Acoustics and Architecture PDF
Acoustics and Architecture PDF
Acoustics and Architecture PDF
OU 158930 >m
OSMANIA UNIVERSITY LIBRARY
Call No. <T?>4 .
Author
Title
This book shduld be returned on or before the date last marked below.
ACOUSTICS AND ARCHITECTURE
ACOUSTICS
AND
ARCHITECTURE
BY
PAUL E. SABINE, PH.D.
Riverbank Laboratories
FIRST EDITION
CHAPTER I
INTRODUCTION 1
CHAPTER II
CHAPTER III
CHAPTER IV v
REVERBERATION THEORETICAL 47
CHAPTER V
REVERBERATION EXPERIMENTAL 66
CHAPTER VI -
CHAPTER VII -
CHAPTER VIII .
CHAPTER IX
ACOUSTICS IN AUDITORIUM DESIGN 171
CHAPTER X
MEASUREMENT AND CONTROL OF NOISE IN BUILDINGS . . 204
CHAPTER XI
THEORY AND MEASUREMENT OF SOUND TRANSMISSION. . . 232
CHAPTER XII
TRANSMISSION OF SOUND BY WALLS 253
CHAPTER XIII
MACHINE ISOLATION 282
APPENDICES 307
INDEX 323
ix
ACOUSTICS
AND
ARCHITECTURE
CHAPTER I
INTRODUCTION
Historical.
Branches of Acoustics.
Velocity of Sound.
/pressure _ JP
> density >p
The pressure and density must of course be expressed in
absolute units. The density of air at C. and a pressure
4 ACOUSTICS AND ARCHITECTURE
compressed, its
temperature This rise of temperature
rises.
makes necessary a greater pressure to produce a given
volume reduction than is necessary if the compression takes
place slowly, allowing time for the heat of compression to
be conducted away by the walls of the containing vessel or
to other parts of the gas. In other words, the elasticity of
air for therapid variations of pressure in a sound wave is
greater than for the slow isothermal changes assumed in
Boyle's law. Laplace showed that the elasticity for the
rapid changes with no heat transfer (adiabatic compression
and rarefaction) is 7 times the isothermal elasticity where
7 the ratio of the specific heat of the medium at constant
is
C = _
yP = -
1. 40 X -
1,012,930
Effect of Temperature.
ct = 1 + (2a)
331.2^: 273
or, if temperature is expressed on the Fahrenheit scale,
-
ct = 1 + t
_32 (26)
331.2^ 491
A simpler though only approximate formula for the velocity
of sound between and 20 C. is
6 ACOUSTICS AND ARCHITECTURE
ct = 331.2 + 0.60*
As an
illustration of the application of the fundamental
cw
fe = /2
-- X 10 10 = 1>M Anr ,
,
\/~
*
\/
*
i 141,400 cm./sec.
1
7,
6
5
|
Jo
\Wtnct31o4MRl
/dhtp velocity*
10 15 20 25 30 35 40 45
Distance in Thousands of Feet
FIG. 1. Variation of amplitude of sound in open air with distance from sou roe.
(AfterL. V. King.)
12
NATURE AND PROPERTIES OF SOUND 13
A, is s
2
(Aco) /A/A = Aco 2 =
Since the tangential speed 2
.
Fx = m = mAu 2
sin ut = mAw 2 sin (o)t + TT) (5)
t^znax.
= Hrae^ 2 - 27T
2
W/ A
2 2
(6)
Wave Motion.
<
R,o
40
FIG. 6.
Fia. 5a. Compreseional plane wave moving to the right.
Fia. 56. Compressional plane wave reflected to the left.
FIG. 6. Stationary wave resultant of Figs. 5a and 56.
of a gas are in a state of thermal agitation, and the pressure which the gas
exerts is due to this random motion of its molecules. Here for the purpose
of our illustration we shall consider stationary molecules held in place by
elastic restraints.
NATURE AND PROPERTIES OF SOUND 17
Jt - -
= A sin
^=A sin 2*f(t
^} (la)
I = -4cu 2
sin w(t - = -4rr 24/2 sin 2wf(t - (7c)
^) ^j
Equation (7o) is the equation of a plane compressional
wave of simple-harmonic type traveling to the right, and
from it the displacement of any particle at any time may
A= c (8)
P
and the rarefaction is at 30 At the end of a complete
.
20 ACOUSTICS AND ARCHITECTURE.
?)= (96)
_
""
A/ ~2
_ 2y P
2
~
2 2
2
J 2w 2pA 2f2~c* pc
3
J = 2! = P.' =J^ = Pl
r
Q 3)
^
pc VyPp Vep
in which
r = Vep and e = yP.
The
expression \/cp has been called the "acoustic resist-
ance" medium. Table II of Appendix A gives the
of the
values of c and r for various media.
Comparison of Eq. (13) with the familiar expression
for the power expended in an electric circuit suggests the
reason for calling the expression r, the acoustic resistance of
a medium. It will be recalled that the power expended W
in a circuit whose electrical resistance is R is given by the
expression
where E
is the electromotive force (e.m.f.) applied to the
"
of Crandall's Theory of Vibrating Systems and Sound"
and to the recently published " Acoustics" by Stewart and
Lindsay.
Equation (13) gives flux intensity in terms of the r.-m.-s.
pressure change and the physical constants of the medium
only. It will be noted that it does not involve the fre-
P+dP
Numerical Values.
The qualitative relationships between the various phe-
nomena that constitute a sound wave having been 'dealt
with, it is next of interest to consider the order of magnitude
of the quantities with which we are concerned. All values
must of course be expressed in absolute units. For this
purpose we shall start with the pressure changes in a sound
of moderate intensity. At the middle of the musical
range, 512 vibs./sec., the r.m.s. pressure increment of a
sound of comfortable loudness would be of the order of 5
bars (dynes per square centimeter), approximately five-
millionths of an atmosphere. From Eq. (13) the flux
intensity J for this pressure is found to be p* -f- 41.5 = 0.6
erg per second per square centimeter or 0.06 microwatt per
square centimeter. From the relation that J = HH; 2 max.
we can compute the maximum particle velocity, which is
found to be 0.17 cm. per second. This maximum velocity
is 27T/A, and the amplitude of vibration is therefore
0.000053 cm. A moment's consideration of the minuteness
of the physical quantities involved in the phenomena of
sound suggests the that are to be encountered
difficulties
in the direct experimental determination of these quantities
and why precise direct acoustical measurements are so
difficult to make. As a matter of fact, it has been only
since the development of the vacuum tube as a means of
amplifying very minute electrical currents that quantitative
work on many of the problems of sound has been possible.
Complex Sounds.
In the preceding sections we have dealt with the case of
sound generated by a body vibrating with simple harmonic
motion. The tone produced by such a source is known as
NATURE AND PROPERTIES OF SOUND 25
FIG. Sa. Disc, pin and slot, and chain mechanism of the Riverbank harmonic
synthesizer.
for the first element is 40:1, while that for the fortieth is
1:1, so that for 40 rotations of the main shaft the disks
revolve 1, 2, 3, 4 ... 40 times respectively. Each disk
carries a pin which moves back and forth in a slot cut in a
sliding member free to move in parallel ways. Each of
these sliding members carries a nicely mounted pulley.
The amplitude of motion of each sliding element can be
adjusted by the amount to which the driving pin is set off
%
Summary.
Starting with the case of a body performing simple
harmonic motion, we have considered the propagation of
this motion as a plane wave in an elastic medium. To
visualize the physical changes that take place when a
sound wave travels through the air, we fix our attention on a
single small region in the transmitting medium. We see
each particle performing a to-arid-fro motion through its
undisplaced position similar to that of the sounding body.
Accompanying this periodic motion is a corresponding
change in its distance from adjacent particles, resulting in
changes in the density of the medium and consequently a
pressure which oscillates about the normal pressure.
Accompanying these pressure changes are corresponding
slight changes in the temperature.
