, USA
(Chairman)
Dr E. Gros, Institute for Hygiene and Occupational Medicine,
University Clinic, Essen, Federal Republic of Germany
Professor L. L. Karagodina, F. F. Erisman Research Institute of
Hygiene, Moscow, USSR (Vice-Chairman)
Professor G. E. Lambert, Médecin Inspecteur du Travail Région Midi-
Pyrénées, Cité Administrative, Toulouse, France
Professor J. B. Ollerhead, Department of Transport Technology,
University of Technology, Loughbrough, Leicester, England
(Rapporteur)
Dr Y. Osada, The Institute of Public Health, Tokyo, Japan
Professor B. Paccagnella, Institute of Hygiene, University of Padua,
Verona, Italy
Dr P. Rey, Institute of Social and Preventive Medicine, University of
Geneva, Geneva, Switzerland
Professor R. Rylander, Department of Hygiene, University of
Gothenburg, Gothenburg, Sweden (Rapporteur)
Professor W. J. Sulkowski, Institute of Occupational Medicine, Lodz,
Poland
Ms A. Suter, Office of Noise Abatement and Control, United States
Environmental Protection Agency, Washington DC, USA (Rapporteur)
Representatives of other organizations
Dr G. H. Coppée, International Labour Organisation, Geneva,
Switzerland
Dr W. Hunter, Commission of the European Communities, Luxembourg
Dr A. Alexandre, Organisation for Economic Co-operation and
Development, Paris, France
Mr L. Nielsen, International Organization for Standardization,
Hellerup, Denmark
Observers
Ms G. Vindevogel, Ministry of Public Health and Family, Brussels,
Belgium
Mr L. Baekelandt, Ministry of Public Health and Family, Brussels,
Belgium
Secretariat
Ms B. Goelzer, Scientist, Office of Occupational Health, World Health
Organization, Geneva, Switzerland
Dr H. W. de Koning, Scientist, Control of Environmental Pollution and
Hazards, World Health Organization, Geneva, Switzerland
(Secretary)
Dr V. Krichagin, Environment and Occupational Health, WHO Regional
Office for Europe, Copenhagen, Denmark
Dr J. Lang, National Institute for Research on Heat and Noise
Technology, Vienna, Austria (Temporary Adviser)
List of abbreviations and symbols used in this document
AI articulation index
c speed of sound
CNEL community noise equivalent level
CNR composite noise rating
f frequency
I sound intensity
Ldn day-night average-sound level
Le aircraft exposure level
Leq equivalent continuous sound pressure level
Lp or SPL sound pressure level
Lp(A) A-weighted sound pressure level
LPN mean peak perceived noise level
NEF noise exposure forecast
NI noiseness index
NIPTS noise-induced permanent threshold shift
NITS noise-induced threshold shift
NITTS noise-induced temporary threshold shift
NNI noise and number index
NPL noise pollution level
p root mean square pressure
p2 mean square sound pressure
P sound power
PNL perceived noise level
SIL speech interference level
SPL or Lp sound pressure level
TNEL total noise exposure level
TNI traffic noise index
WECPNL weighted equivalent continuous perceived noise level
lambda wavelength
ENVIRONMENTAL HEALTH CRITERIA FOR NOISE
A WHO Task Group on Environmental Health Criteria for Noise met
in Brussels from 31 January to 4 February 1977. Dr. H. W. de Koning,
Scientist, Control of Environmental Pollution and Hazards, Division of
Environmental Health, WHO, opened the meeting on behalf of the
Director General and expressed the appreciation of the Organization to
the Government of Belgium for having made available the necessary
financial support for the meeting. On behalf of the Government, the
Group was welcomed by Professor Lafontaine, Director of the Institute
for Hygiene and Epidemiology, Brussels. The Task Group reviewed and
revised the second draft criteria document and made an evaluation of
the health risks from exposure to noise.
The first draft of the criteria document was prepared by a study
group that met in Geneva from 5-9 November 1973. Participants of the
Group included: Dr. T. L. Henderson and Professor G. Jansen (Federal
Republic of Germany); Dr A. F. Meyer (USA); Professor J. B. Ollerhead
(United Kingdom, Rapporteur); Professor P. Rey (Switzerland,
Chairman); Professor R. Rylander (Sweden); Professor W. J. Sulkowski
(Poland); Dr A. Annoni, Mr E. Hellen, and Mr B. Johansson
(Consultant), International Labour Organisation (ILO); Dr A.
Alexandre, Organisation for Economic Co-operation and Development
(OECD); Dr A. Berlin, Commission of the European Communities (CEC);
Professor L. A. Saenz, Scientific Committee on Problems of the
Environment (SCOPE); Mr H. J. Gursahaney, International Civil Aviation
Organization (ICAO); Dr M. Suess, World Health Organization Regional
Office for Europe; and Dr G. Cleary and Dr G. E. Lambert, World Health
Organization, Geneva. Certain sections of the first draft were later
completed with the assistance of Dr A. Alexandre (OECD), Dr D. E.
Broadbent (UK), Professor G. Jansen (FRG), and Professor W. D. Ward
(USA).
The second draft was prepared by the Secretariat after comments
had been received from the national focal points for the WHO
Environmental Health Criteria Programme in Czechoslovakia, Federal
Republic of Germany, Finland, Greece, Japan, New Zealand, Poland,
Sweden, Thailand, United Kingdom, USSR, and USA, and from the
International Labour Organisation, Commission of the European
Communities, the Organisation for Economic Co-operation and
Development, the International Civil Aviation Organization, and the
International Organization for Standardization. Many comments were
also received from individual experts and commercial concerns
including E. I. Du Pont de Nemours & Company, Wilmington, Delaware,
USA, whose contributions are gratefully acknowledged.
The Secretariat particularly wishes to thank Dr D. Hickish, Ford
Motor Company Limited, Brentwood, Essex, England, Dr G. E. Lambert,
Professor J. B. Ollerhead, Professor P. Rey, Professor R. Rylander,
and Ms A. Suter for their most valued help in the final phases of the
preparation of the document.
This document is based primarily on original publications listed
in the reference section and every effort has been made to review all
pertinent data and information available up to 1978. In addition,
reference has often been made to the various publications on noise of
the International Organization for Standardization that include the
international standards for noise assessment (ISO, 1971; 1973a;
1975a). The following reviews and criteria documents have been
referred to: Burns & Robinson (1970), Karagodina et al. (1972), Burns
(1973), NIOSH (1973a), US Environmental Protection Agency (1973a), ILO
(1976), Thiessen (1976), Rylander et al. (1978), and Health and
Welfare, Canada (1979).
Details of the WHO Environmental Health Criteria Programme
including some terms frequently used in the document may be found in
the general introduction to the Environmental Health Criteria
Programme published together with the environmental health criteria
document on mercury (Environmental Health Criteria 1, Mercury, World
Health Organization, Geneva, 1976) and now available as a reprint.
1. SUMMARY AND RECOMMENDATIONS FOR FURTHER STUDIES
1.1 Summary
1.1.1 Introduction
Noise can disturb man's work, rest, sleep, and communication; it
can damage his hearing and evoke other psychological, physiological,
and possibly pathological reactions. However, because of their
complexity, their variability, and the interaction of noise with other
environmental factors, the adverse health effects of noise do not lend
themselves to a straightforward analysis.
Probably the most important issue is the industrial noise
problem, and a need for noise control and hearing conservation
programmes is widely recognized. Road traffic is the main source of
community noise that may disturb large segments of the urban
population. Also of worldwide concern is aircraft noise, which can
significantly affect the mode of life of people living in the vicinity
of airports.
1.1.2 Noise measurement
Sound is produced by the vibration of bodies or air molecules and
is transmitted as a longitudinal wave motion. It is, therefore, a form
of mechanical energy and is measured in energy-related units. The
sound output of a source is measured in watts and the intensity of
sound at a point in space is defined by the rate of energy flow per
unit area, measured in watts per ms. Intensity is proportional to the
mean square of the sound pressure and, as the range of this variable
is so wide, it is usual to express its value in decibels (dB)a.
Because the effects of noise depend strongly upon frequency of sound
pressure oscillation, spectrum analysis is important in noise
measurement.
a decibel = a measure on a logarithmic scale of a quantity such as
sound pressure, sound power, or intensity with respect to a
standard reference value (0.0002 microbars for sound pressure,
10-12W for sound power, and 10-12W/m2 for intensity). Thus,
for example, when the sound intensity increases by a factor of
1.26 (= 100.1), it is said to have increased by 1 decibel (dB);
1 Bel equals 10 dB or a factor of 10 in intensity. The standard
reference values are implied throughout this document unless
otherwise stated.
The perceived magnitude of sound is defined as loudness and its
decibel equivalent is known as the loudness level. The loudness is a
function of both intensity and frequency, and various procedures exist
by which it may be estimated from physical measurements. The simplest
methods involve the measurement of the sound pressure level (SPL)
through a filter or network of filters that represent the frequency
response of the ear. Despite the existence of other slightly more
accurate but more complex techniques, the A-weighted sound pressure
level scale is gaining widespread acceptance and is revommended for
general use.b Whatever procedure is used, such frequency-weighted
measurements are referred to simply as sound (or noise) levels.
Measurements of sound level may be averaged over two distinctly
different periods of time. Steady sound levels and instantaneous
levels of variable sounds are measured on a very short time scale of
1 second or less. Variable sounds can be measured with a much longer
average time, over periods of hours if necessary, and are expressed in
terms of the equivalent continuous sound pressure level (Leq). This
convenient measure of average noise exposure using the A-weighting
correlates reasonably well with many human responses to noise and is
recommended for general use.
Many noise indices have been developed for predicting human
reaction to various noise levels. Some of these incorporate non-
acoustic factors that influence the reaction. Although the use of such
indices is not to be discouraged, it is desirable to adopt a uniform
approach to noise measurement, whenever possible.
1.1.3 Effects of noise
1.1.3.1 Interference with communication
Although there appears to be no firm evidence, it is believed
that interference with speech in occupational situations may lead to
accidents due to inability to hear warning shouts etc. In offices,
schools, and homes, speech interference is a major source of
annoyance. Many attempts have been made to develop a single index of
such interference, based on the characteristics of the masking noise,
that directly indicates the degree of interference with speech
perception. Such indices involve a considerable degree of
approximation. The following are the three most widely used:
b To obtain a single number representing the sound level of a noise
containing a wide range of frequencies in a manner representative
of the ear's response, it is necessary to modify the effects of
the low and high frequencies with respect to the medium
frequencies. The A-filter is one particular frequency weighting
and, when this is used, the resulting sound level is said to be
A-weighted.
