Community Noise

Edited by
Birgitta Berglund & Thomas Lindvall
Stockholm, Sweden, 1995

Editor's Note: This document was published by the World Health Organization (WHO) and placed on this website with permission from the WHO. NPC acknowledges that the WHO holds the copyright to this document.
This is a large document. Please allow some time for it to load completely onto your computer.

Birgitta Berglund
Institute of Environmental Medicine, Karolinska Institute
Department of Psychology, Stockholm University
S-106 91 Stockholm, Sweden

Thomas Lindvall
Institute of Environmental Medicine, Karolinska Institute
S-171 77 Stockholm, Sweden

Center for Sensory Research, Stockholm
1995, B. Berglund & T. Lindvall
ISSN 1400-2817
ISBN 91-887-8402-9

Printed by Jannes Snabbtryck, Stockholm, Sweden, 1995

Berglund, B., & Lindvall, T. (Eds.). Community noise. Archives of the Center for Sensory Research, 1995, 2(1), 1-195.

Abstract. The document critically reviews the adverse effects of community noise, including interference with communication, noise-induced hearing loss, annoyance responses, and effects on sleep, the cardiovascular and psychophysiological systems, performance, productivity, and social behavior. Noise measures or indices based only on energy summation are not enough for the characterization of most noise environments. This is particularly true when concerned with health assessment and predictions. It is equally important to measure and display the maximum values of the noise fluctuations, preferably combined with a measure of the number of noise events, and to assess whether the noise includes a large proportion of low frequency components. For dwellings, recommended guideline values inside bedrooms are 30 dB LAeq for steady-state continuous noise and for a noise event 45 dB LAmax. To protect the majority of people from being seriously annoyed during the daytime, the sound pressure level from steady, continuous noise on balconies, terraces, and in outdoor living areas should not exceed 55 dB LAeq. To protect the majority of people from being moderately annoyed during the daytime, the sound pressure level should not exceed 50 dB LAeq. At nighttime outdoors, sound pressure levels should not exceed 45 dB LAeq, so that people may sleep with bedroom windows open. In schools and preschools, to be able to hear and understand spoken messages in class rooms, the sound pressure level should not exceed 35 dB LAeq during teaching sessions. For hearing impaired children, a still lower level may be needed. The reverberation time in the class room should be about 0.6 s, and preferably lower for hearing impaired children. For assembly halls and cafeterias in school buildings, the reverberation time should be less than 1 s. For outdoor playgrounds the sound pressure level from external sources should not exceed 55 dB LAeq. In hospitals during nighttime, the recommended guideline values for wardrooms should be 30dB LAeq together with 40 dB LAmax. Since patients have less ability to cope with stress, the equivalent soundpressure level should not exceed 35 dB LAeq in most rooms in which patients are being treated, observed or resting. The concern for protecting young people’s hearing during leisure time activities warrants provisional guidelines for concert halls, outdoor concerts and discotheques. It is recommended that patrons should not be exposed to sound pressure levels greater than 100 dB LAeq during a 4-hour period. The same guideline values apply for sounds played back in headphones when converted to equivalent free-field level. To avoid hearing deficits from toys and fireworks, performers and audience should not be exposed to more than 140 dB(peak) of impulsive sounds. Existing large, quiet outdoor areas in parkland and conservation areas should be preserved and the background-to-noise ratio be kept low.

Foreword

This document is prepared for the World Health Organization (WHO) and is a revision of the earlier WHO document “Noise” (WHO Environ- mental Health Criteria 12, Geneva: World Health Organization, 1980) but is expanded largely and supplemented with, i.a., sections on physiology of hearing and related mechanisms, on psychoacoustics, and on mental and behavioral effects of noise. Guidelines for levels of community noise in different environments are also included. The document does not focus on occupational industrial noise.

A draft document of “Community Noise” was prepared by Professor Birgitta Berglund, Stockholm University, and Professor Thomas Lindvall, Karolinska Institute, Stockholm, on behalf of the WHO and the Nordic Noise Group of the Nordic Council of Ministers. Published international and national reviews of community noise have been consulted during the preparation of the document and are listed in the reference list.

A Task Force composed of 18 participants from 9 countries covering three regions of the WHO and two international organizations gathered in the City of Düsseldorf, Federal Republic of Germany, from 24 to 28 November, 1992 (see List of Contributors). The scope and purpose of the meeting were to make an in-depth review of the draft document. Professor Gerd Jansen served as chairperson, Dr. Bernd Rohrmann as vice chairperson and Professor Birgitta Berglund and Professor Thomas Lindvall as rapporteurs. A report on the Task Force Meeting has been published and comprises the recommendations agreed upon (Executive Summary of the Environmental Health Criteria Document on Community Noise.

Copenhagen: World Health Organization, 1993). In this document, these recommendations appear in Chapter 11, Section 1.

After the Task Force Meeting in Düsseldorf, a number of written comments were received and considered in the draft document by the two rapporteurs and Professor Xavier Bonnefoy of the WHO Regional Office for Europe. Before and after the Task Force Meeting, drafts of the document or parts of it were sent out for review among scientists all over the world, including the members of the WHO Task Force, the officers and the chair-and cochairpersons of the International Commission on Biological Effects of Noise (ICBEN), and the members of the Nordic Noise Group.

An external review draft of the document was prepared by Professors Birgitta Berglund and Thomas Lindvall as editors (June 28, 1993) and was presented for comments to all participants at the ICBEN Congress on Noise as a Public Health Problem (Noise & Man ’93) held in Nice, France, July 5 - 9, 1993. A large number of comments were received by the editors during 1993 and 1994, including comments from members of the International Institute of Noise Control Engineering (H. von Gierke, G. Maling). In addition, specific comments have been requested from specialists when the editors felt necessary to fill in obvious gaps in the document.

The editors have tried their best to accommodate all the review comments in the text and to make decisions when conflicting comments have been received. Thus, although the document is the amalgamated result of the work of a large number of persons, the complex and extended work process makes it necessary to declare that the editors are solely responsible for the present text of the document.

Both the World Health Organization’s Regional Office for Europe and Headquarters are grateful to the Nordic Council of Ministers (the Nordic Noise Group) and to the German Government who provided the necessary financial support for the Task Force Meeting 1992. The efforts of all who have helped in the preparation and finalization of the document are gratefully acknowledged.

Participants in the WHO Task Force Meeting, Düsseldorf, Germany, November 24-28, 1992, and Contributors by correspondence as well:

Dr. Sharon M. Abel, Canada
Professor Birgitta Berglund, Sweden (Rapporteur)
Professor Xavier Bonnefoy, WHO, Copenhagen
Professor Gary W. Evans, USA
Dr. Lawrence D. Fechter, USA
Dr. Ian H. Flindell, United Kingdom
Dr. Eric Giroult, WHO, Geneva
Professor Gerd Jansen, Germany (Chairperson)
Dr. Tapani Jauhiainen, Finland
Mr. John A. Lambert, Australia
Professor Thomas Lindvall, Sweden (Rapporteur)
Professor Henryk Mikolajczyk, Poland
Dr. Herbert Müller, CEC
Ms. Sirkka-Liisa Paikkala, Finland
Professor Bernd Rohrmann, Germany (Vice Chairperson)
Professor Shirley J. Thompson, USA
Dr. Michel Vallet, France

Other Main Contributors by correspondence:

