TECHNICAL NOISE SUPPLEMENT
October 1998
ACKNOWLEDGEMENT
The Technical Noise Supplement evolved from a 1991 draft document titled Noise Technical Analysis
Notes ( NoTANs) prepared by the former Division of New Technology and Research ( TransLab). Dick
Wood, a TransLab staff member, and I were the authors of the 1991 draft NoTANs. For various reasons
NoTANs was never finalized. However, I edited, re- wrote, and revised original contents of the draft
NoTANs and incorporated these with new information in this Technical Noise Supplement ( TeNS). Dick
Wood's early contributions to NoTANs were extremely valuable to the development of TeNS, and I
want to sincerely thank him for assisting me in the effort leading to the final completion of TeNS.
I also owe a great debt of gratitude to:
Professor Dean Karnopp of the University of California at Davis, Department of Mechanical and
Aeronautical Engineering, for his constructive review and contributions to TeNS.
l      
l       Joya Gilster, Civil Engineering Student, for her detailed technical review of TeNS.
Keith Jones for his constructive review and comments, and for believing in TeNS and me. His
undying support and enthusiasm for this project is directly responsible for the emergence of TeNS
from NoTAN. Without his support TeNS would not have succeeded.
l      
To all mentioned above, I sincerely appreciate your help.
Environmental Engineering-
Noise, Air Quality, and Hazardous Waste Management Office,
October 1998
http:// www. dot. ca. gov/ hq/ Environmental/ offdocs/ hazdocs/ tens/ ack. htm [ 1/ 19/ 2000 3: 46: 31 PM]
TECHNICAL NOISE SUPPLEMENT
October 1998
N- 1
N- 1000 INTRODUCTION AND OVERVIEW
N- 1100 INTRODUCTION
The purpose of this Technical Noise Supplement ( TeNS) is to provide technical background
information on transportation- related noise in general and highway traffic noise in
particular. It is designed to elaborate on technical concepts and procedures referred to in
the Caltrans Traffic Noise Analysis Protocol ( the Protocol). The contents of this Supplement
are for informational purposes only and unless specifically referred to as such in the
Protocol they are not official policy, standard or regulation. The procedures recommended
in TeNS are in conformance with “ industry standards”.
This document can also be used as a “ stand alone” document for training purposes, or as a
reference for technical concepts, methodology, and terminology needed to acquire a basic
understanding of transportation noise with emphasis on highway traffic noise.
N- 1200 OVERVIEW
TeNS consists of nine sections, numbered N- 1000 through N- 9000. With the exception of
N- 1000 ( this section), each section covers a specific subject of highway noise. A brief
description of the subjects follows.
· N- 2000, BASICS OF HIGHWAY NOISE covers the physics of sound as it pertains to
characteristics and propagation of highway noise, the effects of noise on humans,
and ways of describing noise.
· N- 3000, MEASUREMENTS AND INSTRUMENTATION covers the “ why, where, when,
and how” of noise measurements, and briefly discusses various noise measuring
instruments and operating procedures.
· N- 4000, SCREENING PROCEDURE was developed to aid in determining whether or
not a highway project has the potential to cause a traffic noise impact. If the project
passes the screening procedure, prudent engineering judgment should still be
exercised to determine if a detailed analysis is warranted.
· N- 5000, DETAILED ANALYSIS – TRAFFIC NOISE IMPACTS gives guidance for
studying those projects failing the screening procedure, projects that are
controversial, sensitive, or projects where the net effects of topography and shielding
are complex and ambiguous.
TECHNICAL NOISE SUPPLEMENT
October 1998
N- 2
· N- 6000, DETAILED ANALYSIS - NOISE BARRIER DESIGN CONSIDERATIONS outlines
the major aspects that affect the acoustical design of noise barriers. These include
the dimensions, location, material, and optimization of noise barriers; the acoustical
design of overlapping noise barriers ( to provide maintenance access to areas behind
barriers) and drainage openings in noise barriers. It also points out some pitfalls
and cautions.
· N- 7000, NOISE STUDY REPORTS discusses the contents of noise study reports.
· N- 8000, SPECIAL CONSIDERATIONS covers some special controversial issues that
frequently arise, such as reflective noise, the effects of noise barriers on distant
receivers, and shielding provided by freeway landscaping.
· N- 9000, GLOSSARY provides terminology and definitions common in transportation
noise.
In addition to the above sections the BIBLIOGRAPHY provides a listing of literature used
as a source of information in TeNS.
3
N- 2000 BASICS OF HIGHWAY NOISE
The following sections introduce the fundamentals of sound and provide sufficient detail for
the reader to understand the terminology and basic factors involved in highway traffic noise
prediction and analysis. Those who are actively involved in noise analysis are encouraged
to seek out more detailed textbooks and reference books in order to acquire a deeper
understanding of the subject.
N- 2100 PHYSICS OF SOUND
N- 2110 Sound, Noise, Acoustics
Sound is a vibratory disturbance created by a moving or vibrating source, in the pressure
and density of a gaseous, liquid medium or in the elastic strain of a solid which is capable
of being detected by the hearing organs. Sound may be thought of as mechanical energy of
a vibrating object transmitted by pressure waves through a medium to human ( or animal)
ears. The medium of main concern is air. In absence of any other qualifying statements,
sound will be considered airborne sound, as opposed to, for example, structureborne or
earthborne sound.
Noise is defined as ( airborne) sound that is loud, unpleasant, unexpected or undesired, and
may therefore be classified as a more specific group of sounds. Perceptions of sound and
noise are highly subjective: one person's music is another's headache. The two terms are
often used synonymously, although few would call the sound that emanates from a
highway anything but noise.
Sound ( and noise) is actually a process that consists of three components: 1) the sound
source, 2) the sound path, and 3) the sound receiver. All three components must be present
for sound to exist. Without a source to produce sound, there obviously is no sound.
Likewise, without a medium to transmit sound pressure waves there is also no sound. And
finally, sound must be received, i. e. a hearing organ, sensor, or object must be present to
perceive, register, or be affected by sound or noise. In most situations, there are many
different sound sources, paths, and receivers, instead of just one of each.
Acoustics is the field of science that deals with the production, propagation, reception,
effects, and control of sound. The field is very broad, and transportation related noise and
its abatement covers just a small, specialized part of acoustics.
4
N- 2120 Speed of Sound
When the surface of an object vibrates in air, it compresses a layer of air as the surface
moves outward, and produces a rarefied zone as the surface moves inward. This results in
a series of high and low air pressures waves ( relative to the steady ambient atmospheric
pressure) alternating in sympathy with the vibrations. These pressure waves - not the air
itself - move away from the source at the speed of sound, or approximately 343 m/ s ( 1126
ft/ sec) in air of 20o C. The speed of sound can be calculated from the following formula:
c = 1.401
P
ρ

 

 
( eq. N- 2120.1)
Where:
c = Speed of Sound at a given temperature, in meters per second ( m/ s)
P = Air pressure in Newtons per Square Meter ( N/ m2) or Pascals ( Pa)
ρ = Air density in kilograms of mass per cubic meter ( Kg/ m3)
1.401 = the ratio of the specific heat of air under constant pressure to that of air in
a constant volume.
For a given air temperature and relative humidity, the ratio P/ ρ tends to remain constant in
the atmosphere, because the density of air will reduce or increase proportionally with
changes in pressure. Thus the speed of sound in our atmosphere is independent of air
pressure. However, when air temperature changes, only ρ changes, while P does not. The
speed of sound is therefore temperature dependent, and also somewhat humidity
dependent since humidity affects the density of air. The effects of the latter with regards to
the speed of sound, however, can be ignored for our purposes. The fact that speed of
sound changes with altitude, has nothing to do with the change in air pressure, and is only
caused by the change in temperature.
For dry air of 0o Celsius, ρ = 1.2929 Kg/ m3. At a standard air pressure of 760 mm Hg, the
pressure in Pa = 101,329 Pa. Using eq. N- 2120.1, the speed of sound for standard
pressure and temperature can be calculated:
(. )( )
.
1 401
101329
1 2929
= 331.4 m/ sec, or 1087.3 ft/ sec. From this base value, the variation
with temperature is described by the following equations:
Metric Units ( m/ s): c = 331.4 1+
Tc
273.2 ( eq. N- 2120.2)
5
English Units ( ft/ sec): c = 1051.3 1+
Tf
459.7 ( eq. N- 2120.3)
Where:
c = speed of sound in m/ s ( metric) or ft/ sec ( English)
Tc = Temperature in degrees Celcius ( include minus sign for below zero)
Tf = Temperature in degrees Fahrenheit ( include minus sign for below zero)
The above equations show that the speed of sound increases/ decreases as the air
temperature increases/ decreases. This phenomenon plays an important role in the
atmospheric effects on noise propagation, specifically through the process of refraction,
which is discussed in section N- 2143 ( Meteorological Effects and Refraction).
N- 2130 Sound Characteristics
In its most basic form, a continuous sound can be described by its frequency or wavelength
( pitch) and its amplitude ( loudness).
N- 2131 Frequency, Wavelength, Hertz
For a given single pitch of sound, the sound pressure waves are characterized by a
sinusoidal periodic ( recurring with regular intervals) wave as shown in Figure N- 2131.1.
The upper curve shows how sound pressure varies above and below the ambient
atmospheric pressure with distance at any given time. The lower curve shows how particle
velocity varies above zero ( molecules moving right) and below zero ( molecules moving left).
Particle velocity describes the motion of the air molecules in response to the pressure
waves. It does not refer to the velocity of the waves, otherwise known as the speed of
sound. The distance ( λ) between crests of both curves is the wavelength of the sound.
The number of times per second that the wave passes from a period of compression
through a period of rarefaction and starts another period of compression, is referred to as
the frequency of the wave ( see Figure N- 2131.2).
6
Figure N- 2131.1. Sound Pressure
Frequency is expressed in cycles per second, or Hertz ( Hz). One Hertz equals one cycle per
second. High frequencies are sometimes more conveniently expressed in units of Kilo Hertz
( KHz) or thousands of Hertz. The extreme range of frequencies that can be heard by the
healthiest human ears spans from 16 to 20 Hz on the low end to about 20000 Hz ( or 20
KHz) on the high end. Frequencies are heard as the pitch or tone of sound. High pitched
sounds produce high frequencies, low pitched sounds produce low frequencies. Very- low-frequency
airborne sound of sufficient amplitude may be felt before it can be heard, and is
often confused with earthborne vibrations. Sound below 16 Hz is referred to as infrasound,
while high frequency sound above 20000 Hz is called ultrasound. Both infra- and
ultrasound are not audible to humans. However, many animals can hear or sense
frequencies extending well into one or both of these regions. Ultrasound also has various
applications in industrial and medical processes, specifically in cleaning, imaging, and
drilling.
The distance traveled by a sound pressure wave through one complete cycle is referred to
as the wavelength. The duration of one cycle is called the period. The period is the inverse
of the frequency. For instance, the frequency of a series of waves with periods of 1/ 20 of a
second is 20 Hertz; a period of 1/ 1000 of a second is 1000 Hz, or 1 KHz. Although low
frequency earthborne vibrations, such as earthquakes and swaying of bridges or other
structures are often referred to by period, the term is rarely used in expressing airborne
sound characteristics.
PRESSURE
WAVELENGTH
λ
PARTICLE
VELOCITY
7
Figure N- 2131.2 shows that as the frequency of sound pressure waves increases, their
wavelength shortens, and vice versa. The relationship between frequency and wavelength
is linked by the speed of sound, as shown in the following equations:
λ = cf
( eq. N- 2131.1)
Also: f = c
λ
( eq. N- 2131.2)
and: c = fλ ( eq. N- 2131.3)
Where:
λ = Wavelength ( m or ft)
c = Speed of Sound ( 343.3 m/ s, or 1126.5 ft/ sec at 20o C, or 68o F)
f = Frequency ( Hertz)
In the above equations, care must be taken to use the same units ( distance units in either
meters or feet, and time units in seconds) for wavelength and speed of sound. Although the
speed of sound is usually thought of as a constant, we have already seen that it actually
varies with temperature. The above mathematical relationships hold true for any value of
the speed of sound. Frequency is normally generated by mechanical processes at the
source ( wheel rotation, or back and forth movement of pistons, to name a few), and is
Figure N- 2131.2 - Frequency and Wavelength
λ
Wavelength, λ
Short Wavelength, High Frequency
Long Wavelength, Low Frequency
8
therefore not affected by air temperature. As a result, wavelength usually varies inversely
with the speed of sound as the latter varies with temperature.
The relationships between frequency, wavelength and speed of sound can easily be
visualized by using the analogy of a train traveling at a given constant speed. Individual
boxcars can be thought of as the sound pressure waves. The speed of the train ( and the
individual boxcars) is analagous to the speed of sound, while the length of each boxcar is
the wavelength. The number of boxcars passing a stationary observer each second depict
the frequency ( f). If the value of the latter is 2, and the speed of the train ( c) is 108 km/ hr
( or 30 m/ s), the length of each boxcar ( λ) must be: c/ f = 30/ 2 = 15m.
Using equation N- 2131.1 we can develop a table showing frequency and associated
wavelength. Table N- 2131.1 shows the frequency/ wavelength relationship at an air
temperature of 20o C ( 68o F).
Table N- 2131.1 Wavelength of Various Frequencies
Frequency Wavelength
at 20o C ( 68o F)
( Hz) m ( ft)
16 21 ( 70)
31.5 11 ( 36)
63 5.5 ( 18)
125 2.7 ( 9)
250 1.4 ( 4.5)
500 0.7 ( 2.3)
1000 0.34 ( 1.1)
2000 0.17 ( 0.56)
4000 0.09 ( 0.28)
8000 0.04 ( 0.14)
16000 0.02 ( 0.07)
We can check the validity of Table N- 2131.1 by multiplying each frequency by its
wavelength, which in each case should equal the speed of sound. Notice that, due to
rounding, multiplying frequency and wavelength gives varying results for the speed of
sound in air, which for 20o C should be constant at 343.3 m/ sec ( 1126.5 ft/ sec).
Frequency is an important component of noise analysis. Virtually all acoustical
phenomena are frequency- dependent, and knowledge of frequency content is essential.
Some applications of frequency analysis will be discussed in sections N- 2135 ( A- weighting,
Noise Levels) and N- 2136 ( Octave and third octave Bands, Frequency Spectrums).