Articulation index (AI). This is the most complicated index,
since it takes into account the fact that some frequencies are more
effective in masking speech than others. The frequency range from 250
to 7000 Hz is divided into 20 bands. The difference between file
average speech peak level in each of these bands is calculated and the
resulting numbers combined to give a single index.
Speech interference level (SIL). SIL was designed as a
simplified substitute for the AI. It was originally defined as the
average of the now obsolete octave-band SPLs in the 600-1200,
1200-2400, and 2400-4800 Hz octaves. At the present time, SIL, based
upon the octave band levels at the preferred frequencies of 500,
1000, 2000, and 4000 Hz, is considered to provide a better estimate of
the masking ability of a noise. As SIL does not take the actual
speech level into account, the associated masking effect depends upon
vocal effort and speaker-to-listener distance.
A-weighted sound level. This is also a convenient and fairly
accurate index of speech interference.
It is usually possible to express the relationship between noise
levels and speech intelligibility in a single diagram, based on the
assumptions and empirical observations that, for speaker-to-listener
distances of about 1 m:
(a) speech spoken in relaxed conversation is 100% intelligible in
background noise levels of about 45 dB(A), and can be understood
fairly well in background levels of 55 dB(A); and
(b) speech spoken with slightly more vocal effort can be
understood well, when the noise level is 65 dB(A).
For outdoor speech communication, the "inverse square law"
controls speech transmission over moderate distances, i.e., when the
distance between speaker and listener is doubled, the level of the
speech drops by approximately 6 dB. This relationship is less likely
to apply indoors, where speech communication is affected by the
reverberation characteristics of the room.
In cases where the speech signals are of paramount importance,
e.g., in classrooms or conference rooms, or where listeners with
impaired hearing faculties are involved, e.g., in homes for aged
people, lower levels of background noise are desirable.
1.1.3.2 Hearing loss
Hearing loss can be either temporary or permanent. Noise-induced
temporary threshold shift (NITTS) is a temporary loss of hearing
acuity experienced after a relatively short exposure to excessive
noise. Pre-exposure hearing is recovered fairly rapidly after
cessation of the noise. Noise-induced permanent threshold shift
(NIPTS) is an irreversible (sensorineural) loss of hearing that is
caused by prolonged noise exposure. Both kinds of loss together with
presbyacusis, the permanent hearing impairment that is attributed to
the natural aging process, can be experienced simultaneously.
In the quantification of hearing damage, it is necessary to
differentiate between NIPTS, hearing level (the audiometric level of
an individual or group in relation to an accepted audiometric
standard), and hearing impairment.
NIPTS is the hearing loss (i.e., the reduction of hearing level)
attributable to noise exposure alone, disregarding losses due to
aging. NIPTS occurs typically at high frequencies, usually with a
maximum loss at around 4000 Hz. Noise-induced hearing loss occurs
gradually, usually over a period of years. Once there is considerable
hearing loss at a particular frequency, the rate of loss usually
diminishes. Audiometrically, noise-induced losses are similar to
presbyacusis. Hearing loss due to prolonged excessive noise exposure
is generally associated with destruction of the hair cells of the
inner ear. The severity of hearing loss is correlated with both the
location and the extent of damage in the organ of Corti.
"Hearing impairment" is usually defined as the hearing level at
which individuals begin to experience difficulties in everyday life.
It is assessed in terms of difficulty in understanding speech. The
amount of loss at the speech frequencies has been used as a basis for
compensation and varies from one country to another. The unweighted
average of the losses, in dB, at 500, 1000, and 2000 Hz that is widely
used for assessing noise-induced hearing impairment, is somewhat
misleading since most hearing loss usually occurs at 2000 Hz and
above. Consequently, there is an increased tendency to include the
frequencies of 3000 and 4000 Hz in damage assessment formulae.
Attempts have been made to establish the levels of noise that are
permanently damaging to the ear and to identify individual
susceptibility to NIPTS on the basis of NITTS measurements. However,
the validity of the connection between NITTS and NIPTS has not been
agreed.
There is also some disagreement concerning the relationship
between the relative ear-damaging capacity of the noise level and its
duration. However, the hypothesis that the hearing damage associated
with a particular noise exposure is related to the total energy of the
sound (i.e., the integrated product of intensity and time) is rapidly
gaining favour for practical purposes. Thus, noise should preferably
be described in terms of equivalent continuous sound level, Leq,
measured in dB(A). For occupational noise, the level should be
averaged over the entire 8-h shift (Leq (8-h)).
Available data show that there is considerable variation in human
sensitivity with respect to NIPTS. The hazardous nature of a noisy
environment is therefore described in terms of "damage risk". This may
be expressed as the percentage of people exposed to that environment
who are expected to suffer noise-induced hearing impairment after
appropriate allowance has been made for hearing losses due to other
causes. It is now accepted that this risk is negligible at noise
exposure levels of less than 75 dB(A) Leq (8=h) but increases with
increasing levels. Based on national judgements concerning "acceptable
risk", many countries have adopted industrial noise exposure limits of
85 dB(A) + 5dB(A) in their regulations and recommended practices.
The exposure to ototoxic drugs such as certain aminoglycosidic
antibiotics however, can lower the threshold below which noise can
damage the ear.
It is not yet clear whether the damage risk rules already
mentioned can be extended to the very short durations of impulsive
noise. Available evidence indicates that a considerable risk exists,
when impulsive sound levels reach 130-150 dB, depending upon the
temporal characteristics of the impulse.
Although there is a fairly wide range of individual variability,
especially for high frequency stimuli, the threshold of pain for
normal ears is in the region of 135-140 dB sound pressure level. Aural
pain should always be considered to be an early warning sign of
excessive noise exposure.
Wherever possible, problems of noise control should be tackled at
source, i.e., by reducing the amount of noise produced. An acceptable
alternative is to isolate people from the noise by the use of noise
insulation, including soundproof enclosures, partitions, and acoustic
barriers. If this is not possible, the risk can also be minimized by
limiting the duration of exposure. Only in cases where these control
measures are impracticable should personal ear protection be
considered. These devices can and do provide useful protection but
inherent problems include those of proper fitting and use, and a
degree of discomfort.
If there is any risk of hearing damage, pre-employment and
follow-up audiometric examinations of workers should be carried out to
detect changes in hearing acuity that might indicate possible
development of NIPTS, in order to initiate preventive action.
1.1.3.3 Disturbance of sleep
Noise intrusion can cause difficulty in falling asleep and can
awaken people who are asleep. Detailed laboratory studies of the
problem have been made by monitoring electroencephalographic (EEG)
responses and changes in neurovegetative reactions during sleep.
Studies have indicated that the disturbance of sleep becomes
increasingly apparent as ambient noise levels exceed about 35 dB(A)
Leq. It has been found that the probability of subjects being
awakened by a peak sound level of 40 dB(A) is 5%, increasing to 30% at
70 dB(A). Defining sleep disturbance in terms of EEG changes, the
probability of disturbance increases from 10% at 40 dB(A) to 60% at
70 dB(A). It has also been observed that subjects who sleep well
(based on psychomotoric activity data) at 35 dB(A) Leq complain
about sleep disturbance and have difficulty in falling asleep at
50 dB(A) Leq and even at 40 dB(A) Leq. Weak stimuli that
are unexpected can still interfere with sleep.
Within a population, differences in sensitivity to noise occur
related, for example, to age and sex. Adaptation has been observed
only when noise stimuli are of low intensity. Even though sleep is
more disturbed by noise rich in information, habituation to such noise
has been observed. Based on the limited data available, a level of
less than 35 dB(A) Leq is recommended to preserve the restorative
process of sleep.
1.1.3.4 Stress
Noise produces different reactions along the hypothalamo-
hypophyseal-adrenal axis including an increase in adenocorticotropic
hormone (ACTH) release and an elevation of corticosteroid levels. Some
of these reactions have been elicited in an acute form in laboratory
experiments at rather moderate levels of noise.
Effects on the systemic circulation such as constriction of blood
vessels have been produced under laboratory conditions and a high
incidence of circulatory disturbances including hypertension has been
found in noise-exposed workers. A tendency for blood pressure to be
higher in populations living in noisy areas around airports has been
suggested but no conclusive evidence of this has been presented.
Noise affects the sympathetic division of the autonomic nervous
system. Eye dilation, bradycardia, and increased skin conductance are
proportional to the intensity of noise above 70 dB SPL, without
adaptation to the stimulus.
Other sympathetic disturbances, such as changes is
gastrointestinal motility, can be produced by intense sound. Medical
records of workers have shown that, in addition to a higher incidence
of hearing loss, noise-exposed groups have a higher prevalence of
peptic ulcer; however, a causal relationship has not been established.
More studies are required to determine the long-term health risks
due to the action of noise on the autonomic nervous system.
1.1.3.5 Annoyance
Noise annoyance may be defined as a feeling of displeasure evoked
by a noise. The annoyance-inducing capacity of a noise depends upon
many of its physical characteristics including its intensity, spectral
characteristics, and variations of these with time. However, annoyance
reactions are sensitive to many nonacoustic factors of a social,
psychological, or economic nature and there are considerable
differences in individual reactions to the same noise.
Attempts to define criteria linking noise exposure and annoyance
have led to the development of many methods for the measurement of
both variables. In social surveys, questionaires are used to assess
the annoyance felt by an individual in response to various types of
noise. Much research has been aimed at the definition of suitable
questions through which annoyance reactions could be quantified.
In the search for a suitable noise index, numerous noise and some
nonacoustic variables were assembled in various ways to discover which
combinations were most closely correlated with annoyance reactions.
The resulting diverse indices were given such names as composite noise
rating (CNR), community noise equivalent level (CNEL), noise and
number index (NNI), and noise pollution level (NPL) among many others.
In fact, many experts consider that, in terms of annoyance prediction
ability, there is little practical difference between the various
indices and that an appropriate index should be selected for the
convenience with which it can be measured or calculated. For this
reason, variants of the equivalent continuous A-weighted sound
pressure level (Leq) are being widely adopted for general use. These
are conveniently applied to noise exposure patterns of all kinds, from
multiple sources if necessary, and are reasonably well correlated both
with annoyance and with other specific effects of noise.
Whatever noise scale is used to express noise exposure, it must
be recognized that, at any level of noise annoyance, reactions will
vary greatly because of psychosocial differences. A useful technique
for accommodating the possible extent of individual variation is the
use of a criterion curve showing the percentage of persons who will be
annoyed as a function of noise level.
Such curves have been derived for a variety of noise conditions
but mainly for those concerned with aircraft or road traffic noise. On
the basis of these, it can be concluded that, in residential areas
where the general daytime noise exposure is below 55 dB(A) Leq,
there will be few people seriously annoyed by noise. This is
recommended as a desirable noise exposure limit for the general
community, even though it will be difficult to achieve in many urban
areas. Some residents may consider this level too high, especially as
substantially lower levels currently prevail in many suburban and
rural areas.