Mr. K. Andersson, Sweden
Dr. R.F.S. Job, Australia
Professor A. Axelsson, Sweden
Dr. R. de Jong, The Netherlands
Dr. W. Babisch, Germany
Mr. J. Karlsson, Sweden
Professor A. Cocchi, Italy
Professor T. Kihlman, Sweden
Dr. P. Dickinson, New Zealand
Dr. J. Lambert, France
Professor H.G. Dieroff, Germany
Professor H. Lazarus, Germany
Dr. S.I. Erlandsson, Sweden
Dr. P. Lercher, Austria
Dr. J.M. Fields, USA
Dr. M.E. Lutman, United Kingdom
Dr. L.C. Fothergill, United Kingdom
Dr. G. Maling, USA
Dr. A.M. Garcia, Spain
Dr. H.M.E. Miedema, The Netherlands
Professor H. von Gierke, USA
Dr. H. Møller, Denmark
Professor B. Griefahn, Germany
Professor A. Muzet, France
Professor M. Haider, Austria
Dr. P.O.L. Nilsson, Denmark
Dr. R. Hellman, USA
Dr. E. Öhrström, Sweden
Dr. P.A. Hellström, Sweden
Dr. C. Oliva, Switzerland
Mr. K.E. Hogstad, Norway
Dr. W. Passchier-Vermeer, The Netherlands
Professor T. Houtgast, The Netherlands
Dr. A. Preis, Poland
Dr. S. Hygge, Sweden
Mr. M. Ringheim, Norway
Dr. A. Illényi, Hungary
Mr. S. Solberg, Norway
Dr. E.K. Ishii, USA
Dr. H.G. Stidl, Austria
Professor H. Ising, Germany
Professor W.D. Ward, USA

Contents

1. INTRODUCTION

Almost 25 % of the European population is exposed, in one way or another, to transportation noise over 65 dBA (an average energy equivalent to continuous A-weighted sound pressure level over 24 hours) (Lambert & Vallet, 1994). This figure is not the same all over Europe. In some countries more than half of the population is exposed, in others less than 10 %. When one realizes that at 65 dBA sound pressure level, sleeping becomes seriously disturbed and most people become annoyed, it is clear that community noise is a genuine environmental health problem. In four European countries (France, Germany, Great Britain and the Netherlands; see Lambert & Vallet, 1994) it would seem that road traffic noise is annoying to 20-25 % of the population and railway noise to 2-4 %. In the absence of future ambitious noise abatement policies, the noise environment risks to remain unsatisfactory or even deteriorate.

The acoustic world around us continuously stimulates the auditory system. The brain selects relevant signals from the acoustic input, but the ear and the lower auditory system are continuously receiving stimulation. This is a normal process and does not necessarily imply disturbing and harmful effects. The auditory nerve provides activating impulses to the brain, which enable us to regulate our vigilance and wakefulness necessary for optimum performance.

Problems associated with noise-induced hearing loss go back to the Middle Ages. The workers in certain professions such as blacksmithing, mining, and church bell ringing were known to become deaf after years of work. However, with industrial development, the number of workers exposed to excessive noise increased significantly as has the number of people exposed to other sources of noise such as transportation noise and loud music. In industrialized societies of today, the risk of occupational noise-induced hearing loss mostly is being met by efficient technical and other countermeasures. The occupational health authorities are now much more observant of the problem than before. In developing countries, the risk for much increased rates of occupationally acquired hearing loss have to be met by strong preventive measures in engineering and medicine.

Furthermore, in most countries hearing impairment due to community noise exposure (sociacusis) has become a problem of concern (Glorig, Grings, & Summerfield, 1958; B. Berglund, U. Berglund, & Lindvall, 1984). It has been demonstrated that community noise may have a number of direct adverse effects other than hearing damage. These include adverse effects on communication, performance, and behavior; nonauditory physiological effects; noise-induced disturbance of sleep; and community annoyance.

The indirect or secondary effects of noise are often hard to quantify and satisfactory assessment models are lacking. Often, large-scale epidemiological or social surveys would be required to assess these which involve increased risks of accidents by noise-exposed individuals, reduction in productivity at work, and related effects.

There may be some populations at greater risk for adverse effects of noise. Young children (especially during language acquisition), the blind, and the hearing impaired are examples of such populations.

2. SCOPE

We are constantly exposed to noise in our daily lives. In this document noise exposure outside the industrial work place is called ”community noise” (environmental noise). Main sources of community noise are transportation systems (road, air and rail), industries, construction and public works, and neighborhood. The main indoor sources are ventilation systems, neighbors, office machines, and home appliances. Also leisure activities, such as motor sports, speed boats, and snow scooters, represent important noise sources. Community noise includes all noise sources except noise at the industrial work place.

The scope of this document is to consolidate actual scientific knowledge on the health impacts of community noise and to provide guidance to environmental health authorities and professionals trying to protect people from the harmful effects of noise. The effects of community noise on human beings are ranging from hearing damage to the feeling of annoyance. In noise abatement policy, the effects of noise on different human activities should be taken into consideration. This means that several different guideline values are suggested. Countries are expected to develop their own national and local noise standards in accordance with the amount of noise hazards they are prepared to accept.

Although it is clear that for some levels of noise exposure harmful effects are obvious, in other cases objectivity in the demonstration of health effects is difficult. The effects depend not only on the sound pressure levels but also on the “type” or ”quality” of the noise, on the number of noise events, and on the ”image” of the noise.

Noise control is always more effective and less costly if it is designed at a very early stage of development. It is more expensive to apply noise abatement measures after the noise problem has been realized. In this document, local authorities and national governments may hopefully find guidance for noise control in various type of nonindustrial environments. However, this document does not deal directly with sound pressure levels at the point of noise emission. Thus, it does not give any recommendation for limitation of sound pressure level at the noise source, but instead in the form of guideline values for adverse total noise exposure in the environments.

3. PHYSICAL ASPECTS OF NOISE

Sound is produced by any vibrating body and is transmitted in air only as as a longitudinal wave motion. It is, therefore, a form of mechanical energy and is typically measured in energy-related units. For listeners sound is defined as acoustic energy in the frequency range from 20,000 Hz to below 20 Hz that is typical of the human auditory system. The sound output of a source constitutes its power and the intensity of sound at a point in space is defined by the rate of energy flow per unit area. 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 on a logarithmic scale, in decibels (dB). Sound pressure has the unit Pascal (Pa), while sound pressure level has the unit dB. The effects of noise depend strongly upon frequency of sound-pressure oscillation. Therefore, spectrum analysis is important in noise measurement (see further, e.g., Fahy, 1989).

The perceived magnitude of sound is defined as loudness (e.g., D.M. Green, 1976). The loudness is primarily a function of intensity, frequency and temporal parameters. Various procedures exist by which loudness may be estimated from physical measurements. The simplest methods involve the measurement of the sound pressure level through a filter or network of filters that mimics the frequency response of the auditory system (weighting circuits in sound level meters).Various calculation procedures have also been developed for predicting loudness. Loudness has the unit sone, whereas loudness level has the unit phon. There is a unique relationship between sone and phon at least for levels above 40 phon or 1 sone (ISO 131, 1979a). They are both based on physical measures and should not be confused with the loudness scales that are constructed from reports of perceptions by participants in experiments or field surveys.

3.1 Definitions of Sound and Noise

Physically, sound is produced by mechanical disturbance propagated as a wave motion in air or other media. Physical sound evokes physiological responses in the ear and auditory pathways. These responses can be described and measured using appropriate methods with, for example, physical parameters (like vibratory motion of the eardrum membrane) or with electrophysiological parameters (changes in bioelectric potentials in the sensory and neural tissues). However, not all sound waves evoke auditory-physiological responses, for example, ultrasound has a frequency too high to excite the auditory system and, thus, to evoke sound perception. Psychologically, sound is a sensory perception originating as a mental event evoked by physiological processes in the auditory brain. Other areas of the nervous system are also known to be involved. Thus, it is merely through the perceptual analysis of sounds that the complex pattern of sound waves may be classified as “Gestalts” and labeled noise, music, speech, etc. From a physical point of view there is no difference between the concepts sound and noise, although it is an important distinction for the human listener Noise is a class of sounds that are considered as unwanted. In some situations, but not always, noise may adversely affect the health and wellbeing of individuals or populations. Since long agreed among experts, it is not possible to define noise exclusively on the basis of physical parameters of sound. Instead, it is common practice to define noise operationally as audible acoustic energy that adversely affects, or may affect, the physiological and psychological wellbeing of people.