N- 2132 Sound Pressure Levels ( SPL), Decibels ( dB)
9
Referring back to Figure N- 2131.1, we remember that the pressures of sound waves
continuously changes with time or distance, and within certain ranges. The ranges of these
pressure fluctuations ( actually deviations from the ambient air pressure) are called the
amplitude of the pressure waves. Whereas the frequency of the sound waves is reponsible
for the pitch or tone of a sound, the amplitude determines the loudness of the sound.
Loudness of sound increases and decreases with the amplitude.
Sound pressures can be measured in units of micro Newtons per square meter ( μN/ m2)
called micro Pascals ( μPa). 1 μPa is approximately one- hundredbillionth of the normal
atmospheric pressure. The pressure of a very loud sound may be 200,000,000 μPa, or
10,000,000 times the pressure of the weakest audible sound ( 20 μPa). Expressing sound
levels in terms of μPa would be very cumbersome, however, because of this wide range. For
this reason, sound pressure levels ( SPL) are described in logarithmic units of ratios of actual
sound pressures to a reference pressure squared. These units are called bels, named after
Alexander G. Bell. In order to provide a finer resolution, a bel is subdivided into 10 decibels
( deci or tenth of a bel), abbreviated dB. In its simplest form, sound pressure level in
decibels is expressed by the term:
Sound Pressure Level ( SPL) = 10 Log10 ( 1
0
p
p ) 2
dB ( eq. N- 2132.1)
Where:
P 1 is sound pressure
P 0 is a reference pressure, standardized as 20 μPa
The standardized reference pressure, P 0 , of 20 μPa, is the absolute threshold of hearing in
healthy young adults. When the actual sound pressure level is equal to the reference
pressure, the expression:
10Log 10 ( 1
0
p
p ) 2
= 10Log 10 ( 1) = 0 dB
Note that 0 dB is not the absence of any sound pressure. Instead, it is an extreme value
that only those with the most sensitive ears can detect. Thus, it is possible to refer to
sounds as less than 0 dB ( negative dB), for sound pressures that are weaker than the
threshold of human hearing. For the majority of people, the threshold of hearing is higher
than 0 dB, probably closer to 10 dB.
10
N- 2133 Root Mean Square ( Rms), Relative Energy
Figure N- 2131.1 depicted a sinusoidal curve of pressure waves. The values of the pressure
waves were constantly changing, increasing to a maximum value above normal air pressure
then deceasing to a minimum value below normal air pressure, in a repetitive fashion. This
sinusoidal curve is associated with a single frequency sound, also called a pure tone. Each
successive sound pressure wave has the same characteristics as the previous wave. The
amplitude characteristics of such a series of simple waves can then be described in various
ways, all of which are simply related to each other. The two most common ways to describe
the amplitude of the waves is in terms of the peak sound pressure level ( SPL) and the root
mean square ( r. m. s.) SPL.
The peak SPL simply uses the maximum or peak amplitude ( pressure deviation) for the
value of P 1 in equation N- 2132.1. The peak SPL therefore only uses one value ( the absolute
value of the peak pressure deviation) of the continuously changing amplitudes. The r. m. s.
value of the wave amplitudes ( pressure deviations) uses all the positive and negative
instantaneous amplitudes, not just the peaks. It is derived by squaring the positive and
negative instantaneous pressure deviations, adding these together and dividing the sum by
the number of pressure deviations. The result is called the mean square of the pressure
deviations, and taking the square root of this mean value is called the r. m. s. value. Figure
N- 2133.1 shows the peak and r. m. s. relationship for a sinusoidal wave. The r. m. s. is 0.707
times the peak value.
Figure N- 2133.1 Peak Vs. r. m. s. Sound Pressures
In terms of discrete samples of the pressure deviations the mathematical expression is:
r. m. s. value = √( 1  n( a1
2 + a2
2 + ..... an
2 )/ n) ( eq. N- 2133.1)
Peak Negative Pressure
Peak Positive Pressure
r. m. s. Pressure
PRESSURE
0 Atmospheric Pressure
+
-
a1 a2 an
11
Sound pressures expressed in r. m. s. are proportional to the energy contents of the waves,
and are therefore the most important and often used measure of amplitude. Unless
otherwise mentioned, all SPL’s are expressed as r. m. s. values.
N- 2134 Relationship Between Sound Pressure Level, Relative Energy,
Relative Pressure, and Pressure
Table N- 2134.1 shows the relationship between r. m. s. SPL’s, relative sound energy, relative
sound pressure, and pressure.
Note that SPL’s, Relative Energy, and Relative Pressure are based on a Reference Pressure
of 20 μPa, and by definition all referenced to 0 dB. The Pressure values are the actual
r. m. s. pressure deviations from local ambient atmospheric pressure.
The most useful relationship is that of SPL ( dB) and Relative Energy. Relative Energy is
unitless. Table N- 2134.1 shows that for each 10 dB increase in SPL, the acoustic energy
increases 10- fold. For instance an SPL increase from 60 to 70 dB increases the energy 10
times. Acoustic energy can be thought of as the energy intensity ( energy per unit area) of a
certain noise source, such as a heavy truck ( HT), at a certain distance.
For example, if one HT passing by an observer at a given speed and distance produces an
SPL of 80 dBA, then the SPL of 10 HT’s identical to the single HT would be 90 dBA, if they
all could simultaneously occupy the same space, and travel at the same speed and distance
from the observer.
Since SPL = 10 Log 10 ( P1/ P2) 2, the acoustic energy is related to SPL as follows:
( P1/ P2) 2 = 10SPL/ 10 ( eq. N- 2134.1)
This relationship will be useful in understanding how to add and subtract SPL’s in the next
section.
N- 2135 Adding and Subtracting Sound Pressure Levels ( SPL’s)
Since decibels are logarithmic units, sound pressure levels cannot be added or subtracted
by ordinary arithmetic means. For example, if one automobile produces a SPL of 70 dB
when it passes an observer, two cars passing simultaneously would not produce 140 dB.
In fact, they would combine to produce 73 dB. This can be shown mathematically as
follows.
12
Figure N- 2134.1 - Relationship between Sound Pressure Level, Relative Energy, Relative
Pressure, and Sound Pressure
Sound Pressure
Level, dB
10 Log10 ( 1
0
p
p ) 2
Relative Energy
( 1
0
p
p ) 2
Relative Pressure
( 1
0
p
p )
Sound Pressure,
μPa
P 1
200 dB 1020 1010
140 dB 1014 107
134 dB 108 μPa
130 dB 1013
120 dB 1012 106
114 dB 107 μPa
110 dB 1011
100 dB 1010 105
94 dB 106 μPa
90 dB 109
80 dB 108 104
74 dB 105 μPa
70 dB 107
60 106 103
54 dB 104 μPa
50 dB 105
40 dB 104 102
34 dB 103 μPa
30 dB 103
20 dB 102 101
14 dB 102 μPa
10 dB 101
0 dB 100 = 1 = Ref. 100 = 1 = Ref. P
1 = P
0 = 20 μPa
Note: P 0 = 20 μPa = Reference Pressure
The sound pressure level ( SPL) from any one source observed at a given distance from the
source may be expressed as 10log10( P 1 / P 0 ) 2 ( see eq. N- 2132.1) The SPL from two equal
sources at the the same distance would therefore be:
SPL = 10log10 [( P 1 / P 0 ) 2+( P 1 / P 0 ) 2] = 10log10[ 2( P 1 / P 0 ) 2].
13
This is can be simplified as 10log10( 2)+ 10log10( P 1 / P 0 ) 2. Because the logarithm of 2 is
0.301, and 10 times that would be 3.01, the sound of two equal sources is 3 dB greater
than the sound level of one source. The total SPL of the two automobiles would therefore
be 70 + 3 = 73 dB.
Adding and Subtracting Equal SPL’s - The previous example of adding the noise levels of
two cars, may be expanded to any number of sources. The previous section discussed the
relationship between decibels and relative energy. The ratio ( P 1 / P 0 ) 2 is the relative
( acoustic) energy portion of the expression SPL = 10log10( P 1 / P 0 ) 2, in this case the relative
acoustic energy of one source. This must immediately be qualified with the statement that
this is not the acoustic power output of the source. Instead, the expression is the relative
acoustic energy per unit area received by the observer. We may state that N identical
automobiles, or other noise sources, would yield an SPL of:
SPL( Total) = SPL( 1) + 10log10( N) ( eq. N- 2135.1)
in which: SPL( 1) = SPL of one source
N = number of identical sources to be added ( must be ≥ 0)
Example: If one noise source produces 63 dB at a given distance, what would be the noise
level of 13 of the same sources combined at the same distance?
Solution: SPL( Total) = 63 + 10log10( 13) = 63 + 11.1 = 74.1 dB
Equation N- 2135.1 may also be rewritten as:
SPL( 1) = SPL( Total) - 10log10( N) ( eq. N- 2135.2)
This form is useful for subtracting equal SPL’s.
Example: The SPL of 6 equal sources combined is 68 dB at a given distance. What is the
noise level produced by one source?
Solution: SPL( 1) = 68 dB - 10log10( 6) = 68 - 7.8 = 60 dB
In the above examples, adding equal sources actually constituted multiplying one source by
the number of sources. Conversely, subtracting equal sources was performed by dividing
the total. For the latter, we could have written eq. N- 2135.1 as SPL( 1) = SPL( Total) +
10log10( 1/ N). The logarithm of a fraction yields a negative result, so the answers would
have been the same.
14
The above excercises can be further expanded to include other useful applications in
highway noise. For instance, if one were to ask what the respective SPL increases would be
along a highway if existing traffic were doubled, tripled and quadrupled ( assuming that
traffic mix, distribution, and speeds would not change), we could make a reasonable
prediction using equation N- 2135.1. In this case N would be the existing traffic ( N= 1), N= 2
would be doubling, N= 3 tripling, and N= 4 quadrupling the existing traffic. Since the
10log10( N) term in eq. N- 2135.1 represents the increase in SPL, we can solve N for N= 2,
N= 3, and N= 4. The results would respectively be: + 3 dB, + 4.8 dB, and + 6 dB.
The question might also come up what the SPL decrease would be if the traffic would be
reduced by a factor of two, three, or four. In this case N = 1/ 2, N= 1/ 3, and N = 1/ 4,
respectively. Applying the 10log10( N) term for these values of N would result in - 3 dB, - 4.8
dB, and - 6 dB, respectively.
The same problem may come up in a different form. For instance, if the traffic flow on a
given facility is presently 5000 vehicles per hour ( vph) and the present SPL is 65 dB at a
given location next to the facility, what would the expected SPL be if future traffic increased
to 8000 vph? Solution: 65 + 10log10( 8000/ 5000) = 65 + 2 = 67 dB.
The N value may thus represent an integer, a fraction, or a ratio. However, N must always
be greater than 0! Taking the logarithm of 0 or a negative value is not possible.
Adding and Subtracting Unequal Noise Levels. If noise sources are not equal, or if
equal noise sources are at different distances, the 10log10( N) term cannot be used. Instead,
the SPL’s have to be added or subtracted individually, using the SPL and relative energy
relationship in section N- 2134 ( eq. N- 2134.1). If the number of SPL’s to be added is N, and
SPL( 1), SPL( 2), ..... SPL( n) represent the 1st, 2nd, and nth SPL, respectively, the addition is
accomplished by:
SPL( Total) = 10log10[ 10SPL( 1)/ 10+ 10SPL( 2)/ 10 + ......... 10SPL( n)/ 10] ( eq. N- 2135.3).
The above equation is the general equation for adding SPL’s. The same equation may be
used for subtraction also ( simply change the “+” to “-” for the term to be subtracted.
However, the result between the brackets must always be greater than 0!
For example, find the sum of the following sound levels: 82, 75, 88, 68, 79. Using
eq. 2135.3, the total SPL is:
SPL = 10 Log 10 ( 1068/ 10 + 1075/ 10 + 1079/ 10 + 1082/ 10 + 1088/ 10) = 89.6 dB
15
Adding SPL’s Using a Simple Table - When combining sound levels, the following table
may be used as an approximation.
Table N- 2135.1 Decibel Addition
When Two Decibel Add This Amount
Values Differ By: to the Higher Value: Example:
0 or 1 dB 3 dB 70+ 69 = 73
2 or 3 dB 2 dB 74+ 71 = 76
4 to 9 dB 1 dB 66+ 60 = 67
10 dB or more 0 dB 65+ 55 = 65
This table yields results within ± 1 dB of the mathematically exact value and can easily be
memorized. The table can also be used to add more than two SPL’s. First, sort the list of
values, from lowest to highest. Then, starting with the lowest values, combine the first two,
add the result to the third value and continue until only the answer remains.
Example: find the sum of the sound levels used in the above example, using Table N-
2135.1. First, rank the values from low to high:
68 dB
75 dB
79 dB
82 dB
88 dB
?? dB Total
Using table 2135.1 add the first two noise levels. Then add the result to the next noise
level ............, etc.
a. 68 + 75 = 76,
b. 76 + 79 = 81,
c. 81 + 82 = 85,
d. 85 + 88 = 90 dB ( For comparison, using eq. 2135.3, the
total SPL was 89.6 dB).
Two decibel addition rules are important. First, when adding a noise level with another
approximately equal noise level, the total noise level rises 3 dB. For example doubling the
traffic on a highway would result in an increase of 3 dB. Conversely, reducing traffic by
one half, the noise level reduces by 3 dB. Second, when two noise levels are 10 dB or more
apart, the lower value does not contribute significantly (< 0.5 dB) to the total noise level.
16
For example, 60 + 70 dB ≈ 70 dB. The latter means that if a noise level measured from a
source is at least 70 dB, the ambient noise level without the target source must not be
more than 60 dB to avoid risking contamination.
N- 2136 A- Weighting, Noise Levels
Sound pressure level alone is not a reliable indicator of loudness. The frequency or pitch of
a sound also has a substantial effect on how humans will respond. While the intensity
( energy per unit area) of the sound is a purely physical quantity, the loudness or human
response depends on the characteristics of the human ear.
Human hearing is limited not only to the range of audible frequencies, but also in the way
it perceives the sound pressure level in that range. In general, the healthy human ear is
most sensitive to sounds between 1,000 Hz - 5000 Hz, and perceives both higher and lower
frequency sounds of the same magnitude with less intensity. In order to approximate the
frequency response of the human ear, a series of sound pressure level adjustments is
usually applied to the sound measured by a sound level meter. The adjustments, or
weighting network, are frequency dependent.
The A- scale approximates the frequency response of the average young ear when listening
to most ordinary everyday sounds. When people make relative judgements of the loudness
or annoyance of a sound, their judgements correlate well with the A- scale sound levels of
those sounds. There are other weighting networks that have been devised to address high
noise levels or other special problems ( B- scale, C- scale, D- scale etc.) but these scales are
rarely, if ever, used in conjunction with highway traffic noise. Noise levels for traffic noise
reports should be reported as dBA. In environmental noise studies A- weighted sound
pressure levels are commonly referred to as noise levels.