Criteria relating noise exposure and complaint potential have
found widespread application for environmental control purposes in
some countries. However, the scientific basis for such criteria is
rather fragmentary and surveys have indicated that the correlation
between noise exposure and individual complaint behaviour is low. This
may be explained in terms of the strong influence of psychosocial
factors.
1.1.3.6 Effects on performance
The effect of noise on the performance of tasks has mainly been
studied in the laboratory and, to some extent, in work situations,
but, there have been few, if any, detailed studies of the effects of
noise on human productivity in real-life situations. It is evident
that when a task involves auditory signals of any kind, noise at an
intensity sufficient to mask or interfere with the perception of these
signals will interfere with the performance of the task.
Noise can act as a distracting stimulus, depending on how
meaningful the stimulus might be, and may also affect the psycho-
physiological state of the individual. A novel event, such as the
start of an unfamiliar noise will cause distraction and interfere with
many kinds of tasks. Impulsive noise (such as sonic booms) may produce
disruptive effects as the result of startle responses which are more
resistant to habituation.
Noise can change the state of alertness of an individual and may
increase or decrease efficiency.
Performance of tasks involving motor or monotonous activities is
not always degraded by noise. At the other extreme, mental activities
involving vigilance, information gathering, and analytical processes
appear to be particularly sensitive to noise. It has been suggested
that, in industry, the most likely indicator of the effects of noise
on performance would be an increase in accidents attributable to
reduced vigilance.
1.1.3.7 Miscellaneous effects
Certain noises, especially impulsive ones, may induce a startle
reaction. This consists of contraction of the flexor muscles of the
limbs and the spine, a contraction of the orbital which can be
recorded as an eye blink, and a focusing of attention towards the
location of the noise. The startle reflex to acoustic stimulation has
been observed in the 27-28 week fetus in utero as a change in the
pulse rate.
It has been suggested that observed noise-induced equilibrium
effects are due to the noise stimulating the vestibular apparatus, the
receptors of which are part of the inner ear structure.
Although there is no clear evidence of a direct relationship
between noise and fatigue, noise can be considered as an environmental
stress which, in conjunction with other environmental and host
factors, may induce a chronic fatigue that could lead to non-specific
health disorders.
1.1.4 Summary of recommended noise exposure limits
The equivalent continuous A-weighted sound pressure level Leq
is recommended for use as a common measure of noise exposure. The
measurement period should be related to the problem under study, for
example in the case of occupational noise, Leq (8-h) would be
measured for a complete 8-h shift.
For the working environment, there is no identifiable risk of
hearing damage in noise levels of less than 75 dB(A) Leq (8-h). For
higher levels, there is an increasing predictable risk and this must
be taken into account when setting occupational noise standards.
In other occupational and domestic environments, acceptable noise
levels can be established on the basis of speech communication
criteria. For good speech intelligibility indoors, background noise
levels of less than 45 dB(A) Leq are required.
At night, sleep disturbance is the main consideration and
available data suggest a bedroom noise limit of 35 dB(A) Leq.
Data from surveys of community noise annoyance lead to the
recommendation that general daytime outdoor noise levels of less than
55 dB(A) Leq are desirable to prevent any significant community
annoyance. This is consistent with speech communication requirements.
At night, a lower level is desirable to meet sleep criteria; depending
upon local housing conditions and other factors this would be in the
order of 45 dB(A) Leq.
1.2 Recommendations for Further Studies
Considerable research aimed at improving the scientific basis and
application of environmental health criteria for noise is in progress
in many countries. However, there are certain areas where present
national and international efforts do not appear adequate. Thus,
further studies should include:
(a) The identification of long-term health effects due to high
level industrial noise and lower level general environmental noise.
The potential contribution of noise stress to the general morbidity of
the population, the ability of people to adapt to environmental noise,
and the possibilities of noise-induced disease must be established not
only for the working population, but also for the more vulnerable
population segments, including the elderly, pregnant women, people
undergoing medication, particularly with ototoxic drugs such as
salicylates, quinine, and certain antibiotics, and those generally
under stress. The possibility that the disturbance of sleep by noise
can result in definite health impairment should be examined as part of
these investigations.
(b) Studies on young people over many years prior to, and during,
occupational noise exposure to find out to what extent changes in
hearing acuity during adolescence are attributable to normal growth or
to environmental conditions, to learn about noise susceptibility in
childhood, and to obtain data on the progressive effects of noise
(including high-level music and other leisure-time sounds) on the
"normal" hearing level of the population. Monitoring of the total
noise exposure of these groups over the whole observation period would
be part of these studies. Similar studies in nonindustrialized
countries would be of particular value.
(c) Work on the development of sensitive hearing tests and on
tests to evaluate the problem of individual susceptibility to noise,
since pure tone audiometry is only a crude technique for measuring
hearing acuity and for detecting pathological damage.
(d) Longitudinal studies of communities exposed to major changes
in environmental noise to refine existing dose-response (noise-
annoyance) relationships and to include the effects of adaptation and
societal changes on public reaction to noise. Attention should be
given to the study of the response of specially vulnerable segments of
the population.
The methods of study should be internationally uniform, as far as
is feasible, to allow pooling of data and broader interpretation of
the results.
2. PROPERTIES AND MEASUREMENT OF NOISE
Noise is considered as any unwanted sound that may adversely
affect the health and well-being of individuals or populations.
Physically, sound is a mechanical disturbance propagated as a
wave motion in air and other elastic or mechanical media such as water
or steel.
Physiologically, sound is an auditory sensation evoked by this
physical phenomenon. However, not all sound waves evoke an auditory
sensation: for example, ultrasound has a frequency too high to excite
the sensation of hearing.
The physical properties and perception of sound or noise are
expressed and measured in different concepts and units.
2.1 Physical Properties and Measurements
Sound waves involve a succession of compressions and rarefactions
of an elastic medium such as air. These waves are characterized by the
amplitude of pressure changes, their frequency, and the velocity of
propagation. The speed of sound (c), the frequency (f), and
the wavelength (lambda), are related by the equation
lambda = c/f
A mechanical energy flux accompanies a sound wave, and the rate
at which sound energy arrives at, or passes through, a unit area
normal to the direction of propagation is known as the sound
intensity, I. In a free sound field, the sound intensity is related
to the root mean square a sound pressure, p, and the density of the
medium, q, by the expression
p2
I =
q c
Sound intensity is normally measured in watts per square metre
(W/m2). The total sound energy emitted by a source per unit time is
known as the sound power, P, and is measured in watts.
Sound intensities of practical interest cover a very large range
and are therefore measured on a logarithmic scale. The relative
intensity level of one sound with respect to another is defined as 10
times the logarithm (to the base 10) of the ratio of their
a The square root of the mean value of the squares of the
instantaneous values of a quantity. For a periodic variation, the
mean is taken over one period.
intensifies. Levels defined in this way are expressed in decibels
(dB). Any acoustic quantity that is related to sound energy, e.g.,
power, intensity, or mean square pressure, may be expressed as a
decibel level. To establish an absolute level, a reference value must
be agreed. Thus, the sound pressure level of a sound with a mean
square sound pressure p2 is:
Table 1. Table for combining intensity levels
Excess of stronger Add to the stronger to get
component combined level
d8 d8
0 3.0
1 2.5
2 2.0
3 1.8
4 1.5
5 1.2
6 1.0
7 0,8
8 0.7
9 0.6
10 0.5
p 2
Lp = 10 log10 ( ) dB
pref
where the reference pressure pref has an internationally agreed
value of 20 micropascals (µPa) (ISO, 1959). The reference values for
sound power level and sound intensity level are 10-12 watts and
10-12 W/m2, respectively (ISO, 1963). Sound levels are expressed in
decibels (dB) relative to the international standard reference
quantities, unless otherwise stated (dB re: 20 µPa).
Whereas sound intensities or energies are additive,b sound
pressure levels (SPL) (in decibels) have to be first expressed as mean
square pressures, and then added. The summation of sound pressure
levels can be easily performed by using the following equation:
Lp1 Lp2 Lp3
Lp = 10 log10 [ + + .... ] dB
1010 1010 1010
b Such combinations of decibel values may be simplified by using
Table 1.
A simple example will illustrate the use of this equation. If two
sound sources of 80 dB SPL each have to be combined, then
L = 10 log10 [108 + 108]
= 10 log10 2 + 80 = 10 X 0.301 + 8 = 83 dB
It is only when two sources generate similar levels that there is
a significant increase in level when the sources are combined. The
example just quoted gave a 3 dB increase. If there is any difference
in the original, independent levels, the combined level will exceed
the higher of the two levels but by less than 3 dB. When the
difference between the two original levels exceeds 10 dB, the
contribution of the quieter source to the combined noise level is
negligible.
Sound is measured with a microphone that generates a voltage
proportional to the acoustic pressure acting upon it. This signal can
be measured and analysed using conventional electronic
instrumentation. A sound level meter is usually a portable, self-
contained instrument incorporating a microphone, amplifiers, a
voltmeter and attenuators, the whole of which carl be calibrated to
read sound pressure levels directly. Intensity levels and power levels
can be derived from sound pressure level measurements if required.
The sound at a given location can be completely described in
terms of the history of the sound pressure fluctuation. If this
fluctuation is periodic, its fundamental frequency is the number of
repetitions per second, expressed in hertz (Hz). Most real periodic
cycles are quite complex and consist of a component at the fundamental
frequency and components at multiples of this basic frequency, known
as harmonics.
The simplest kind of sound, known as a pure tone, has a
sinusoidal pressure cycle that is completely defined in terms of a
single frequency and pressure amplitude (a more precise definition
would also include phase which effectively defines the starting point
in time, but this is usually of little or no interest).
Pure tones are relatively rare -- perhaps the nearest
approximation is the sound of a tuning fork. Most musical sounds are
periodic but contain many harmonics. Analytically these may be
expressed as a sum of harmonically related components. This assembly
is known as the frequency spectrum of the sound, and it specifies how
the energy in the periodic sound is concentrated at certain discrete
frequencies. The frequency distribution of sound energy is measured by
electronic filters.
Although some kinds of machinery produce sound that is largely
periodic, most noise is nonperiodic, i.e., the sound pressure does not
oscillate with time in any regular or predictable way. Such sound is
said to be random. Examples of random sound include the roar of a jet
engine, the rumble of distant traffic, and the hiss of escaping steam.
