136 Chapter 6
Fig. 6-11 is an Envelope* Time Curve of a large,
reverberant room, measured using time delay spectrom-
etry. The left side of the graph represents t0 or the begin-
ning of the measurement. Note that this is the point in
time what the signal leaves the loudspeaker. If the micro-
phone represents the observer in this system, the
observer would not hear anything until t(0) +t(x) where
t(x) is the time of takes for the sound to leave the loud-
speaker and arrive at the observation point. In this
measurement, that time is 50 ms because the loudspeaker
was about 56 ft (17 m) away from the microphone. This
first arrival is known as the direct sound because it is the
sound energy that first arrives at the listener or micro-
phone, before it reflects off of any surface. A careful
examination of this graph shows a small gap between the
direct sound and the rest of the energy arriving at the
microphone. This is known as the initial time gap (ITG)
and it is a good indicator of the size of the room. In this
room it took about 50 ms for the sound to travel from the
loudspeaker to the microphone then another 40 ms
(90 ms total) for sound to leave the speaker and bounce
off of some surface to arrive at the microphone. There-
fore, in this room the ITG is about 40 ms wide.
Fig. 6-12 is an enlargement of the first 500 ms of
Fig. 6-11. The ITG can be clearly seen and is about
40 ms long. The sound then takes about 130 ms or so to
build up to a maximum at around 270 ms. Fig. 6-12
shows that the sound then decays at a fairly even rate
over the next 4 s till the level falls into the noise floor.
If we perform a Schroeder integration^28 of the energy
then measure the slope and extrapolate down to 60 dB
below the peak, we see the reverb time of this room to
be on the order of 6.8 s at 500 Hz, Fig. 6-13.
It is useful to take a look at the ETC of a small room
for comparison, Fig. 6-14.
Careful examination of Fig. 6-14 reveals a room
dominated by strong discrete reflections that start
coming back to the observation point within a few ms of
the direct sound. By 30 ms after the direct sound, the
energy has decayed into the noise floor.
Since there is no significant diffuse or reverberant
field in acoustically small rooms, equations having
reverb time as a variable are not appropriate.
It is important to understand that small rooms must
be treated differently with respect to frequency.
Consider Fig. 6-15. These boundaries should not be
understood as absolute or abrupt. They are meant to
serve as guidelines and the transitions from one region
to another are actually very gradual.
Region 1 is the region from 0 Hz up to the first mode
associated with the longest dimension. In this region
there is no support from the room at all, and there is not
much one can do to treat the room. Region 2 is bounded* The term Energy Time Curve was suggested by
Heyser and adopted by Crown and Gold Line,
respectively, in their Time Delay Spectrometry
software. Recently Davis and Patronis have
suggested that Envelope Time Curve is a better
label for this graph.Figure 6-11. ETC of a large reverberant church. Measurement courtesy of Jim Brown.