Middleton W.M. (ed.) Reference Data for Engineers: Radio, Electronics, Computer and Communications

Подождите немного. Документ загружается.

40-8

REFERENCE DATA FOR ENGINEERS

MEASUREMENT

SOUND

Sound Level Meters

A Sound Level Meter (SLM) is an instrument de-

signed to measure directly sound-pressure levels while

incorporating frequency weighting networks which ap-

proximate the inverse of equal loudness contours at

various levels. The current American National Standard

Specification for Sound Level Meters (ANSI S1.4A-

1985) designates three alternate frequency weighting

networks (Fig.

which correspond to the inverse of an

experimentally determined moderate level “equal noisi-

ness contour” and a “flat,” or unweighted, measure.

(In the latter case, the measurement bandwidth is

determined by the microphone and electrical network

characteristics

Standard SLM also complies with

International Electro-technical Commission (IEC)

651.*

An SLM measures the “quasi-root-mean-square”

(rms) value of a time-varying sound pressure. Table

lists the rms detector time constant and corresponding

label as given in ANSI S1.4A-1985. In addition, some

SLMs incorporate a peak-hold (p) which captures the

instantaneous largest sound pressure level.

An integrating-averaging SLM provides a “true-

integrating” detector circuit in lieu of the quasi-rms

detector. These meters are built in accordance with

IEC-804. Typically, they measure time-average sound

level (also called equivalent sound level) and sound

exposure level.

In text, complete description must be given for any

measurement made (e&,

dB B-weighted slow). In

equations, SPLs are abbreviated as

LA,

for A-weighted

fast,

for B-weighted impulse,

LPc

for C-weighted

Definitions for these terms are given

the next section.

FREQUENCY

HERTZ

Fig.

IEC

Standard

and

weighting curves for Sound

Level Meters.

(EPA.

Public Health and Welfare Criteria for

Noise.

550/9-73-002.

2-2.)

TABLE

SOUND

LEVEL METER TIME CONSTANTS

Label RMS Detector Time Constant

Slow

(S)

Fast

(F)

Impulse

(I)

0.1

peak, etc. Maximum should not be confused with peak.

Peak refers to an instantaneous measure, while maxi-

mum refers to a detected level [e.g., max A-weighted

slow

(LAFrnax)l.

Measurement and/or

Laboratory Microphones and

Their Calibration

Current measurement and/or laboratory microphones

are

generally capacitor, electret, or ceramic (piezoelec-

tric) types. These are chosen because of their flat

frequency response, low noise, and stability. One-inch

and 1/2-inch microphones are the most common with

smaller sizes also available. Capacitor microphones

require a polarization voltage but are not potentially as

susceptible to aging or temperature effects. All three are

very high impedance devices requiring a FET follower

close coupled for impedance transformation. For a

given size, the ceramic microphone is on the order of 20

dB less sensitive than the capacitor or electret micro-

phone.

Calibration is normally performed with a close-

coupled single-frequency acoustical source. The acous-

tical signal may be generated mechanically (pistons

driven by a cam and motor) or electromechanically.

Typical frequencies are

1000

Hz and

250

Hz. Other

frequencies, high pressure, and reciprocity calibrators

are also available as needs dictate, but are generally not

required.

Sound Level

In recent years it has become customary to measure

most environmental or industrial work-place (occupa-

tional) noise with the A-weighting.

In the United

States, A-weighted SPL is the commonly designated

sound level. (Special high-amplitude impulse sounds,

such

sonic booms or explosions, which can notice-

ably vibrate structures are measured with C-weight-

ing.)

Sound Power Level

One is usually concerned with the acoustic power

radiated by some source of noise or radiated purpose-

fully by a transducer such as a loudspeaker. Typically,

these sources have a directivity pattern

which

function

angle. Radiated power is the normal compo-

nent of the intensity integrated over an imaginary

surface encompassing the source. (Sufficiently far from

ELECTROACOUSTICS

TABLE

ACOUSTIC POWER

AND

SOUND POWER LEVELS OF TYPICAL NOISE SOURCES*

(A- WEIGHTED)

40-9

Total Power

(A-weighted)

(Watts)