The energy of random sound is distributed continuously over a range of
frequencies instead of being concentrated at discrete values, so that
its frequency spectrum may be depicted as a curve of energy density
plotted against frequency.
Frequency is related, but not identical, to the subjective pitch.
Any periodic sound has a tonal character that can be ascribed a
particular musical note. The note is basically defined by the
fundamental frequency of the sound. For example, the note A above
middle C on the piano has a fundamental frequency of 440 Hz. On the
other hand, random sound has no distinct pitch, being characterized as
a nondescript rumbling, rushing, or hissing noise, or low and high
frequency noises depending upon the range of frequencies present.
Human hearing is sensitive to frequencies in the range of about
16-20 000 Hz (the "audiofrequency range"). The audible frequency range
is covered by 10 octave bands. An octave is the frequency interval the
upper limit of which is twice the lower limit. The so-called
"preferred frequencies" at the centres of the standardized octave
bands are spaced at octave intervals from 16 to 16 000 Hz (ISO,
1975a). It should be noted that the limits of the octave bands are
f/square root 2 and f square root 2, where f is the centre
frequency. The octave band level at a particular centre frequency is
the level of the sound measured when all acoustic energy outside this
band is excluded. One-third octave band filters, widely used for
noise assessment purposes, subdivide each octave interval into three
parts and provide a more complete description of the sound spectrum.
In order to measure sound pressure level, the mean square
pressure must be averaged over a certain period of time. For steady
sounds, the choice of averaging time is immaterial providing that it
is long compared with the time period of sound pressure fluctuations.
Standard sound level meters normally incorporate "fast" and "slow"
response settings corresponding to averaging times of approximately
0.1 and 1.0 second respectively (IEC, 1973a) (section 2.2.4).
Impulsive noise consists of one or more bursts of sound energy,
each of a duration of less than about one second (ISO, 1973a). Sources
of impulsive noise include impacts of all kinds, e.g., hammerblows,
explosions, and sonic booms. These may be heard singly or, as in the
case of a stamping press, repetitively. To characterize such sounds
acoustically, it is necessary to estimate the peak sound pressures
together with the duration, rise time, repetition rate, and the number
of pulses. The mean square pressure of such sounds may change so
rapidly that it cannot be measured with a conventional sound level
meter, even using the "fast response" (0.1 sec) setting. For more
accurate measurements, a 35-millisecond averaging time is specified
for standard "impulse" sound level meters (IEC, 1973b). The averaging
time of the inner ear is very short (about 30 microseconds) and some
new impulse sound level meters have "peak hold" settings with an
averaging time of 20 microseconds.
2.2 Sound Perception and its Measurement
2.2.1 Loudness and loudness level
The physical magnitude of a sound is given by its intensity and
its subjective or perceived magnitude is called its loudness. Loudness
depends on both intensity and frequency and the average quantitative
relationship between these factors has been deduced by experiment (see
for example Fletcher & Munson, 1933; Stevens, 1955).
The basic unit of loudness is the sone which is defined as the
loudness of a 1000 Hz pure tone heard at an SPL of 40 dB re: 20 µPa
under specified listening conditions (ISO, 1959). Two sones equal
twice the loudness of one sone and so on. For sound at a particular
frequency, at least over a significant fraction of the practical
intensity range, loudness is proportional to some power of the sound
intensity. This is the power law of loudness which is in general
accordance with the Weber-Fechner law (Stevens, 1957b). In the mid
audiofrequency range, the exponent in the power law is such that a
twofold change in loudness corresponds to a tenfold change in
intensity, i.e., a 10 dB change in level (Stevens, 1957a). At low
frequencies, loudness changes more rapidly with changes in level. This
is demonstrated in Fig. 1, which shows a standard set of equal
loudness contours for pure tones (Robinson & Dadson, 1956; ISO, 1961),
each line showing how the SPL of the tone must be varied to maintain a
constant loudness. Each curve, in fact, corresponds to a particular
loudness in phons. The loudness of a sound, in phons, is, by
definition, equal to the SPL of that 1000 Hz tone which is equally
loud -- again under specified listening conditions (ISO, 1959). For
practical purposes, the relationship between the phon and sone scales
may be expressed as:
phon = 40 + log2 (sone)
2.2.2 Calculation and measurement of loudness level
Ideally, sound measurement meters should give a reading equal to
loudness in phons bu it is difficult to achieve this objective,
because the human perception processes are complex. Nevertheless,
procedures have been developed and adopted as international standards
(ISO, 1975b) but, as they are too complex to be incorporated into a
simple measurement meter, they are rarely used in practice, except
where the highest possible precision is required.
For most practical purposes, a much simpler approach is used. A
filter is used to weight sound pressure level measurements as a
function of frequency, approximately in accordance with the frequency
response characteristics of the human ear, i.e., energy at low and
high frequencies is de-emphasised in relation to energy in the mid-
frequency range. Most precision sound level meters incorporate three
selectable filters labelled A, B, and C (IEC, 1973a) and sometimes
D-filter (see section 3.7.2) (IEC, 1973b), the characteristics of
which are illustrated in Fig. 2. The A, B and C filters are intended
to match the ear-response curves at low, moderate, and high loudness
respectively. However, extensive experience has shown that the
A-filter usually provides the highest correlation between physical
measurements and subjective evaluations of the loudness of noise.
Levels on the A-scale are also measured in decibel units and are
commonly expressed as dB(A), a convention that is used throughout this
document.
The A-weighting is used for sound measurements in a variety of
situations, as it is widely accepted that the A-weighted sound
pressure level, Lp(A), is a reasonably reliable and readily measured
estimate of loudness (Botsford, 1969; Young & Peterson, 1969). It must
be emphasized that this in only true for broadband sounds with no
spectral concentrations of energy, in which case Lp(A) is typically
some 10 decibel units lower than loudness in phons. For narrow
frequency range sounds, considerable care must be exercised in the
interpretation of A-weighted sound pressure level readings, since they
may not accurately reflect the loudness of the sound. It should be
noted that the A-scale has been adopted so generally that sound levels
frequently quoted in the literature simply in dB are in fact
A-weighted levels. Furthermore, many general purpose sound level
meters are restricted solely to A-weighted measurements (IEC, 1961).
2.2.3 Sound level and noise level
The phrase "noise level' is widely used by laymen to describe the
severity of an environmental noise. In acoustics, the word "level"
should be reserved for all quantities expressed on a decibel scale. In
this document, as is now common practice in many countries, the
phrases "sound level" and "noise level" refer to decibel scales that
account for human hearing characteristics (the A-weighted SPL scale
being the most widely used). Care should be exercised to distinguish
between sound pressure level, sound power level, sound intensity
level, and sound or noise level.
2.2.4 The time factor
Sounds can appear to be steady to the human ear because the
auditory averaging time is inherently long, much longer than the
acoustic cycle times. Similarly, sound level measurements can be made
to appear steady by selecting a suitably long averaging time. On
precision sound level meters the "slow" value is appreciably longer
than the auditory averaging time and is used to obtain a steady
reading, when the signal level audibly fluctuates at a rapid rate. The
"fast" response time is of the same order as that of the ear.
Sound level fluctuations, which can be smoothed out by the use of
the slow response setting, are usually ignored for noise assessment
purposes. However, difficulties arise when "slow response" readings
vary significantly with time, as they do in many environments. Often,
such level fluctuations are small but in some situations, for example,
near to roads and airports, the fluctuations can be measured in tens
of dB; the rate of fluctuation can also vary widely.
2.2.5 Noise exposure scales
In many noise indices that are well correlated with the
subjective effects of interest, various underlying acoustic and
nonacoustic factors have been combined in different ways. These
composite indices are discussed in section 3.7 and the present section
is restricted to the question of the physical measurement of noise.
The basic objective of measurement is to quantify overall noise
exposure in the simplest possible terms. The physical characteristics
of a noise which, on the basis of intuition and laboratory experiment,
might be expected to influence its subjective effects include the
following: loudness level (recognizing average and peak values
together with impulsive characteristics where appropriate); total
noise "dose"; level fluctuation amplitudes; and rates of fluctuation.
Clearly, the acoustic variables alone have many dimensions; the
following two procedures are commonly used to measure some of them.
2.2.6 Equivalent continuous sound pressure level
To measure an average sound level the meter averaging time is
extended to equal the period of interest T, which may be an
interval of seconds, minutes, or hours. This gives the equivalent
continuous sound pressure level (Leq) derived from the mathematical
expression:
1 T Lp(A)(t)
Leq = 10 log10 integral db(A)
T o 1010dt
Because the integral is a measure of the total sound energy during the
period T, this process is often called "energy averaging". For
similar reasons, the integral term representing the total sound energy
may be interpreted as a measure of the total noise dose. Thus, Leq
is the level of that steady sound which, over the same interval of
time, contains the same total energy (or dose) as the fluctuating
sound.
Equivalent continuous sound level is gaining widespread
acceptance as a scale for the measurement of long-term noise exposure.
For example, it has been adopted by the International Organization for
Standardization for the measurement of both community noise exposure
(ISO, 1971) and hearing damage risk (ISO, 1975c). It also provides a
basis for more elaborate composite noise indices discussed in
subsequent sections including the day-night sound level (Ldn)
(section 3.7.3.3).
Following the introduction of jet aircraft into commercial
service, it was suggested that the then existing loudness scales were
inadequate for aircraft noise rating purposes. An alternative scale of
perceived noise level (PNL) was developed, with units dB(PN) (Kryter,
1959). This was derived from the loudness level procedure of Stevens
(1956) on the grounds that the attribute of perceived noisiness
defined as the "unwantedness" of the sound was different and more
relevant to aircraft noise than loudness. In fact, the only difference
between the calculations involved was the use of different frequency
response curves. As research progressed towards legislation for
aircraft noise emission control (US Federal Aviation Regulations,
1969; ICAO, 1971), the perceived noise level scale was modified to
include special weightings for "discrete frequency components", i.e.,
irregularities in the spectrum caused by the noticeable periodic
components of engine fan and compressor noise, and the duration of the
sound (Kryter & Pearsons, 1963). This modified quantity, known as
effective perceived noise level, is expressed in dB(EPN).
Because PNL could not be measured with a simple meter, a parallel
development was the D-weighting filter, with characteristics based on
an equal noisiness (rather than an equal loudness) frequency response
curve (IEC, 1976). This filter is available on some sound level meters
and is intended for aircraft noise monitoring purposes.
2.2.7 Level distribution
A widely used method of recording the variations in sound level
is that of level distribution analysis, sometimes called statistical
distribution analysis. This yields a graph of the percentage of the
total time (T) for which any given sound level is exceeded; such
information can be summarized by reading specific levels from this
graph. For example L10, L50, and L90, the levels exceeded for
10%, 50%, and 90% of the time, are frequently used as measures of
typical peak, average, and background levels, respectively.