Viewing the progress of the plane wave through the air,
we see all the foregoing changes advancing from particle
to particle with a velocity equal to the square root of the
ratio of the elasticity to the density. In time, each set of
conditions is repeated at any point in the medium once in
each cycle. In space, the conditions prevailing at any
instant at a given point are duplicated at points distant
from by any integral number of wave lengths.
it
=
-Arin^ + ^ (17)
d+r
= - - sin + =
A[sin (* Jj (* *JJ
wt sin 1 (18)
-2^|cos
In Fig. 6, the particle positions at quarter-period intervals
are shown, each displacement being the resultant of the
two displacements due to the direct and reflected waves
shown 5a and 56 respectively. One notes imme-
in Figs.
diately that particles P P-20, and P 40 remain stationary
,
= 8p7r
2
A /2 sin
2 2
co* sin 2
c
sin 2
O
ax = (21)
c
2
mc
~f
i - <m
T,
unity.
Referring to Eq. 18, we may write in the case of an
absorbent barrier
? (d+r) = - - k sin +
A[sin co(< 5) co(*
\(l
- k) sin wt cos
- (1 + k) cos *>* sin 1 (23o)
L c c J
amplitude is A (1 k).
SUSTAINED SOUND IN AN INCLOSURE 39
angles to the axis of the tube, the tube becomes a room, and
the character of the standing wave system becomes much
more complicated. We no longer assume that the particle
motion occurs in a single direction parallel to the axis of
the tube. As a result, the simple distribution of nodes and
loops in the one-dimensional case gives place to an intricate
pattern of sound intensities, a pattern which may be
radically altered by even slight changes in the position of
reflecting surfaces in the room and of the source of sound.
The single series of natural or resonance frequencies
obtained for the one-dimensional case by putting integral
values 1, 2, 3, etc., for m in Eq. (22) is replaced by a trebly
infinite series with the number of possible modes -of vibra-
tion greatly increased. The simplest three-dimensional
case would be that of the cube. For the sake of com-
parison, the first five natural frequencies of a tube 28 ft.
1
TAYLOR, H. O., Phys. Rev., vol. 2, p. 270, 1913.
40 ACOUSTICS AND ARCHITECTURE
he film.
Figure 14 is a reproduction of one of the films. Dis-
ances along the film are proportional to distances along the
piral. The width of the light band produced by the image
REVERBERATION (THEORETICAL)
In Chap. Ill, we have noted that with a sustained source
of sound within an inclosure, there is set up an intensity
pattern in which the intensity of the sound energy varies
from point to point in a complicated manner and that in the
absence of dissipative forces, the vibrational energy of the
air within the inclosure tends to persist after the source has
ceased. This prolongation of sound within an inclosed
space is the familiar phenomenon of reverberation which
has come to be recognized as the most important single
factor in the acoustic properties of audience rooms.
Reverberation in a Tube.
We shall now proceed to the consideration of this
phenomenon, the one-dimensional case of a plane
first in
FIG. 19. Tube with partially absorbent ends and source at one end.
c
Summing up the entire series, we have
El
Total energy = [1 + (1
- a) + (1
- a)
2
+ (1
- a)
3
-
(1
- a)"-
1
]
1,000
900
800
XTOO
5 600
500
c 400
300
200
100
1 2 3456 7 8 9 10 11 12 13 14 IS 16 17 ia 19 20
vf
FIG. 20. Growth of average intensity of sound in tube with partially absorbent
ends. (1) Emission and absorption assumed continuous. (2) Discontinuous
emission and absorption.
dl E E
'ft F V (24)
50 ACOUSTICS AND ARCHITECTURE
E caI 1
~
___
V~ I
whence
---
I = Ml -e l
(25)
After the sound has been fed into the tube for a time
long enough for the intensity to reach the final steady
state, let us assume that the source is stopped. Consider
first the case assuming that the absorption of energy at the
IOK.
= ~ (286)
or
r. -
52 ACOUSTICS AND ARCHITECTURE
___ -
a b
Fia. 21. Tube with end reflection replaced by image sources.
I 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20
t'f
FIG. 21a. Decay of sound in a tube. (1) Average intensity, absorption
assumed continuous. (2) Intensity at source end, absorption discontinuous.
(3) Intensity at end away from source, absorption discontinuous.
1 = 1,- + (1
- a) + (1
- a)
2
+ (1
- a)
8
...
^[1
(1
- a)"-']
1 - )" _ 1
-
(1
- ~ " (1 )"
1 (1 a) a
and
- - m " _ (i
_ /( i
_ a)"
).)
and
= (1
- a) (30)
rT
log. 7!
- log e 7 = -~- log. (1
- OL) (31)
- log e (1
- a) = 13.8
steady-state intensity
where
i,
faces
<* 2 , 3,
/i
am
etc.,
whose areas are
-
= --
-ifr
Si,
+
S2 S3
,
2>S>2
-
+ a3S3
_
A
further simplification is effected if we can express
p, the mean free path, in terms of measurable quantities.
As a result of his earlier experiments, Sabine arrived at a
tentative value for p of 0.62 F He recognized that this
3
*.
expression does not take account of the fact that the mean
free path will depend upon the shape as well as the size of
the room and, subsequently, as a result of experiment put
his equations into a form in which p is involved in another
constant k. Franklin first showed in a theoretical deriva-
tion of Sabine's reverberation equation 1 that
4F
1
For the derivation of this relationship, see Franklin, Phys. Rev., vol. 16,
p. 372, 1903; Jaeger, Wiener Akad. Ber. Math-Nature Klasse, Bd. 120, Abt.
Ho, 1911; Eckhardt, Jour. Franklin Inst., vol. 195, pp. 799-814, June, 1923;
Buckingham, Bur. Standards Sci. Paper 506, pp. 456-460.
REVERBERATION (THEORETICAL) 59
/! = (33c)
v x
ac
c a
pSIoT
dl _
~ ~~
ad ~
_ acl
dt Sp ~47
log*
-- - ^
1,000,000
J
= acT Q
(35)
(1 a )
2
. Hence the total energy in the steady state
*
- - aa - =
^[1 +
and
(1 ) + (1 )
2
(1 .)-]
g (36)
T
11
-
~ J&
REVERBERATION (THEORETICAL} 61
Ep ~~ Ep
stops, and there will be drops of <*)>
7^(1 ~7y
(1 a )
2
, etc., at the beginning of each succeeding p/c
interval. The total diminution in the intensity at the
end of the time T = np/c will be
+ (1
- aa + ) (1
- aa + )
2
(1
- aj' +
|?[1
(1
- a,,)"-
1
]
and we have
-
- a a )\
7 = 7i - Ep(\ (1
= /l(1
~ a] =
^KT^O--- j y
/,(!
- a.)?
Taking the logarithm of both sides of this equation, we
have
log e
T
y =
-y
i'T
log, (1
- aa - )
-~
r^T
log, (1
- a) (37)
sion S log e f 1
-gj.
It must be remembered, however,
that the absorption coefficients of materials now extant
are all derived on the older theory, so that to shift would
involve a recalculation of all absorption coefficients based
on the later formula.
Very recently, apparatus has been devised for making direct measure-
1
ments of this character. In the Jour. Soc. Mot. Pict. Eng., vol. 16, No. 3,
p. 302, Mr. V. A. Schlenker describes a truck-mounted acoustical laboratory
for studying the acoustic properties of motion-picture theaters, in which an
oscillograph is used for recording the decay of sound within a room. The
equipment required is elaborate. The spark chronograph of E. C. Wente
described in Chap. VI has also been used for this purpose.
64 ACOUSTICS AND ARCHITECTURE
REVERBERATION (EXPERIMENTAL)
40 80 120160200240
y Meters of, Cushions
40 80 120160200240280320360400
- + Lc)
T l (Lr (40)
ac T = K V,
C (42)
The fact that the difference in time for one and two
pipes is very nearly one-half of the difference for one and
four pipes suggests that the
difference in times is propor-
tional to the logarithm of the
ratio of the number of pipes,
a relation that is clearly
brought out by graph 1 of
Fig. 25.