3.2 Characteristics of Sound and Noise

Sound waves involve a succession of compressions and refractions of an elastic medium such as air. These waves are characterized by the amplitude of sound pressure changes, their frequency, and the velocity of propagation. The speed of sound (c), the frequency (f), and the wavelength (l) are related by the equation

l = c/f (1)

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). Sound intensity can be defined in any direction, often as a vector. In a free sound field, the sound intensity is related to the root mean square of the sound pressure (p), the static mass density of the medium (r), and the speed of sound in the medium (c).

p 2
I = ___ (2)
rc

The total sound energy emitted by a source per unit time is known as the sound power and is measured in Watts (W). Sound intensity (Eq. 2) is normally measured in Watts per square meter (W/m 2 ). Sounds are described by means of time-varying sound pressure, p(t). Compared to the magnitude of the atmospheric pressure, the temporal variations in sound pressure, caused by sound are extremely small. The values of sound pressure between 10 -5 and 10 2 Pa (or Newton per square meter, N/m 2 , according to Système International d’Unités, SI) are relevant for the human listener. Since the range of this variable is so wide, it is usual to express its value on a logarithmic scale in dB. Sound intensity level is defined as 10 times the logarithm (to the base 10) of the ratio of the sound intensity of a target sound to the sound intensity of another sound (or alternatively, the sound pressure level as 20 times the logarithm of the ratio of their sound pressures). Any acoustic quantity that is related to sound energy, for example, power or mean square pressure, may be expressed as a dB-value. To establish an absolute level, a reference value must be agreed. Thus, the sound pressure level (Lp) of a sound expressed in dB-values depends on the mean square sound pressure (p 2 ) such that

Lp = 10 log10 [p/pref ] 2 (3)

where the reference pressure pref has an internationally agreed value of 2 . 10 -5 N/m 2 (often given in micropascal, 20 mPa). The corresponding standardized reference values for sound power level and sound intensity level are 10 -12 W and 10 -12 W/m 2 , respectively. Unless otherwise stated in

Table 1. How to combine two sound pressure levels expressed in dB.

this document, sound pressure levels are expressed in the unit dB relative to the international standard reference quantities (i.e., dB re 20 mPa). Whereas sound intensities or energies are additive, sound pressure levels have first to be expressed as mean square pressures, then added, and then transferred to a sound-pressure-level value again. However, this assumes uncorrelated sources. The summation of sound pressure levels can be performed by using the equation:

Lp = 10 log10 [10 Lp1/10dB + 10 Lp2/10dB + 10 Lp3/10dB +..... ] (4)

A simple example will illustrate the use of this equation. If two sound sources of 80 dB in sound pressure level are combined, then the sound pressure level of the resulting sound will become 83 dB:

L = 10 log10 [10 8 + 10 8 ] = 10 log10 [2 . 10 8 ] = 10 [log10 2 + log10 10 8 ] = 10 [0.3 + 8] = 83

It is only when two sources generate similar levels that the combined output will result in a significant increase in level above the louder noise. The example just quoted gave a 3-dB increase. If there is any difference in the original, uncorrelated 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 always exceeds 10 dB, the contribution of the softer source to the combined sound pressure level may be viewed as negligible. The results of such combinations of decibel values may be found in a simplified manner by using Table 1.

Sound is measured with the aid of a microphone that generates a voltage proportional to the acoustic pressure acting upon it. This signal can be measured and analyzed using conventional electronic instrumentation. A sound level meter is usually a portable, self-contained instrument incorporating a microphone. The microphone should be calibrated so that sound pressure levels may be determined in accordance with reference pressure. If certain prerequisites are known (e.g., sound field) intensity levels and power levels can be derived from sound pressure level measurements.

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 base 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 at a given time. 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. The frequency spectrum of a sound is not restricted to harmonic frequencies; it is discrete for periodic signals and continuous for nonperiodic signals. For examle, the frequency spectrum may specify how the energy in the periodic sound is concentrated at certain discrete frequencies. The frequency distribution of sound energy is measured by electronic filters or with the aid of a computer by calculation. Although some kinds of machinery produce sound that is largely periodic, much sound perceived as noise is nonperiodic, that is, 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 perception named 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 (e.g., Small, 1970). 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 as low and high frequency noises depending upon the range and proportion of frequencies present. Human hearing is sensitive to frequencies in the range from about 20,000 Hz to below 20 Hz (the “audiofrequency range”). Downwards there is no established limit; frequencies down to at least 2 Hz can be detected by the ear (B. Berglund, Hassmén, & Job, 1994). Sound components lower than 16 Hz are named infrasound and those higher than 20,000 Hz ultrasound. The human hearing has a very “narrow” range of sensibility at infrasound frequencies. Whereas the sensibility range within the audiofrequency range is 120 to 140 dB, the sensibility from barely perceptable to pain is 30 to 40 dB at infrasound frequencies.

The audible frequency range is technically 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 centers of the standardized octave bands are spaced at octave intervals from 16 to 16,000 Hz (ISO 266, 1975a). The octave band level at a particular center 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 detailed 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 (time window). For steady-state sounds, the choice of averaging time is immaterial provided that it is long enough 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 125 ms and 1 s, respectively (IEC 651, 1979; Brüel, 1977).

Impulsive noise consists of one or more bursts of sound energy, each of a duration of less than about 1 s (ISO 2204, 1979b). Sources of impulsive noise include impacts of all kinds, for example, hammer blows, explosions, and sonic booms. These may be heard as single events or, as in the case of a stamping press, repetitively. The averaging time of the inner ear is very short (about 30 ms). To characterize impulsive 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” (0.125 s) setting. For somewhat more accurate measurements, a shorter averaging time is specified for standard “impulse” sound level meters (an averaging time of 35 ms when the level is rising and of 1500 ms when it is decreasing; IEC 651, 1979). The peak level (“peak”) is the level of the instantaneous peak and it is much higher than the “impulse” level.

3.3 Sound Pressure Levels and their Measurement

3.3.1 Loudness and Loudness Level

The physical magnitude of a sound is given by its intensity. The subjective or perceived magnitude is called its loudness. Primarily, loudness depends on intensity, frequency and duration (see, e.g., H. Fletcher & Munson, 1933; S.S. Stevens, 1955; Zwislocki, 1960, 1969). Binaural sound is perceived to be twice as loud as monaural sound (H. Fletcher & Munson, 1933; Hellman & Zwislocki, 1963); everyday sound exposure is typically binaural. That is one reason why knowledge from laboratory experiments may not always be generalizable to environmental conditions. Owing to the complexity of operation of the human auditory system, it is not possible to design an objective sound measuring apparatus for all types of noise to give results which are fully comparable with those obtained by subjective methods (IEC 651, 1979).

Figure 1. Normal equal-loudness contours for pure tones (From: ISO 226, 1987a; D.W. Robinson & Dadson, 1956).