2.3 Sources of Noise
2.3.1 Industry
Mechanized industry creates the most serious of all large scale
noise problems, subjecting a significant fraction of the working
population to potentially hazardous noise levels. This noise is due to
machinery of all kinds and often increases with the power of the
machines. The characteristics of industrial noise vary considerably,
depending on specific equipment. Rotating and reciprocating machines
generate sound that is dominated by periodic components; air moving
equipment tends to generate broad-band random sounds. The highest
noise levels are usually caused by components or gas flows that move
at high speed [e.g., fans, steam pressure relief valves) or by
operations involving impacts (e.g., stamping, riveting, road
breaking). In industrial areas, the noise usually stems from a wide
variety of sources, many of which are of a complex nature.
Machinery noise generation mechanisms are reasonably well
understood and the technical requirements for low noise output in new
machinery can usually be specified. The difficulty of reducing the
noisiness of existing equipment is a serious obstacle to the
improvement of working environments.
2.3.2 Road traffic
The noise of road vehicles is mainly generated from the engine
and from frictional contact between the vehicle and the ground and
air. In general, road contact noise exceeds engine noise at speeds
higher than 60 km/h. The level of noise from traffic is correlated
with the traffic flow rate, the speed of the vehicles, and the
proportion of heavy vehicles, which, together with motorcycles, tend
to be about twice as loud as motor cars.
Special problems arise in areas where the traffic movements
involve a change in engine speed and power, such as at traffic lights,
hills, and intersecting roads.
2.3.3 Rail traffic
Trains generate a relatively low frequency noise but variations
are present depending upon the type of engine, wagons, and rails.
Impact noises are generated in stations and marshalling yards because
of shunting operations. The introduction of high speed trains has
created special noise patterns, especially when such trains pass over
bridges or other structures that cause amplification of the noise. At
speeds of around 200 km/h, the proportion of high frequency sound
energy increases and the sound is perceived to be similar to that of
overflying jet aircraft. Furthermore, with increasing speed the onset
of the noise is more sudden than with conventional trains. Thus,
severe noise problems have been created in countries where high speed
trains operate, notably in Japan.
2.3.4 Air traffic
Aircraft operations have caused severe community noise problems.
Introduction of the early turbojet transport aircraft led to a surge
of community reactions against commercial airports, and more research
has been devoted to aircraft noise than to any other environmental
noise. The noise generation is related to air velocity, which is an
important feature for aircraft and aircraft engines. Fast moving
bodies such as propellers and compressor blades, as well as jet
exhaust gases are very efficient sources of noise.
Aircraft noise is characterized by a wide frequency range with
the periodic components of rotating machinery noise (fans, propellers,
and rotors) superimposed on a general broadband background noise. For
jet aircraft, the periodic components tend to be more dominant on
landing than on take-off when the broadband exhaust noise
predominates. For aircraft with quiet engines, noise from the hull may
become dominant when landing.
Aircraft noise control depends critically on the reduction of
engine component and gas velocities. The high by-pass ratio turbo-fan
engines of newer aircraft with components operating at significantly
lower speeds have resulted in a reduction in aircraft noise levels,
and offer considerable promise of less noisy airports, as they
gradually replace older equipment.
2.3.5 Sonic booms
The sonic boom is a shock wave system generated by an aircraft,
when it flies at a speed slightly greater than the local speed of
sound. The shock wave extends from an aircraft throughout supersonic
flight in a roughly conical shape. At a given point, the passage of
the shock wave causes an initial sudden rise in atmospheric pressure
followed by a gradual fall to below the normal pressure and then a
sudden rise back to normal. These pressure fluctuations, when
recorded, appear in their typical form as so-called N-waves. When they
occur with a separation greater than about 100 milliseconds, the sonic
boom has a characteristic double sound. Rise times from less than 0.1
to 15 milliseconds and durations up to 500 milliseconds have been
recorded for typical sonic booms generated by military or civilian
aircraft.
Low intensity sonic booms with longer rise times are perceived as
a noise similar to distant thunder. As the rise time increases, the
noise becomes progressively sharper and attains a "dry cracking"
character. An aircraft in supersonic flight trails a sonic boom that
can be heard over more than 50 km on either side of its ground track
depending upon the flight altitude and the size of the aircraft
(Warren, 1972).
2.3.6 Construction and public works
Building construction and earth works are activities that cause
considerable noise emissions. A variety of sounds is present from
cranes, cement mixers, welding, hammering, boring, and other work
processes. Construction equipment is often poorly silenced and
maintained, and building operations are frequently carried out without
considering the environmental noise consequences.
2.3.7 Indoor sources
Indoor noise originates from a variety of sources such as air
conditioners, waste disposal units, and furnaces. Noises from outdoor
sources also penetrate through windows and weaknesses in building
structures, although with some attenuation. Within a building, noise
is transmitted from room to room through ventilation ducts and through
the building structure itself. Of particular interest is the low
frequency sound emitted by ventilation or air conditioning equipment.
This noise, which often has discreet frequencies, can be generated by
fans, vibrations in conducting ducts, or at air outlets.
2.3.8 Miscellaneous sources
Apart from the major categories of noise already identified,
which affect a large number of people in the community, many other
sources of noise can be important in individual cases. Firing ranges,
sports fields, and pleasure grounds are examples of fixed sources,
while noises from garbage collection and power-operated lawn-mowers
are other examples of machine-produced noise that can interfere with
man's comfort and rest. Neighbourhood noise also includes noise from
domestic animals, farm equipment, boats, and the sirens of emergency
vehicles.
3. EFFECTS OF NOISE
3.1 Noise-induced Hearing Loss
3.1.1 Hearing impairment
Normal hearing is regarded as the ability to detect sounds in the
audiofrequency range (16-20 000 Hz) according to established
standards. However, individual hearing ability in man varies. Some of
these variations may be attributed to the effects of different
environmental influences (Roberts & Bayliss, 1967); in industrialized
countries, women generally have better hearing than men (Kylin, 1960;
Dieroff, 1961; Gallo & Glorig, 1964).
As a rule, hearing sensitivity diminishes with age, a condition
known as presbyacusis (Glorig & Nixon, 1962). Consequently,
corrections for aging should be considered when examining data on
hearing loss caused by noise exposure. However, the literature
reflects controversy concerning the degree to which cumulative effects
of noise exposure in everyday life may contribute to eventual hearing
loss (socioacusis), thus obscuring the effect due to aging alone.
Moreover, there is considerable variation between individuals in both
the amount and rate of hearing loss due to aging. The general pattern
of progression of presbyacusis has been quite well-established, and
data are available in numerous reference sources (US National
Institute for Occupational Safety and Health, 1972; US Environmental
Protection Agency, 1973a, 1974). Loss of hearing sensitivity due to
aging occurs mainly at the higher audiometric frequencies and is
almost invariably bilateral (i.e., in both ears).
3.1.1.1 Hearing level, noise-induced threshold shift, and hearing
impairment
In order to discuss the effects of noise on hearing, it is
necessary to differentiate between hearing level, noise-induced
threshold shift (NITS), and hearing impairment.
Hearing level refers to the audiometric threshold level of an
individual or group in relation to an accepted audiometric standard
(ISO, 1975d) and is sometimes termed "hearing loss". Noise-induced
threshold shift is the quantity of hearing loss attributable to noise
alone, after values for presbyacusis (including socioacusis) have been
subtracted. These values may differ slightly according to where and
how the presbyacusis data were collected (see for example Hinchcliffe,
1959; Gallo & Glorig, 1964; Spoor, 1967; US National Centre for Health
Statistics, 1975).
Hearing impairment is generally referred to as the hearing level
at which individuals begin to experience difficulty in leading a
normal life, usually in relation to understanding speech. Hearing
impairment has been defined in the USA as an arithmetic average of
26 dB or more hearing loss at the frequencies, 0.5, 1, and 2 kHz (the
definition is currently being revised); in Poland, it is defined as
30 dB or more at 1, 2, and 4 kHz (after age correction), and in the
United Kingdom, it is 30 dB or more at 1, 2, and 3 kHz. It should be
noted that a damage risk criterion of 30 dB at 1, 2, and 4 kHz may be
more protective than a criterion of 26 dB at 0.5, 1, and 2 kHz,
because hearing loss at high frequencies is usually greater than the
loss at 500 Hz.
3.1.1.2 Noise-induced temporary threshold shifta
A person entering a very noisy area may experience a measurable
loss in hearing sensitivity but recover some time after returning to a
quiet environment. This phenomenon can be measured as a shift in
audiometric thresholds, and is called noise-induced temporary
threshold shift (NITTS).
Recovery from NITTS depends on the severity of the hearing shift,
individual susceptibility, and the type of exposure. If recovery is
not complete before the next noise exposure, there is a possibility
that some of the loss will become permanent. Information on NITTS has
been used for two purposes: first, to predict noise levels that could
be permanently damaging to the ear, and second, to attempt to predict
individual susceptibility to hearing loss caused by excessive noise.
Measurements of NITTS are made by comparing pre- and post-exposure
audiograms. The extent of NITTS, for the same exposure, varies
considerably between individuals. Recovery can take hours, days, or
even weeks after exposure. It should be noted that NITTS can be
experienced by individuals who already suffer from permanent noise-
induced hearing losses. Thus, when assessing permanent damage,
sufficient recovery time in the quiet should be allowed before
audiometry.
It would appear from recent investigations that the relationship
between NITTS and the noise-induced permanent threshold shift (NIPTS)
is very uncertain and that damage-risk criteria should be based on
epidemiological rather than on NITTS data.
3.1.1.3 Noise-induced permanent threshold shift
The typical pattern of NIPTS usually involves a maximum loss at
around 4000 Hz. Because the loss is sensorineural, it is seen in both
air and bone conduction audiograms. Noise-induced hearing loss is not
an abrupt process but occurs gradually, usually over a period of
years. The rate and extent of loss depends on the severity and
duration of the noise exposure, but individual susceptibility also
a Sometimes called auditory fatigue.
seems to have a considerable effect on the rate of progression. Noise-
induced losses are rather similar to losses due to aging and the two
types of losses are difficult, if not impossible, to distinguish.
Fig. 3 shows the progression of noise-induced hearing loss observed in
workers with increasing duration of exposure to high noise levels
(Johansson, 1952).
The first stages of noise-induced hearing loss are often not
recognized because they do not impair speech communication ability. As
the loss becomes greater, difficulty may be encountered particularly
in noisy locations.
Hearing of important sounds other than speech, such as door
bells, telephones, or electronic signals, may also be impaired. With
further loss in hearing, speech communication may be severely
affected.