Four Organ experiment Let us now assume that the
1. Room bare .
log^,
1
= A(T, - 0) (45o)
t
^=
dt
-AI
A1
Integrating, we have
- log e 7 +C= AT.
When T is 0, that is, at the moment of cut-off of the source,
/ = /i. So that - log I\ + C = e
whence
log./!- log,/ = AT
If T =TI, the time required for the reverberant sound to
decrease to the threshold intensity i, we have
aa -o
= change of intensity per unit intensity per reflection
A = change of intensity per unit intensity per second
Then A = maa where ra is the average number of
reflections per second which any single element of the
sound undergoes in its passage back and forth across the
room during the decay process. In terms of the mean free
path, already referred to in Chap. IV,
A = ~a a = - -
(46)
1
We may now express the results of the four-organ
experiment terms of the total absorbing power as
in
measured by the cushion and open-window experiments.
We have
o f) t
whence
aTl = SP log. k (47)
C if
Power of Source.
Let E
be the number of units of sound energy supplied
by the source to the room in each second. The rate then
at which the sound energy density / is increased by E is
E/V. The rate at which the energy decreases due to
absorption is AI. The net rate of increase while the source
is operating is therefore given by the equation
dl _E _ Al
A
~3JL V
The steady-state intensity I\ is that at which the rates
of absorption and emission are just balanced, so that
/J T P 1
and
AT, = log,'
1
= log,
~ (48)
and
log.f
= AI + 7
! log. VA
if
c %
aT = kV log, 10
6 = k X 13.8V
where T n
is the time required for a sound of initial intensity
10 6 X to decrease to the threshold intensity,
i and k is a
new constant whose precise value is desired.
One compelled to admire the skill with which Sabine
is
Let w equal the open- window area, and T'\ the time under
the open-window condition. Then
(51)
lo *
aT l
(52)
(a + w)I",
The p = 4F'/S is
theoretical derivation of the relation
given in Appendix B. The close agreement between the
experimental and theoretical relations furnishes abundant
evidence of the validity of the general theory of reverbera-
tion as we have it today. We shall hereafter use the value
of K = 0.162, corresponding to the theoretically derived
value of p, where V is expressed in cubic meters and a is
expressed in square meters of perfectly absorbing surface.
If these quantities are expressed in English units, the
reverberation equation becomes
aT Q = 0.0494F
Complete Reverberation Equation.
In Chap. IV, the picture of the phenomenon of reverbera-
tion was drawn from the simple one-dimensional case of
sound in a tube and extended by analogy to the three-
dimensional case of sound within a room, considering the
building up and decay of the sound. In the preceding
paragraphs, we have followed the experimental work, as
separate processes, by which the fundamental principles
were established and the necessary constants were evaluated.
We shall now proceed to the formal derivation of a single
equation giving the relation between all the quantities
involved.
The underlying assumptions are as follows :
I t
= A(l - e-") = -(\ - e- A ') (54)
IT
= e
AT or K(l -- -
Ajrf
e-")
VIr T
= e
AT
AT = log e (l
- -*) (55a)
= 4F log, .
(57)
i.o
0123401234
Time From Start Time From Cuf-off
Fio. 26. Statistical growth and decay of sound in a room absorption assumed
Continuous.
Summary.
In the present chapter we have followed Sabine's attack
on the problem of reverberation from the experimental
side to the point of a precise evaluation of the constant
in his well-known equation. This equation has been shown
to be a special case of the more general relation given in
Eq. (56). He expressed all the relations with which we
are concerned in terms of this constant K
with correction
REVERBERATION (EXPERIMENTAL) 85
I *- Ir -
a = = 1
f (58a)
J. d *-
d
methods. It is not
easy to subject the question to direct
experiment. In view of the fact that there is not a very
close agreement between measurements made by the two
methods, we shall for the sake of clarity refer to coefficients
obtained by measurements in tubes as " stationary-wave
coefficients." Coefficients obtained by reverberation meth-
ods we shall call "reverberation coefficients."
M = B (1 + ft)
N = B (I
- ft)
Adding,
M + N = 2B
Subtracting,
M-N= 2ftJ5
whence
88 ACOUSTICS AND ARCHITECTURE
By definition,
a = 1
1 - Ir
,
- 1
1 -P
,2 - !
1 - M ~ N2
Td
or
"
(M + TV)
2
PARIS, E. T., Proc. Phys. Soc. London, vol. 39, No. 4, pp. 269-295, 1927.
1
2
DAVIS and EVANS, Proc. Roy. Soc., Ser. A, vol. 127, pp. 89-110, 1930.
MEASUREMENT OF ABSORPTION COEFFICIENTS 89
- Ni
Af
+ i
a = 1 - (60)
Nn = N + -AXM (62)
rigid, and massive walls. The size of the tube that can be
used is limited, however, since sharply defined maxima and
minima are difficult to obtain in tubes of large diameter.
Davis and Evans found that with a pipe 30 cm. in diameter,
radial vibrations may be set up for frequencies greater
than 1 ,290 cycles per second. This of course would vitiate
the assumption of a standing-wave system parallel to the
axis of the tube, limiting the frequencies at which absorp-
tion coefficients can be measured in a tube of this size.
To
amplifier ana/
potentiometer
To oscillafor
ot (63)
^60
3
<
5 0.40
0.70
packing of the hair felt. The three curves were all obtained
upon the same sample of material but with different degrees
of packing. They are of a great deal of significance in the
light which they throw upon the discrepancy of the results
obtained by different observers in the measurement of
absorption coefficients of materials of this sort. Both
thickness and degree of packing produce very large differ-
ences in the absorption coefficients, so that it is safe to say
that differences in the figures quoted by different authorities
are due in part at least to actual differences in the samples
tested but listed under the same description.
96 ACOUSTICS AND ARCHITECTURE
alll-a
GO
> :
oI
COCOrHiOQQ
rt CD CO O5 O5 "*<
OQCO
COt*-
COCO
t*-iO
<N r-
o
t-H
odddod do do do
8:
O '"^ O5 00 O O 00 CO
'
o d -
do o' o d d
i
i
>
o o o o dd d dd
X X
<x> o>
1=! "S
:l
;
11
s a a a s s
o o o o o
as..x : -
;-S
c3
O
.
.. CO O5 O iO C ts 1C
^ ' '
W2 1
dt
98 ACOUSTICS AND ARCHITECTURE
Sound Chamber.
Any empty room with highly reflecting walls and a
sufficiently long period of reverberation may be used as a
sound chamber. The calibration of a sound chamber
amounts simply to determining the total absorbing power
room in its standard condition for tones covering the
of the
desired frequency range. This standard condition should
be reproducible at will. For this reason, whatever furnish-
ing it may have in the way of apparatus and the like should
be kept fixed in position and should be as non-absorbent as
possible. If methods depending upon the threshold of
18 air space
Section A-A
Plan and section of Riverbank sound rhamber
A
Plcin
feei
012345
Section
A -A
a = on + ^~ (64)
Computing E/i for the pipe of the large organ and the
particular observer, we can compute a for the sound
chamber in its standard condition.
104 ACOUSTICS AND ARCHITECTURE
TABLE VI. SOUND-CHAMBER CALIBRATION, FOUR-ORGAN PIPE METHOD
* Relative
humidity 80 per cent,
t Relative humidity 63 per cent.
1
For the variation in the absolute sensitivity of normal ears see FLETCHER,
"Speech and Hearing," p. 132, D. Van Nostrand Company, 1929; KRANZ,
F. W., Phys. Rev., vol. 21, No. 5, May, 1923.
MEASUREMENT OF ABSORPTION COEFFICIENTS 105
atmospheric absorption
energy. High-frequency sound is more strongly absorbed
20 30 40 50 60 70- 80 90 100
per cent to 6.52 sec., relative humidity 100 per cent with
fog present. Knudsen is inclined to ascribe this marked
increase in absorption with fog in the air to the presence
of moisture on the wall. The author questions this
explanation in view of the magnitude of the effect, par-
ticularly at the low frequencies. The decrease from 12.65
to 6.52 sec. calls for a doubling of the coefficient of absorp-
tion, if we assume the effect to be due only to changes in
MEASUREMENT OF ABSORPTION COEFFICIENTS 107
0004
--
-^0003
=
0.002
0001
EO 30 40 50 60 v 70
Relaiive Humidity (21 to 22 C)
Pencentage
Fia. 396. Values of m (sec toxt) as a function of relative humidity. (After
Knudsen.)