The basic unit of loudness is the sone which is defined as the loudness of a 1,000 Hz pure tone heard at a sound pressure level of 40 dB re. 20 µPa under specified listening conditions (ISO 131, 1979a). Two sone 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 psychophysical “power law” of loudness, often referred to as Stevens’s law, which is in general in accordance with the Weber fraction for just noticeable differences (S.S. Stevens, 1957b; S.S. Stevens, 1961a). In the mid audiofrequency range, the exponent of the power function is such that a twofold change in loudness corresponds to a tenfold change in intensity, that is, a 10 dB change in sound pressure level (S.S. Stevens, 1957a). At low frequencies, loudness changes more rapidly with changes in sound pressure level. This is demonstrated in Fig. 1, which shows a standard set of equal sound pressure level contours for pure tones (D.W. Robinson & Dadson, 1956; ISO 226, 1987a). Each line shows how the sound pressure level of the tone must be varied to maintain a constant loudness (cf. the equal-loudness contours by H. Fletcher & Munson, 1933, 1957, 1958). Each iso-phon curve, in fact, represents a particular loudness level expressed in the unit phon. In other words, any tone that is perceived equally loud to a 1,000-Hz tone assumes the same phon-value as the 1,000-Hz tone. At 1,000 Hz, the phon value is identical to the dB-value. Thus, loudness level is expressed as a 1,000-Hz loudness equivalent in sound pressure level and determined under specified listening conditions (ISO 131, 1979a). For practical purposes (ISO 131, 1979a), the relationship between the scales of loudness (S, in sone) and loudness level (P, in phon) may be expressed as follows, for loudness level larger than 40 phon:

S = 2 (P-40)/10 (5)

This equation shows that (perceived) loudness doubles for an increase of 10 phon. It also reflects the definition of sone which states that the (perceived) loudness of a 1000-Hz tone at 40 phon is 1 sone.

3.3.2 Calculation and Measurement of Loudness Level

Ideally, meters for sound pressure measurements should give a reading equal to loudness in phon. This objective is difficult to achieve, because the intervening human perceptual processes are complex. Nevertheless, such procedures have been developed and adopted as international standards (ISO 532, 1975b). Until recently they have been too complex to be incorporated into a simple measuring instrument, and, therefore, they are rarely used in practice. Presently, these techniques are being implemented in modern digital equipment.

For most practical purposes, a much simpler approach is used. The A-weighting curve is used to weight sound pressure levels as a function of frequency, approximately in accordance with the frequency response characteristics of the human auditory system for pure tones. That is, energy at low and high frequencies is de-emphasized in relation to energy in the mid-frequency range. Most precision sound level meters incorporate three selectable weighting circuits, the A-, B-, and C-weightings (IEC 651, 1979) and sometimes a D-filter (IEC 537, 1976). The characteristics of these weighting curves are illustrated in Fig. 2. The A-, B- and C-filters were intended to match the auditory-system response curves at low, moderate, and high loudness levels, respectively. Sound pressure levels in the weighted scales are measured in decibel units and are often expressed by indicating which weighting was used, for example, dBA.

The D-weighting curve is based on the so-called 40-noy curve according to Karl Kryter and is described in the now withdrawn IEC 537 (1976) “Frequency weighting for the measurement of aircraft noise”. Whereas the equal-loudness contours were established for pure tones, equal-noisiness contours were based on noise bands. The unit of measurement here is noy. The rationale for constructing the equal-noisiness contours was that higher frequencies tended to be more annoying than lower frequencies although they were equally loud (Kryter, 1959, 1970, 1985, 1994). However, also the lower frequencies at the other end of the audiofrequency range tend to be more annoying (Goldstein, 1994).

Figure 2. Standard A, B, C, and D filter characteristics for sound level meters (IEC 179, 1973a; IEC 179a, 1973b).

The weighting curves, A to D, have broader applications than, for example, for evaluating the risk of damage to hearing and the sound pressure level of traffic noise. The efforts to describe the effect of noise in the simplest possible way, that is, in terms of a one-figure value, have resulted in a number of proposals for weighting, and, apart from the weighting curves A, B, C, and D, also to various noise indices (e.g., Kryter, 1985, 1994), for example, the Noise Rating numbers (NR) and Noise Criteria (NC). However, the weighting curves were all developed for stationary or quasi-stationary sound exposures and may, therefore, easily give rise to more or less serious errors in other community-noise applications.

The weighting curves A, B and C are a compromise between the American and German standards of the mid 1950’s so that the tolerance limits of the new curves included the nominal values of both standards. The A-curve was based on the 40 phon equal loudness contour and was recommended for use for loudness levels between 20 to 55 phon. The A-weighting is widely used for sound level measurements in a variety of situations. For sounds of narrow frequency range, 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-filter has been adopted so generally that sound pressure levels frequently quoted in the literature simply in dB are in fact A-weighted levels. Furthermore, many older general purpose sound level meters are restricted solely to A-weighted sound pressure level measurements.

King (1941) was perhaps the first to suggest a calculation method for predicting (perceived) loudness from octave band analysis of complex sounds. Many years later, two different calculation methods for loudness were developed and standardized (ISO R532, 1966; ISO 532, 1975b): Method A (according to Stanley S. Stevens) using 1/1 octave analysis and Method B (according to Eberhard Zwicker) using 1/3 octave analysis data.

3.4 The Time Factor

Sounds can appear to be steady to human hearing 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. In precision sound level meters, the “slow” response time (1.0 s) is appreciably longer than the auditory averaging time and is used to obtain a steady reading, when the signal level fluctuates at a rapid rate. The “fast” response time is considered to be of similar order as that of the auditory system (0.125 s). However, in noise assessment, sound level fluctuations are usually ignored and, therefore, the “slow” response time is commonly applied. Difficulties arise when these 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. For impulsive sound, often the time-weighting “impulse” is used (0.035 s).

Series of sound events or intermittent sound are described in various ways: Percentiles of the occurrence, per cent in excess of defined sound pressure level, Noise and Number Index (NNI), etc. The dynamic characteristics of noise measurements are described in detail in IEC 651 (1979) with, i.a., integration time, bandwidth and handling of short signal impulses. It should be emphasized that sound pressure level as measured with meter setting “impulse” is based on loudness level, originating from perceptual measurements. Therefore, in order to assess the risk for damage to hearing the instantaneous “peak” may instead be measured directly and not, for example, the maximum sound pressure level. The IEC-standard for sound level meters adduces four different classes of accuracy 0, 1, 2 and 3, where 0 describes the most accurate instrument.

For a determination of the equivalent continuous sound pressure level (see section 3.5.1), for example by A-weighting over the period T hours (LAeq,T), the instrument should be used in accordance with IEC 804 (1985).

3.5 Noise Exposure Scales

In many noise measures and indices that are correlated with perception or other effects of interest, various underlying acoustic and nonacoustic (physical) properties have been combined in different ways. The basic objective of measurement is then 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 perception include the following: loudness level (recognizing average and peak values together with impulsive characteristics where appropriate), total noise “dose”, amplitudes of level fluctations, rates of fluctuation, number of noise events and duration of events, and duration of total noise exposure. Clearly, the acoustic stimulation alone have many dimensions; the following three procedures are most commonly used to measure some of them.

3.5.1 Equivalent Continuous Sound Pressure Level

To measure an average sound pressure level the meter averaging time is extended to equal the period of interest, T, which may be an interval in seconds, minutes, or hours. This gives a dB-value in Leq which stands for equivalent continuous sound pressure level; or according to a forthcoming standard by IEC should be named the “time average level”. It is derived from the following mathematical expression in which A-weighting has been applied:

T [L pA(t) / 10dB]
LAeq,T = 10 log10 [(1/T) 0þ 10 dt ] (6)

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 as the fluctuating sound of interest, has the same mean square sound pressure, usually applied as an A-frequency weighting (Eq. 6). The interval of time must be stated.