100

000

1000

100

1.0

0.1

0.01

0.001

0.0001

0.000

001

0.000

0001

0.000

000

0.000

000

001

Power Level

(dB re

lo-'*

Watts)

170

160

150

140

130

125

120

110

100

Source

Ram jet

Turbojet engine with afterburner

Turbojet engine, 7000 lb thrust

4-propeller airliner

75-piece orchestra

Peak rms level in

li8-second intervals

Pipe organ

Small aircraft engine

Large chipping hammer

Piano

BBb tuba

[

Blaring radio

Centrifugal ventilating fan (13

000

cfm)

4-fOOt loom

Auto on highway

Vanaxial ventilating fan (1500 cfm)

Voice-shouting (average long-time rms)

Peak rms level in

1/8-second intervals

Voice-conversational level

(average long-time rms)

Voice-very soft whisper

*Space average sound pressure level at

meters

Power level

for

source radiating into a hemisphere.

the source, the intensity is proportional to the pressure

squared.) Table

lists the PWL for some common

noise sources.

DEFINITIONS

The following definitions are in accordance with

ANSI

S12.9-1988

and ANSI

S12.7-1986.

Instantaneous

sound

pressure-Total instantaneous

pressure at a point in the presence of a sound wave

minus the static pressure at that point. Unit: pascal (Pa).

Sound

power (level)-(A,) Sound power: The radiat-

ed acoustic power in a stated frequency band. Unit: watt

(W). (B.) Sound power level: Ten times the common

logarithm

the ratio of the

sound

power to the

reference sound power,

Wo,

lo-''

watts. Unit:

decibel (dB).

log

WIWO

where

is the radiated acoustic power in watts and

is the reference power of

lo-'*

watts.

Sound

pressure (level)-(A.) Sound pressure: Root

mean square of the instantaneous sound pressures

stated frequency band and during a specified time

interval, unless another time-averaging process is indi-

cated. Unit: pascal (Pa).

(B.)

Sound pressure level: Ten

times the common logarithm of the square of the ratio

of the sound pressure to the reference sound pressure,

po,

micropascals. Unit: decibel (dB).

Sound (Level)-(A.) (Frequency-weighted) sound

pressure: Root mean square of the instantaneous sound

pressure which is filtered (frequency-weighted) with a

standard frequency characteristic (e.g.,

etc.)

and exponential time weighted in accordance with the

standardized characteristics slow

(S),

fast (F), impulse

(I),

or peak, both weightings as specified in ANSI

S1.4A-1985. The time weighting should be specified.

The frequency weighting should be specified; other-

wise, A-weighting will be understood. Unit: pascal

(Pa). (B.) Sound level: Ten times the common loga-

rithm of the square of the ratio of the frequency-

weighted (and time-averaged) sound pressure to the

reference sound pressure,

po,

of 20 micropascals. Unit:

decibel (dB).

NOTE:

A, C, or octave band sound levels

are all sound levels having different frequency weight-

ings which must be specified.

Time-average

sound

(level)-(A.) Time-average (fre-

quency-weighted) sound pressure:

(1)

square root of the

quotient of the time integral of frequency-weighted

squared instantaneous sound pressures divided by the

time period of integration in seconds; or (2) square root

of the quotient of the sound exposure, in pascal-squared

seconds, in a specified time period, divided by the time

period of integration in seconds. Unit: pascal (Pa). (B.)

Time-average sound level: Ten times the common

logarithm of the square

the ratio of time-average

(frequency-weighted) sound pressure to the reference

pressure, po, of 20 micropascals. Unit: decibel (dB).

NOTES:

(1.)

Time-average sound level is also termed

equivalent-continuous sound level or equivalent-contin-

uous

frequency-weighted sound pressure level.

(2.)

octave band time-average sound levels are all

time-average sound levels having different frequency

weightings which must be specified.

Peak

sound

pressure (level)-(A.) Peak sound pres-

sure: For any specified time interval, the maximum

absolute value of the instantaneous sound pressure in

that interval. Unit: pascal (Pa). (B.) Peak sound pres-

sure level: Ten times the common logarithm of the

square of the ratio of the peak sound pressure to the

reference sound pressure of 20 micropascals

Unit:

decibel (dB)

Sound

pressure (level)-(A.) Maximum sound pres-

sure: For any specified time interval, the maximum

value of the sound pressure in that interval. Unit: pascal

(Pa).