E= kC*
where A; is a constant.
1
Recent experiments show that this relation holds over only a limited
range for most commercial types of dynamic loud speakers.
MEASUREMENT OF ABSORPTION COEFFICIENTS 109
aci
whence
-
n
a - 4yrio ge
_ Ct loge c,i _
~ gVriogio(CiVCi)-|
J 2
TL T, - T* \
-
cL T,-T t \
(65)
4 "7
y
/ 1-5/2 vi
Z bs/sec.
01234-56780
FIG. 40. Difference in reverberation time as a function of logarithm of loud-
speaker current ratio.
Time in Seconds
FIG. 41. Sound-chamber calibration using Rayleigh disk. (After Watson.)
\Loudspeaker
of their curves are plotted against the time. In Fig. 46, the
absorbing powers of the sound chamber at the Bureau of
Standards as determined both by oscillograph and by the
ear method are plotted as a function of the frequency of
the sound. In summing up the results of their research
with the oscillograph, Chrisler and Snyder state that for
Amplifier
FIG. 47. Meyer and Just's apparatus for recording decay of sound intensity.
Fia. 48. Decay of sound intensity recorded with apparatus of Meyer and Just-
1
Electr. Nachr-Techn., vol. 5, pp. 293-300, 1928.
MEASUREMENT OF ABSORPTION COEFFICIENTS 117
previous paragraph.
whence
acl l = 4E
If now an absorbing area be brought into the room, the
total absorbing power becomes a', and the steady-state
intensity 7'i, so that
a'c/'i = 4E
and
Subtracting,
a! - a = A~ f/1
(66)
a[ ]
uuf^J
Motor *sariv&
variable inductance
TABLE IX
ABSORPTION COEFFICIENT
OF SAMPLE /
Steady) t a 9'
W t'
8.19 473 8.QS
ra, and m 1
being the slopes of the straight lines representing
the experimental points without and with the absorbent
material present. (2) Assuming equal acoustic outputs
for equal loud-speaker currents, o! maybe computed from
the values of T and T", the times for any given current
value before and after the introduction of the absorbent
material, and a, the absorbing power of the empty room.
Thus if E be the acoustical power of the source, assumed
for the moment to be the same for a given current under
the two room conditions, we have
aT = 7.70 lo glo
*-f
= 7.7 lo glo
7 (67a)
CiCt' L
f
4/T J
a'T = 7.70 logio -=F-. = 7.7 logio -^ (676)
a ci %
-)
We note, however, that if the straight lines of Fig. 50
be extrapolated, their intersection falls on the axis of zero
time. This means that equal currents in the loud-speaker
set up equal steady-state intensities throughout the room,
whereas the assumption of equal acoustical powers for a
given current implies a lower intensity in the more absorbent
room. We are forced to the conclusion, therefore, that the
sound output of a loud-speaker operating at a fixed amplitude
is not independent of the room conditions. The data
here presented indicate that for a fixed amplitude of
the source the output of the speaker is directly proportional
to the absorbing power; that is, E'/a' = E/a, I = /', and
hence
=
'
^ (69)
" Collected
Papers on Acoustics," Harvard University Press, p. 279,
1
1922.
122 ACOUSTICS AND ARCHITECTURE
the contrary, this is just the sort of source whose emitting power varies
with the position in which it is placed in the room. On the other hand,
an organ pipe is able within certain limits to adjust itself automatically
to the reaction due to the interference system. We may say briefly that
the best standard source of sound is one in which the greatest percentage
of emitted energy takes the form of sound.
gl Cii
Loud-speaker, variable current, a
1
= 1
15.4^^^ _ |
(
Summary.
Wehave seen that the standing-wave method gives
coefficients of absorption fornormal incidence only and
motion node of the standing-
for samples placed always at a
wave system. Moreover, with materials whose absorption
is due to inelastic flexural vibrations, the small-scale
measurements on rigidly mounted samples fail to give
the values that are to be expected from extended areas
having a degree of flexural motion. On the other hand,
reverberation coefficients are deduced on the assumptions
made and verified in the reverberation theory as it is
applied to the practical problems of architectural acoustics.
We have seen also that all of the reverberation methods
now in use go back to the determination of the rate of
decay of sound in a reverberation chamber and the effect
of the absorbent material on this rate of decay and that
very nature of the case, the precision of such measure-
in the
ments is not great. The oscillograph method is equally
laborious and requires repeated measurements in order to
eliminate the error due to the irregularities in the decay
curve resulting from interference. Finally, we have noted
that there is a certain degree of uncertainty as to the
of the pore channels, (6) their depth, and (c) the ratio of
perforated to unperforated area of the surface. Rayleigh's
analysis is for normal incidence and leads to the conclusion
that the absorption increases approximately as the square
root of the frequency. For
a given ratio of unperforated
to perforated area, the ab-
sorption low frequencies
at
increases, though not linearly
200 400 fiOO 3200 6400
with the radius of the pores,
Frequency
considered as cylindrical
Fia. 51. Theoretical absorption of
a closely packed porous material of tubes. the radius of pores
If
great thickness. (After Crandall.) be greater than 0.01 cm., the
assumptions made in the theory do not hold. For a
coarse-grained structure, the thickness required to produce
a given absorption at a given frequency is greater than
with a fine-pored material. Crandall 2 has worked out
the theoretical coefficients of absorption of an ideal wall
of closely packed honeycomb structure (i.e., one in which
the ratio of unperforated to perforated area is small),
the diameter of the pores being 0.02 cm. The thick-
1
RAYLEIGH, "Theory of Sound," vol. II, pp. 328-333.
,Phil. Mag., vol. 39, p. 225, 1920.
PARIS, E. T., Proc. Roy. Soc., Ser. A, vol. 115, 1927.
CRANDALL, "Theory of Vibrating Systems and Sound," p. 186, D. Van
Nostrand Company, 1926.
2
CRANDALL, op. tit., p. 189.
ROUND-ABSORPTION COEFFICIENTS OF MATERIALS 129
Thickness in Inches
Fio. 52a. Absorption coefficients of asbestos hair as a function of thickness.
assumed.
0,11
go.io
050
0,08
0.07
0.06
Fig. 52a. The former are for a stiff pressed board with a
fairly impervious surface made of wood fiber. The lower
curve is for a %-in. thickness, while the upper curve is for
the same material %e in. thick. In contrast to the felt,
the thinner more flexible material shows the higher absorp-
tion. In absorption due to flexural vibration, the density,
stiffness, and damping coefficient of the material affect
the absorption coefficient. The mathematical theory of
the process has not yet been worked out. The almost
uniform value of the coefficients for different frequencies
in Fig. 56 indicates the effect of damping in decreasing the
effects due to resonance.
"0133456789 Area
FIG. 57. Absorbing power per unit area as a function of area.
*
It will be noted that there is a marked variation in
the values of the coefficients for the open window. The
data for Room 15 show a decrease in the apparent absorb-
ing power as the area of the individual openings is increased,
quite in agreement with the results obtained in this
laboratory.
One an explanation of these facts in the phenomenon
finds
of diffractionand the screening effect of an absorbent area
upon adjacent areas In the reverberation theory, we
assume a random distribution of the direction of propaga-
tion of sound energy. Thus, on the average, two-thirds of
the energy is being propagated parallel to the surface of
the absorbent material. Neglecting diffraction, in such a
distribution, only that portion traveling at right angles
134 ACOUSTICS AND ARCHITECTURE
FIG. 58. Sound pulse incident, upon a barrier with an opening. A-B marks the
limits of the opening.
FIG. 60. Absorbing power per square unit of distributed material plotted as a
function of ratio of absorbent area to total area of pattern. (After Parkinson.)