Figure N- 2136.1 shows the A- scale weighting network that is normally used to approximate
human response. The zero dB line represents a reference line; the curve represents
frequency- dependent attenuations provided by the ear’s response. Table N- 2136.1 shows
the standardized values ( ANSI S1.4, 1983). The use of this weighting network is signified
by appending an " A" to the sound pressure level as dBA, or dB( A).
The A- weighted curve was developed from averaging the statistics of many psycho- acoustic
tests involving large groups of people with normal hearing in the age group of 18- 25 years.
The internationally standardized curve is used world wide to address environmental noise
and is incorporated in virtually all environmental noise descriptors and standards. Section
N- 2200 covers the most common of these, applicable to transportation noise.
17
Figure N- 2136.1 A- Weighting Network
- 60
- 50
- 40
- 30
- 20
- 10
0
10
16 31.5 63 125 250 500 1K 2K 4K 8K 16K
Center Frequency, Hertz
Table N- 2136.1 “ A”- Weighting Adjustments for 1/ 3 Octave Center Frequencies
Frequency, Hz “ A” - Weighting, dB Frequency, Hz “ A” - Weighting, dB
16 - 56.7 630 - 1.9
20 - 50.5 800 - 0.8
25 - 44.7 1000 0
31.5 - 39.4 1250 + 0.6
40 - 34.6 1600 + 1.0
50 - 30.6 2000 + 1.2
63 - 26.2 2500 + 1.3
80 - 22.5 3150 + 1.2
100 - 19.1 4000 + 1.0
125 - 16.1 5000 + 0.5
160 - 13.4 6300 - 0.1
200 - 10.9 8000 - 1.1
250 - 8.6 10K - 2.5
315 - 6.6 12.5K - 4.3
400 - 4.8 16K - 6.6
500 - 3.2 20K - 9.3
Source: American National Standards Institute ( ANSI S1.4 ( 1983).
Sound level meters used for measuring environmental noise have an A- weighting network
built in for measuring A- weighted sound levels. This is accomplished through electronic
filters, also called band pass filters. As the name indicates, each filter allows the passage of
a selected range ( band) of frequencies only, and attenuates its sound pressure level to
modify the frequency response of the sound level meter to approximately that of the A-weighted
curve and the human ear.
A range of noise levels associated with common in- and outdoor activities are shown in
Table N- 2136.2. The decibel scale is open- ended. As was discussed previously, 0 dB or
dBA should not be construed as the absence of sound. Instead, it is the generally accepted
threshold of best human hearing. Sound pressure levels in negative decibel ranges are
inaudible to humans. On the other extreme, the decibel scale can go much higher than
shown in Table N- 2136.2. For example, gun shots, explosions, and rocket engines can
18
reach 140 dBA or higher at close range. Noise levels approaching 140 dBA are nearing the
threshold of pain. Higher levels can inflict physical damage on such things as structural
members of air and spacecraft and related parts. Section N- 2301 discusses the human
response to changes in noise levels.
Table N- 2136.2 - Typical Noise Levels
COMMON OUTDOOR
ACTIVITIES
NOISE LEVEL
dBA
COMMON INDOOR
ACTIVITIES
--- 110--- Rock Band
Jet Fly- over at 300 m ( 1000 ft)
--- 100---
Gas Lawn Mower at 1 m ( 3 ft)
--- 90---
Diesel Truck at 15 m ( 50 ft), Food Blender at 1 m ( 3 ft)
at 80 km/ hr ( 50 mph) --- 80--- Garbage Disposal at 1 m ( 3 ft)
Noisy Urban Area, Daytime
Gas Lawn Mower, 30 m ( 100 ft) --- 70--- Vacuum Cleaner at 3 m ( 10 ft)
Commercial Area Normal Speech at 1 m ( 3 ft)
Heavy Traffic at 90 m ( 300 ft) --- 60---
Large Business Office
Quiet Urban Daytime --- 50--- Dishwasher Next Room
Quiet Urban Nighttime --- 40--- Theater, Large Conference
Quiet Suburban Nighttime Room ( Background)
--- 30--- Library
Quiet Rural Nighttime Bedroom at Night, Concert
--- 20--- Hall ( Background)
Broadcast/ Recording Studio
--- 10---
Lowest Threshold of Human
Hearing
--- 0--- Lowest Threshold of Human
Hearing
N- 2137 Octave and Third Octave Bands, Frequency Spectra
Very few sounds are pure tones ( consisting of a single frequency). To represent the
complete characteristics of a sound properly, it is necessary to break the total sound down
into its frequency components; that is, determine how much sound ( sound pressure level)
comes from each of the multiple frequencies that make up the sound. This representation
of frequency vs sound pressure level is called a frequency spectrum. Spectrums ( spectra)
usually consist of 8 to 10 octave bands, more or less spanning the frequency range of
human hearing ( 20- 20,000 Hz) . Just as with a piano keyboard, an octave represents the
frequency interval between a given frequency and twice that frequency. Octave bands are
internationally standardized and identified by their " center frequencies" ( actually geometric
means).
19
Because octave bands are rather broad, they are frequently subdivided into thirds to create
1/ 3- octave bands. These are also standardized. For convenience, 1/ 3- octave bands are
sometimes numbered from band No. 1 ( 1.25 Hz third- octave center frequency, which
cannot be heard by humans) to band No. 43 ( 20000 Hz third- octave center frequency).
Within the extreme range of human hearing there are 30 third- octave bands ranging from
No. 13 ( 20 Hz third- octave center frequency), to No. 42 ( 16,000 Hz third- octave center
frequency).
Table N- 2137.1 shows the ranges of the standardized octave and 1/ 3- octave bands, and
band No’s.
Frequency spectra are used in many aspects of sound analyses, from studying sound
propagation to designing effective noise control measures. Sound is affected by many
different frequency- dependent physical and environmental factors. Atmospheric
conditions, site characteristics, and materials and their dimensions used for sound
reduction are some of the more important examples.
Sound propagating through the air is affected by air temperature, humidity, wind and
temperature gradients, vicinity and type of ground surface, obstacles and terrain features.
These factors are all frequency dependent.
The ability of a material to transmit noise depends on the type of material ( concrete, wood,
glass, etc.), and its thickness. Different materials will be more or less effective at
transmitting noise depending on the frequency of the noise. See section N- 6110 for a
discussion of Transmission Loss ( TL) and Sound Transmission Class ( STC).
Wavelengths serve to determine the effectiveness of noise barriers. Low frequency noise,
with its long wavelengths, passes easily around and over a noise barrier with little loss in
intensity. For example, a 16 Hz noise with a wavelength of 21 m ( 70 ft) will tend to pass
right over a 5 m ( 16 ft) high noise barrier. Fortunately, A- weighted traffic noise tends to
dominate in the 250 to 2000 Hz range with wavelengths in the order of 0.2 - 1.4 m ( 0.6 -
4.5 ft). As will be discussed later, noise barriers are less effective at lower frequencies, and
more effective at higher ones.
20
Table N- 2137.1 Standardized Band No’s, Center Frequencies, 1/ 3 Octave and Octave
Bands, and Octave Band Ranges
Band No. Center Frequency,
Hz
1/ 3- Octave Band
Range, Hz
Octave Band
Range, Hz
12 16 14.1 - 17.8 11.2 - 22.4
13 20 17.8 - 22.4
14 25 22.4 - 28.2
15 31.5 28.2 - 35.5 22.4 - 44.7
16 40 35.5 - 44.7
17 50 44.7 - 56.2
18 63 56.2 - 70.8 44.7 - 89.1
19 80 70.8 - 89.1
20 100 89.1 - 112
21 125 112 - 141 89.1 - 178
22 160 141 - 178
23 200 178 - 224
24 250 224 - 282 178 - 355
25 315 282 - 355
26 400 355 - 447
27 500 447 - 562 355 - 708
28 630 562 - 708
29 800 708 - 891
30 1000 891 - 1120 708 - 1410
31 1250 1120 - 1410
32 1600 1410 - 1780
33 2000 1780 - 2240 1410 - 2820
34 2500 2240 - 2820
35 3150 2820 - 3550
36 4000 3550 - 4470 2820 - 5620
37 5000 4470 - 5620
38 6300 5620 - 7080
39 8000 7080 - 8910 5620 - 11200
40 10K 8910 - 11200
41 12.5K 11.2K - 14.1K
42 16K 14.1K - 17.8K 11.2K - 22.4K
43 20K 17.8 - 22.4
Source: Bruel & Kjaer Pocket Handbook - Noise, Vibration, Light, Thermal Comfort; September 1986
Figure N- 2137.1 shows a conventional graphic representation of a typical octave- band
frequency spectrum. The octave bands are depicted as having the same width, even though
each successive band should increase by a factor of two when expressed linearly in terms of
one Hertz increments.
21
Figure N- 2137.1 - Typical Octave Band Frequency Spectrum
FREQUENCY SPECTRUM
0
10
20
30
40
50
60
70
80
90
100
31.5 63 125 250 500 1K 2K 4K 8K 16K
Center Frequency, Hertz
Sound Pressure Level, dB
A frequency spectrum can also be presented in tabular form. For example, the data used to
generate Figure N- 2137.1 is illustrated in tabular form in Table N- 2137.2.
Table N- 2137.2 Tabular Form of
Octave Band Spectrum
Octave Band
Center Frequency, Hz
Sound Pressure
Level, dB
31.5 75
63 77
125 84
250 85
500 80
1000 ( 1K) 75
2000 ( 2K) 70
4000 ( 4K) 61
8000 ( 8K) 54
16000 ( 16K) 32
Total Sound Pressure Level = 89 dB
Often, we are interested in the total noise level, or the summation of all octave bands.
Using the data shown in Table N- 2137.2 we may simply add all the sound pressure levels,
as was explained in section N- 2135 ( Adding and Subtracting Decibels). The total noise level
for the above octave band frequency spectrum is 89 dB.
The same sort of charts and tables can be compiled from 1/ 3- octave band information. For
instance, if we had more detailed 1/ 3- octave information for the above spectrum, we could
construct a third octave band spectrum as shown in Figure N- 2137.2 and Table 2137.2.
22
Note that the total noise level does not change, and that each subdivision of three 1/ 3-
octave bands adds up to the total octave band shown in the previous example.
Figure N- 2137.2 - Typical 1/ 3- Octave Band Frequency Spectrum
FREQUENCY SPECTRUM
0
20
40
60
80
100
25 50 100 200 400 800 1.6K 3.2K 6.3K 12.5K
Center Frequency, Hertz
dB
Frequency spectrums are usually expressed in linear, unweighted sound pressure levels
( dB). However, they may also be A- weighted by applying the adjustments from Table N-
2136.1. For example, the data in Table N- 2137.2 can be “ A”- weighted as follows ( rounded
to nearest dB) as shown in Table N- 2137.3.
Table N- 2137.2 Tabular Form of
Octave Band Spectrum
1/ 3- Octave Band
Center Frequency, Hz
Sound Pressure
Level, dB
1/ 3- Octave Band
Center Frequency, Hz
Sound Pressure
Level, dB
25 68 800 71
31.5 69 1000 ( 1K) 70
40 72 1.25K 69
50 72 1.6K 68
63 72 2K 65
80 73 2.5K 61
100 76 3.2K 58
125 79 4K 55
160 81 5K 53
200 82 6.3K 52
250 80 8K 50
315 79 10K 39
400 77 12.5K 31
500 75 16K 25
630 73 20K 20
Total Sound Pressure Level = 89 dB
Table N- 2137.3 Adjusting Linear
Octave Band Spectrum to A- weighted Spectrum
23
Octave Band
Center Frequency, Hz
Sound Pressure
Level, dBA
31.5 75 - 39 = 36
63 77 - 26 = 51
125 84 - 16 = 68
250 85 - 9 = 76
500 80 - 3 = 77
1000 ( 1K) 75 - 0 = 75
2000 ( 2K) 70 + 1 = 71
4000 ( 4K) 61 + 1 = 62
8000 ( 8K) 54 - 1 = 53
16000 ( 16K) 32 - 7 = 25
Total Sound Pressure Level = 89 dB( Lin), and 81.5 dBA
The total A- weighted noise level now becomes 81.5 dBA, compared with the linear noise
level of 89 dB. In other words, the original linear frequency spectrum with a total noise
level of 89 dB sounded to the human ear as having a total noise level of 81.5 dBA.
However, a linear noise level of 89 dB with a different frequency spectrum, could have
produced a different A- weighted noise level, either higher or lower. The reverse may also be
true. Actually, there are theoretically an infinite amount of frequency spectrums that could
produce either the same total linear noise level or the same A- weighted spectrum. This is
an important concept, because it can help explain a variety of phenomena dealing with
noise perception. For instance, some evidence suggests that changes in frequencies are
sometimes perceived as changes in noise levels, even though the total A- weighted noise
levels do not change significantly. Sec. N- 8000 ( Special Problems) deals with some of these
phenomena.
N- 2138 White Noise, Pink Noise
White noise is noise with a special frequency spectrum that has the same amplitude ( level)
for each frequency interval over the entire audible frequency spectrum. It is often
generated in laboratories for calibrating sound level measuring equipment, specifically its
frequency response. One might expect that the octave or third- octave band spectrum of
white noise would be a straight line. This is, however, not true. Beginning with the lowest
audible octave, each subsequent octave spans twice as many frequencies than the previous
ones, and therefore contains twice the energy. This corresponds with a 3 dB step increase
for each octave band, and 1 dB for each third octave band.
24
Pink noise, in contrast, is defined as having the same amplitude for each octave band ( or
third- octave band), rather than for each frequency interval. Its octave or third- octave band
spectrum is truly a straight, “ level” line over the entire audible spectrum. Pink noise
generators are therefore conveniently used to calibrate octave or third- octave band
analyzers.
Both white and pink noise sound somewhat like the static heard from a radio that is not
tuned to a particular station.
N- 2140 Sound Propagation
From the source to the receiver noise changes both in level and frequency spectrum. The
most obvious is the decrease in noise as the distance from the source increases. The
manner in which noise reduces with distance depends on the following important factors:
• Geometric Spreading from Point and Line Sources
• Ground Absorption
• Atmospheric Effects and Refraction
• Shielding by Natural and Manmade Features, Noise Barriers, Diffraction, and
Reflection
N- 2141 Geometric Spreading from Point and Line Sources
Sound from a small localized source ( approximating a " point" source) radiates uniformly
outward as it travels away from the source in a spherical pattern. The sound level
attenuates or drops- off at a rate of 6 dBA for each doubling of the distance ( 6 dBA/ DD).
