Page 295 - The Master Handbook Of Acoustics

270 CHAPTER THIRTEEN

general, the smoothness of the decay increases as frequency is
increased. The reason for this, as explained in Chap. 7, is that the
number of modes within an octave span increases greatly with fre-
quency, and the greater the mode density, the smoother their average
effect. Beats in the decay are greatest at 63 Hz and 125 Hz. The decays
of Fig. 7-10 indicate that the diffusion of sound in this particular stu-
dio is about as good as can be achieved by traditional means. It is the
beat information on the low-frequency reverberation decay that
makes possible a judgment on the degree of diffusion prevailing.
Reverberation-time measuring devices that yield information only on
the average slope and not the shape of the decay pass over informa-
tion that most consultants consider important in evaluating the dif-
fuseness of a space.

Exponential Decay

A truly exponential decay is a straight line on a level vs. time plot,
and the slope of the line can be described either as a decay rate in
decibels per second or as reverberation time in seconds. The decay
of the 250-Hz octave band of noise pictured in Fig. 13-2 has two
exponential slopes. The initial slope gives a reverberation time of
0.35 second and the final slope a reverberation time of 1.22 seconds.
The slow decay that finally takes over once the level is low enough
is probably a specific mode or group of modes encountering low
absorption either by striking the absorbent at grazing angles or strik-
ing where there is little absorption. This is typical of one type of
nonexponential decay, or stated more precisely, of a dual exponen-
tial decay.
Another type of nonexponential decay is illustrated in Fig. 13-3.
The deviations from the straight line connecting the beginning and
end of the decay are considerable. This is a decay of an octave band of
noise centered on 250 Hz in a 400-seat chapel, poorly isolated from an
adjoining room. Decays taken in the presence of acoustically coupled
spaces are characteristically concave upward, such as in Fig. 13-3, and
often the deviations from the straight line are even greater. When the
decay traces are nonexponential, i.e., they depart from a straight line
in a level vs. time plot, we must conclude that true diffuse conditions
do not prevail.