(B.)

Maximum sound pressure level: Ten times the

common logarithm of the square of the ratio of the

maximum sound pressure to the reference sound pres-

sure of

micropascals. Unit: decibel (dB).

Sound

exposure (level)-(A.) Sound exposure: Time

integral of squared, frequency-weighted instantaneous

sound pressure. The time period of integration should

be specified when the sound exposure of the back-

ground noise is a significant contributor to the total

sound exposure, when the data may be useful for

identifying a source, or when the time period of

integration is otherwise useful. The frequency weight-

ing should be specified; otherwise, A-weighting will be

understood. Unit: pascal-squared second (Pa’s).

NOTES:

(1.)

One pascal-squared second is called a

sound exposure unit.

(2.)

In principle, sound exposure

is independent

the time period of integration. How-

ever,

practice, a measurement of sound exposure may

be limited by the square of the equivalent sound

pressure

the background noise multiplied by the time

period of integration. In such a case, the time period of

integration should be specified.

Some instruments

for measurement of sound exposure include a threshold

sound level below which the instrument does not

accumulate contributions to the integral. If the thresh-

old sound level is above the level of the background

noise, then the time period of integration need not be

stated, but the threshold should be stated. (4.) For most

transient sounds such as an aircraft flyby or a vehicle

pass-by

etc., the A-frequency weighting is preferred.

For short-duration, high-energy impulsive sounds, the

C-frequency weighting is preferred, and measurements

are to be made in accordance with ANSI S12.4-1986.

(B.) Sound exposure level: Ten times the common

logarithm of the ratio of sound exposure to the reference

sound exposure

(Eo)

400

micropascal-squared sec-

onds

[400

(~Pa)~s]. Unit: decibel (dB).

NOTES:

(1.)

a given measurement time period of

seconds, the

sound exposure level

(LET)

is related to the time-average

sound (pressure) level

(LpT)

as follows:

where

is the reference duration of

second.

(2.)

C, or octave-band sound exposure levels are all sound

exposure levels having different frequency weightings

which must be specified.

ENVIRONMENTAL NOISE

Current methods to assess environmental noise em-

body several relatively simple concepts. A fundamental

concept is loudness. The human ear is more sensitive to

the middle frequencies than it is to very low or very high

frequencies.*

for purposes of noise assessment, the

sound measured by a microphone, like the sound

“measured” by the human ear, is filtered to remove

much of the low and high frequencies. The International

Organization for Standardization (ISO) and the Ameri-

can National Standards Institute (ANSI) recommend

the A frequency-weighting curve to accomplish this

filtering, and in the USA (as in most countries), the

Federal Environmental Protection Agency (USEPA),

the Federal Aviation Administration (FAA), etc., have

adopted the use of this A frequency-weighting curve, a

curve which removes much of the low and high frequen-

cies (references

and

3).t

Middle frequencies

are

about middle

to four octaves

above middle

the piano.

The Illinois property-line-noise-source regulations go

further and divide the sound spectrum into octave bands

to create an environment which is more acceptable to

communities than would be an environment regulated only by

the A-weighting cum. These regulations go a second step

further and divide the sound spectrum into

113

octave bands

in order to regulate “pure tones” which may be particularly

bothersome.

40-11

The first rule of current environmental noise assess-

ment makes use of the concept

loudness. In effect, it

states: All else being equal, a louder noise is more

annoying to a community than is a quieter noise. In

other words, all else being equal, community annoy-

ance grows as loudness increases.

Environmental noise regulations typically take one of

two forms. Either they are specified as simple SPL

limits at some measurement locations or they are

specified by time-integrated quantities. The former are

similar in concept to “speed limits” or “discharge

limits” and are relatively easy

enforce. They may

take the form of property line limits such as

many

decibels (A-weighted, slow) or may even be expressed

as octave band or 1/3-octave band limits. (Some SLMs

include an octave band filter set; others have octave and

1/3-octave filter sets available as attachments. The

filters are specified by ANSI

1.1 1- 1966 and IEC

225

For vehicular sources, the limits are usually specified in

terms of a maximum A-weighted fast (or slow) level at a

reference distance (typically

m).