3.1.1.4 Incidence of noise-induced permanent hearing loss
The prevalence of hearing loss among workers in noisy industries
has been recognized since ancient times, and excessively loud noises
are popularly described as deafening. Clinical observations of noise-
induced hearing loss have been reported for more than a century, but
it is only recently that the problem has been studied intensively. It
has been suggested that even though people exposed to intense noise
frequently experience a substantial noise-induced temporary threshold
shift, sometimes accompanied by tinnitus (ringing in the ears), the
fact that very often such symptoms seem to disappear within a short
time may lead them to believe that no permanent damage has occurred.
However, neither the subjective loudness of a noise, nor the extent to
which the noise causes discomfort, annoyance, or interference with
human activity, are reliable indicators of its potential danger to the
hearing mechanism.
As there is considerable variation among individuals, it is very
difficult to identify a safe limit of noise exposure that can be
applied for all ears.
Most current knowledge of hearing loss due to noise has been
obtained from industrial surveys. There is also evidence that non-
industrial exposure to noise can be harmful. Results of several
studies have confirmed that high levels of "rock and roll" and similar
music can produce considerable temporary threshold shift and even
permanent threshold shift. Audiograms of "pop-musicians" typically
show losses at 400 Hz in both ears (Kowalczuk, 1967). It has also been
shown that men and women are equally at risk of hearing damage, when
exposed to over-amplified music (Fletcher, 1972). Other non-
occupational activities that can contribute to hearing loss include
shooting and motorcycling.
3.1.2 Relation between noise exposure and hearing loss
In the normal auditory process, sound vibrations in the air
travel through the ear canal and cause the eardrum to vibrate. The
vibrations are then transmitted by the bones of the middle ear to the
sensory organ of the inner ear (cochlea). Here they are transduced by
hair cells into nerve impulses and transmitted to the brain, where
they are perceived as sound or noise.
Blasts and other intense or explosive sounds can rupture the
eardrum or cause immediate damage to the structures of the middle and
inner ear, while hearing loss due to prolonged noise exposure is
generally associated with destruction of the hair cells of the inner
ear. The severity of noise-induced hearing loss depends on both the
location and the extent of damage in the organ of Corti, which, in
turn, depend on the intensity and frequency of the sound stimulus. The
higher the frequency, the nearer the point of maximum displacement of
the basilar membrane is to the base of the cochlea where the basilar
membrane is narrowest. This point is shifted towards the apex of the
cochlea as the stimulus frequency decreases. The maximum stimulation
of cells occurs at the point of maximum displacement. A large part of
the upper cochlea is responsive to low frequency stimulation and loss
of hair cells can be quite extensive without significant loss in low
frequency sensitivity. On the other hand, much more localized portions
of the basal region of the cochlea are responsible for high frequency
sound sensation and loss of hair cells in these lower portions results
in significant losses of high frequency sensitivity (Miller, 1971a).
The number of hair cells damaged or destroyed increases with
increasing intensity and duration of noise and, in general,
progressive loss of hair cells is accompanied by progressive loss of
hearing.
Even though numerous experiments have been performed with
animals, the mechanisms involved in the destruction of the Corti organ
are not completely clear, although several explanations have been
proposed. For example, mechanical stresses could destroy cells,
repeated circulatory troubles through vascular contractions could
deprive cells of an appropriate blood supply; an increase in local
temperature could damage proteins, and repeated stimuli could exhaust
the metabolic supply of cells. Various theories have been reviewed by
Ward (1973).
An important fact is that noise-induced hearing loss is of a
neural type involving irreversible injury to the inner ear.
Furthermore, such losses are almost always bilateral.
3.1.2.1 Laboratory studies
Laboratory studies on temporary and permanent hearing loss and on
the anatomy of the noise-damaged inner ear have been carried out on a
number of animal species. Temporary hearing loss studies on human
subjects have included a variety of noise exposure patterns, including
noises of different spectra, interrupted noise patterns, and short-
duration noise exposures. In extrapolating the results of such studies
to permanent hearing loss in man, it has always been necessary to
consider: (a) temporary versus permanent threshold shift in man; (b)
permanent threshold shift in man versus permanent threshold shift in
animals; and (c) anatomical damage in animals versus permanent
threshold shift in man. However, it should be noted that a thorough
knowledge of such relationships has not been necessary. For example,
in using animals to study the cumulative effects of noise, it has not
been necessary to assume that the absolute sensitivity of animals and
man to noise is the same, but merely that the relative sensitivity of
animals to alternative noises of specified temporal patterns is
similar to that of man.
Experimental studies have resulted in the following general
observations:
(a) There is considerable variability among individuals in
susceptibility to temporary hearing loss, the rate at which temporary
hearing loss approaches its asymptotic level, and the rate of
recovery.
(b) Temporary hearing losses in man are most pronounced at
frequencies slightly above the predominant frequency of the noise
stimulus.
(c) In most cases, the rate of increase of, and subsequent
recovery from, temporary hearing loss is different for impact noises
and for steady noise. NITTS from impulse noise increases more slowly
than NITTS from steady noise (Ward et al., 1961) and recovery is
slower (Cohen et al., 1966).
(d) In general, the equal energy rule (section 3.1.3) has been
found to be compatible with experimental results for uninterrupted
exposures to steady noise. However, it may not always be the best
predictor of NITTS with regard to the audiometric frequency since it
tends to overestimate NITTS below 2000 Hz and underestimate losses
above 2000 Hz (Yamamoto et al., 1968). Although NITTS from interrupted
noise may be overestimated (Ward, 1970), it is thought that the rule
gives a good prediction of NIPTS from interrupted noise (Burns &
Robinson, 1970).
(e) Audiograms of persons exhibiting temporary hearing loss in
laboratory studies tend to be similar to those of persons exposed to
comparable noise over a period of several years (Nixon & Glorig,
1961).
3.1.2.2 Occupational hearing loss
Several reports have been published on the subject of
occupational hearing loss (Atherley et al., 1967; Burns & Robinson,
1970; King, 1971; Robinson, 1971; Stone et al., 1971; Baughn, 1973;
Burns, 1973; Paschier-Vermeer, 1974; Sulkowski, 1974).
All these studies were cross-sectional audiometric studies and
many incorporated surveys of noise exposure. Specific occupational
groups were usually studied, including workers in heavy industry,
shipyards, textiles, jet-cell test rooms, foundries, transportation,
and forestry. Some definition of hearing impairment was generally
applied in order to define a percentage of people with hearing loss.
Audiograms were usually compared with so-called "normal" thresholds.
In this respect, presbyacusis was often accounted for. In many cases,
efforts were made to screen the data to exclude those persons who had
previously held noisy jobs, possible nonoccupational noise exposures,
and otological abnormalities. In some studies, such persons were
purposely included in order to provide a realistic estimate of hearing
levels in a typical noise-exposed population.
Virtually every study revealed that workers exposed to intense
noise daily, for several years, showed noise-induced hearing loss
fitting the classic pattern. Considerable hearing loss was rare at
lower frequencies but frequent at higher frequencies.
In the studies for which noise exposure levels were known, a
clear relationship was generally seen between increasing incidence of
hearing loss and increasing noise level. In groups exhibiting
considerable noise-induced hearing loss, the variation of audiometric
thresholds was generally higher than in groups not exposed to noise.
Cases of sudden deafness occurring after long-term exposure to noise,
without previous impairment, have been reported in Japan (Kawata &
Suga, 1967) and may indicate special susceptibility.
Taking into account duration of exposure and age as well as other
pathological conditions, Rey (1974) found that the proportion of
workers with noise-induced deafness (defined as 25 dB average loss at
0.5, 1, and 2 kHz) was as high as 60% in the metal industry (noise
levels equal to and above 95 dB(A)). Cohen et al. (1970) compared the
mean hearing levels of exposed workers with those of a control group
for several noise intensities and several durations of exposure and
found that noise levels between 85 and 88 dB(A) could be harmful to
the ear, and that, even at 75 dB(A), there was some loss of hearing.
According to two other studies performed in industry, there is a
definite risk of hearing damage associated with prolonged exposure to
noise levels between 85 and 90 dB(A) (Roth, 1970; Martin et al.,
1975).
Fig. 4 compares the percentages of workers with hearing
impairment as a function of age for unexposed groups and for groups
exposed to occupational noise levels of 85, 90, and 95 dB(A) (NIOSH,
1973b). In this case, hearing impairment is defined as an average
hearing loss greater than 25 dB(A), at frequencies of 1, 2, and 3 kHz.
3.1.2.3 Factors that may influence the incidence of noise-induced
permanent threshold shift
Certain people who live in remote and generally quiet areas of
the world have been found to have unusually acute hearing in
comparison with members of urban populations in corresponding age
groups (Rosen et al., 1962). However, it is not clear whether such
audiometric differences are due to the lack of noise exposure alone.
Differences in the patterns of hearing found between communities that
are widely separated geographically and culturally may result from
cultural, dietary, and genetic factors and differences in general
environment (Rosen et al., 1962; Rosen & Rosen, 1971).
Although it has been suggested that older people are more
susceptible to NIPTS (Kryter, 1960), there is no clear experimental
evidence that this is so (Kupp, 1966; Nowak & Dahl, 1971). Indeed,
studies by Schneider et al. (1970) and Davis (1973) indicate that
there is probably no causal relationship between age and
susceptibility to NIPTS, at least in people of working age.
There is some controversy in the literature as to whether
pathological changes in the middle ear protect the inner ear from
noise-induced damage, or whether they may instead increase the chance
of noise-induced hearing loss. Some authors have expressed the view
that in cases of middle ear damage, bone conduction becomes more
effective and that the defence action of the middle ear muscles is
impaired (Mounier-Kuhn et al., 1960; Ward, 1962; Dieroff, 1964; Mills
& Lilly, 1971). In contrast, others have reported cases where noise-
induced hearing loss was less in damaged ears than in normal ears
(Johansson, 1952).
Variation in individual susceptibility to noise-induced permanent
hearing loss is illustrated by observations from surveys of
occupational hearing loss, which indicate that workers from the same
noisy environment display radically different audiograms, and that
some workers, even after many years of exposure to noise, show little
or no sign of noise-induced hearing loss.
Factors causing such differences in individual susceptibility
could include fatigue of the acoustic reflex, anatomical differences
in the structure of the middle and inner ear, the functional status of
the autonomic system, and latent vitamin B deficiency (Kawata, 1955).