Equivalent continuous sound pressure level is gaining widespread acceptance as a scale for the measurement of long-term noise exposure. For example, it has been adopted by the ISO for the measurement of both community noise exposure (ISO 1996, 1982, 1987,a, 1987c) and hearing damage risk (ISO 1999, 1990). It also provides a basis for more elaborate composite noise indices discussed in subsequent sections of this document including the day-night weighted sound pressure level (Ldn).

Following the introduction of jet aircraft into commercial service, it was suggested that the then existing loudness scales were inadequate for aircraft noise assessment purposes. An alternative scale of Perceived Noise Level (PNL) was, therefore, developed, with the unit dB(PN) (Kryter, 1959, 1985, 1994). This scale was derived using the equal loudness level procedure of S.S. Stevens (1956, 1972) but instead based on the attribute of perceived noisiness (defined as the “unwantedness” of the sound) that was considered different and more relevant to aircraft noise than loudness. In fact, the only difference between the calculations involved is the use of different frequency response curves. As research progressed towards legislations for aircraft noise emission control (U.S. Federal Aviation Regulations, 1969; see also ICAO, 1993), the perceived noise level scale was modified to include special weightings for “discrete frequency components”, that is, 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, the D-weighting filter constituted a parallel development. Its characteristics were based on an equal noisiness (rather than an equal loudness) frequency response curve (IEC 537, 1976; D-weighting now withdrawn by IEC). This weighting circuit is available in some sound level meters and is intended for aircraft noise monitoring purposes.

The equivalent continuous sound pressure level, expressed in dB LAeq,T is unsuitable as a measure of value for predicting long-term adverse effects, that is., owing to the fact that the distribution over time of exposure does not appear and that the temporal profile is not stated. A number of proposals for corrections due to time have been presented: number of events, time of day, statistical distribution, number of heavy vehicles passing, Noise Number Index, etc. (e.g., Kryter, 1985; Zwicker & Fastl, 1990).

3.5.2 Level Distribution

A widely used method of recording the variations in sound pressure level is that of level distribution analysis, sometimes called statistical distribution analysis. This yields a graph of the percentage of the total time for which any given sound pressure 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, or 90 % of the time, are frequently used as average measures typical for the “maximum”, “median”, and “background” levels, respectively. The same statistical approach is used to describe the distribution of loudness values in N5, (Fastl, 1993), N10 (Berry & Zwicker, 1986) or N50 (Watts, 1991).

3.5.3 Limitations of A-Weighted SPL as a Measure of Loudness or Annoyance

As pointed out by Hellman and Zwicker (1982), A-weighted SPL was first introduced into a sound level meter in 1936. Due to its simplicity and convenience, the A-weighting has become a popular and often useful frequency weighting also for assessing the perceived magnitude of noise. However, for many years international commissions have been aware that dBA is an overall value which may simulate neither the spectral selectivity of human hearing nor its nonlinear relation to sound intensity. Thus, if sounds with different spectral envelopes are compared (e.g., various community noises), the dBA-value obtained may be an inaccurate indicator of human subjective response. Human hearing is possible to simulate much better via computer software or/and signal processors.

In the past, sound pressure level has been measured widely by A-weighting. At the same time, both in the laboratory and in the field evidence has accumulated that A-weighting predicts loudness and annoyance ofcommunity noise rather poorly. Not only does A-weighted sound pressure level underestimate the impact of the low-frequency components of noise (Goldstein, 1994), but it is also strongly dependent on the exposure pattern with time. For sounds that exceed 60 dB the reliability of A-weighting decreases. Moreover, A-weighted sound pressure level neither considers the effects of mutual masking among the components in a complex sound nor the asymmetry of masking patterns produced in the auditory system (Zwicker & Fastl, 1990). Yet, despite these well-known limitations, the A-weighted sound pressure level is widely used in practice.

The A-filter is unrepresentative of the loudness of sounds containing a mixture of noises and tonal components. In such cases, A-weighted sound pressure level is less suitable for the prediction of loudness or annoyance. That is also true for noise containing most of its energy in the low-frequency range of 15-400 Hz. It may then underpredict perceived loudness by 7 to 8 dBA, relative to a 1,000 Hz target noise (Kjellberg & Goldstein, 1985). The reason is that loudness increases due to bandwidth increase and that spectrum shape is not accounted for to a satisfactory degree by the A-filter (cf. Zwicker, 1987). A decrease in A-weighted sound pressure level can result in a corresponding increase in loudness (Hellman & Zwicker, 1982) or annoyance. This clearly reveals the shortcoming of using overall SPL, either unweighted or A-weighted, as an indicator of loudness and annoyance.

4. TYPES OF ENVIRONMENTAL NOISE

4.1 Introduction

Noise is a problem that affects everybody. Noise is likely to continue as a major issue well into the next century. To understand noise we must understand the different types of noise, where noise comes from, the effect of noise on humans and the various ways we have of measuring both the sound as a cause of noise and the noise effects. This chapter describes the various types of noise that can affect the community and offers some basic definitions used in measuring sound for assessing the expected effects. Sound is produced by a mechanical disturbance spreading out as a wave motion in the air at a speed of about 330 m/s. Acoustic waves entering the ear evoke a physiological response which causes nerve impulses to be transmitted to the brain. The brain interpret these impulses so that they can be perceived as sound.

Noise is unwanted sound and thus implicitly refer to a subjective classification of sound. Sound can have a range of different physical characteristics, but it only becomes noise when it has an undesirable physiological or psychological effect on people. Nevertheless, it is important to understand the physical characteristics of sound since these characteristics determine the various ways we have of measuring and describing sound. The main physical characteristics are: sound pressure level, sound frequency, type of sound, and variation in time. Typical sound pressure levels range from about 20 dB LAeq in a very quiet rural area to between 50 and 70 dB LAeq in towns during the day time, to 90 dB LAeq or more in noisy factories and discotheques to well over 120 dB LAmax near to a jet-aircraft at take-off.

An audiofrequency is associated with the perception of the pitch of a tonal sound. Sound frequency is measured by the number of repeated cycles of the sound wave in one second (c/s or Hz) and the audible frequency range is 20-20,000 Hz. An idling diesel engine can produce large amounts of low frequency sound in the range of 20 to 150 Hz, whereas a warning siren usually produces a medium to high frequency sound typically around 2,000 Hz. The sound design of warning signals is based on the fact that the human auditory system is most sensitive in the middle range of frequencies from 1,000 to 4,000 Hz. Sound pressure level weighted with A-, B-, and C-filters in sound level meters is intended to take into account part of the differential frequency sensitivity.

Type of sound describes the particular features of a sound which makes it possible for a listener to identify it. The ability to identify the source is very important in determining community annoyance. These features can include tonal and harmonic qualities, impulsiveness, the relative balance of high and low frequencies and the steadiness or irregularity of the sound. There are a whole range of physical measurements which can express these different features in a more or less appropriate way for noise impact predictions.

Sound pressure levels normally vary with time. Rapid fluctuation in sound pressure level over less than 1 s can contribute to impulsiveness. Moving sound sources such as overflying aircraft or road vehicles produce a time-varying sound pressure level over event periods of typically 10 to 100 s. Noise from fixed installations, such as ventilation systems, can often be steady for much of the day, but may drop at night. The maximum A-weighted sound pressure level is described by the quantity LAmax and depends on the time constant set in the measurement instrument (“slow” or “fast”). Peak level is the peak of the instantaneous sound pressure oscillation in the time domain, not the maximum value of the effektive sound pressure (prms) or of the sound pressure level. Peak level is commonly expressed in dB by calculation (Eq. 3).