Airports, highways, railways, and other large, dis-

tributed time-varying sources are usually regulated and

assessed by the use of time-integrated measures. Exten-

sive research over the past three decades has shown that

these time-integrated measures better correlate with

human and community response than do the single

event limits described above, but they are more difficult

to use for purposes

enforcement, Time-integrated

measures incorporate the second and third rules of

current noise assessment procedures.

The second rule of environmental noise assessment

deals with the duration or repetition of a sound.

Basically, it states: For the same loudness, a longer-

duration sound is more annoying than a shorter-dura-

tion sound, and sounds repeated many times are more

annoying than are sounds repeated only a few times.

The third rule of noise assessment deals with the time

of day. It basically states: All else being equal, sounds

during some period(s) of the day result in greater overall

net community annoyance than do these same sounds

during other periods of the day.

In terms of noise assessment these three rules imply:

We use a filter (or filters) proportional to hearing

to measure the “loudness,”*

We add up all the sound an event or operation

creates to account for the duration as well as the

loudness,

We add in each event or operation separately to

account

for

repetition, and,

We count each event or operation at night as

equal to ten daytime events.

The Illinois property-line-noise-source octave-band level

spectra are very similar to the inverse of equal loudness

spectra and thus yield results monotonically related to loud-

ness.

To do all of this, we predict or measure the sound

exposure of each event or operation. Sound exposure

can be thought of as a measure of the loudness

multiplied by the time it takes for the event or operation

to occur, Le., loudness times duration.

Next, we add up the sound exposure of each individu-

event or operation in order to develop the total sound

exposure. In this fashion, we account for

all

of the

noise. This point is very important and frequently

misunderstood. Each event or operation in and of itself

can be annoying, can interfere with communications,

etc. But the noise impact to a community experiencing

or 100 of these events each day

far greater than is

the impact to a community experiencing just one of

these events.

Finally, we multiply the sound exposure of each event

or operation which occurs at night by ten, or we divide

its criterion sound exposure by ten before assessing the

noise. In this way, we apply a nighttime adjustment.

This, too, is a very important factor in assessing noise.

Two measures, defined below, are typically used to

quantify noise impact. One is sound exposure

(SE),

and

the other is time-average sound. These two are related

by the time period of measurement. Sound exposure is

proportional to the

total

sound energy during the time

period, and time-average sound is the average sound

during this same time period. Sound exposure and

time-average

sound

level

are

usually expressed in deci-

bels with reference to the standard reference pressure,

p,,

pPa or the standard reference sound exposure,

SE,,

400

(pPa)*-s.

In the United States, with the help of the National

Academy

Sciences (NAS), the

Environmental

Protection Agency (USEPA) created the day-night

sound level (DNL) as the descriptor of choice for

composite environmental noise assessment (references

and

3).*

This descriptor, DNL, is really a measure of

total day-night sound exposure

described in ANSI

S12.9-1988 (reference

4).?

In reality, measures such as

DNL model the relation: Average community response

is proportional to total sound exposure received by that

community during a typical day. The underlying mea-

sure in DNL is total day-night sound exposure

(DNSE)

which is given by:

DNSE

Total

SED,

(10)(Total

SEN,^^^)

and

The best summary of references

and

is:

Protective

Noise Levels-Condensed Version

EPA

Levels Document.

U.S. Environmental Protection Agency,

EPA

550/9-79-100,

November

1978.

Sound exposure is exactly analogous to the concept

light exposure as used in photography. It is the product of the

sound intensity times the duration of the sound. Sound

exposure is discussed in Chapter

reference

and a

formal definition based on

ANSI

S12.9-1988

is given in

Chapter

of reference

40-12

DNL

log [DNSE/(ToSE,)]

where

400,

the number of seconds in a day

This

the actual way we calculate or measure DNL.

ASSESSMENT OF

NOISE

WORKPLACE-INDUSTRIAL

A Sound Level Meter (SLM) set on the A-scale can

be used to assess industrial noise for hearing hazard.

For enforcement purposes, the

Department of Labor

(DOL) has set

(A)

as the limit for eight hours of

exposure per day. Most industrial noise is not uniform

and continuous, and

the regulations of DOL include

procedures for determining the equivalent exposure.