To some extent, the ear is protected from damage by the middle
ear reflex or stapedius reflex. The contraction of the stapedius
muscle changes the movement of stapes which increases the impedance of
the conductive mechanisms. The amount of sound energy delivered to the
inner ear is reduced by about 15-20 dB at low and middle frequencies
(Miller, 1961). The effectiveness of the middle ear reflex as a
protective device varies with the intensity and the spectrum of the
sound. In normal ears, the onset of the reflex occurs at sound levels
of 75-90 dB. In man, the muscle contraction subsides very quickly
after the onset of the sound for frequencies above 3000 Hz, while for
lower frequencies, the contraction can last for a considerable time
(Johansson et al., 1967). Impulsive sounds or sounds with a sudden
onset can penetrate the ear without stimulating the protective
mechanism, because of a time lag in the muscle contraction.
Furthermore, the reflex action weakens with fatigue and thus provides
little protection against prolonged steady sounds. The fact that its
effectiveness also varies considerably among individuals may be
related to variations in individual sensitivity to certain sounds.
Measurements of NITTS have been used to investigate the
protection provided by the stapedius reflex. In patients with
peripheral facial palsy including unilateral stapedius muscle
paralysis, the NITTS after low frequency noise exposure was
significantly greater in the affected ear than in the unaffected ear
(Zakrisson, 1974). However, results of animal studies, in which the
stapedius muscle was severed, contradict these findings (Steffen et
al, 1963; Ferris, 1966).
3.1.2.4 Combined effects of intensity and duration of noise exposure
Most data concerning the long-term hazard of noise are related to
occupational exposure. There is a shortage of information about short-
term exposures, and very little information concerning exposures
lasting longer than 8 h. In order to predict the effects of long-term
noise exposure, investigators have been obliged to extrapolate the
results of field observations and laboratory investigations of NITTS.
It is difficult to establish limits for safe noise exposure, since
predictions using different methods of extrapolation conflict with
each other. The following is a brief review of the bases of some of
the methods used to integrate the combined effects of intensity and
duration.
The equal temporary effects rule is the hypothesis that the NIPTS
due to long-term, daily, steady-state noise exposure is equal to the
average NITTS produced by the same daily noise in healthy young ears
(Ward et al, 1958, 1959). In a later study, Ward (1960) suggested that
metabolic insufficiency induced in the hearing organ by noise might
underlie both the temporary and permanent hearing defects caused by
excessive noise. NITTS studies also tend to support the observation
(reflected in industrial studies of NIPTS) that for a given length of
exposure, frequently interrupted noise is less harmful than continuous
steady-state noise of the same level (Ward et al, 1959; Miller et al.,
1963).
An extension of this theory is that NIPTS is unlikely, if there
is complete recovery from the NITTS before the beginning of the next
day's exposure. An early occupational noise criterion was based on
this assumption (Kryter et al., 1966).
The equal energy rule is the theory that the hazard to hearing is
determined by the total sound energy (the integrated product of sound
intensity and duration) entering the ear each day. This rule has
natural appeal, since the exposure dose is quite simple to assess and,
according to epidemiological data, is reasonably well correlated with
the accumulated physical damage. The rule allows a 3-db increase in a
steady sound level for each halving of the duration (Burns & Robinson,
1970; Ward & Nelson, 1971; US Environmental Protection Agency, 1973b;
Martin, 1976). However, it should be noted that the range of sound
duration covered by this rule might be limited by the need for
protection against possible damage by high level, short duration,
impulsive sounds (section 3.1.3).
Various other theories are based, to a certain extent, on the
equal temporary effect hypothesis. Such criteria are usually
identified by the change in sound level that is necessary for each
doubling of the exposure duration, e.g., the "5-dB rule" means that
the level must be 5 dB less for each doubling of the exposure
duration. The rules most frequently quoted in the literature are:
(a) 3 dB rule: equal energy rule incorporated in ISO standard
1999 (ISO, 1975c);
(b) 5 dB rule: purported to partially compensate for typical
interruptions and intermittency and used in the 1969 Walsh-Healey
Public Contracts Act in the USA (Federal Register, 1969);
(c) 4 dB rule: purported to be more reliable for protection at
higher frequencies than the 5 dB rule and used by the United States
Air Force (US Air Force, 1973); and
(d) 6 dB equal pressure rule, a more conservative criterion
suggested by some research workers (US Department of Health, Education
and Welfare, 1972).
None of the rules (a) to (d), account for a reordering of the
noise exposure pattern, i.e., the predicted risk is independent of the
order in which a sequence of sounds is experienced, even if this
sequence includes periods of quiet. Thus, there is some conflict
between these rules and the equal temporary effect hypothesis.
To simplify different damage risk criteria, noise exposure
histories are frequently expressed as equivalent 8-h continuous
levels. For example, using the equal energy (3 dB) rule, an exposure
of 88 dB for 4 h could be expressed as an equivalent level of 85 dB.
3.1.2.5 Estimation of hearing impairment risk
The hearing loss that may result from noise exposure, can be
expressed in terms of probable NIPTS, or hearing impairment. For
example, the percentage of people who will suffer an NIPTS of 5 dB
(the smallest amount measurable) at the most sensitive frequency
(4000 Hz) may be defined as a function of an equivalent 8-h level
(Fig. 5). From this diagram, an 8-h equivalent level of 75 dB(A) can
be identified as the limit for protection against significant NIPTS
(ISO, 1975c). Since it is often impractical to reduce occupational 8-h
equivalent noise levels to 75 dB(A), practical criteria for "safe"
levels have been based upon less stringent definitions of hearing
impairment or hearing handicap. For example, "damage-risk" has been
defined as the percentage of a population with a given amount of
hearing impairment after corrections have been made for those people
who would "normally" incur losses from causes other than noise
exposure. Table 2 shows the percentage risk and the total percentage
with impaired hearing resulting from various levels of noise and years
of exposure (ISO, 1975c).
3.1.2.6 The importance of high-frequency hearing
It is common practice to assess hearing handicap for compensation
purposes, and even for prevention purposes, in terms of the ability to
understand "everyday" speech. According to the ISO definition (ISO,
1975c), hearing handicap begins with a 25 dB loss averaged for the
frequencies 500, 1000, and 2000 Hz. However, in most languages, speech
includes energy at higher frequencies and therefore good high
frequency hearing is important for speech intelligibility, especially
when listening conditions are less than optimal (i.e., in background
noise or when the speech is disorted in some way) (Kryter et al.,
1962; Harris, 1965; Niemeyer, 1987; Acton, 1970; Kuzniarz, 1974;
Antansson, 1975). Under good listening conditions, impaired hearing
may not diminish speech intelligibility because of the redundancy
(multiplicity of cues) of speech (section 3.2.1). This redundancy is
reduced in noisy conditions or when the speech is muffled, the accent
or the message is unfamiliar, or when these constraints occur in
combination.
Table 2. Percentage of exposed people with impaired hearing as a function of occupational
noise level (Leq (8-h) dB(A))x after different periods of exposure
Period of exposure
Occupational (years)
noise level Cause of impairment
Leq 8-h dB(A) 0 5 10 15 20 25 30 35 40 45
<80 (a) All causes y 1 2 3 5 7 10 14 21 33 50
(b) Occupational noise 0 0 0 0 0 0 0 0 0 0
85 (a) All causes 1 3 6 10 13 17 22 30 43 57
(b) Occupational noise 0 1 3 5 6 7 8 9 10 7
90 (a) All causes 1 6 13 19 23 26 32 41 54 65
(b) Occupational noise 0 4 10 14 16 16 18 20 21 15
95 (a) All causes 1 9 20 29 35 39 45 53 62 73
(b) Occupational noise 0 7 17 24 28 29 31 32 29 23
100 (a) All causes 1 14 32 42 49 53 58 65 74 83
(b) Occupational noise 0 12 29 37 42 43 44 44 41 33
105 (a) All causes 1 20 45 58 65 70 76 82 87 91
(b) Occupational noise 0 18 42 53 58 60 62 61 54 41
Table 2 (contd)
Period of exposure
Occupational (years)
noise level Cause of impairment
Leq 8-h dB(A) 0 5 10 15 20 25 30 35 40 45
110 (a) All causes 1 28 58 76 85 88 91 93 95 95
(b) Occupational noise 0 26 55 71 78 78 77 72 62 45
115 (a) All causes 1 38 74 88 94 94 95 96 97 97
(b) Occupational noise 0 36 71 83 87 84 81 75 64 47
x Based on: ISO (1975c).
y The values in row (a) for Leq < 80 dB(A) are estimates of the percentage of people with
hearing impairment caused by factors other than occupational noise exposure and should be
subtracted from row (a) in all cases to obtain row (b) the percentages of people with
impairment attributable to occupational noise. Impairment is defined as a loss of 25dB or
more averaged for the frequencies 500, 1000, and 2000 Hz.
Example: Out of a group of people exposed to an occupational noise level of 95 dB(A) Leq (8-h)
for 25 years, 39% will exhibit hearing impairment. However, 10% (see c) would have had
impaired hearing without exposure to occupational noise. Thus the risk of occupational noise
damage is 29%.
The use of a simple, unweighted average at 500, 1000, and 2000 Hz
for assessing noise-induced hearing handicap is restrictive because
most hearing loss occurs at higher frequencies. Consequently, the
frequencies 3000 Hz and 4000 Hz are included in damage-risk formulae
by some countries.
3.1.3 Effects of impulsive noise
At present, most know]edge of hearing loss due to impulsive noise
comes from studies of the effects of gunfire (see for example Coles et
al., 1968) with some limited data from industrial situations (Dieroff,
1974; Ceypek & Kuzniarz, 1974). Important properties of impulsive
noise exposure include the peak SPL, duration, rise and decay times,
type of wave form, repetition rate, spectrum, and number of impulses.
The present state of knowledge is that a hazard exists and,
accordingly, that ear protection should be worn when impulsive noises,
measured with appropriate instrumentation, exceed an SPL of 140 dB for
more than 5 milliseconds regardless of rise time, spectrum, or the
presence of oscillatory transients. Higher peak levels may be
tolerable for durations of less than 5 milliseconds. Levels in excess
of 165 dB SPL, even for short durations, are likely to cause cochlear
damage (Acton, 1967; Burns & Robinson, 1970). It should be noted that
the response time of the acoustic reflex (section 3.1.2.3) is of the
order of 100-300 milliseconds, which is too long to give any
protection against such short duration sound (Coles et al., 1968;
Coles & Rice, 1970).
Although it is not common practice to extend the equivalent 8-h
sound level criteria down to impulsive durations, the recent studies
of Rice & Martin (1973) and Martin (11976) suggest that the criteria
based on the equal energy rule, may be applicable to high-intensity
impulsive noise (Fig. 6).