The equivalent continuous sound pressure level is described by the quantity LAeq,T (Eq. 6). In practice, adjustments (often called “penalty” factors) for various sound characteristics are sometimes introduced into noise indices, for example, for impulsiveness, tonal components, low frequency, and different time period (day/night). Such indices which are based on LAeq,T are sometimes called “rating scales” (Schultz, 1982a).

4.2 Sources of Noise

4.2.1 Machinery Noise, Noise from Industrial Plants

Mechanized industry creates serious noise problems, subjecting a significant fraction of the working population to potentially harmful sound pressure levels of noise. It is responsible for high noise emissions indoors as well as outdoor of plants. In industrialized countries it has been estimated that 15-20 % or more of the working population is affected by sound pressure levels of 75-85 dBA. 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 tonal and harmonic components; air moving equipment tends to generate sounds with a wide frequency range. The highest sound pressure 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 mechanical 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 complex nature. Various types of machinery are involved and they represent artificial noises which are of concern because they may contain predominantly low or high frequencies as well as tonal components, they may be impulsive and also present unpleasant and disruptive temporal sound patterns.

Machinery that moves air are of special interest because it usually creates noise with a large component of low frequencies. Unlike noise containing mainly higher frequencies, low-frequency noise is less attenuated by walls or other structures and it can cross great distances with little energy loss due to atmospheric and ground attenuation.

In residential areas, noise may stem from mechanical devices (e.g., heat pumps and ventilation systems, traffic) as well as voices, music and other kinds of noises generated by neighbors (e.g., lawn movers, vivid parties, and other social activities). Due to low-frequency characteristics, noise from ventilation systems in residential buildings may cause considerable concern even at low and moderate sound pressure levels.

Sound generation mechanisms of machinery are reasonably well understood. The technical requirements for low noise output in new machinery can usually be specified but the noise declaration of machinery, which describes the noise output of the machine, is not yet used efficiently. The noise declaration should preferably be used for selecting and purchasing the machine which is least noisy. The difficulty of reducing the sound output and the noisiness of existing equipment is a serious obstacle to the improvement of working environments (e.g., jack hammer or shooting range). Machinery should preferably be silenced at the source. Noise from fixed installations such as factories or construction sites, heating pumps and ventilation system plants on roofs, can also affect the nearby communities. To reduce the sound output from such sources, either the use of quieter plant and equipment is encouraged, or through zoning to separate industrial land uses from the more noise-sensitive residential areas. As last resorts, insulation or restriction of operation time may be used.

4.2.2 Transportation Noise

4.2.2.1 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 sound pressure level from traffic can be predicted from the traffic flow rate, the speed of the vehicles, the proportion of heavy vehicles, and the nature of the road surface. Special problems can arise in areas where the traffic movements involve a change in engine speed and power, such as at traffic lights, hills, and intersecting roads.

4.2.2.2 Rail traffic

Railway noise depends primarily on the speed of the train but variations are present depending upon the type of engine, wagons, and rails. Impact noises can be generated in stations and marshaling-yards because of shunting operations. The introduction of high speed trains has created special noise problems. At speeds greater than 250 km/h, the proportion of high frequency sound energy increases and the sound can be perceived as similar to that of overflying jet aircraft.

4.2.2.3 Air traffic

Aircraft operations have caused severe community noise problems over the past 20 to 30 years. The introduction of the early turbojet transport aircraft led to a surge of community reactions against commercial and military airports. More research has been devoted to aircraft noise than to any other environmental noise problem (B. Berglund, Lindvall, & Nordin, 1990). The main mechanism of noise generation in the early turbojet aircraft was the turbulence created by the jet exhaust mixing with the surrounding air. This noise source has been significantly reduced in modern high by-pass ratio turbo-fan engines which surround the high velocity jet exhaust with a lower velocity airflow generated by the fan. The fan itself can be a significant noise source, particularly during landing and taxiing operations. Fan noise can be controlled to a certain extent by providing acoustic absorption in the fan cowling. There is some concern over the possible use of advanced multi-bladed turbo-prop engines in the future, as these engines can produce relatively high levels of tonal noise. Aircraft takeoffs are known to produce intense noise including vibration and rattle but also landings cause noise annoyance especially when reverse thrust is applied. In general, larger and heavier aircrafts produce more noise than lighter aircrafs. The smaller aircraft types as used for private business, flying training and leisure purposes can cause particular noise problems near to general aviation airports. Leisure flying at weekends can cause difficulties because nearby residents are more likely to be at home. Airports hosting many helicopters often create a specifically severe noise problem.

4.2.2.4 Sonic booms

The sonic boom is a shock wave system in air 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 ms, the sonic boom has a characteristic double sound. High intensity sonic booms can damage property. Lower intensity sonic booms can cause a startle response in people as well as animals. The startle response is a secondary effect due to the sudden and unexpected exposure. The sonic boom can be heard as a very loud and boomy sound.

An aircraft in supersonic flight trails a sonic boom that can be heard up to 50 km on either side of its ground track depending upon the flight altitude and the size of the aircraft (C.H.E. Warren, 1972).

4.2.3 Construction Noise, Public Works Noise and Military Noise

Building construction and earth works are activities that can 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 sometimes carried out without considering the environmental noise consequence. Street services such as garbage disposal and street cleaning can cause considerable disturbance if carried out at sensitive times of day.

In certain instances, military activities may be an important noise source such as noise produced by heavy vehicles (tanks), helicopters, and small and large fire arms. Noise from military airfields may present particular problems compared to civil airports, for example, if used for training interrupted landings and takeoffs (so-called touch down).

4.2.4 Building Services Noise

Building service noise can affect people both inside and outside the building. Ventilation and air conditioning plants and ducts, heat pumps, plumbing systems, and lifts, for example, can compromise the internal acoustic environment and upset nearby residents.

4.2.5 Domestic Noise

Noise from neighbors is often one of the main causes of noise complaints. These complaints are largely due to the inconsiderate or thoughtless use of powered domestic appliances (vacuum cleaners, washing machines, lawn mowers, etc.), systems for music reproduction, TV sets, or hobby activities. Substantial societal problems, more infrequent but nonetheless important, are caused by disturbing noise emanating from neighbours and their social activities.

4.2.6 Noise from Leisure Activities

The possibilities of using powered machines in leisure activities are increasing all the time. For example, motorracing, off-road vehicles, motorboats, water skiing, snowmobiles, etc., can all contribute significantly to loud sound pressure levels in previously quiet areas. Shooting activities not only have considerable potential for disturbing nearby residents, but can also damage the hearing of those taking part. Even tennis play and church bell ringing can lead to noise complaints.

Discotheques and rock concerts may exceed hearing damage risk criteria for the musicians, employees and the audience. This sometimes applies also to outdoor concerts. Careful attention to the design of the building can substantially eliminate neighborhood noise problems caused by discotheques. But, there can still be a noise problem outdoors due to customers arriving and leaving.

The general problem of access to leisure activity sites often adds to the road traffic noise problems in particular areas.

5. EXPOSURE TO ENVIRONMENTAL NOISE

Sound waves travel from source to receiver through a variety of media. Outdoors it will be through the atmosphere and will then be influenced by wind turbulence and gradients, air temperature, ground reflections, etc. The amplitude, the spectrum as well as the duration of the sound will be affected. For instance, the sound will be attenuated by air absorption, fog, rain or snow, barriers such as walls and buildings, and by ground effects. However, under certain circumstances, attenuation may not take place, for example, wet snow on ground or at night for thin growth of trees and shrubs.

Indoors noise may travel through the air and the structure of the building and be modified by the sound insulation of walls and windows, the reverberation time of the space, and the design and surface materials of the room. A frequent problem is the transmission of sound from one dwelling to another or even between rooms in the same dwelling. Many dwellings are not adapted to the large diversities in social activities among age generations or to time of the day. Aberrant social behavior is a well recognized noise problem in multi-flat dwellings. Noise from the ventilation system is a common source of complaints.