*
CHHIMLEH, V L ,
Jour Aeons Soc. Amer,, p. 126, July, 1930.
l
Jour. of Acou*. Soc. Awcr., Vol. 2, No. 1, p. 127, July, 1930.
SOUND-ABSORPTION COEFFICIENTS OF MATERIALS 143
These values are more nearly in accord with those that are
universally used in practice.
It is. apparent from the foregoing that the sound-absorb-
Total 909
Units
1. Absorbing power of bare room, 0.3^842,000*
(assuming wood seats 0.3) .... 2 630
,
_T
r '-
0.5 X 842,000 = 2 - 91s .
14,468
*
On account heavy padding of the boxes, the total opening of the boxes into the
of the
main body of the room was considered as an area from which no sound was reflected.
t Seats with absorbing power of 0.3 are assumed in the formula for the bare room; hence
the deduction of 3 from 2.6, the absorbing power of the seats used.
*
The added absorbing power per person is assumed to be 4 6. The absorbing power of
the person less 2 6, the absorbing power of the seat, which he is assumed to replace, is 2 0.
The value
of the upholstered chairs in minimizing the
audience upon the reverberation time is well
effect of the
Eyring, Jour. Acous. Soc. Amer., vol. 3, No. 2, p. 181, October, 1931.
2
Jour. Soc. Mot. Pict. Eng., vol. 15, pp. 528-548, October, 1930.
152 ACOUSTICS AND ARCHITECTURE
Compute the total absorbing power of the space under the balcony.
than the absorption supplied by treating the opening as a
If this is less
Units
1. Absorbing power of bare room O.S^?" . .
2,400
2. Boxes, 8 by 112 ft. X 1.0 .............. 896
3. Stage, 1,800 sq. ft. X 0.44 .............. 792
4. Carpets, 2,000 sq. ft. X 0.25 ............. 500
5. Seats, 2,450 X (2.6
- 0.3) .......... 5,650
6. Wall drapes, 2,400 sq. ft. X 0.44 ......... 1 ,050
7. Balcony opening, 15 by 112 ft. X 1.0.. .. 1,680
8. Balcony opening, 15 by 114 ft. X 1.0 .... 1,710
1
"Collected Papers on Acoustics," p. 71.
REVERBERATION AND THE ACOUSTICS OF ROOMS 155
Reverbe*,*,- seconds
articulation for the auditorium.
FIG. 62. Relation between
1 he CUrveS OI I Ig. 52 1
1
Jour. Acous. Soc. Amer., vol. 1, No. 1, p. 57. 1929.
156 ACOUSTICS AND ARCHITECTURE
Percentage Intelligibility,
Articulation Per Cent
70 98
80 99
90 99 +
00083V,-
(9.1 -
.
r. -
,
log, a).
* 0.05V
Computed by formula TQ =
0-1)
t Reverberation considered too low
t Reverberation considered too high.
w l' 2
integrated over the total time
of decay shall have a constant
value. This implies that both
the maximum intensity and
Volume in Cubic Feet
the duration of a sound con-
PIG. 64. Optimum reverberation
tribute to the magnitude OI times proposed by Watson and
Llfshltz -
1
BAGENAL, HOPE, and BURSAR GODWIN, Jour. R.I.B.A., vol. 36, pp.
56-763, Sept. 21, 1929.
160 ACOUSTICS AND ARCHITECTURE
-ASlog e (1 Ota)
have preferred to use the latter. As we have seen, this
gives lower values for the reverberation time than does the
earlier formula, the ratio between the two increasing with
the average value of the absorption coefficient. Therefore,
there is a considerable difference between the two, partic-
ularly for the full-audience conditions. As long as we
adhere to a single formula both in setting up our criterion
of acoustical excellence on the score of reverberation and
1
Since the above was written, the author has been informed by letter that
Professor Knudsen's recommended reverberation times are based on the
Eyring formula. This fact materially lessens the difference between his
conclusions and those of the writer.
REVERBERATION AND THE ACOUSTICS OF ROOMS 163
-1 _ logjoa) (7Q)
1
Jour. Soc. Mot. Pict. Eng., vol. 45, No. 2, p. 157.
REVERBERATION AND THE ACOUSTICS OF ROOMS 167
Early radio broadcasting studio. All surfaces were heavily padded with absorb-
ent material.
is enhanced.
On this point, it may be said that phonograph records
of large orchestras are often made in empty theaters.
REVERBERATION AND THE ACOUSTICS OF ROOMS 169
Frequency Coefficient
128 46
256 61
512 82
1,024 82
2,048 64
4,096 60
coved portions at the side are farther back than when the
source is located at the front of the stage. In ''Collected
Papers," Professor Sabine shows the effect of a cylindrical
1
- i I
i
Height 63.5ft.
Computed volume 1,242,400 cu. ft.
cu. ft. per seat. A balcony extending around the entire room and the
relatively low ceiling height for the horizontal dimension yields a low value
of the volume-absorbing power ratio.
8 The case of one of the best known and most beautiful of theaters in New
York City may be cited as an example of undesirable acoustical results from
elliptical contours.
182 ACOUSTICS AND ARCHITECTURE
Ticket
Office
-"
irMain Floor Plan
Room
Ann
FIG. 72. Paraboloidal plan and section of the Hill Memorial Auditorium,
Arbor, Michigan.
184 ACOUSTICS AND ARCHITECTURE
lmf>05t
Grade 100-0-
FIG. 73. Ray reflections from parabolic surface, Hill Memorial Auditorium.
ABSORBING POWER
Units
__.
Empty room including seals, 0.3 \/V* .
1,480
Stage, 36 by 20 X 0.44 316
Units
Empty room, *\/V* 1,330
Stage, 36 by 20 X 0.44. 316
800 persons X 4.3 . . . .
3,440
Total 5,086
Seated in the wood seats, the audience adds 4.3 units per
person and in the upholstered seat 3.0 units per person, so
that we have:
Absorbing power
5,300 sq. ft. of paneling accounts for only about 5.5 per
cent of the total absorbing power of the Leipzig Gewand-
haus when the audience is present. Substituting plastered
surfaces for the paneled area would not make a perceptible
difference in the reverberation time, nor could it change
the quality of tone in any perceptible degree, in a room of
this size. It is possible that in relatively small rooms, in
which the major portion of the bounding surfaces are of a
resonant material, a real enhancement of tone might result.
In larger rooms and with only limited areas, the author
inclines strongly to the belief that the effect is largely
psychological.
, For the stage floor, and perhaps in a somewhat lesser
degree for stage walls in a concert hall, a light wood con-
struction with an air space below would serve to amplify
the fundamental tones of cellos and double basses. These
instruments are in direct contact with the floor, which
would act in a manner quite similar to that of a piano
sounding board. This amplification of the deepest tones
of an orchestra produces a real and desired effect. One
may note the effect by observing the increased volume of
tone when a vibrating tuning fork is set on a wood table
1
top.
1
Of interest on some recent work by Eyring (Jour. Soc. Mot.
this point is
Pict. Eng., vol. 15, No. 4, p. 532). At points near a wall made of fKe-in.
ply-wood panels in a small room, two rates of decay of reverberant sound
were observed. The earlier rate, corresponding to the general decay in the
room, was followed by a slower rate, apparently of energy, reradiated from
the panels. This effect was observed only at two frequencies. Investiga-
tion proved that the panels were resonant for sound of these frequencies and
that the effect disappeared when the panels were properly nailed to supports
at the back.
196 ACOUSTICS AND ARCHITECTURE
above this, the choir loft in the rear. Speakers who use a
speaking voice of only moderate power and a sustained
FIG. 78. Sound pulse in modified cruciform plan. Hearing conditions much
better in portion represented by lower half of photograph.
0, Pulpit; 1, direct sound; 2, reflected from side wall A; 3, diffracted from P\
4, reflected and focused from curved wall of sanctuary.
Interior of Riverside Church, New York. (Henry C. Pclton, Allen & Collens
Associates, Architects.)
Decibel Scale.