These procedures incorporate a 5-dB trading rule in that

halving the exposure time allows the noise limit to rise

dB.* The present limits

are

given in Table

an alternative to using an SLM to measure

industrial noise exposure, several personnel dosimeters

are available that directly measure a worker's dose

(usually as a percent of the eight-hour limit). These

devices directly incorporate the 5-dB trading rule in

their electronics.

Exposure to impulsive or impact noise should not

exceed

140

dB peak sound pressure level.

Most European countries

use

a 3-dB or

equal

energy

rule, and the USAF

uses

a 4-dB rule.

TABLE

PERMISSIBLE

NOISE

EXPOSURES*

Sound Level,

dBA Slow Response Duration per Day, Hours

6 92

100

102

105

110

less

115

When the daily noise exposure is composed of two or

more periods of noise exposure of different levels, their

combined effect should be considered, rather than the indi-

vidual effect of each. If the sum of the following fractions:

CIIT,

C21T2

Cnic

exceeds unity, where

indi-

cates the total time of exposure at the nth noise level and

indicates the total time of exposure permitted at that level,

then

the

mixed exposure should be considered to exceed the

limit value.

REFERENCE DATA FOR ENGINEERS

SOUND

ENCLOSED

ROOMS*

Indoors, acoustics is of concern for a variety of

reasons. Auditoria, broadcast studios, etc., require low

background noise levels and "good" acoustical design;

offices require varying background levels and communi-

cations capabilities depending on use; and in the

industrial workplace, conservation of hearing and abili-

ty to communicate

are

usually the primary concerns.

In any room, the reverberant SPL increases as the

acoustic power entering the room increases and decreas-

es as sound absorption in the room increases and/or as

sound power flows from the room. Sources of sound

may be internal to rooms, such as machinery or

loudspeakers, or may be external. In the latter case, the

sound flows through walls, windows, ventilation ducts,

etc., from other building spaces or components, or

from outdoors. In either case, the sound field in a room

is composed of two parts, the direct field and the

reverberant field. The direct field of a source is

where,

(2@c) (rad/m),

pressure (N/m2),

speed

sound

(ds),

density of air (kg/m3)

source strength (m3/s*rad),

directivity (rad)?,

distance from the source (m).

Near a source, the sound field may be very compli-

cated; the pressure and velocity may not be in phase,

and hence the power radiated is not proportional to the

pressure squared. This is particularly a problem when

determining the noise power radiated by machinery in

the industrial environment.

By substitution in equation

the direct field intensi-

(Id)

of a source can be written as

WQ/4m2

(Eq.

10)

where,

radiated source power (watts),

See,

for example, reference

2.5, 2.6,

2.8.

The directivity factor of a source is the ratio of the sound

pressure squared, at some fixed distance and specified direc-

tion (such as the axis of a loudspeaker), to the mean-square

sound pressure at the same distance averaged over all direc-

tions from the source. The distance must be great enough

that

the

sound appears to diverge spherically from the

effective acoustic center of the source.

Sound-intensity meters are a valuable aid in

these

situations, as they indicate the direction of the intensity.

40- 13

directivity (rad),

distance from the source (meters).

The total intensity

(I)

in an enclosed room (direct

plus reverberant field) is

WQ/4m2

4WlR

(Eq.

11)

The quantity

is the room constant defined by

Sa(l

(Eq. 12)

where,

total surface area of the room,

average room absorption coefficient.

Fig. 3 illustrates the sound fields in a room. In terms

of the sound pressure level,

SPL

PWL

log

(Q/4m2

4/R)

(Eq. 13)

where,

0.2 dB if metric units are used,

10.5 dB if English units are used.

Standing Sound Waves

Resonant conditions in sound studios cause standing

waves by reflections from opposing parallel surfaces,

such as ceiling-floor and parallel walls, resulting in

serious peaks in the

reverberation-timelfrequency

curve. Standing sound waves in a room can be consid-

ered comparable to standing electrical waves in an

improperly terminated transmission line where the

transmitted power is not fully absorbed by the load.

In properly proportioned rooms, resonances can be

effectively reduced

and

standing waves practically elim-

inated by introducing numerous absorbing surfaces

disposed obliquely. Thus, large-order reflections can be

DISTANCE IN METERS

Fig.