The extent of the noise problem is large. In the EU countries about 40 % of the population are exposed to road traffic noise with an LAeq,T exceeding 55 dB daytime and 20 % are exposed to levels exceeding 65 dB (Lambert & Vallet, 1994). Taking all exposure to transportation noise together about half of the EU citizens are estimated to live in zones which do not ensure acoustic comfort to residents. More than 30 % are exposed at night to noise levels exceeding 55 dB LAeq which are disturbing to sleep. It is no surprise that annoyance to community noise is widespread among citizens: in some EU-countries 20-25 % are being annoyed by road traffic, 2-15 % by aircraft, and 2-4 % by railway noise (Lambert & Vallet, 1994).

Until now the introduction of noise emission standards for vehicles have had limited impact on the exposure to road traffic noise (Sandberg, 1993). Traffic planning and correction policies may diminish the number of people exposed to the very high community noise levels (>70 dB LAeq) but the number exposed to moderately high levels (55-65 dB LAeq) continues to increase in industrialized countries.

A substantial growth in air transport in Europe is expected in the future; in the U.K. by 50-80 % in passenger movements over ten years. General aviation noise at regional airports will increase (Large & House, 1989). However, at the same time jet aircrafts may become 8 to 12 dB quieter due to regulation. An outlook for exposure to noise has been made by OECD (1991). The number of noise sources is expected to increase and is likely to be accompanied by a deterioration of the noise environment. At the same time, it is expected that the public will become more aware of noise pollution and also be protected from noise problem. The OECD (1991) identifies the following four factors of increasing importance in the future: (1) Expanding use of increasingly numerous and powerful sources of noise.

(2) Wider geographical dispersion of noise sources together with greater individual mobility and spread of leisure activities.

(3) Increasing spread of noise over time particularly in the early morning, evenings and weekends.

(4) Increasing public expectations which are closely linked to increases in incomes and in education levels.

The OECD (1991) report forecasts (a) a strengthening of present noise abatement policies and their applications, (b) a further sharpening of emission standards, (c) a coordination of noise abatement measures and transport planning, particularly designed to reduce mobility, and (d) a coordination of noise abatement measures with urban planning.

High-level noise exposures giving rise to noise-induced hearing deficits are by no means restricted to occupational situations. Such levels can also occur in concerts, discotheques, motor sports, shooting ranges, and leisure activities. Other sources are also important such as music played back in headphones and impulse noise from toys and fireworks. It has also been argued that community noise exposure would be a contributing factor to hearing deficits with increasing age. The existence of such a “sociacusis” waits for final scientific verification since so many other factors and agents are also influencing hearing.

The acoustics of a space designed for speech must primarily ensure clarity and intelligibility. Therefore it is important to design spaces for optimum reverberation time and spatial-temporal aspects including the time delay between the direct and first reflected sound.

Planners need to know the likely effects on the noise pollution in a community of introducing a new noise source as well as increasing the level of an existing source (Diamond & Rice, 1987). There are a number of models to predict annoyance due to a combination of noise sources, such as models of energy summation, of source addition, of source difference, of response summation and response inhibition, and of the (subjectively) dominant source (e.g., Vos, 1992a). Policy makers, when considering applications for new developments, must take into account maximum levels, equivalent levels, frequency of occurrence, and operating time of the major noise sources.

6. ANATOMY, PHYSIOLOGY, AND PSYCHOPHYSICS OF THE AUDITORY SYSTEM

The auditory system is a complex comprising the outer ear, middle ear, cochlea of the inner ear, the connection to the brain through the auditory nerve, and pathways within the brain. Detailed descriptions of the auditory system are found in articles by, for example, Flock (1971), Pickles (1982), and Møller (1983).

6.1 The Outer and the Middle Ear

The outer ear collects sound waves through the auricle (pinna) and the external acoustic meatus which ends with the tympanic membrane (eardrum). The auricle is particularly important for high-frequency directional hearing. The collected sound waves causes a resonance vibration of the eardrum. Transmission is nonlinear which may lead to frequency specific effects. In the middle ear the vibration is transmitted by a chain of three bones (malleus, incus, stapes), (cf. Møller, 1961). The ossicles connects with the inner ear through a window in the cochlea known as the “oval window”. An alternate sound conduction pathway to the inner ear is via bone conduction involving the mastoid bone and the skull.

The so-called middle ear muscle reflex plays a significant role in the effect of noise particularly in regard to masking, loudness, and auditory fatigue. This aural reflex is mediated by two small muscles in the middle ear, the tensor tympani and stapedius, which are attached to the small bones (malleus and stapes) that connect the ear drum with the cochlea. When intense sound occur (above 80 dB), or objects touch the external ear canal, these muscles contract pulling the stapes and tympanic membrane towards the middle ear cavity putting increased resistance to movement into the ossicular chain. This protects the ossicles from excessive movement and damage to themselves and the cochlea. But there is a finite delay in this occurring and impulsive sounds, with a rapid rise time, may be too quick for the reflex to come into operation. The ear then responds in a different way and is more susceptible to damage. For protective criteria, the “peak” level is used. As stated before this is unweighted.

The aural reflex is more responsive to broad band sounds than to pure tones and more responsive to lower frequencies than to higher, and is most readily activated and maintained by intermittent, intense impulses (Borg & Courter, 1989). The middle ear muscle contraction increases the impedance of the middle ear resulting in an attenuated input of sound energy through the cochlea.

6.2 The Cochlear Mechanisms

The cochlea contains the organ of Corti which is located between two fluid-filled chambers. Impulses arise as a result of pressure waves displacing the organ of Corti in response to vibration produced at the oval window of the cochlea. The organ of Corti contains sensory cells, which convert the pressure wave into ionic and electric events which constitute a nerve impulse. The sensory cells have hair-like projections (stereocilia). There are two groups of hair cells located on the basilar membrane of the organ of Corti. The inner hair cells serve as the pre-synaptic sensory receptors which connect to the afferent Type I spiral ganglion nerve cells. The outer hair cells, which are more commonly damaged by noise and ototoxic agents, are believed to serve as an amplification system due to their contractile properties, and their efferent innervation.

According to von Békésy (1960), a particular region of the basilar membrane responds by maximal vibration depending on the frequency of the sound. When the stereocilia of the inner hair cells are bent there is an initiation of action potentials in the sensory nerve endings. The brain interprets the impulses from the place of maximal stimulation of the organ of Corti as a particular pitch of sound (place pitch). This localization is enhanced by the inhibitory effect of centrifugal nerve signals and feedback circuits in the central pathways. Up to a certain frequency range, nerve impulses are time-locked with periods of sound wave (rate pitch). When the sound intensity increases, an ever larger region of the basilar membrane will become involved and more hair cells are being activated. Prolonged exposures to intense sounds may cause degenerative changes in the organ of Corti.

6.3 The Auditory Pathway and the Brain

Neural information is conducted by means of the acoustic (8th) nerve from the organ of Corti to the brain. The pathway to the cerebral cortex involves synaptic relays and the transmission of acoustic information to the cortex of the brain is rather complicated. A number of nuclei have been identified that are connected to form complex integrated systems. The auditory pathway projects on the auditory cortex of the temporal lobes of the brain. Many aspects of auditory processing take place in the cochlea, peripheral auditory nerve and brainstem. Advanced analysis of acoustic stimuli involving recognition and interpretation of sounds occurs in the auditory association cortex. At certain levels of the auditory pathway there are links between the two sides of the brain. Thus, a lesion on one side of the brain is often insufficient to be detected in audiometric testing. The discrepancy in time of arrival of the stimulus in the left and right ear and the inter-aural sound level difference, which is encoded pri-marily at the level of the brainstem, mainly determines the direction and distance of the sound source.