The enormous range of intensities covered
by ordinary
auditory experience suggests the desirability of a scale
whose readings correspond to ratios rather than to differ-
ences of intensity. This suggests a logarithmic scale, since
Phys. Rev., vol. 21, No.
1
5, May, 1923.
206 ACOUSTICS AND ARCHITECTURE
TABLE XVI
ia
logarithm of 10 is 13, and the difference in intensity levels
between a painfully loud and a barely audible sound of
this frequency is 13 bels or 130 db.
Sensation Level.
The term " sensation level is used to denote the intensity
' '
= k (constant)
-j-
f>2
40
-80
dOO
p
(\j
ooog ~
<3*CPoo<-> Q
JP
2
S Q
Q =
Fr?qU5 nV W * **o-
FIG. 81. Each ourved line shows the intensity
levels at different frequencies
which sound as loud as a 1,000-cycle tone of the indicated intensity level above
threshold. (After Kingsbvry.)
Loudness of Noises.
ffotse fobe
measured enters
same ear
FIG. 82. Buzzer audiometer.
1
GALT, R. H., Noise Out of Doors, Jour. Acous. Soc. Amer., vol. 2, No 1,
I 2 3 4 b
Time in Fifth Seconds
- .
(1
- a)
1
]
m .
Total,
energy
&J =
ca
-
= 4A NeV
^- = 4
aSc
NeV A - ac
_ON
(72)J
^
MEASUREMENT AND CONTROL OF NOISE 219
greater coefficient.
220 ACOUSTICS AND ARCHITECTURE
ac
ac
whence
L= ?_' = Z
/' a T'o
Reduction (db)
= 10 log -=
a
10 log ^
1 o
Time Seconds
ogous to that used in the in
Quieting of Hospitals.
Nowhere is the necessity for quiet greater than in the
hospital, and perhaps no type of building, under the usual
conditions, is more apt to be noisy. The accepted notions
of sanitation call for walls, floors, and ceiling with hard, non-
1
How
to Achieve Quiet Surroundings in Hospitals, Modern Hospital,
vol. 32,Nos. 3 and 4, March and April, 1929; Practical Methods of Making
the Hospital Quiet, Hospital Progress, March, 1931; Are Acoustical Materials
a Menace in the Hospital? Jour. Acous. 8oc. Amer., vol. 2, No. 1, July, 1930;
Correct Type of Hardware, Hospital Management, February, 1931.
MEASUREMENT AND CONTROL OF NOISE 229
1
Some Studies on the Absorption of Noise in Ventilating Ducts, Jour.
Heating, Piping, and Air Conditioning, Amer. Soc. Heating Ventilating Eng. t
January, 1931.
230 ACOUSTICS AND ARCHITECTURE
FIG. 89.
"
ments for a single tone are described in his Collected
Papers." The method used was based on the reverberation
theory, already employed in absorption measurements.
The experimental arrangements are shown in Fig. 90
taken from the " Collected Papers." Sound was produced
by organ pipes located in the constant-temperature room
of the Jefferson Physical Laboratory. The test panels were
mounted in the doorway constituting the only means of
entrance into the room, so that when the panel was in place
the experimenter had to be lowered into the room by means
1
"Acoustics of Buildings," p. 179, George Bell & Sons, 1927.
MEASUREMENT OF SOUND TRANSMISSION 235
loge
- = At
A = = TT/
/r sp 4V
If be the time required for the intensity in the source room
ti
, /i = acti
Iog <
T 4F
Let the reduction factor of the partition, be the ratio
k,
at any moment of the average intensity of the sound in
the source room to the intensity at the same moment on the
farther side of the partition and close to it. Then when
the sound heard through the partition has just reached
the threshold intensity, the intensity / in the source room
is given by the relation
I = Id
ft 4F
whence, by subtraction, we have
log e k = ~(ti -
and
logic t
= - W
~g ~(ti (73)
FIG 91. Room for sound transmission measurements at the Bureau of Standards.
y^w#^^
Feet
Fia. 93. National Physical Laboratory rooms for sound transmission measure-
ments. (After Davis and Littler.)
^ UP
Kreuger.
JSl^ ^3 t
ta
t I
Audiometer receiver]
with offset cap
Fiu. 96. Method of sound transmission measurements used by Meyer and Just.
hence,
T
Jt =
E, = SJ =
t (74)
1
Bur. Standards Sci. Paper 506.
248 ACOUSTICS AND ARCHITECTURE
In the steady state, the wall acts for the receiving room
as a sound source whose acoustic output is J? 2 and the ,
whence
1 IlS
energy r = 0.001,
- = 1,000, log
- = 3, and the transmis-
T T
" "
sion loss is 30 db. The relation between
reduction f actor
as measured and transmission loss as just defined remains to
be considered.
1
Jour. Acous. Soc. Amer., vol. 11, No. 1, p. 129, July, 1930.
MEASUREMENT OF SOUND TRANSMISSION 249
T. L. = 10 log (k .
^] = loftog k + log ~1 (77)
\ fib/ L #2 J
IV ~~~
logio
- = logic k + lo glo - - logic
- ^/] (78)
T (l
2 \ ttaKi/
Frequency Correction
128 2 5
256 1.2
512 0.5
1,024 0.5
2,048 1 3
Average 1
cl
f =
1
10 log 10 log
*2 %1 (79)
40 = 10 log -, log
- = 4, and r = 0.0001
For door:
25 = 10 log -
, log
- = 2.5, and r -= 0.0032
Statement of Results.
Reference has already been made to the fact that due
to resonance, the reduction of sound by a given wall may
vary markedly with slight variations in pitch. For this
reason, tests at a single frequency or at a small number of
frequencies distributed throughout the frequency range
may be misleading, and difficult to duplicate under slightly
altered conditions of test, such, for example, as variations
in the size of the test panel. It appears, however, that
for most of the constructions studied there is a general
similarity in the shape of the frequency-reduction curve,
253
254 ACOUSTICS AND ARCHITECTURE
Sound-proof Doors.
" "
Various types of nominally sound-proof doors are
now on the market. These are usually of heavy construc-
tion, with ingenious devices for closing the clearance
cracks.
Table XVIIa presents the results of measurements made
on a number of doors of this type. It is interesting to note
the increase in the sound reduction with increasing weight,
and this regardless of whether the increased weight is due
to the addition of lead or steel sheets incorporated within
the door or simply by building a door of heavier con-
struction. The significance of this will be considered in a
later section.
TABLE XVII. SOUND REDUCTION BY DOORS AND WINDOWS
Average
Number Description reduction,
decibels
Porous Materials.
1
For a full account of the study of the transmission of sound by porous
materials, see Amcr. Architect, Sept. 28, Oct. 12, 1921.
TRANSMISSION OF SOUND BY WALLS 257
3, p. 177, 1927.
1
Phil. Mag., vol.
258 ACOUSTICS AND ARCHITECTURE
whence
= +
log Y
* 2
log (r 0)
and
R = 10 log (r + qt)
that this may also hold true for the composite structures
just considered. Accordingly, in Fig. 98 we have plotted
the reduction against the logarithm of the number of
layers for the paper and felt, for the sheet iron and air
space, and for the sheet iron, felt, and air space. (In
Sabine's experiments, >^-in. felt was placed in a 1-in.
space between the metal sheets.) We note, in each case,
a linear relation between the reduction and the logarithm
of the number of layers, which, in any one type of con-
k Y-
= aw b
where w is the weight per unit area of the composite
structure, and a and b are empirical constants whose
values are functions of the frequency of the sound and also
of the materials and arrangement of the elements of which
the wall is built. The values of 6 for the single frequency
512 vibs./sec. for the three walls are as follows:
260 ACOUSTICS AND ARCHITECTURE
Continuous Masonry.
" "
By continuous masonry we shall mean single walls,
as contrasted with double walls, of clay or gypsum tile,
either hollow or solid, of solid plaster laid on metal lath
and channels, and of brick or concrete. These include
most of the common types of all-masonry partitions.
Figure 99 shows the general similarity of the curves
TRANSMISSION OF SOUND BY WALLS 261
Frequency
finished construction.