Description

the sound field around a

sound

source

a reverberant room.

(From Ginn,

Architectural

Acoustics.

Briiel and Kjaer,

1978.

avoided by breaking them up into numerous smaller

reflections. The object is to prevent sound reflection

back to the point of origin until after several absorbing

reflections.

The most desirable ratios of dimensions for broadcast

studios are given in Fig.

Reverberation Time

Reverberation time varies with frequency and is

measured by the time required for a sound, when

suddenly interrupted, to die away or decay to a level

decibels below the original sound. Measurement of

reverberation time is the easiest way to find the average

absorption coefficient

(Z)

and its associated value of

reflectivity (1

in an acoustic environment. Con-

versely, absorption coefficients and room sizes are used

to predict the reverberation time for new construction or

remodeling.

Measurement

Reverberation Time

The reverberation time of an enclosed space that

already exists is an important quantity that is relatively

easy to measure. The degree of accuracy required is

determined by the use to which the data will be put. The

majority of day-to-day reverberation measurements are

taken with quite satisfactory results by using the inter-

rupted-noise method (Fig.

5).

The Schroeder-Kuttruff method (Fig.

employs a

2.7-ms rectangular pulse that is used to excite a

standard %-octave bandpass filter. The resulting “ring-

ing”

the filter provides a statistically reliable signal

that allows highly repeatable decay recordings and

excellent resolution of the EDT.

Calculation

Reverberation Time

The Sabine equation (equation

for room dimen-

sions

feet, equation 14a for room dimensions in

meters) is commonly used to calculate reverberation

time,

RTbo

0.049VIS

(Eq.

14)

RT&

0.161VlS

(Eq. 14a)

where,

is the time in seconds required for a sound to

decay

dB,

is the volume

the room,

is the boundary surface area,

is the average absorption coefficient.

The value

is:

(s,a,

s2a2

. . .

s,a,)/S

(Eq. 15)

40-14

REFERENCE DATA FOR ENGINEERS

103

IO'

LOW CEILING

lo3

104

lo4

IO4

lo4

105

IO5

106

VOLUME

CUBIC FEET

Fig.

Preferred

room

dimensions based

2L'3

ratio. Permissible deviation

percent.

(Courtesy Acoustical Society

America

and

RCA.)

where,

RT60

O.O49V/[-S

ln(1

CU)]

(Eq. 16)

with room dimensions in feet

a,,

cyz,

etc., are the absorption values for the

boundary areas with which they are associated,

s,a,

is the total absorption of the people,

furniture, etc., present in the room.

Note that

SCU

can be replaced by

the total

absorption in the room. This concept is useful when

considering the effects of surfaces plus objects

people

the room.

the limiting case, the Sabine equation predicts a

finite reverberation time in a room with

100%

absorp-

tion present, and for true absorption values in excess

0.63, this equation can give

values in excess of 1.0

(100% absorption). The Norris-Eyring equation (equa-

tion 16 or 16a) gives

values from

for true

absorption values when calculated from actual

RT60

measurements.

s2,

etc., are boundary surface areas,

Ri"60

0.161V/[-S ln(1

E)]

(Eq. 16a)

with room dimensions in meters.

If the value of

RT60

is measured and the correspond-

ing value of

is calculated, insertion of this value of

into the expression -In(

converts the

value

into a Sabine

For example, -ln(l

0.63)

0.99.

Absorption Coefficients

When tables of absorption values

are

examined, it is

important to know which formula was used in determin-

ing the numerical values. If the values were obtained

from the Sabine equation, be sure to use the Sabine

POWFR

MICROPHONE

Fig.

The

interrupted-noise method.

ELECTROACOUSTICS

POWFR

40-15

. .

-. .

REVERBERATION

1/3-OCTAVE

lI3-OCTAVE

FILTER SET PROCESSOR FILTER SET

TEST

LEVEL-METER SWITCH LOUDSPEAKER

MICROPHONE

GRAPHIC

LEVEL

RECORDER

Fig. 6. The Schroeder-Kuttruff method.

equation variations consistently for any further manipu-

lations of the data. If the Norris-Eyring equation was

used, remain consistent in its use for any further

manipulations of the data. The Sabine formula is used

in measuring and specifying most building materials.