This decrease, due to the geometric spreading of the energy over an ever increasing area, is
referred to as the inverse square law. Doubling the distance increases each unit area,
represented by squares with sides “ a” in Figure N- 2141.1, from a2 to 4a2.
Since the same amount of energy passes through both squares, the energy per unit area at
2D is reduced 4 times from that at distance D. Thus, for a point source the energy per unit
area is inversely proportional to the square of the distance. Taking 10 log10 ( 1/ 4) results in
a 6 dBA reduction ( for each doubling of distance). This is the point source attenuation rate
for geometric spreading.
25
Figure N- 2141.1 Point Source Propagation ( Spherical Spreading)
As can be seen in Figure N- 2141.2, based on the inverse square law the change in noise
level between any two distances due to the spherical spreading can be found from:
dBA2 = dBA1 + 10 Log10 [( D1/ D2)] 2 =
= dBA1 + 20 Log10 ( D1/ D2) ( eq. N- 2141.1)
Where:
dBA1 is the noise level at distance D1, and
dBA2 is the noise level at distance D2
Figure N- 2141.2 Change in Noise Level with Distance Due to Spherical Spreading
However, highway traffic noise is not a single, stationary point source of sound. The
movement of the vehicles makes the source of the sound appear to emanate from a line
( line source) rather than a point when viewed over some time interval ( see Figure N- 2141.3).
This results in cylindrical spreading rather than the spherical spreading of a point source.
a
a
a
a
2a
2a
2a
2a
Area = a2
Area = 4a2
D
2D
POINT SOURCE
Source Rec. 2
D1 D2
Rec. 1
26
Since the change in surface area of a cylinder only increases by two times for each doubling
of the radius instead of the four times associated with spheres, the change in sound level is
3 dBA per doubling of distance. The change in noise levels for a line source at any two
different distances due to the cylindrical spreading becomes:
dBA2 = dBA1 + 10 Log10 ( D1/ D2) ( eq. N- 2141.2)
Where:
dBA1 is the noise level at distance D1, and conventionally the known noise level
dBA2 is the noise level at distance D2 , and conventionally the unknown noise
level
Note: the expression 10 Log10 ( D1/ D2) is negative when D2 is greater than D1, positive
when D1 is greater than D2, and the equation therefore automatically accounts for the
receiver being farther out or closer in with respect to the source ( Log10 of a number less
than 1 gives a negative result; Log10 of a number greater than 1 is positive, and Log10 ( 1) =
0).
Figure N- 2141.3 Line Source Propagation ( Cylindrical Spreading)
N- 2142 Ground Absorption
Most often, the noise path between the highway and the observer is very close to the
ground. Noise attenuation from ground absorption and reflective wave canceling adds to
the attenuation due to geometric spreading. Traditionally, the access attenuation has also
been expressed in terms of attenuation per doubling of distance. This approximation is
done for simplification only, and for distances of less than 60 m ( 200 feet) prediction results
based on this scheme are sufficiently accurate. The sum of the geometric spreading
attenuation and the excess ground attenuation ( if any) is referred to as the attenuation rate,
a
a
b 2b
Area = ab
Area = 2ab
D
2D
Line Source
27
or drop- off rate. For distances of 60 m ( 200 feet) or greater the approximation causes
excessive inaccuracies in predictions. The amount of excess ground attenuation depends
on the height of the noise path and the characteristics of the intervening ground or site. In
practice this excess ground attenuation may vary from nothing to 8- 10 dBA or more per
doubling of distance. In fact, it varies as the noise path height changes from the source to
the receiver and also changes with vehicle type since the source heights are different. The
complexity of terrain is another factor that influences the propagation of sound by
potentially increasing the number of ground reflections. Only the most sophisticated
computer model( s) can properly account for the interaction of soundwaves near the ground.
In the mean time, for the sake of simplicity two site types are currently used in traffic noise
models:
1. HARD SITES - These are sites with a reflective surface between the source and the
receiver such as parking lots or smooth bodies of water. No excess ground
attenuation is assumed for these sites and the changes in noise levels with distance
( drop- off rate) is simply the geometric spreading of the line source or 3 dBA/ DD
( 6dBA/ DD for a point source).
2. SOFT SITES - These sites have an absorptive ground surface such as soft dirt,
grass or scattered bushes and trees. An excess ground attenuation value of 1.5
dBA/ DD is normally assumed. When added to the geometric spreading results in
an overall drop- off rate of 4.5 dBA/ DD for a line source ( 7.5 dBA/ DD for a point
source).
The combined distance attenuation of noise due to geometric spreading and ground
absorption in the above simplistic scheme can be generalized with the following formulae:
dBA2 = dBA1 + 10 Log10 ( D1/ D2) 1 + α ( Line Source) ( eq. N- 2142.1)
dBA2 = dBA1 + 10 Log10 ( D1/ D2) 2 + α ( Point Source) ( eq. N- 2141.2)
where: α is a site parameter which takes on the value of 0 for a hard site and 0.5
for a soft site.
The above formulae may be used to calculate the noise level at one distance if the noise
level at another distance is known. The “ α scheme” is just an approximation. It is used in
older versions of the FHWA Highway Traffic Noise Prediction Model. Caltrans research has
shown that for average traffic and “ soft site” characteristics, the α scheme is fairly accurate
within 30 m ( 100 ft) from a typical highway. Between 30 - 60 m ( 100 - 200 ft) form a
28
highway, the algorithm results in average over predictions ( model predicted noise levels
higher than actual) of 2 dBA. At 60 - 150 m ( 200 - 500 ft) over predictions average about 4
dBA.
Following are some typical examples of distance adjustment calculations using equations
N- 2141.1 and N- 2141.2:
1. The maximum noise level of truck passing by an observer is measured to be 83 dBA
at a distance of 25 m. What is the maximum noise level at 62 m if the terrain is
considered a soft site?
Solution: The truck is a point source; α for a soft site = 0.5. Hence, at 62 m the
noise level is:
83 dBA + 10 Log10 ( 25/ 62) 2 + 0.5= 83 + (- 9.9) = 73.1 dBA. ( eq. N- 2141.2)
2. The average noise level from a two- lane highway is 65 dBA at a receiver located 50
m from the centerline. The ground between the highway and receiver is a grassy
field. What noise level can be expected for a receiver 20 m from the centerline of the
same highway?
Solution: The two- lane highway may be considered a line source ( a series of moving
point sources). The site parameter α is 0.5 ( grassy field is a soft site). Hence, at 20
m the estimated noise level is:
65 dBA + 10 Log10 ( 50/ 20) 1 + 0.5 = 65 + (+ 6.0) = 71 dBA ( eq. N- 2141.1)
Notice that in the first example the known noise level was closer to the highway
than the unknown one; in the second example the reverse was true.
3. The average noise level from a single truck passby, measured from the time the
truck can first be heard ( above the ambient noise) to the time that the truck’s noise
dips below ambient noise, is 62 dBA at a distance of 35 m. What is the average
noise level of the truck at 50 m, if the the site is hard?
Solution: In this case the line source formula should be used. The difference
between example 1 and this example is that in 1 the maximum noise level was
29
measured. The maximum noise level is an instantaneous noise level, occurring at
one location only: presumably the closest point to the observer. In this example the
noise was an average noise level, i. e. the truck noise was measured at many
different locations representing the entire passby and therefore a series of point
sources that may be represented by a line source. Hence, eq. N- 2141.1 should be
used with α = 0. The answer is 60.5 dBA at 50 m.
Table N- 2142.1 shows a simple generalization regarding the use of point or line source
distance attenuation equations for various source types, instantaneous noise and time-averaged
noise levels.
Sec. N- 5500 contains additional discussions on how to use the appropriate drop- off rate in
the noise prediction models.
Table N- 2142.1 Use of Point and Line Source Distance Attenuation Equations.
NOISE LEVEL AT STATIONARY RECEIVERS
SOURCE TYPE INSTANTANEOUS
( Usually maximum)
TIME- AVERAGED
Single, Stationary Point
Source ( e. g. idling truck,
pump, machinery)
Use Point Source Equation
( eq. N- 2142.2)
Use Point Source
Equation
( eq. N- 2142.2)
Single, Moving Point
Source ( e. g. moving
truck):
Use Point Source Equation
( eq. - N 2142.2)
Use Line Source Equation
( eq. N- 2142.1)
Series of Point Souces on
a Line, Stationary or
Moving: ( e. g. highway
traffic)
Use Line Source Equation
( eq. N- 2142.1)
Use Line Source Equation
( eq. N- 2142.1)
N- 2143 Atmospheric Effects and Refraction
Research by Caltrans and others has shown that atmospheric conditions can have a
profound effect on noise levels within 60 m ( 200 ft) from a highway. Wind has shown to be
the single most important meteorological factor within approximately 150 m ( 500 ft), while
vertical air temperature gradients are more important over longer distances. Other factors
such as air temperature and humidity, and turbulence, also have significant effects.
Wind. The effects of wind on noise are mostly confined to noise paths close to the ground.
The reason for this is the wind shear phenomenon. Wind shear is caused by the slowing
down of wind in the vicinity of a ground plane due to friction. As the surface roughness of
30
the ground increases, so does the friction between the ground and the air moving over it.
As the wind slows down with decreasing heights it creates a sound velocity gradient ( due to
differential movement of the medium) with respect to the ground. This velocity gradient
tends to bend sound waves downward in the same direction of the wind and upward in the
opposite direction. The process, called refraction, creates a noise " shadow" ( reduction)
upwind from the source and a noise " concentration" ( increase) downwind from the
source. Figure N- 2143.1 shows the effects of wind on noise. Wind effects on noise levels
along a highway are very much dependent on wind angle, receiver distance and site
characteristics. A 10 km/ hr ( 6 mph) cross wind can increase noise levels at 75 m ( 250 ft)
by about 3 dBA downwind, and reduce noise by about the same amount upwind. Present
policies and standards ignore the effects of wind on noise levels. Unless winds are
specifically mentioned, noise levels are always assumed to be for zero winds. Noise
analyses are also always made for zero wind conditions.
Wind also has another effect on noise measurements. Wind " rumble" caused by friction
between air and a microphone of a sound level meter can contaminate noise measurements
even if a windscreen is placed over the microphone.
Limited measurements performed by Caltrans in 1987 showed that wind speeds of about 5
m/ s produce noise levels of about 45 dBA, using a 1/ 2 inch microphone with a wind
screen. This means that noise measurements of less than 55 dBA are contaminated by
wind speeds of 5 m/ s. A noise level of 55 dBA is about at the low end of the range of noise
levels routinely measured near highways for noise analyses. FHWA document No. FHWA-DP-
45- 1R, titled “ Sound Procedures for Measuring Highway Noise: Final Report”, August
1981, recommends that highway noise measurements should not be made at wind speeds
above 12 mph ( 5.4 m/ s). A 5 m/ s criterion for maximum allowable wind speed for routine
highway noise measurements seems reasonable and is therefore recommended. More
information concerning wind/ microphone contamination will be covered in the noise
measurement section N- 3000 of this Appendix.
31
Figure N- 2143.1 - Wind Effects on Noise Levels
Wind turbulence. - Turbulence also has a scattering effect on noise levels, which is
difficult to predict at this time. It appears, however, that turbulence has the greatest effect
on noise levels in the vicinity of the source.
Temperature gradients - Figure N- 2143.2 shows the effects of temperature gradients on
noise levels. Normally, air temperature decreases with height above the ground. This is
called the normal lapse rate, which for dry air is about - 1o C/ 100 m. Since the speed of
sound decreases as air temperature decreases, the resulting temperature gradient creates a
sound velocity gradient with height. Slower speeds of sound higher above the ground tend
to refract sound waves upward in the same manner as wind shear does upwind from the
source. The result is a decrease in noise. Under certain stable atmospheric conditions,
however, temperature profiles are inverted, or temperatures increase with height either
from the ground up, or at some altitude above the ground. This inversion results in
speeds of sound that temporarily increase with altitude, causing noise refraction similar to
that caused by wind shear downwind from a noise source. Or, once trapped within an
elevated inversion layer, noise may be carried over long distances in a channelized fashion.
Both ground and elevated temperature inversions have the effect of propagating noise with
less than the usual attenuation rates, and therefore increase noise. The effects of vertical
temperature gradients are more important over longer distances.
Wind Velocity
Upwind
Noise
Decrease
Downwind
Noise
Increase
32
Figure N- 2143.2 - Effects of Temperature Gradients on Noise
a. No Temperature Gradient - Reference
( Speed of sound stays same with altitude)
b. Normal Lapse Rate - Noise Decrease
( Speed of sound decreases with altitude)
c. Temperature Inversion - Noise Increase
( Speed of sound increases with altitude)
Speed of Sound Speed of Sound
Speed of Sound
Speed of Sound Speed of Sound
Speed of Sound
Source
Ground
Source
Ground
Source
Ground
33
Temperature and humidity - Molecular absorption in air also reduces noise levels with
distance. Although this process only accounts for about 1 dBA per 300 m ( 1000 ft) under
average conditions of traffic noise in California, the process can cause significant longer
range effects. Air temperature, and humidity affect molecular absorption differently
depending on the frequency spectrum, and can vary significantly over long distances, in a
complex manner.
Rain. - Wet pavement results in an increase in tire noise and a corresponding increase in
frequencies of noise at the source. Since the propagation of noise is frequency dependent,
rain may also affect distance attenuation rates. On the other hand, traffic generally slows
down during rain, decreasing noise levels and lowering frequencies. When wet, different
pavement types interact differently with tires than when they are dry. These factors make
it very difficult to predict noise levels during rain. Hence, no noise measurements or
predictions are made for rainy conditions. Noise abatement criteria and standards do not
address rain.
N- 2144 Shielding by Natural and Man- made Features, Noise Barriers,
Diffraction, and Reflection
A large object in the path between a noise source and a receiver can significantly attenuate
noise levels at that receiver. The amount of attenuation provided by this “ shielding”
depends on the size of the object, and frequencies of the noise levels. Natural terrain
features, such as hills and dense woods, as well as manmade features, such as buildings
and walls can significantly alter noise levels. Walls are often specifically used to reduce
noise.