In the column headed "Relative
stiffness" are given the steady pres- for homogeneous partitions.
sure in pounds per square foot over the entire face of the
walls that produce a yielding of 0.01 in. at the middle
point of the wall.In each case, the test partition was 6 by
8 ft. built solidly into the opening. Gypsum plaster was
used throughout this series of tests.
55
40
- Continuous
/ masonry plaster, file and'brick
2- Cinder, slag and haydife concrete
30 30 40 80 90 100
Weight per Sq. Ft
FIG. 100. Riverbank measurements on (1) plaster, tile, and brick; (2) cinder,
slag, and haydite concrete.
The
fact of the predominating part played by the weight
per unit area in the reduction of sound produced by walls
of masonry material was first stated by the writer in 1923. x
No generalization beyond the actual facts of experiment
was made. Since that time, work in other laboratories
has led to the same general conclusion in the case of other
Weight/Area
Fio. 101. Reduction as a function of weight per unit area. Results from
different laboratories compared.
per square foot and also results for glass, wood, and steel
partitions are shown.
We note that from the Bureau of Standards tests the
points for the extremely light materials of paper, fabric,
aluminum, and fiber board show a linear relation between
the average reduction and the logarithm of the weight per
unit area. The heavier, stiffer partitions of lead, glass,
and steel as well as the massive masonry constructions
tested by the Bureau show, in general, lower reductions
than are called for by the extrapolation of the straight line
for the very light materials.
Taken by themselves, the Bureau of Standards figures
for heavy masonry constructions, ranging in weight from
30 to 100 Ib. per square foot, do not show any very definite
correlation between sound reduction and weight. On the
whole, the points fall closer to the Riverbank line for
continuous masonry than to the extension of the Bureau
of Standards line for light materials.
Figure 101 gives also the results on homogeneous struc-
tures ranging in weight from 10 to 50 Ib. per square foot,
as reported by Heimburger. 1 As was indicated in Chap.
XI, in Kreuger's tests, a loud-speaker horn was placed
close to the panel and surrounded by a box, and the
intensities with and without the test panels intervening
were measured. This method will give much higher
values of the reduction than would be obtained if the whole
face of the panel were exposed to the action of the
sound. It is worth noting, however, that the slope of the
line representing Kreuger's data is very nearly the same
as that for the Riverbank measurements.
Here, again, the data presented in Fig. 101 lead one to
doubt whether there is a single numerical formula connect-
ing the weight and the sound reduction for homogeneous
structures that will cover all sorts of materials. Equation
(81) gives an approximate statement of the relation
between weight and reduction for all-masonry construc-
tions. The Bureau of Standards findings on light septa
1
Amer. Architect, vol. 133, pp. 125-128, Jan. 20, 1928
268 ACOUSTICS AND ARCHITECTURE
gypsum
Rock lath r -
"*4"batten \- <
:
57
cr/es- 7'apart
24" fe
3
/4"C-/2"O.C. &$* /Vo30 flat expJath
Efcfc
^
^t Gypsum plaster
Ce/ofex (Loose)
3-5$
^ P/asfer on tile
60r
8 10 20 30 40 50 60 60 100
FIG. 108. Various constructions compared with masonry walls of equal weight.
General Conclusions.
plaster, and brick the reductions will fall very close to the
straight line plotted. Wood-stud construction with gyp-
sum plaster falls on this line (No. 62). Lime plaster on
wood studs and gypsum plaster on fiber-board plaster
bases on wood studs give somewhat greater reductions than
continuous masonry of equal weight (Nos. 63, 68, 66).
Glass and steel show greater reductions than masonry of
equal weights (Nos. 1 and 10). The superiority of lime
over gypsum plaster seems to be confined to wood-stud
constructions. The Bureau of Standards reports the results
of tests in which lime and gypsum plasters were applied to
identical masonry walls of clay tile, gypsum tile, and
brick. In each case, two test panels were built as nearly
alike as possible, one being finished with lime plaster. In
each case, the panel finished with the gypsum plaster
showed slightly greater reduction than a similar panel
finished with lime plaster. The difference, however, was
not sufficiently great to be of any practical importance. 1
These facts bring out the fact referred to earlier, that the
sound insulation afforded by partitions is a matter of struc-
tural properties, rather than of the properties of the
materials comprising the structure.
2. The reduction afforded by double construction is
II I I I 1 1 1 I I I I
1.1 1 II _L
~04 0506 0810 20 30 40 5060 ao 10 20 30 40 5060 80100
Weight/Area
FIG. 109. Comparison of results obtained at different laboratories.
MACHINE ISOLATION
In every large modern building, there is usually a certain
amount of machine installation. The operation of venti-
For a full theoretical treatment the reader should consult a recent text
1
Fia. 110. The mass m moves under the action of the elastic restoring force of the
spring and the frictional resistance in the dash pot.
= (83)
MACHINE ISOLATION 285
For the present purpose, the most useful form for the
solution of Eq. (83) is
(84)
sn
where "
k, called the damping coefficient/' is defined by
the equation k = r/2m, and coi is defined by the equation
(85)
1.0 Sees,
.
Graph of damped sinusoidal oscillation. Ao/Ai Ai/A?
w2 o
Case I
|^-
< 2 (small damping)
r2 s
Case II -2
= (critical damping)
T2 S
Case III
4^-2
> (large damping)
case
given
Here o>
by the equation
^o
L*. = Ao = ^ \4fc^F
i
W _ W2
/
--
Practical problems of machine isolation come under
I, and the particular solution under this condition is
(87)
1
WOOD, A. B., "A Textbook of Sound," p. 37.
288 ACOUSTICS AND ARCHITECTURE
Inertia Damping.
We consider now the question of the effect of mass
upon the amplitude of vibration of asystem under the
action of an impressed periodic force. In the expression
for the amplitude given in Eq. (87), it would appear that
increasing m
will always decrease A since appears in ,
m
the denominator of the right-hand member of the equation.
It must be remembered, however, that is involved in m
both k( r/2m) and co (= ^s/m).Putting in these
values, Eq. (87) may be thrown into the form
(88)
(s
- <mo> 2 ) 2
It is produced on A by increasing m
clear that the effect
will depend upon whether this increases or lowers the
absolute value the expression (s
of wco 2 ) 2 . If co
2 =
sjm < co
2
,
then increasing m
decreases (s mco 2
)
2
,
decreas-
ing the denominator of the fraction and hence increasing
the value of A Q In other words, if the driving frequency
.
Graphical Representation.
The foregoing discussion will perhaps be clarified by
reference to Fig. 112. Here are shown the relative ampli-
tudes of vibration of a machine weighing 1,000 Ib. for a
fixed value of the amplitude and frequency of the impressed
force. These are the familiar resonance curves plotted so
290 ACOUSTICS AND ARCHITECTURE
0.1 02 0.4
FIG. 112. Amplitudes of forced vibrations with a fixed driving frequency and
varying values of the stiffness of the resilient mounting. The upper abscissae
are the ratios of the natural frequencies to the fixed driving frequency
5 = Trr/VSw = fc//o = Ao/Ai, in Fig. 111.
Transmission of Vibrations.
Let us suppose that the mass m of Fig. 110 is a machine
which due to unbalance or some other cause exerts a
periodic force on its mounting. Suppose that the maximum
value of the force exerted by the machine on the support
is FI and that this transmits a maximum force F 2 to the
r = (89)
\4A 2co 2 + (co
2 - co
2
)
2
where co =
27r/ and co
= 2?r/, / and / being the natural ,
_ 2 -
_ 2
__ 2
R - 2
CO C0 CO
*
_ l
I
<0
08 10 15
Log.S
FIG. 113. Transmissibility of resilient mounting. The upper abscissae are the
ratios of the natural frequencies to the driving frequency.
then
1
not even approximately true for a material like rubber,
This statement is
J .- ~\l U\J*S
\
-I
2ir md
As an example, let us take the case of a motor-driven
fan weighing 8,000 lb., with a base 3 by 4 ft. Suppose
that a light-density cork is to be used as the isolating
medium. For 2-in. thickness, a material of this sort
shows a compression of about 0.2 in. under a loading of
5,000 lb. per square foot. If the cork is applied under
the entire base of the machine, then we shall have the
natural frequency of the machine so mounted
f = 3 13 /^OQQ"^"^ = 19 l
,
= 6
ft = t 5,000
~
XA
/ /
8,000 X 0.2
from which we find A 1.17 sq. ft. This gives a loading
of about 6,800 lb. per square foot. A larger area of material
is perhaps desirable. We can keep the same spring factor
by using a larger area of a thicker material. Thus if we
double the thickness, we should need twice the area, giving
a loading of only 3,400 lb. per square foot and the same
transmission.