Hence, the Sabine equation must be used unless the

absorption values are converted (see above). Table

gives absorption coefficients of some typical building

materials.

Optimum Reverberation Time

Optimum, or most desirable, midfrequency

(500-

1000 Hz) reverberation time varies with room size and

use, such as music, speech, etc. (see Fig.

7).

INDOOR DESIGN NOISE

LEVELS AND SPEECH

COMMUNICATION

Preferred Speech

Interference Levels/

Speech Interference Levels

Noise interference with speech is usually a masking

process. For many noises, the measurement and calcu-

lation of articulation index can be simplified further by

the use of a three-band analysis.

The bands chosen are

the octave bands centered on

500,

1000, and

2000

Hz.

The arithmetic average of these three

SPLs

in these

three bands gives the quantity called the three-band

preferred octave speech-interference level

(PSIL).

One

can use this level

for

determining when telephone use or

speech communication is easy or difficult and what

changes in levels are required to improve the situation

(Figs.

and

9).

For satisfactory intelligibility of difficult speech ma-

terial, maximum permissible values of speech-inter-

ference levels for men with average voice strengths are

given in Fig.

which is an extension by Webster

of Beranek’s work.* It is assumed in this chart that

there are no reflecting surfaces nearby, that the speak-

er is facing the listener, and that the spoken material

is not already familiar to the listener. For example,

the speech-interference level in a factory might be

dB, which is high. The chart indicates that the

two people must ordinarily be no more than two feet

apart in order to be understood satisfactorily. If the

words spoken are carefully selected and limited in

number, intelligible speech will be possible at greater

distances.

If a number of conversations are to be held in the

same reverberant room, the procedure is more compli-

cated. This chart cannot be used on the basis of the

background-noise level before the conversations are in

progress, because a given conversation will be subject

to interference from the noise produced by all the other

conversations. The general procedure for calculating a

speech-interference level under those conditions has not

been completely worked out.

Not only background, equipment, and other occu-

pant noise affects speech intelligibility. In a live room

(too much reverberation), speech syllables are smeared

by reflected noise at the other end, even though the

noise is not very great. That is, if the distance from

speaker (or loudspeaker) to listener is too great or the

room too reverberant, then the speaker’s own sound

energy may make the speech unintelligible.

a rule of

thumb, the direct field from the speaker (or loud-

speaker) must exceed his (its) own reverberant field by

in order to maintain good speech intelligibility and

prevent feedback.

Beranek,

“The Design

Speech Communications

Systems.”

Proc.

ZEEE,

Vol. 35, No. 9, Sept. 1947, pp.

880-890.

Webster,

C. “Effects

Noise

Speech Intelligibili-

ty.”

Proc. Conference Noise as a Public Health Hazard,

Washington, D. C., June 13-14, 1968, ASHA Rpt. 4.

Washington, DC: The American Speech and Hearing Associ-

ation, 1969, pp. 49-73.

TABLE

10.

COEFFICIENTS OF

GENERAL

BUILDING

MATERIALS

AND

FURNISHINGS

Complete tables of coefficients of the various materials that normally constitute the interior finish of rooms may be found in

the various books

architectural acoustics. This short list will be useful in making simple calculations.

Coefficients

Materials

125

250

500

1000

2000

4000Hz

Brick, unglazed

Brick, unglazed, painted

Carpet, heavy, on concrete

Same,

oz hairfelt or

foam rubber

Same,

with impermeable latex

backing

oz hairfelt or

foam rubber

Concrete block, coarse

Concrete block, painted

Fabrics (draperies)

Light velour,

oz per

yd,

Medium velour,

per sq yd,

Heavy velour,

oz per

yd,

hung straight, in contact with wall

draped to half area

Concrete or terrazzo

Linoleum, asphalt, rubber, or cork

Wood

Wood parquet in asphalt

concrete

Large panes of heavy plate glass

Ordinary window glass

Gypsum board,

Yz"

nailed to

2Xl's,

16"

O.C.