Trees and Vegetation - For a vegetative strip to have a noticeable effect on noise levels it
must be dense and wide. A stand of trees with a height that extends at least 5 m ( 16 ft)
abve the line of sight between source and receiver, must be at least 30 m ( 100 ft) wide and
dense enough to completely obstruct a visual path to the source to attenuate traffic noise
by 5 dBA. The effects appear to be cumulative, i. e. a 60 m ( 200 ft) wide stand of trees
would reduce noise by an additional 5 dBA. However, the limit is generally a total
reduction of 10 dBA. The reason for the 10 dBA limit for any type of vegetation is that
sound waves passing over the tree tops (“ sky waves”) are frequently refracted back to the
surface, due to downward atmospheric refraction caused by wind, temperature gradients,
and turbulence.
Landscaping - Caltrans research has shown that ordinary landscaping along a highway
accounts for less than 1 dBA reduction. Claims of increases in noise due to removal of
vegetation along highways are mostly spurred by the sudden visibility of the traffic source.
34
There is evidence of the psychological " out of sight, out of mind" effect of vegetation on
noise.
Buildings - Depending on the site geometry, the first row of houses or buildings next to a
highway may shield the second and successive rows. This is often the case where the
facility is at- grade or depressed. The amount of noise reduction varies with house or
buildig sizes, spacing of houses or buildings, and site geometry. Generally, for an at- grade
facility in an average residential area where the first row houses cover at least 40% of total
area ( i. e. no more than 60% spacing) , the reduction provided by the first row is reasonably
assumed at 3 dBA, and 1.5 dBA for each additional row. For example, behind the first row
we may expect a 3 dBA noise reduction, behind the second row 4.5 dBA, third row 6 dBA,
etc. For houses or buildings “ packed” tightly, ( covering about 65- 90% of the area, with 10-
35% open space), the first row provides about 5 dBA reduction. Successive rows still
reduce 1.5 dBA per row. Once again, and for the reason mentioned in the above vegetation
discussion, the limit is 10 dBA. For these assumptions to be true, the first row of houses
or buidings must be equal to or higher than the second row, which should be equal to or
higher than the third row, etc.
Noise Barriers - Although technically any natural or man- made feature between source
and receiver that reduces noise is a noise barrier, the term is generally reserved for either a
wall or a berm that is specifically constructed for that purpose. The acoustical design of
noise barriers is covered in sections N- 4000 ( Traffic Noise Model) and N- 6000 ( Acoustical
Barrier Design Considerations). However, it is appropriate at this time to introduce the
acoustical concepts associated with noise barriers. These principles loosely apply to any
obstacle between source and receiver.
Referring to Figure N- 2144.1, when a noise barrier is inserted between a noise source and
receiver, the direct noise path along the line of sight between the two is interrupted. Some
of the acoustical energy will be transmitted through the barrier material and continue to the
source, albeit at a reduced level. The amount of this reduction depends on the material’s
mass and rigidity, and is called the Transmission Loss.
The Transmission Loss ( TL) is expressed in dB and its mathematical expression is:
TL = 10log10( Ef/ Eb) ( eq. N- 2144.1)
where: Ef = the relative noise energy immediately in front ( source side) of the barrier
Eb = The relative noise energy immediately behind the barrier ( receiver side)
35
Figure N- 2144.1 - Alteration of Sound Paths After Inserting a Noise Barrier Between Source
and Receiver.
Note that Ef and Eb are relative energies, i. e. energies with reference to the energy of 0 dB
( see section N- 2134). As relative energies they may be expressed as any ratio ( fractional or
percentage) that represents their relationship. For instance if 1 percent of the noise energy
striking the barrier is transmitted, TL = 10log10( 100/ 1)= 20 dBA. Most noise barriers have
TL’s of 30 dBA or more. This means that only 0.1 percent of the noise energy is
transmitted.
The remaining direct noise ( usually close to 100 percent) is either partially or entirely
absorbed by the noise barrier material ( if sound absorptive), and/ or partially or entirely
reflected ( if the barrier material is sound reflective). Whether the barrier is reflective or
absorptive depends on its ability to absorb sound energy. A smooth hard barrier surface
such as masonry or concrete is considered to be almost perfectly reflective, i. e. almost all
the sound striking the barrier is reflected back toward the source and beyond. A barrier
surface material that is porous with many voids is said to be absorptive, i. e. little or no
sound is reflected back. The amount of energy absorbed by a barrier surface material is
expressed as an absorption coefficient α, which has a value ranging from 0 ( 100%
reflective) to 1 ( 100% absorptive). A perfect reflective barrier ( α= 0) will reflect back virtually
all the noise energy ( assuming a transmission loss of 30 dBA or greater) towards the
opposite side of a highway. If we ignore the difference in path length between the direct
and reflected noise paths to the opposite ( unprotected) side of a highway, the maximum
expected increase in noise will be 3 dBA.
If we wish to calculate the noise increase due to a partially absorptive wall we may use eq.
N- 2144.1. Ef in this case is still the noise energy striking the barrier, but Eb now becomes
DIRECT TRANSMITTED
DIFFRACTED
REFLECTED
SOURCE
NOISE
BARRIER
RECEIVER
ABSORBED
A
SHADOW
ZONE
36
the energy reflected back. For example, a barrier material with an α of 0.6 absorbs 60% of
the direct noise energy and reflects back 40%. To calculate the increase in noise on the
opposite side of the highway in this situation the energy loss from the transformation of the
total noise striking the barrier to the reflected noise energy component is 10log10( 100/ 40)=
4 dBA. In other words, the energy loss of the reflection is 4 dBA. If the direct noise level of
the source at a receiver on the opposite side of the highway is 65 dBA, the reflective
component ( ignoring the difference in distances traveled) will be 61 dBA. The total noise
level at the receiver is the sum of 65 and 61 dBA, or slightly less than 66.5 dBA. The
reflected noise caused an increase of 1.5 dBA at the receiver.
Referring back to Figure N- 2144.1, we have discussed the direct, transmitted, absorbed,
and reflected noise paths. These represent all the variations of the direct noise path due to
the insertion of the barrier. Of those, only the transmitted noise reaches the receiver behind
the barrier. There is, however, one more path, which turns out to be the most imported
one, that reaches the receiver. The noise path that before the barrier insertion was directed
towards “ A” is diffracted downward towards the receiver after the barrier insertion.
In general, diffraction is characteristic of all wave phenomena ( including light, water, and
sound waves). It can best be described as the “ bending” of waves around objects . The
amount of diffraction depends on the wavelength and the size of the object. Low frequency
waves with long wavelengths approaching the size of the object, are easily diffracted.
Higher frequencies with short wavelengths in relation to the size of the object, are not as
easily diffracted. This explains why light, with its very short wavelengths casts shadows
with fairly sharp, well defined edges between light and dark. Sound waves also “ cast a
shadow” when they strike an object. However, because of their much longer wavelengths
( by at least a half dozen or so orders of magnitude) the “ noise shadows” are not very well
defined and amount to a noise reduction, rather than an absence of noise.
Because noise consists of many different frequencies that diffract by different amounts, it
seems reasonable to expect that the greater the angle of diffraction is, the more frequencies
will be attenuated. In Figure N- 2144.1, beginning with the top of the shadow zone and
going down to the ground surface, the higher frequencies will be attenuated first, then the
middle frequencies and finally the lower ones. Notice that the top of the shadow zone is
defined by the extension of a straight line from the noise source ( in this case represented at
the noise centroid as a point source) to the top fo the barrier. The diffraction angle is
defined by the top of the shadow zone and the line from the top of the barrier to the
receiver. Thus, the position of the source relative to the top of the barrier determines the
extent of the shadow zone and the diffraction angle to the receiver. Similarly, the receiver
37
location relative to the top of the barrier is also important in determinig the diffraction
angle.
From the previous discussion, three conclusions are clear. First, the diffraction
phenomenon depends on three critical locations, that of the source, the top of barrier, and
the reciver. Second, for a given source, top of barrier and receiver configuration, a barrier
is more effective in attenuating higher frequencies than lower frequencies ( see Figure N-
2144.2). Third, the greater the angle of diffraction, the greater the noise attenuation is.
Figure N- 2144.2 - Diffraction of Sound Waves
The angle of diffraction is also related to the path length difference ( δ) between the direct
noise and the diffracted noise. Figure N- 2144.3 illustrates the concept of path length
difference. A closer examination of this illustration reveals that as the diffraction angle
becomes greater, so does δ. The path length difference is defined as δ = a+ b- c. If the
horizontal distances from source to receiver and source to barrier, and also the differences
in elevation between source, top barrier and receiver are known, a, b, and c can readily be
calculated. Assuming that the source in Figure N- 2144.3 is a point source, a, b, and c are
calculated as follows:
a = [ d1 ( ) ]
2
2 1
+ h − h 2
b =( ) d h 2 2 2 2
+
c =( d2 h)
1
+ 2
Source
Barrier
High Frequencies
Low Frequencies
38
Figure N- 2144.3 - Path Length Difference Between Direct and Diffracted Noise Paths.
Highway noise prediction models use δ in the barrier attenuation calculations. Section N-
5500 covers the subject in greater detail. However, it is appropriate to include the most
basic relationship between δ and barrier attenuation by way of the so- called Fresnel
Number ( N0). If the source is a line source ( such as highway traffic) and the barrier is
infinitely long, there are an infinite amount of path length differences. The path length
difference ( δ0) at the perpendicular line to the barrier is then of interest.
Mathematically, N0 is defined as:
N0 = 2( δ0/ λ) ( eq. N- 2144.2)
where: N0 = Fresnel Number determined along the perpendicular line
between source and receiver ( i. e. barrier must be perpendicular to
the direct noise path)
δ0 = δ measured along the perpendicular line to the barrier
λ = wavelength of the sound radiated by the source.
According to eq. N- 2131.1, λ = c/ f , and we may rewrite eq. N- 2144.2:
N0 = 2( fδ0/ c) ( eq. N- 2144.3)
where: f = the frequency of the sound radiated by the source
c = the speed of sound
c
a
b
h1
h2
d1
d
d2
SOURCE
RECEIVER
TOP
BARRIER
PATH LENGTH DIFFERENCE ( δ) = a+ b- c
Diffraction Angle
39
Note that the above equations relate δ0 to N0. If one increases, so does the other, and
barrier attenuation increases as well. Similarly, if the frequency increases, so will N0, and
barrier attenuation. Figure N- 2144.4 shows the barrier attenuation ΔB for an infinitely long
barrier, as a function of 550 Hz ( typical “ average” for traffic).
Figure N- 2144.4 - Barrier Attenuation ( ΔB) vs Fresnel Number ( N0), for Infinitely Long Barriers
There are several “ rules of thumb” for noise barriers and their capability of attenuating
traffic noise. Figure N- 2144.5 illustrates a special case where the top of the barrier is just
high enough to “ graze” the direct noise path, or line of sight between source and receiver.
In such an instance the noise barrier provides 5 dBA attenuation.
Figure N- 2144.5 - Direct Noise Path “ Grazing” Top Barrier Results in 5 dBA Attenuation
Another situation, where the direct noise path is not interrupted but still close to the
barrier, will provide some noise attenuation. Such “ negative diffraction” ( with an associated
- 25
- 20
- 15
- 10
− 5
.01 .1 1 10 100
N0
ΔΒ,
dB
SOURCE
NOISE
BARRIER
DIRECT, “ GRAZING” RECEIVER
ATTENUATION: 5 dBA
40
“ negative path length difference and “ negative Fresnel Number”) generally occurs when the
direct noise path is within 1.5 m ( 5 ft) above the top of barrier for the average traffic source
and receiver distances encountered in near highway noise environments. The noise
attenuation provided by this situation is between 0 - 5 dBA: 5 dBA when the noise path
approaches the grazing point and near 0 dBA when it clears the top of barrier by
approximately 1.5 m ( 5 ft) or more.
Figure N- 2144.6 - “ Negative Diffraction” Provides Some Noise Attenuation
The aforementioned principles of barriers loosely apply to terrain features ( such as berms,
low ridges, as well as other significant manmade features). The principles will be discussed
in greater detail in sections N- 5500 and N- 6000.
N- 2200 EFFECTS OF NOISE; NOISE DESCRIPTORS
N- 2210 Human Reaction to Sound
People react to sound in a variety of ways. For example, rock music may be pleasant to
some people while for others it may be annoying, constitute a health hazard and/ or disrupt
activities. Human tolerance to noise depends on a variety of acoustical characteristics of
the source, as well as environmental characteristcs. These factors are briefly discussed
below:
1. Level, variability in level ( dynamic range), duration, frequency spectrums and time
patterns of noise. Exposures to very high noise levels can damage hearing. A high
level is more objectionable than a low level noise, and intermittent truck peak noise
levels are more objectionable than the continuous level of fan noise. Humans have
better hearing sensitivities in the high frequency region than in the low. This is
reflected in the A- scale ( section N- 2136) which de- emphasizes the low frequency
SOURCE
NOISE
BARRIER
NEGATIVE DIFFRACTION RECEIVER
ATTENUATION: < 5 dBA
41
sounds. Studies indicate that the annoyance or disturbance correlates with the A-scale.
2. The amount of background noise present before the intruding noise. People tend to
compare an intruding noise with the existing background noise. If the new noise is
readily identifiable or considerably louder than the background or ambient, it
usually becomes objectionable. An aircraft flying over a residential area is an
example.
3. The nature of the work or living activity that is exposed to the noise source. Highway
traffic noise might not be disturbing to workers in a factory or office, but the same
noise might be annoying or objectionable to people sleeping at home or studying in
a library. An automobile horn at 2: 00 a. m. is more disturbing than the same noise
in traffic at 5: 00 p. m.
N- 2211 Human Response to Changes in Noise Levels
Under controlled conditions in an acoustics laboratory, the trained healthy human ear is
able to discern changes in sound levels of 1 dBA, when exposed to steady, single frequency
(“ pure tone”) signals in the mid- frequency range. Outside of such controlled conditions, the
trained ear can detect changes of 2 dBA in normal environmental noise. It is widely
accepted that the average healthy ear, however, can barely perceive noise level changes of 3
dBA.
Earlier, we discussed the concept of " A" - weighting and the reasons for describing noise in
terms of dBA. The human response curve of frequencies in the audible range is simply not
linear, i. e. humans do not hear all frequencies equally well.
It appears that the human perception of loudness is also not linear, neither in terms of
decibels, nor in terms of acoustical energy. We have already seen that there is a
mathematical relationship between decibels and relative energy. For instance, if one source
produces a noise level of 70 dBA, two of the same sources produce 73 dBA, three will
produce about 75 dBA, and ten will produce 80 dBA.