It is apparent from the foregoing that the successful
use of a material like cork or rubber involves a knowledge
of the stress-strain characteristics of the material and the
296 ACOUSTICS AND ARCHITECTURE
0.10 0.20 030 0.40 050 0.60 0.70 0.80 0.90 1.00
Deformation in Inches
30 70
Natural Frequency
FIG. 115. Natural frequencies for various loadings on the light density cork of
Fig. 114.
FIG. 116. Effect of area of rubber pads upon the natural frequency for various
loadings. (Hull and Stewart.)
From known
properties of steel, springs may be
the
designed having any desired characteristics over a fairly
wide range. From the standpoint of predictability of
performance and the control of spring factor to meet
any desired condition, spring mounting is advantageous.
Because of the relatively low damping, in comparison with
organic materials, the amplitude of vibration of the machine
and the transmission to the floor are large when the machine
is operating at or near the resonance speed. The trans-
1
Elastic; Supports for Isolating Rotating Machinery, Trans. A. I. E. E.,
vol. 50, pp. 1063-1068, September, 1931.
MACHINE ISOLATION 299
Results of Experiment.
(a)
(c) (d)
FIG. 117. (a) Vibration of machine with solid mounting, (b) Vibration of
floor with solid mounting, (c) Vibration of machine mounted on U. S. G. 500 Ib.
clip, (d) Vibration of floor machine mounted on U. S. G. 500 Ib. clip.
(c) (d)
"""*
FIG. 118. (a) Vibration of machine with solid mounting, (fe) Vibration of
floor with solid mounting, (c) Vibration of machine on 1-in. high-density cork;
loading 1,000 Ib. per square foot, (d) Vibration of floor, machine mounted on
1-in. high-density cork; loading 1,000 Ib. per square foot, (c) Vibration of
machine on 2-in. low-density cork; loading 3,000 Ib. per square foot. (/) Vibra-
tion of floor, machine mounted on 2-in. low-density cork; loading 3,000 Ib. per
square foot.
General Conclusions.
REFERENCES
HULL, E. H.: Influence of Damping inthe Elastic Mounting of Vibrating
Machines, Trans. A. S. M. E., vol.
53, No. 15, pp. 155-165.
and W. C. STEWART: Elastic Supports for Isolating Rotating
Machinery, Trans. A. I. E. E. t vol. 50, pp. 1063-1068, September, 1931.
KIMBALL, A. L.: Jour. Acous. Soc. Amer., vol. 2, No. 2, pp. 297-304, October,
1930.
NICKEL, C. A.: Trans. A. I. E. E. p. 1277, 1925.
t
307
308 ACOUSTICS AND ARCHITECTURE
TABLE II. COEFFICIENTS OF VOLUME ELASTICITY, DENSITY, VELOCITY
OF SOUND, AND ACOUSTIC RESISTANCE
v = 4F
~s
FIG. 1.
2"
IdS fi . .
f ad r-
= IcdS
- (4)'
I sin (f>
cos <pd<f> \ \
dr A
v
j
47T Jo JO Jo ^
The total energy that is incident per second on a unit-
area is therefore /C/4, and on the entire bounding surface
S is IcS/4.
Now we can find an expression for this same quantity
in terms of the mean free path of our supposed unit energy
APPENDIX B 311
312
APPENDIX C 313
1
Measurements made by timing duration of audible sound from organ-pipe source.
Krnpty-room absorbing power measured by variable source methods (loud-speaker). River-
bank Laboratory tests.
314 ACOUSTICS AND ARCHITECTURE
TABLE II. COEFFICIENTS OF ABSORPTION OF MATERIALS. (Continued)
APPENDIX C 315
317
318 ACOUSTICS AND ARCHITECTURE
TABLE II. NOISE IN BUILDINGS*
Level above
Threshold,
Location and Source Decibels
Boiler factory . . 97
Subway, local station with express passing 95
Noisy factories 85
Very loud radio in home 80
Stenographic room, large office. .. 70
Average of six factories 68
Information booth, large railway station . . 57
Noisy office or department store . . . .57
Moderate restaurant clatter .... 50
Average office . . . 47
Noises measured in residence .45
...
. .
Average residence. . . 32
Quietest residence measured . 22
"
1
Taken from City Noise," Department of Health, New York City.
APPENDIX E
Within the past year, a considerable amount of research has been done
in an attempt to standardize methods of measurement of absorption coeffici-
ents. Diverse results obtained by different laboratories on ostensibly the
same material have led to considerable confusion in commercial applica-
tion. Early in 1931, a committee was appointed by the Acoustical Society
of America to make an intensive study of the problem with a view to estab-
A full report of the Bureau of Standards investigation has not as yet been
published, and further study of the sources of disagreement is still in progress
at the Riverbank Laboratories. The data of the foregoing table serve,
however, to indicate the degree of congruence in results that is to be expected
in methods of measurements thus far employed.
Method la in the same as the Bureau of Standards method
Table II is
of this sort. In Table II, we note rather wide divergence at certain fre-
quencies between the results of ear observations and those in which a
microphone WPS employed. Very recent work at both tho Bureau of Stand-
ards and the Riverbank Laboratories points to the possibility that in some
cases, the rate of decay of reverberant sound is not uniform. If this is
true, then the measured value of the coefficient of absorption will depend
upon the initial intensity of the sound and the range of intensities over which
the time of decay is measured. In such a case, uniformity of results can be
obtained only by adoption by mutual agreement of a standard method and
carefully specified test conditions. This is tho procedure frequently fol-
lowed where laboratory data have to be employed in practical engineering
problems. The measurements of thermal insulation by materials is a caso
in point. Pending such an establishment of standard methods and specifica-
tions, the data on absorption coefficients given in Table II (Appendix C)
may be taken as coefficients measured by the method originally devised and
used by W. C. Sabiiie. The data given out by the Bureau of Standards for
1931 and later were obtained by the method outlined in the second para-
graph of this section.
INDEX
Audiometer, buzzer type, 213
Auditorium Theater, Chicago, 185
Absorbents, commercial, 143, 313
Absorbing power, of audience, 140 B
defined, 58
empirical formula for, 147 Bagenal and Wood, 194
of seats, 140, 312 Balcony recesses, allowance for, 151
total and logarithmic decrement, Buckingham, E. A., 58, 247
74 Bureau of Standards, absorption
unit of, 69 measurements, 90
Absorption, due to flexural vibra-
tion, 130
due to porosity, 128
Carnegie Hall, New York, 186
uniform, 169 Chrisler, V. L., 89, 122, 142, 237
Absorption coefficient, definition of, Chrisler and Snyder, 115, 266
86
Churches, Gothic,, 197
effect of area on, 131
various types of, 196
of edges on, 134
Concert halls, 161
of thickness on, 92, 129
Condensation, in compressional
of mounting on, 137
wave, 20
of quality of test tone on, 138
Cork, for machine isolation, 296,
of spacing on, 136
300, 302
impedance measurement of, 93
Crandall, I. B., 23, 93, 128
measurement of, by different
Cruciform plan, reflections in, 199
methods, 119, 319 Curved shapes, allowable, 176
of small areas, 135
defects caused by, 171
stationary-wave measurement of,
87 D
two meanings of, 86
Absorption coefficients of materials, Damping, due to inertia, 289
313-316 on transmission, 293
effect of,