3.1.4 Infrasound and ultrasound
Frequencies below 16 Hz are referred to as infrasonic
frequencies. Perception of sound from 100 Hz down to about 2 Hz is a
mixture of aural and tactile sensations. For example, frequencies
around 10 Hz, can cause discomfort through a modulation of the vocal
cords. Reactions caused by extremely high levels of infrasound can
resemble those of mild stress reaction and may include bizarre
auditory sensations, describable as pulsation and flutter. High levels
of infrasound can cause resonance responses in various organs in the
human body, although the long-term effects of such stimulation are not
known (Johnson, 1973).
The effects of high intensity ultrasound (above 20 kHz and 105 dB
SPL), which will be discussed in a separate document, are reported to
be similar to those observed during stress. However, these effects may
be partly due to associated high (but less than ultrasonic) frequency
sound (Acton, 1967). Although it is usually accepted that levels below
105 dB SPL have no adverse effects, there is evidence from one
experiment, that physiological changes can occur at lower levels
(98-102 dB) (Lisickina, 1968).
3.2 Interference with Communication
3.2.1 Masking and intelligibility
The interference of noise with speech communication is a process
in which one of two simultaneous sounds renders the other inaudible.
The ratio of a given desired signal (speech, music) to that of the
interfering noise will determine whether or not the signal can be
perceived. The higher the level of the masking noise and the more
energy it contains at speech frequencies, the greater will be the
percentage of speech sounds that are inaudible to the listener.
An important aspect of communication interference in occupational
situations is that the failure of workers to hear warning signals or
shouts may lead to injury. Although cases do not appear to have been
documented in the literature, there is anecdotal evidence of such
occurences.
In the last half century, knowledge concerning the masking of
simple signals such as pure tones, narrow bands of noise, and even
isolated phonemes of speech has increased considerably. Empirical
relationships are available that permit accurate prediction of the
audibility for a normal-hearing listener of a particular speech sound
in the presence of a specified noise (Webster, 1969, 1974; Kryter,
1970). However, communication is almost never carried on by means of
single acoustic signals, but rather by a rapid sequence of different
speech sounds, the overall intensity and spectral distribution of
which are constantly shifting; in fact, the same word, when repeated,
may be quite different acoustically. Furthermore, even when the
masking noise is judged to be steady, the energy in different
frequency regions fluctuates from moment to moment.
Most of the sentences of ordinary discourse can be understood
fairly well, even when a large number of individual speech sounds are
masked, because of the redundancy of speech. Even when a particular
sound is masked or even omitted, the word or sentence in which it
occurs may be correctly perceived because the remaining sounds are
sufficient to convey the meaning. However, the interpretation required
to compensate for the masking effect is an additional strain on the
listener.
Other characteristics of the communication process may affect the
effectiveness of communication, when additional sounds are present.
Examples of such factors are the familiarity of the listener with the
dialect or accent of the speaker, the presence of reverberation, the
importance and familiarity of the message, distance from speaker to
listener, the motivation of the listener, and any hearing loss that
may produce a degradation in the perceived sound. Thus, the
relationship between the spectrum, level, and temporal characteristics
of a masking noise and the "intelligibility" of ordinary speech, i.e.,
the proportion of speech correctly understood is very complex. Much
research has involved the measurement of intelligibility of nonsense
syllables and of isolated words in phonetically-balanced lists. Based
upon work with real sentences, conversion charts have been constructed
to transform scores involving only words to approximate expected
scores for sentences of ordinary speech. For example, when 75% of the
items on a list of isolated words are correctly perceived, about 95%
of the key words in a sentence of ordinary discourse will be correctly
heard (Kryter, 1970). Sentence intelligibility refers to the
percentage of key words that are perceived correctly in a series of
sentences.
3.2.2 Speech interference indices
Many attempts have been made to develop a single index based on
the characteristics of the masking noise that directly indicates the
degree of interference with speech perception. Naturally, such indices
involve considerable degrees of approximation. The three most common
indices are: the articulation index (AI), speech interference level
(SIL), and the A-weighted sound pressure level (Lp(A)).
3.2.2.1 Articulation index
The AI (French & Steinberg, 1947; Kryter, 1962) is the most
complicated of these indices, since it takes into account the fact
that some frequencies are more effective than others in masking
speech. Frequencies below 250 Hz and above 7000 Hz are not included,
as they are not considered to contribute to the intelligibility of
speech. The frequency range from 250 to 7000 Hz is divided into 20
bands, each of which contributes 5% to the total intelligibility. In
order to determine the AI for a particular noise, the difference in dB
between the average speech level and the average noise level in each
of these 20 bands is calculated, and the resultant numbers are
combined to give a single index. Essentially, this process predicts
how much masking of individual speech sounds will occur and then
integrates this information.
Although the AI is an accurate index for the prediction of the
effects of noise on speech intelligibility, it is complicated to use
and difficult for the layman to interpret. Thus, simplified procedures
for estimating the AI from weighted measurements of octave-band levels
have been developed (Kryter, 1962).
3.2.2.2 Speech interference level
The SIL was designed as a simplified substitute for the AI
(Beranek 1947). Contributions to intelligibility by the lowest and
highest frequencies have been omitted to a greater extent than for the
AI. A modern version of the SIL is the arithmetic average of the sound
pressure levels in the three octave bands centred at the preferred
frequencies 500, 1000, and 2000 Hz (abbreviated SIL 0.5, 1, and 2).
Many variations of SIL in terms of the specific octave bands to be
averaged have been suggested. For example, SIL (0.25, 0.5, 1, 2)
includes the 250 Hz band. At the present time, the US National
Standards Institute recommends SIL (0.5, 1, 2, 4) as providing the
best estimate of the masking ability of a noise.
3.2.2.3 A-weighted sound pressure level
The simple A-weighted SPL is also a useful index of speech
interference. The A-weighting process emphasizes the middle
frequencies, as do the AI and SIL, but does not omit the lowest and
highest frequencies completely.
Experiments have shown that the AI is more accurate than any of
the SILs or the A-weighted SPL in predicting the speech-masking
ability of a large variety of noises. For noises of practical
importance however, A-weighted SPL and SIL continue to be used, as the
advantage of accuracy in the AI does not outweigh the ease of
measurement of the first two indices. Comparisons of SILs and
A-weighted SPLs show that, on average, the SIL is about 10 decibels
lower than the A-weighted SPL for the same degree of interference
(Klump & Webster, 1963; Kryter, 1970), although for unusual noises the
average difference could vary substantially.
3.2.3 Perception of speech out-of-doors
Measurements indicate that, during relaxed conversation in the
home, the speech level is approximately 55 dB(A) (Kryter, 1970;
Pearsons et al., 1976), and that as the noise levels increase, people
tend to raise their voices to overcome the masking effect. The
so-called "normal effort" voice resembles a "stage" voice, and is used
when people are given a prepared text to read (Korn, 1954), or when
they wish to project their voices. Since everyday speech is spoken at
a reasonably predictable level, it is possible to express many of the
empirical relationships between background noise level and speech
intelligibility in a single graph, as in Fig. 7 (US Environmental
Protection Agency, 1974).
This figure, which is applicable to outdoor conditions, is based
on the assumptions and empirical observations that:
(a) at a distance of 1 m from the speaker, relaxed conversation
occurs at a voice level of approximately 56 dB(A) and normal and
raised voices at levels of approximately 66 dB(A) and 72 dB(A),
respectively; and
(b) for 100% sentence intelligibility the speech level should
exceed the noise level by 10 dB(A). When the speech level is 10 dB(A)
lower than the noise level, intelligibility falls to 95%. Because of
the redundancy of speech, 95% intelligibility usually permits reliable
although not necessarily comfortable conversation. The location of the
curves in Fig. 7 may shift in certain circumstances, although it is
difficult to predict to what extent spatial factors may facilitate or
impair speech communication in noise. Lower noise levels may be
required, if the speaker does not enunciate clearly or if the speaker
and the listener use different dialects. People with hearing
impairment may need more favourable speech-to-noise ratios depending
on the variation of speech-to-noise ratio with frequency.
Adequate communication in higher noise levels than those
indicated in Fig. 7 can occur, if the messages are restricted, e.g.,
when only numbers are being transmitted. Lipreading or observing
facial or manual gestures may also improve communication. If the noise
source is clearly localized at a position different from that of the
speaker, speech communication may be possible in higher noise levels
than those indicated in Fig. 7.
Intermittent and impulsive noises as well as noises fluctuating
in level will provide various degrees of masking. Again, the
redundancy of speech means that an isolated short burst of noise is
unlikely to produce much disruption in the communication process;
however, the likelihood of disruption increases with increasing
duration and frequency of occurrence of the noise bursts.
The detailed characteristics of noises are also important. While
the A-weighted SPL is an adequate index of the speech-interfering
quality of many noises, others may require a more detailed analysis.
This is true of noises that are dominated by either low or high
frequencies, e.g., the rumble of distant traffic or the hiss of
compressed air. For unusual noises, the AI should be calculated for a
reliable prediction of speech intelligibility.
3.2.4 Indoor speech communication
The relationships shown in Fig. 7 apply only to outdoor (free
field) communications, as they depend on the applicability of the
inverse square law. Relationships indoors are different because of
reverberations caused by reflections from the walls, floor, ceiling,
and objects in a room. Instead of decreasing 6 dB for each doubling of
distance, the sound level of the speech or the noise may drop by only
1 or 2 dB. There is no simple formula that will predict speech
interference indoors. Instead, it is usual to set standards on the
basis of the average noise levels that have been judged in the past to
be acceptable in similar settings.
For example, Fig. 8 (US Environmental Protection Agency, 1974)
shows the estimated sentence intelligibility, at speaker-listener
distances greater than 1 m, as a function of A-weighted SPL in the
reverberant conditions found in a typical living room. This shows that
for 100% intelligibility, which is considered desirable for indoor
listening conditions, a background noise level of less than 45 dB(A)
is required.
3.3 Pain
Aural pain is induced, when the tympanic membrane tissue is
stretched by large amplitude sound pressures. Under extreme
conditions, the membrane can rupture (Hirsch, 1968).
Although there is a fairly wide range of individual variability
especially for high frequency stimuli (von Gierke et al., 1953), the
threshold of pain for normal ears is in the region of 110-130 dB. The
threshold for physical discomfort is in the region of 80 dB (Spreng,
1975).
In abnormal ears, for example in cases of inflammation, pain may
be caused in the eardrum or middle ear by sound levels of about 80-90
dB SPL. By comparison, people without eardrums may feel no sensation
of pain at sound levels of up to 170 dB SPL.
A second type of aural symptom occurs as a result of abnormal
function in the cochlea. Certain sensorineural disorders, and most
frequently noise-induced hearing losses, are accompanied by a
condition called auditory recruitment. Re