Human perception is complicated by the fact that it has no simple correlation with
acoustical energy. Two noise sources do not " sound twice as loud" as one noise source.
Based on the opinions of thousands of subjects tested by experts in the field, however,
some approximate relationships between changes in acoustical energy and corresponding
human reaction have been charted. The results have been summarized in Table N- 2211.1,
which shows the relationship between changes in acoustical energy, dBA and human
perception. The table shows the relationship between changes in dBA ( ΔdBA), relative
42
energy with respect to a reference of a ΔdBA of 0 ( no change), and average human
perception. The factor change in relative energy relates to the change in acoustic energy.
Figure N- 2211.1Relationship Between Noise Level Change, Factor Change in Relative Energy,
and Perceived Change
Perceived Change
Noise Level
Change,
ΔdBA
Change in
Relative
Energy,
10± ΔdBA/ 10
Perceived Change in
Percentage,
( 2± ΔdBA/ 10- 1) x 100%
Descriptive Change in Perception
+ 40 dBA 10,000 x Sixteen Times as Loud
+ 30 dBA 1,000 x Eight Times as Loud
+ 20 dBA 100 x + 300 % Four Times as Loud
+ 15 dBA 31.6 x + 183 %
+ 10 dBA 10 x + 100 % Twice as Loud
+ 9 dBA 7.9 x + 87 %
+ 8 dBA 6.3 x + 74 %
+ 7 dBA 5.0 x + 62 %
+ 6 dBA 4.0 x + 52 %
+ 5 dBA 3.16 x + 41 % Readily Perceptible Increase
+ 4 dBA 2.5 x + 32 %
+ 3 dBA 2.0 x + 23 % Barely Perceptible Increase
0 dBA 1 0 % REFERENCE ( No change)
- 3 dBA 0.5 x - 19 % Barely Perceptible Reduction
- 4 dBA 0.4 x - 24 %
- 5 dBA 0.316 x - 29 % Readily Perceptible Reduction
- 6 dBA .25 x - 34 %
- 7 dBA 0.20 x - 38 %
- 8 dBA 0.16 x - 43 %
- 9 dBA 0.13 x - 46 %
- 10 dBA 0.10 x - 50 % Half as Loud
- 15dBA 0.0316 x - 65 %
- 20 dBA 0.01 x - 75 % One Quarter as Loud
- 30 dBA 0.001 x One Eighth as Loud
- 40 dBA 0.0001 x One Sixteenth as Loud
Section N- 2133 mentioned that the r. m. s. value of the sound pressure ratio squared ( P 1 / P 2
is proportional to the energy content of sound waves ( acoustic energy).
Human perception is displayed in two columns ( percentage and descriptive). The
43
percentage of perceived change is based on the mathematical approximation that the factor
change of human perception relates to ΔdBA by the following:
Factor Change in Perceived Noise Levels = 2± ΔdBA/ 10 ( eq. N- 2211.1)
According to the above approximation, the average human ear perceives a 10 dBA decrease
in noise levels as half of the original level ( 2± ΔdBA/ 10 = 2 - 10/ 10 = 0.5). By subtracting 1 and
multiplying by 100 the result will be in terms of a percentage change in perception, where a
positive (+) change represents an increase, and a negative (-) change a decrease. The
descriptive perception column puts into words how the percentage change is perceived.
N- 2220 Describing Noise
Noise in our daily environment fluctuates over time. Some of the fluctuations are minor,
some are substantial; some occur in regular patterns, others are random. Some noise
levels fluctuate rapidly, others slowly. Some noise levels vary widely, others are relatively
constant. In order to describe noise levels, we need to choose the proper noise descriptor or
statistic.
N- 2221 Time Patterns
Figure N- 2221.1 is a graphic representation of how noise can have different time patterns
depending on the source. Shown are noise level vs. time patterns of four different sources:
a fan ( a), a pile driver ( b), a single vehicle passby ( c), and highway traffic ( d).
Figure N- 2221.1 - Different Noise Level Vs. Time Patterns
The simplest noise level time pattern is that of constant noise ( a), which is essentially a
straight and level line. Such a pattern is characteristic of stationary fans, compressors,
pumps, air conditioners, etc. At each instant the noise level is about the same for a fixed
observer. A single measurement taken at random, would suffice to describe the noise level
a. Constant ( e. g. fan) b. Impacts ( e. g. pile driver)
c. Single Vehicle Pass by d. Traffic
TIME
dBA
44
at a specific distance. The minimum and maximum noise level would be nearly the same
as the average noise level.
Other noise level vs. time patterns are more complicated. For instance, to describe the pile
driving noise ( b), noise samples need to include the instantaneous “ peaks” or maximum
noise levels. In our environment, there are a whole range of noises of many different
patterns in addition to the ones shown in Figure 2220.1. The levels may be extremely short
in duration such as a single gun shot ( transient noise), or intermittent such as the pile
driver, or continuous as was the case with the fan. Traffic noise along major highways
tends to lie somewhere between intermittent and continuous ( d). It is characterized by the
somewhat random distribution of vehicles, each of which emits a pattern such as shown in
( c).
N- 2222 Noise Descriptors
To choose the proper noise descriptor, we have to know the nature of the noise source and
also how we want to describe it. Are we interested only in the maximum levels, the average
noise levels, the percentage of time above a certain level, or the levels that are exceeded
10%, 50% or 90% of the time? How can we compare the noise of a fast flying jet aircraft -
loud but short in duration - with a slower but quieter propeller airplane? It is easy to see that
the proper descriptor depends on the spatial distribution of noise source( s), duration,
amount of fluctuation, and time patterns.
There are dozens of descriptors and scales which have been devised over the years to
quantify community noise, aircraft fly- overs, traffic noise, industrial noise, speech
interference, etc. The descriptors shown in Table N- 2222.1 are the ones most often
encountered in traffic, community, and environmental noise. There are many more
descriptors, but they are not mentioned here. The word " LEVEL", abbreviated L, is
frequently used whenever sound is expressed in decibels relative to the reference pressure.
Thus, all of the descriptors shown in Table N- 2222.1 have " L" as part of the term.
All Caltrans highway traffic noise analysis should be done in terms of worst noise hour
Leq( h). If a noise analysis requires other descriptors ( to satisfy city or county requirements)
then see section N- 2230 for a discussion of descriptor conversions.
Table N- 2222.1. Common Noise Descriptors.
45
NOISE DESCRIPTOR DEFINITION
LMAX ( Maximum Noise Level) The highest instantaneous noise level during a
specified time period. This descriptor is sometimes
referred to as “ peak ( noise) level”. The use of “ peak”
level should be discouraged because it may be
interpreted as a non- r. m. s. noise signal ( see sec. N-
2133 for difference between peak and r. m. s.)
LX ( A Statistical Descriptor) The noise level exceeded X percent of a specified time period.
The value of X is commonly 10. Other values of 50 and 90 are
sometimes also used. Examples: L10, L50, L90.
Leq ( Equivalent Noise Level. Routinely used by Caltrans and
FHWA to address the worst noise hour ( Leq( h)).
The equivalent steady state noise level in a stated period of time
that would contain the same acoustic energy as the time varying
noise level during the same period.
Ldn ( Day - Night Noise Level. Used commonly for describing
community noise levels).
A 24- hour Leq with a " penalty" of 10 dBA added during the night
hours ( 2200 - 0700). The penalty is added because this time is
normally sleeping time.
CNEL ( Community Noise Equivalent Level. A common
community noise descriptor; also used for airport noise).
Same as the Ldn with an additional penalty of 4.77 dBA, ( or 10
Log 3), for the hours 1900 to 2200, usually reserved for relaxation,
TV, reading, and conversation.
SEL ( Single Event Level. Used mainly for aircraft noise; it
enables comparing noise created by a loud, but fast overflight,
with that of a quieter, but slow overflight).
The acoustical energy during a single noise event, such as an
aircraft overflight, compressed into a period of one second,
expressed in decibels.
N- 2223 Calculating Noise Descriptors
The following formulae and examples may be used to calculate various noise descriptors
from instantaneous noise vs time data.
Lx - The Lx, a statistical descriptor, signifies the noise level that is exceeded x% of the time.
This descriptor was formerly used in highway noise ( before the Leq). The most common
value of x was 10, denoting the level that was exceeded 10% of the time. Hence, the L10
descriptor will be used as an example to represent the Lx family of calculations. The
following instantaneous noise samples ( Table 2223.1) shown as a frequency distribution
( dBA vs number of occurrences), will serve to illustrate the L10 calculation.
The total No. of samples taken at 10 second intervals was 50. For the L10 we therefore
need to find the 5 highest values ( 10% of 50). These are exceeding the L10. In the above
46
data set, we can simply count down from the top. The “ boundary” of the top 10 % lies at 76
dBA. Therefore the L10 lies at 76 dBA. The L50 would be 66 dBA ( 25 occurrences from the
top), etc.
Table N- 2223.1 Noise Samples for L10 Calculation
Noise Level, dBA Occurrences ( Sampling Interval 10 seconds)
( Each X is one occurrence)
Total No. of
Occurrences
80 0
79 0
78 X 1
77 X 1
76 X X X 3
75 X X 2
74 X X 2
73 X X 2
72 0
71 X X X 3
70 X 1
69 X X 2
68 X X X X X 5
67 X X 2
66 X X X X 4
65 X X X X X X X 7
64 X X X X X 5
63 X X X 3
62 X X X 3
61 X X 2
60 X X 2
Total No. of Samples 50
Leq - The Leq descriptor is a special sort of average noise level. Instead of averaging decibel
levels, the energy levels are averaged. The Leq is also called an energy- mean noise level.
The instant noise levels over a certain time period are energy- averaged by first converting
all dBA values to relative energy values. Next, these values are added up and the total is
divided by the number of values. The result is average ( relative) energy. The final step then
is to convert the average energy value back to a decibel level. Section N- 2135, equation N-
2135.3 showed the method of adding the energy values. This equation can be expanded to
yield an Leq:
Leq = 10log10[( 10SPL( 1)/ 10+ 10SPL( 2)/ 10 + ... 10SPL( n)/ 10)/ N] ( eq. N- 2223.1)
Where: SPL( 1), SPL( 2), SPL( n) = the 1st, 2nd, and nth noise level
47
N = number of noise level samples
Example: Calculate the Leq of the following noise instantaneous samples taken at 10-
second intervals:
Time dBA
10: 00: 10 60
10: 00: 20 64
10: 00: 30 66
10: 00: 40 63
10: 00: 50 62
10: 01: 00 65
Solution ( using eq. N- 2223.1 with above data):
Leq = 10log10[( 1060/ 10+ 1064/ 10+ 1066/ 10+ 1063/ 10+ 1062/ 10+ 1065/ 10)/ 6] =
= 10log10( 14235391.3/ 6) = 63.8 dBA
Usually, longer time periods are preferred. Using the sampling data in the L10 example
( Table N- 2231.1) the following equation ( discussed in sec. N- 2135) can be used to add the
dBA levels for each set of equal noise levels ( occurrences):
SPL( Total) = SPL( 1) + 10log10( N) ( eq. N- 2135.1)
in which: SPL( 1) = SPL of one source
N = number of identical noise levels to be added ( in this case
number of occurrences of each noise level)
Next we can use eq. N- 2135.3 to add the sub totals:
SPL( Total) = 10log10[ 10SPL( 1)/ 10+ 10SPL( 2)/ 10 + ......... 10SPL( n)/ 10] ( eq. N- 2135.3).
The resulting total noise level is 87.5 dBA, which must then be energy averaged to get the
Leq. This may be accomplished by the following equation:
Leq = 10log10[ 10SPL( TOTAL)/ 10/ N] ( eq. N- 2223.2)
Where N = the total number of samples, in this case 50.
The final result is Leq = 10log10[ 1087.5/ 10/ N] = 70.5 dBA. Calculation procedures are
shown in Table N- 2223.2.
48
Table N- 2223.2 - Noise Samples for Leq Calculation
Noise Level,
dBA
Occurrences ( Sampling Interval 10 seconds)
( Each X is one occurrence)
No. of
Occurrences
( N)
Total Noise Levels
dBA + 10log10( N)
80 0
79 0
78 X 1 78
77 X 1 77
76 X X X 3 80.8
75 X X 2 78
74 X X 2 77
73 X X 2 76
72 0
71 X X X 3 75.8
70 X 1 70
69 X X 2 72
68 X X X X X 5 75
67 X X 2 70
66 X X X X 4 72
65 X X X X X X X 7 73.5
64 X X X X X 5 71
63 X X X 3 67.8
62 X X X 3 66.8
61 X X 2 64
60 X X 2 63
Totals 50 87.5
Leq = 10 Log10 [( 108.75)/ 50] = 70.5 dBA
Ldn - The Ldn descriptor is actually a 24 hour Leq, or the energy- averaged result of 24 1-
hour Leq‘ s, with the exception that the night- time hours ( defined as 2200 - 0600 hours) are
assessed a 10 dBA “ penalty”. This attempts to account for the fact that nighttime noise
levels are potentially more disturbing than equal daytime noise levels.
Mathematically this “ day- night” descriptor is expressed as:
Ldn = 10 Log10 [( 1
24 ) 10
1
24
i =
 Leq( h) i + Wi/ 10] ( eq. N- 2223.3)
where: Wi = 0 for day hours ( 0700 - 2200)
Wi = 10 for night hours ( 2200 - 0700)
L eq ( h) i = L eq ( for the ith hour)
49
To calculate an Ldn accurately, we must have 24 successive hourly Leq‘ s, representing one
typical day. The hourly values between 2200- 0700 ( 9 hourly values) must first be weighted
by adding 10 dBA. An example is shown in Table N- 2223.3.
The energy average calculated from the 9 weighted and 15 unweighted hourly Leq‘ s is the
L dn . Once the hourly data is properly weighted, the Ldn can be calculated as an Leq ( in this
case a weighted 24 hour Leq). We may use eq. N- 2223.1 with the weighted data. The
resulting Ldn is 65 dBA.