There are also descending, efferent neurons which provide feedback circuits, producing the possibility of inhibition. The central nervous system also controls part of the initiation of nerve impulses in the organ of Corti. The transmission of neural data to the brain from the sensory hair cells isnot just a simple relay to the cortex of the brain. At all steps of this pathway a complex processing takes place which is important to a number of soundqualities such as perceived intensity, perceived pitch, speech feature analysis, and noise identification. The feedback inhibition is especially important for auditory sharpening. Connections to the reticular activating system of the mid-brain are particularly important for the arousal function. It is assumed that the central inhibition suppresses the background noise when one is concentrating on a particular acoustic signal. The auditory system also has connections to motor and autonomic centers of the brain.

6.4 Psychoacoustics

The perceptual attributes of simple sounds mainly include loudness, pitch, timbre, and temporal extent. These correspond to combinations of levels of stimulus intensity, frequency, and duration. The relationship between the physical and perceptual attributes of sound are explored in psychophysical experiments.

6.4.1 Detection of Sounds

The absolute threshold refers to the physical intensity or air pressure of a sound, which elicits a sensation on a specified portion of the occasions on which it occurs (usually 50 %) whereas on the other occasions no sensations are experienced. To understand threshold psychophysics one must make a distinction between traditional psychophysical theory and contemporary information processing theory. Traditional psychophysics puts its main emphasis on the effect of the stimulus (e.g., sound) and on threshold values dividing a physical continuum (sound intensity) into those values that elicit a sensation and those that do not.

It is important to note that the proponents of traditional theory believed the boundary between these values to be fixed at any one moment (the absolute momentary threshold). The traditional model is analogous to the case when sensory transducers function like “smoke detectors”. The basic point made in this regard by contemporary information processing theorists is that there really are no such cutoffs (stimulus threshold values). The Signal Detection Theory argues that the observer always interprets his sensory experiences and decides whether they are caused by a certain stimulus or other factors (e.g., spontaneous neural activity). The decision is determined by the observer’s experience with the stimulus situation and his attitudes (e.g., D.M. Green & Swets, 1966; Baird & Noma, 1978). The perception of traffic noise would be different for an automobile manufacturer and a citizen who does not like automobiles. Therefore, the outcome of a threshold test will depend partly on the stimulus value and partly on the observer’s response criterion. This limits the usefulness of statements of psychophysical threshold results expressed purely in physical stimulus terms. The observation is particularly important in field research where it is impossible to control variables that affect the response criterion. This must be considered in any development of methodology of threshold measurement.

In spite of the theoretical ambiguity of the threshold concept, the threshold of hearing has become a conventional and useful measure of hearing sensitivity and impairment. Since in everyday life sounds are nearly always well above normal threshold, hearing threshold level is primarily useful as a predictor variable. In fact, it predicts auditory performance on speech tasks as well as many other tasks remarkably well even though its effect is only indirect (King, Coles, Lutman, & D.W. Robinson, 1992).

6.4.2 Psychophysical Relationship for Loudness

The prediction of loudness from the physical analysis of different complex sound sources has long been a goal of applied psychoacoustics. Several loudness-evaluation procedures have been proposed (see e.g., Scharf, 1978). The most common are based on weightings of the complete frequency spectrum and are applied as filters in sound level meters. Others are calculation methods for predicting the loudness of complex sounds, and they are usually based on loudness summation of continuous octave or fractional octave bands (e.g., Mark VI, S.S. Stevens, 1961a; see Mark VII, S.S. Stevens, 1972) or critical bands (e.g., Zwicker, 1958). A compilation of studies and a comparison of methods for evaluating (perceived) loudness (or perceived noisiness or annoyance) was performed by Scharf, Hellman and Bauer (1977) and Scharf and Hellman (1979, 1980). The analysis include those procedures that relies on spectral sound properties; among these, S.S. Stevens’s (1961a) Mark VI was found to be the best predictor. The ISO has recommended Mark VI as a method for calculating the loudness of complex sounds (ISO R532, 1966; ISO 532, 1975).

As a rule of thumb, people agree that when moderately intense single component sound such as a tone or a band of noise is raised in intensity by about 10 dB, it sounds twice as loud. This is consistent with the psychophysical power function (S.S. Stevens, 1957a, 1957b) that relates loudness to sound energy,

Y = a I n (7)

where Y stands for loudness perceived and I stands for physical sound intensity and a is a multiplicative constant related to the units of measurement. The exponent n is approximately 0.3 for tones and narrow-band noise (S.S. Stevens, 1975) and somewhat lower (Å0.2) for various community noises (e.g., B. Berglund, U. Berglund, & Lindvall, 1976). Loudness not only depends on sound energy but also on frequency and other physical parameters. At moderate levels, low-frequency sounds (those below 900 Hz) are judged to be less loud than high-frequency sounds (those between 900 to 5,000 Hz) when sounds are of equal physical intensity. The frequency weighting function, referred to as A-weighting, was developed to simulate this effect at low sound levels and for pure tones. It is well known that with the use of this weighting it is necessary to use different level limits for different types of sources. Not only the source itself but also the listener’s attitude is of importance.

The A-weighting function is widely used to obtain index measures of community noise. One should realize, however, that a single weighting function used for various sound pressure levels cannot reflect the perception or other adverse effects of different noises. For example, two sources of community noise that are equal in dBA may differ substantially in loudness (e.g., B. Berglund, U. Berglund, & Lindvall, 1975a, 1976; Goldstein, 1994). The loudness of a complex sound is the sum of the loudnesses of the individual components only if these are widely spaced on the frequency continuum and about equally loud (e.g., Marks, 1978). When the components are not widely spaced, or differ greatly in loudness, mutual inhibition and perceptual interference result in the total loudness being less than the sum of the loudnesses of the components. This knowledge has led to the development of Stevens’s and Zwicker’s procedures for calculating total loudness from physical sound measures (S.S. Stevens, 1961a, 1972; Zwicker & Scharf, 1965). Several studies have been conducted to evaluate the adequacy of these procedures. They do not hold uniformly for all types of stimuli; they are especially weak in predicting loudness of sounds with strong tonal components, discontinuous spectra, and impulsive time structures. For example, one experiment showed that while Stevens’ Mark VI accurately predicted loudness of white noises, it failed to predict the loudness of power line noise (B. Berglund, U. Berglund, & Lindvall, 1986a). This study seems to be the only one that actually tried to use the Mark VI formula with loudnesses of octave bands to predict the loudness of a real community noise.

The conclusion is that the equal loudness contours based on broad-band noise often are not applicable to community noises. However, Zwicker’s procedure (ISO 532, 1975) has been shown to be able to handle tonal components reasonably well, better than Stevens’s procedure (Hellman, 1991). Furthermore, Zwicker’s procedure has been demonstrated to give accurate predictions of (percepeived) loudness for various kinds of community noises, and in addition, surprisingly good performance was shown for complex impulse noises such a shots from rifles or sounds from cannons (B. Berglund, U. Berglund, & Lindberg, 1986).

For many years, regulatory authorities have concentrated on loudness as the sole component responsible for annoyance. In such a case, noise control would be relatively straightforward. However, there are other psychological and physical characteristics of complex sounds that may be more relevant, for example, the intrusiveness of sound (Fidell, Teffeteller, Horonjeff, & D.M. Green, 1979; Fidell & Teffeteller, 1981; Preis, 1987), their sharpness (e.g., Aures, 1985) and fluctuation strength (e.g., Zwicker & Fastl, 199