Marble or glazed tile

Openings

Floors

tile on concrete

Glass

Stage, depending on furnishings

Deep balcony, upholstered seats

Grills, ventilating

Plaster, gypsum or lime, smooth

finish on tile or brick

Plaster, gypsum or lime, rough finish

on lath

Same,

with smooth finish

Plywood paneling,

thick

Water surface, as in a swimming pool

Air, Sabins per

1000

cubic feet

50%

0.03

0.01

0.02

0.08

0.36

0.10

0.03

0.07

0.14

0.01

0.02

0.15

0.04

0.18

0.35

0.29

0.01

0.013

0.14

0.28

0.008

0.03

0.01

0.06

0.24

0.27

0.44

0.05

0.04

0.31

0.35

0.01

0.03

0.11

0.04

0.06

0.25

0.10

0.01

0.015

0.10

0.22

0.008

0.03 0.01

0.02 0.02

0.14 0.37

0.57 0.69

0.39 0.34

0.31 0.29

0.06 0.07

0.11 0.17

0.49 0.75

0.55

0.72

0.015 0.02

0.03 0.03

0.10 0.07

0.07

0.06

0.04 0.03

0.18

0.12

0.05 0.04

0.01 0.01

0.25-0.75

0.15-0.50

0.02 0.03

0.50-

.OO

0.06 0.05

0.06 0.04

0.17 0.09

0.013 0.015

0.9

0.05

0.02

0.60

0.71

0.48

0.39

0.09

0.24

0.70

0.02

0.03

0.06

0.02

0.07

0.02

0.04

0.10

0.020

2.3

0.07

0.03

0.65

0.73

0.63

0.25

0.08

0.35

0.60

0.65

0.02

0.07

0.02

0.04

0.09

0.02

0.05

0.03

0.11

0.025

7.2

Absorption of Seats and Audience

Values given are in Sabins per square foot of seating area per unit.

125

250

500Hz 1OOOHz

2000

4000

Audience, seated in upholstered seats,

Unoccupied cloth-covered upholstered

Unoccupied leather-covered upholstered

Wooden pews, occupied, per sq ft of

Chairs, metal or wood seats,

per

ft of floor area

0.60 0.74

0.88

0.96 0.93 0.85

seats, per sq ft of floor area

0.49 0.66

0.80

0.88

0.82

0.70

seats, per

ft of floor area

0.44 0.54

0.60 0.62 0.58 0.50

floor area

0.57 0.61

0.75 0.86 0.91 0.86

each, unoccupied

0.15 0.19

0.22 0.39 0.38 0.30

ELECTROACOUSTICS

QUIET

40-17

/,IIII1

I I

CHURCHES

CONCERT

STUbIOS OPEN PLAN

[SPEECH) OFFICES

IIIII

Illll

III,~

100

Z10

500

ROOM

VOLUME

CUBIC

FEET

1000

Fig.

Optimum midfrequency reverberation times.

Interior Design Noise Levels

Table 11 contains a list of planning levels for activi-

ties conducted in interior spaces.* The (interior) plan-

ning levels for

exterior noise and

for

interior equipment

In offices or office-type environments, Preferred Noise

Criteria Curves (PNC) are frequently used to select and

specify

the

ambient.

that is

not

continuously operated

are given in terms of

values.

Continuous noise sources,

for example

ventilating systems or other mechanical equipment,

emit steady-state noise that is measured in terms of

L,.

The sources must be considered separately. As indicat-

ed in the right-hand column of Table

11,

permissible

values are

and 10 dB less than

Le,

values for the same

activity.

Interior steady-state noise levels more than

5-10

below the levels specified in Table 11 are not desirable.

Annoyance will actually increase with the lowered

background noise levels because individuals will hear

intruding sounds that normally would be masked by the

steady-state noise. Occasionally, where adequate noise

insulation cannot be provided, increasing the continu-

ous

background noise levels over the values shown in

Table 11 will provide better masking of intruding

intermittent sounds. For such occasions, the character-

istics of both the intruding noises and the background

noise should be considered during the design of the

facility.

SOUND

ISOLATIONP

Sound entering a room from external areas can be

controlled by quieting the source, by adding absorption

to the source room (if the source

indoors) to lower the

the A-weighted noise level produced by the

ventilation or mechanical systems (or other interior noise

sources) which operate more

less continuously. The

value for design should be the noise level produced in the

space during the time of occupancy while the equipment is at

the typical mode of operation.

-F

See, for example, references 2.1 through 2.11.