Table N- 2223.3 Noise Samples for Ldn Calculations
Begin
Hour
Leq( h),
dBA
Weight,
dBA
Weighted
Noise,
dBA
Begin
Hour
Leq( h),
dBA
Weight,
dBA
Weighted
Noise,
dBA
00: 00 54 + 10 64 12: 00 65 0 65
01: 00 52 + 10 62 13: 00 65 0 65
02: 00 52 + 10 62 14: 00 63 0 63
03: 00 50 + 10 60 15: 00 65 0 65
04: 00 53 + 10 63 16: 00 65 0 65
05: 00 57 + 10 67 17: 00 63 0 63
06: 00 62 + 10 72 18: 00 64 0 64
07: 00 65 0 65 19: 00 62 0 62
08: 00 63 0 63 20: 00 60 0 60
09: 00 64 0 64 21: 00 58 0 58
10: 00 66 0 66 22: 00 57 + 10 67
11: 00 66 0 66 23: 00 55 + 10 65
CNEL - The CNEL is the same as the Ldn EXCEPT for an additional weighting of almost 5
dBA for the evening hours of 1900 - 2200. The equation is essentially the same as eq. N-
2223.3, with an additional definition of Wi= 10Log10( 3), which is 4.77. Calculations for
CNEL are done similarly to those for Ldn. The result is normally about 0.5 dBA higher
than that of an Ldn using the same 24- hour data. Following is the equation for the CNEL.:
CNEL = 10 Log10 [( 1
24 ) 10
1
24
i =
 Leq( h) i + Wi/ 10] ( eq. N- 2223.4)
Where: Wi = 0 for day hours ( 0700 - 1900)
Wi = 10log 10 ( 3) = 4.77 for evening hours ( 1900 - 2200)
50
Wi = 10 for night hours ( 2200 - 0700)
L eq ( h) i = L eq ( for the ith hour)
The above 24- hour data used in the Ldn example, yields a CNEL of 65.4 dBA, as compared
to 65.0 dBA for the Ldn.
SEL - The SEL is useful in comparing the acoustical energy of different events involving
different source characteristics. For instance, the over flight of a slow propeller driven
plane may not be as loud as a jet aircraft, but the former is slower and therefore lasts
longer than the jet noise. The SEL makes a noise comparison of both events possible,
because it combines the effects of time and level. For instance, the Leq of a steady noise
level will remain unchanged over time. It will be the same when calculated for a time
period of 1 second or 1000 seconds. The SEL of a steady noise level, however, will keep
increasing, because all the acoustical energy within a given time period is included in the
reference time period of one second. Since both values are energy- weighted they are
directly related to each other by time as shown in the following equations:
SEL = Leq( T)+ 10log10( T) ( eq. N- 2223.5)
Leq( T) = SEL+ 10log10( 1/ T) = SEL- 10log10( T) ( eq. N- 2223.6)
where: T = the duration of the noise level in seconds.
Example: The Leq of a 65- second aircraft over flight is 70 dBA. What is the SEL?
Solution ( using eq. N- 2223.2): SEL = Leq( 65 sec)+ 10log10( 65) = 70+ 18.1 = 88.1 dBA.
N- 2230 Conversion Between Noise Descriptors
Although Caltrans exclusively uses the Leq descriptor, there are some times that
comparisons need to be made with local noise standards, most of which are in terms of Ldn
or CNEL. Twenty- four hour noise data are often not available. Following is a methodology
that allows a reasonably accurate conversion of the worst hourly noise level to a Ldn or
CNEL.
N- 2231 Leq To Ldn/ CNEL, and Vice Versa.
51
The previous section showed that the L dn is defined as an energy- averaged 24- hour L eq with
a night- time penalty of 10 dBA assessed to noise levels between the hours of 2200 and
0700 ( 10: 00 pm and 7: 00 am). If traffic volumes, speeds and mixes were to remain
constant throughout the entire 24 hours, and if there were no night time penalty, there
would be no peak hour and each hourly L eq would equal the 24- hour L eq . Hourly traffic
volumes would then be 100%/ 24, or 4.17% of the average daily traffic volume ( ADT). Peak
hour corrections would not be necessary in this case. Let this be the REFERENCE
CONDITION.
To convert Peak Hour L eq to L dn , at least two corrections must be made to the above
reference condition. First, we must make a correction for peak hour traffic volumes
expressed as a percentage of the ADT. Secondly, we must make a correction for the night-time
penalty of 10 dBA. For this we need to know what fraction of the ADT occurs during
the day and what fraction at night. Depending on the accuracy desired and information
available, other corrections can be made for different day/ night traffic mixes and speeds.
These will not be discussed here.
The first correction for peak hour can be expressed as:
10 Log 10
4.17
P
where :
P = Peak Hour volume % of ADT
The second correction for night time penalty of 10 dBA is:
10 Log 10 ( D + 10N)
where :
D and N are day and night fractions of ADT ( D + N = 1)
To convert from PEAK HOUR L eq to L dn :
Ldn = Leq ( h) pk + 10 Log10
4.17
P + 10 Log10 ( D + 10N) ( eq. N- 2231.1)
To convert L dn to PEAK HOUR L eq :
Leq( h) pk = Ldn - 10 Log10
4.17
P - 10 Log10 ( D + 10N) ( eq. N- 2231.2)
Where:
52
L eq ( h) pk = Peak Hour L eq
P = Peak Hour volume % of ADT
D = Day- time fraction of ADT
N = Night- time fraction of ADT
Note: ( D + N) must equal 1
Example: The peak hour L eq at a receiver near a freeway is 65.0 dBA; the peak hour traffic
is 10% of the ADT; the day- time traffic volume is .85 of the ADT; the night- time
traffic volume is .15 of the ADT. Assume that the day and night- time heavy
truck percentages are equal and traffic speeds do not vary significantly. What is
the estimated L dn at the receiver?
Solution:
L dn = 65.0 + 10 Log 10
4.17
10 + 10 Log 10 ( 0.85 + 1.50)
= 65.0 + (- 3.8) + 3.70
= 64.9 dBA
Note that in the above example, which is a fairly typical case, the L dn is approximately
equal to the L eq ( h) pk . The rule of thumb is that L dn is within +/- 2 dBA of the L eq ( h) pk under
normal traffic conditions.
The values in the following Table N- 2231.1 can also be used in equations N- 2231.2 and N-
2231.3. Notice that the “ peak hour %” term of the equation always yields a negative value,
while the weighted “ day/ night split” always yields a positive value. The difference between
the two is the difference between the L eq ( h) pk and the L dn .
Table N- 2231.1 - Leq/ Ldn Conversion Factors
53
P, % 10Log10( 4.17/ P) D N 10 Log10( D+ 10N)
5 - 0.8 0.98 0.02 + 0.7
6 - 1.6 0.95 0.05 + 1.6
7 - 2.3 0.93 0.07 + 2.1
8 - 2.8 0.90 0.10 + 2.8
9 - 3.3 0.88 0.12 + 3.2
10 - 3.8 0.85 0.15 + 3.7
11 - 4.2 0.83 0.17 + 4.0
12 - 4.6 0.80 0.20 + 4.5
13 - 4.9 0.78 0.22 + 4.7
14 - 5.3 0.75 0.25 + 5.1
15 - 5.6 0.73 0.27 + 5.4
17 - 6.1 0.70 0.30 + 5.7
20 - 6.8 0.68 0.32 + 5.9
0.65 0.35 + 6.2
0.63 0.37 + 6.4
0.60 0.40 + 6.6
Figure N- 2231.1 shows the difference between L eq ( h) pk and L dn graphically. For example if
P is 10% and D/ N = 0.85/ 0.15, the L dn ≈ L eq ( h) pk .
Figure N- 2231.1 - Relationship Between Ldn and Leq( h) pk
If CNEL is desired, the Ldn to CNEL corrections ( Δ) in Table N- 2231.2 may be used.
1.00 0.95 0.90 0.85 0.80 0.75 0.70 0.65 0.60 D
0.00 0.05 0.10 0.15 0.20 0.25 0.30 0.35 0.40 N
40
35
30
25
20
15
10
5
0
0 dBA
- 1 dBA
- 2 dBA
- 3 dBA
+ 1 dBA
+ 2 dBA
+ 3 dBA
P %
D/ N SPLIT
Ldn = Leq( h) pk +
54
Table N- 2231.2 - Ldn/ CNEL Corrections ( Δ); must be added to Ldn to obtain CNEL.
d E
( CNEL = Ldn + Δ)
Δ
0.80 0.05 0.3
0.79 0.06 0.4
0.78 0.07 0.5
0.77 0.08 0.5
0.76 0.09 0.6
0.75 0.10 0.7
0.74 0.11 0.7
0.73 0.12 0.8
0.72 0.13 0.8
0.71 0.14 0.9
0.70 0.15 0.9
The values shown assume a fixed night time fractional traffic contribution of 0.15 ( D/ N
split of .85/. 15 for Ldn). The remaining day time traffic contribution of .85 is further
subdivided into day ( d) and evening ( E) hours. In each instance, d+ E = 0.85.
N- 2240 Negative Effects on Humans
The most obvious negative effects of noise are physical damage to hearing. Other obvious
effects are the interference of noise with certain activities, such as sleeping, conversation,
etc. Less obvious, but nevertheless very real, are the stress effects of noise. A brief
discussion of each of the topics follows.
N- 2241 Hearing Damage.
A person exposed to high noise levels can suffer hearing damage. The damage may be
gradual or traumatic. These are described as follows:
1. Gradual. Sustained exposure to moderately high noise levels over a period of time
can cause gradual hearing loss. It starts out as a temporary hearing loss, such as
immediately after a loud rock concert. The hearing usually restores itself within a
few hours after exposure, although not quite to its pre- exposure level. This is also
called a temporary threshold shift. Although the permanent deterioration may be
negligible, it will become significant after many repetitions of the exposure. At that
time, it is labeled permanent hearing damage. The main causes of permanent
damage are daily exposure to industrial noise. Transportation noise levels
experienced by communities and the general public are normally not high enough to
produce hearing damage.
2. Traumatic. Short and sudden exposure to an extremely high noise level, such as a
gun shot or explosion at very close range can cause a traumatic hearing loss. Such
a loss is very sudden and can be permanent.
Hearing damage is preventable by reducing the exposure to loud noise. This can be done
by quieting the source, shield the receiver by a barrier, or having the receiver wear proper
55
ear protection. Occupational exposure to noise is controlled by the Occupational Safety
and Health Agency ( OSHA), and is based on a maximum allowable noise exposure level of
90 dBA for 8 hours. For each halving of the exposure time, the maximum noise level is
allowed to increase 5 dBA. Thus, the maximum allowable noise exposure ( 100 %) is 90
dBA for 8 hours, 95 dBA for 4 hours, 100 dBA for 2 hours, 105 dBA for 1 hour, 110 dBA
for 30 minutes, and 115 dBA for 15 minutes. Dosimeters, worn by workers in noisy
environments, can measure noise during the workday in percentages of the maximum daily
exposure.
N- 2242 Interference with Activities.
Activities most affected by noise include rest, relaxation, recreation, study and
communications. Although most interruptions by noise can be considered annoying, some
may be considered dangerous. An example would be the inability to hear warning signals
or verbal warnings in noisy industrial situations, or in situations involving workers next to
a noisy freeway. Figure N- 2242.1 gives an estimate of the speech communication that is
possible at various noise levels and distances.
Figure N- 2242.1 - Interference of Conversation due to Background Noise
dBA
BACK
GROUND
NOISE
120
110
100
90
80
70
60
50
40
0 1.5 3 4.5 6 7.5 9 10.5
TALKER TO LISTENER DISTANCE, m
CONVERSATION
IMPOSSIBLE
CONVERSATION
DIFFICULT
CONVERSATION
POSSIBLE
NEAR - NORMAL
CONVERSATION ( Normal Voice)
( Shout)
( Maximum Vocal Effort)
56
For instance, if the talker to listener distance is 6 m, normal conversation can be conducted
with the background level at about 50 dBA. If the background level is increased to 60 dBA,
the talker must either raise his/ her voice, or decrease the distance to the listener to 3 m.
N- 2243 Stress Related Diseases
There is ample evidence that noise can cause stress in humans, and thus may be
responsible for a host of stress- related diseases, such as hypertension, anxiety, heart
disease, etc. Although noise is probably not the sole culprit in these diseases, it can be a
contributor. The degree of how much noise contributes to stress related diseases, depends
on noise frequencies, their band widths, noise levels, and time patterns. In general, higher
frequency, pure tone, and fluctuating noise tend to be more stressful than lower frequency,
broad band, and constant- level noise.
TECHNICAL NOISE SUPPLEMENT
October 1998
N- 57
N- 3000 MEASUREMENTS & INSTRUMENTATION
Noise measurements play an important role in noise analysis and acoustical design of noise
attenuation for transportation projects. This section covers recommendations on why,
where, when, and how noise measurements should be taken. A brief discussion on
available instrumentation is also included. Because of the variety of sound
instrumentation, coverage of equipment setup and operational procedures has been kept at
a general level. For greater detail, manufacturers' manuals should be consulted.
The noise analyst should be aware of the importance as well as the limitations of noise
measurements. As is the case with all field work, quality noise measurements are relatively
expensive. They take time, personnel and equipment. The noise analyst should therefore
carefully plan the locations, times, duration, and number of repetitions of noise
measurements before actually taking the measurements. A conscientious effort should be
made during the measurements to document site, traffic and meteorology and other
pertinent factors discussed in this section.
The contents of this section are consistent with methods described in the Federal Highway
Administration ( FHWA) document FHWA- DP- 45- 1R, “ Sound Procedures for Measuring
Highway Noise: Final Report”, August 1981, and FHWA- PD- 96- 046, “ Measurement of
Highway - Related Noise”, May 1996.
N- 3100 PURPOSES OF NOISE MEASUREMENTS
There are five major purposes for measuring transportation noise. These purposes are to:
1. Determine existing ambient and background noise levels
2. Calibrate noise prediction models
3. Monitor construction noise levels for compliance with Standard Specifications,
Special Provisions, and Local Ordinances
4. Evaluate the effectiveness of abatement measures such as noise barriers
5. Perform special studies and research
Ambient and background noise and model calibration measurements are routinely
performed by the Districts. Construction noise monitoring is also frequently done by the
Districts. Some Districts conduct before- and- after noise abatement measurements.
Special studies and noise research measurements, however, are done rarely by the Districts
and are often contracted out to consultants with Caltrans oversight.
TECHNICAL NOISE SUPPLEMENT
October 1998
N- 58
Where, when, and how noise measurements are performed depends on the purpose of the
measurements. The following sections discuss the reasons for the above measurements,
what they include, and how the results are used.
N- 3110 Ambient and Background Noise Levels
Ambient noise levels are all- encompassing noise levels at a given place and time, usually a
composite of sounds from all sources near and far, including specific sources of interest.
Typically, ambient noise levels include highway plus community noise levels. Ambient noise
levels are measured for the following reasons:
· To assess highway traffic noise impacts for new highway construction or
reconstruction projects. Existing ambient noise levels provide a baseline for
comparison to predicted future noise levels. The measurements are also used to
describe the current noise environment in the area of the proposed project. This
information is reported in appropriate environmental documents. Generally, the
noise resulting from the natural and mechanical sources and human activity,
considered to be usually present, should be included in the measurements.
· To prioritize retrofit noise barrier sites along existing freeways. The measured noise
levels are part of a formula used to calculate a priority index. Prioritization is
required by Section 215.5 of the Streets and Highways Code. The measured noise
levels are also used to design retrofit noise barriers.
· To investigate citizens' traffic or construction noise complaints. Noise
measurements are usually reported in a memo to the interested party or parties,
with recommendations for further actions or reasons why further actions are not
justified.
Background noise is the total of all noise in a specific region without the presence of noise
sources of interest. Typically, this would be the noise generated within the community,
without the highway, and is usually measured at locations away from the highway where
highway noise does not contribute to the total noise level. Background noise levels are
typically measured to determine the feasibility of noise abatement and to insure that noise
reduction goals can be achieved. Noise abatement cannot reduce noise levels below
background. Section N- 6160 discusses the importance of background noise levels.
Depending on the situation, the noise sources to be measured may typically include
highway traffic, community noise, surface street traffic, train noise, and sometimes airplane
noise ( when project is near an airport).
TECHNICAL NOISE SUPPLEMENT
October 1998
N- 59
N- 3120 Model Calibration
Noise measurements near highways or other transportation corridors are routinely used to
calibrate the computer models by comparing calculated noise levels with actual ( measured)
noise levels. The calculated levels are modeled results obtained from traffic counts and
other parameters recorded during the noise measurements. The difference between
calculated and measured noise levels may then be applied to calculated future noise levels
assuming site conditions will not change significantly, or modeled existing noise levels ( see
sections N- 5400 and N- 5330). Obviously, model calibration can only be performed on
projects involving existing highways.
N- 3130 Construction Noise Levels
These measurements are frequently done by Districts to check for the contractor's
compliance with the standard specifications and special provisions of a transportation
construction project, and with local ordinances.
N- 3140 Performance of Abatement Measures
Before- and- after abatement measurements can be used to evaluate the performance of
noise barriers, building insulation, or other abatement options. The measurements provide
a " systems check" on the design and construction procedures of the abatement. Although
these measurements are done occasionally by some Districts, they are not part of a routine
program.
N- 3150 Special Studies and Research
These measurements are usually done by the NT, M& R. They may involve District
assistance and generally involve noise research projects. Setups are usually complex and
include a substantial amount of equipment and personnel positioned at many locations for
simultaneous noise measurement. The studies generally require more sophisticated
equipment than that used for routine noise studies.
N- 3200 MEASUREMENT LOCATIONS
The selection of measurement locations requires a considerable amount of planning and
foresight by the noise analyst. A fine balance must be achieved between a sufficient
amount of quality locations on one side, and the cost in person hours on the other. Good
TECHNICAL NOISE SUPPLEMENT
October 1998
N- 60
engineering judgment must be exercised in site selection; experience makes this task
easier.
There are many tools available in the search for quality noise measurement sites.
Preliminary design maps ( 50 scale geometrics), cross sections, aerial photographs, and field
survey data are all helpful sources of information; however, noise measurement sites
should only be selected after a thorough field review of the project area.
N- 3210 General Site Recommendations
Following are some general site requirements common to all outside noise measurement
sites:
1. Sites must be clear of major obstructions between source and receiver, unless they
are representative of the area of interest; reflecting surfaces should be more than 10
feet from the microphone positions.
2. Sites must be free of noise contamination by sources other than those of interest.
Avoid sites located near barking dogs, lawn mowers, pool pumps, air conditioners,
etc., unless it is the express intent to measure these sources.
3. Sites must be acoustically representative of areas and conditions of interest. They
must either be located at, or represent locations of human use.
4. Sites must not be exposed to prevailing meteorological conditions that are beyond
the constraints discussed in this chapter. For example, in areas with prevailing high
wind speeds avoid selecting sites in open fields.
More detailed considerations will be discussed in the next section.
N- 3220 Measurement Site Selection
For the purpose of this document, a distinction will be made between receivers ( including
sensitive receivers) and noise measurement sites. Receivers are all locations or sites of
interest in the noise study area. Noise measurement sites are locations where noise levels
are measured. Unless an extremely rare situation exists when a noise measurement site is
used for a specialized purpose, all noise measurement sites may be considered receivers.
However, not all receivers are noise measurements sites.
For the purposes of describing existing noise levels at selected receivers, measured noise
levels are normally preferred. Restricted access, or adverse site conditions may force the
selection of noise measurement sites at locations that are physically different from, but
TECHNICAL NOISE SUPPLEMENT
October 1998
N- 61
acoustically equivalent to the intended receivers. In some cases measurements are not
feasible at all. In such cases the existing noise levels must be modeled. This can only be
accomplished along an existing facility.
Generally, there are more modeled receivers than noise measurement sites. It is far less
expensive to take noise measurements at selected, representative receivers and model
results for the rest. Nevertheless, there needs to be an adequate overlap of measurement
sites and modeled receivers for model calibration and verification.
The following factors should be considered when selecting noise measurement sites.
N- 3221 Site Selection By Purpose of Measurement
Noise measurement sites should be selected according to the purpose of the measurement.
For example, if the objective is to determine noise impacts of a highway project, sites
should be selected in regions that will be exposed to the highest noise levels generated by
the highway after completion of the project. The sites should also represent areas of human
use.
Conversely, if the objective is to measure background community noise levels, the sites
should be located in areas that represent the community, without influence from the
highway. These measurements are often necessary for acoustical noise barrier design ( see
section 6150) and to document before project noise levels at distant recievers. Past
controversies concerning unsubstantiated increases in noise levels at distant receivers,
attributed to noise barriers could have readily been resolved if sufficient background noise
receivers would have been selected ( see Section N- 8200) after the project has been built.
Classroom noise measurements ( Street and Highways Code Section 216), or receivers
lacking outside human use, require inside - as well as outside - noise measurements in
rooms with worst noise exposures from the highway. Measurements should generally be
made at a point in a room, hall or auditorium where people would be impacted by
infiltrating noise from the sources of interest. These are typically desks, chairs, or beds
near windows. Several sensitive points may have to be tested and results averaged. No
measurements should be made within 3- 4 feet of a wall. It is also important to take
measurements in the room in its typical furnished condition. If windows are normally
open, take measurements with windows open and closed. Fans, ventilation, clocks,
appliances, telephones, etc. should be turned off. People should preferably vacate the room
or be extremely quiet.
TECHNICAL NOISE SUPPLEMENT
October 1998
N- 62
Model calibration measurements usually require sites to be near the highway, preferably at
receivers or acoustical equivalents to the receivers. ( See Model Calibration Section N- 5400
for additional details).
Sites for construction noise monitoring are dictated by standard specifications, special
provisions, and local ordinances, which detail maximum allowable noise levels at a
reference distance: e. g. Lmax 86 dBA at 15 m ( 50 ft), or other requirements.
Before- and- after measurements for evaluations of noise barriers and other abatement
options, and measurements for special studies or research are non- routine and require a
detailed experimental design. Coordination with the NT, M& R is advisable.
N- 3222 Site Selection By Acoustical Equivalence
Noise measurement sites should be representative of the areas of interest.
Representativeness in this case means acoustical equivalence. The concept of acoustical
equivalence incorporates equivalences in noise sources, distances from these sources,
topography, and other pertinent parameters.
The region under study may need to be subdivided into subregions in which acoustical
equivalence can generally be maintained. Boundaries of each subregion must be estimated
by one or more of the previously mentioned acoustical parameters. Also, in cases where
measurements are being taken for more than one purpose, separate sub- regions may be
defined by each purpose. The size of regions or subregions may vary from small to large.
For example, noise abatement for a school may cover one small region ( the school), while a
noise study for a large freeway project may range from one large region to many subregions.
The number of measurement sites selected within each region or subregion under study
depends on its size, number of receivers, and remaining variations in acoustical
parameters. Obviously, the more conscientiously an effort is made to define acoustical
subregions, the less sites are needed within each subregion. The minimum number of sites
recommended for each region or subregion is two.
Figure N- 3222.1 shows an example of receiver and noise measurement site selections for
an at- grade freeway widening and noise barrier project. Also shown are alternate noise
measurement sites to be used if the selected receivers are not accessible, or otherwise not
suitable for noise measurement locations. Only sites near the freeway are shown.
Background noise measurement sites would typically be off the map, further away from the
freeway. Actual site selection would depend on field reviews and more information not
shown on the map.
TECHNICAL NOISE SUPPLEMENT
October 1998
N- 63
N- 3223 Site Selection By Geometry
In addition to being an important consideration in determining acoustical equivalence,
topography - or site geometry - plays an important role in determining locations of worst
exposure to highway noise.
Sometimes, those receivers located farther from a highway may be exposed to higher noise
levels, depending on the geometry of a site. One typical example is a highway on a high
embankment, where the first tier receivers may be partially shielded by the top of the fill.
Unshielded second or third tier receivers may then be exposed to higher noise levels, even
though their distances from the source are greater. This concept is shown in Figure N-
3223.1. Another common situation involves a close receiver shielded by the top of cut, and
an unshielded receiver farther from the source.
Y
A
W
E
E
R
F
= FIRST ROW RESIDENCE
= NOISE MEASUREMENT SITE = ALTERNATE NOISE MEASUREMENT SITE
= VACANT LOTS
Figure N- 3222.1 - Typical Measurement Sites
TECHNICAL NOISE SUPPLEMENT
October 1998
N- 64
Figure N- 3223.2 - illustrates the effects of site geometry on selection of highest noise
exposure. The unshielded Receiver 1 shows a higher noise level than Receiver 2, although
the latter is closer to the freeway.
Numerous other examples can be generated in which the nature of terrain and natural or
man- made obstructions cause noise levels at receivers closer to the source to be lower than
those farther away. This concept is an important consideration in impact analysis, where
usually the noisiest locations are of interest.
( PLAN VIEW)
Receiver 1
( unshielded)
65 dBA
Receiver 2
( partially shielded)
60 dBA
FREEWAY
At- Grade
Section
Cut
Section
Figure N- 3223.2 - Receiver Partially Shielded by Top of Cut Vs. Unshielded Receiver
Figure N- 3223.1 - Receiver Partially Shielded by Top of Embankment
FREEWAY
EMBANKMENT
SOURCES
RECEIVERS
( CROSS SECTION)
1
TECHNICAL NOISE SUPPLEMENT
October 1998
N- 65
N- 3300 MEASURING TIMES, DURATION, AND NUMBER
OF REPETITIONS
N- 3310 Measuring Times
FHWA 23 CFR Part 772 requires that traffic characteristics which yield the worst hourly
traffic noise impact on a regular basis be used for predicting noise levels and assessing
noise impacts. Therefore, if the purpose of the noise measurements is to determine a future
noise impact by comparing predicted noise with measured, the measurements must reflect
the highest existing hourly noise level that occurs regularly. In some cases, weekly and/ or
seasonal variations need to be taken into consideration. In recreational areas, weekend
traffic may be greater than on week days and, depending on the type of recreation, may be
heavily influenced by season.
Measurements made for retrofit noise barrier projects also require noise measurements
during the highest traffic noise hour.
The noise impact analysis for classrooms, under the provisions of the Streets and Highways
Code Section 216, requires noise measurements to be made " at appropriate times during
regular school hours ...." and sets an indoor noise limit of 52 dBA, L eq ( h), from freeway
sources. Therefore, noise measurements for schools qualifying for school noise abatement
under Section 216 need to be made during the noisiest- traffic hour during school hours.
Noise from school children often exceeds traffic noise levels. In order to avoid contaminated
measurements it is often necessary to evacuate class rooms for the duration of the
measurements, or take measurements during vacation breaks.
Noise measurements for model calibration do not have to be made during the highest noise
hour, but it is desirable to have about the same estimated traffic mix ( heavy truck
percentages of the total volume) and traffic speeds as during the noisiest hour. Accurate
traffic counts and meteorological observations ( see Section N- 3600) must be made during
these measurements.
Noise monitoring for background community noise levels should preferably be done during
the expected time of the highest noise level from the highway, even though the
measurements are taken at sites that are far enough removed from an existing highway to
not be contaminated by it. The reason for this is that the background levels will later be
added to predicted near- highway noise levels.
TECHNICAL NOISE SUPPLEMENT
October 1998
N- 66
Noise monitoring for investigating citizen's complaints may have to be done at a mutually-agreed-
on time. Frequently, these measurements are taken before or after normal working
hours, as dictated by the nature of the complaint.
Construction monitoring is performed during operation of the equipment to be monitored.
This may require night work on some construction projects.
Unless other times are of specific interest, before and after noise abatement ( e. g. noise
barrier) measurements to verify noise barrier performance should preferably be done during
the noisiest hour. There are several reasons for this. First, noise barriers are designed for
noisiest hour traffic characteristics, which probably include highest truck percentages, and
second, to minimize contamination by background noise. Traffic should be counted during
these measurements. If before and after traffic conditions are different, measurements
should be normalized or adjusted to the same conditions of traffic ( see section N- 3312).
The nature of special studies and research projects dictate the appropriate times for those
measurements.
N- 3311 Noisiest Hour For Highway Traffic
The peak traffic hour is generally NOT the noisiest hour. During rush hour traffic, vehicle
speeds and heavy truck volumes are often low. Free flowing traffic conditions just before or
after the rush hours often yield higher noise levels. Preliminary noise measurements at
various times of the day are sometimes necessary to determine the noisiest hour.
If accurate traffic counts and speeds for various time periods are available, the noisiest
hour may be determined by using the prediction model.
Experience based on previous studies may also be of value in determining the noisiest hour
for a particular facility.
N- 3312 Adjusting Other- Than- Noisiest Hour
For the sake of efficiency, highway traffic noise measurements are often not made when the
highest hourly traffic noise levels occur. These measurements may be adjusted upward to
noisiest hour levels by using the prediction model. To make the adjustments, traffic must
be counted and speeds determined simultaneously with the noise measurements. The
following procedure must be followed:
1. Take noise measurements and count traffic simultaneously during each
measurement. Although lane- by- lane traffic counts yield the most accurate results
it is usually sufficient to count traffic by direction ( e. g. east bound and west bound).
TECHNICAL NOISE SUPPLEMENT
October 1998
N- 67
Separate vehicles in the three vehicle groups used by the model ( autos, medium
trucks, and heavy trucks). Obtain average traffic speeds ( both directions). These
may be obtained by radar or by driving a test vehicle through the project