808 Chapter 24
35 ms. We can now analyze the sound at each of the
three points:
- The loudspeaker is 6 dB louder than the source and
5 ms later. The overall sound is heard as originating
at the source but 6 dB louder. - The loudspeaker is 2 dB louder than the source and
20 ms later. The overall sound is heard as origi-
nating at the source but 2 dB louder. - The loudspeaker is the same level as the source but
35 ms later. At this point the delay is too long and a
distinct echo is heard.
In reality, the coverage pattern of the loudspeaker
should be chosen to ensure that the level of the sound is
sufficiently attenuated outside the area where the delay
works effectively.
24.2.3 Reverberation Synthesis
Reverberation is the result of many reflections of the
original sound. The general pattern of events, as shown
in Fig. 24-8 is that there is first a direct sound, followed
by a short gap, referred to as the initial time gap (ITG).
Next come the first distinct early reflection echoes
caused by sound bouncing off surfaces near either the
source or the listener. Thereafter the reflected sounds
start to generate their own second, third and higher-order
reflections and the energy level settles down to a constant
decay rate. This decay rate is related to the distances
traveled and the amount of absorption in the room.
Delay is used as the basis for reverberation synthesis
because it provides a convenient method for storing the
signal and releasing it at a later time, much as reflec-
tions from the surfaces of a room arrive at the listener at
a later time than the direct sound. Typical applications
for synthetic reverberation include the enhancement of
program material in the production of recordings, the
introduction of special effects in live entertainment pro-
ductions, and compensation for poor or lacking natural
reverberation in entertainment spaces.
Requirements for good reverberation synthesis are
essentially the same as for an acoustically well-designed
hall. There are many parameters that need to be consid-
ered to help achieve realism in reverberation simulation.
- Distance from Source. The perception of distance is
controlled primarily by the relative energy levels of
the direct components and the decay components. - Room size. The perceived room size is controlled by
the delay time from the first grouping of direct sound
and early reflections to the start of the decay tail and
by the length of the decay tail. The requirements for
the decay tail are the same as for an acoustically
well- designed room. A relatively smooth decay rate
is desirable, with longer decay times at lower
frequencies than at higher frequencies to simulate the
high-frequency losses as sound travels through the
air.- Brightness. The spectral balance of the decay tail
determines the character of the reverberation. A lot of
high- frequency roll-off simulates a room with a lot
of absorption from carpets, curtains, or designed
absorption devices and gives a dark sound. Less
height frequency roll-off simulates a hard-surfaced
room such as the inside of a stone church, giving a
bright sound. - Character. The smoothness of the decay determines
the character of the sound. A room with large
opposing flat surfaces will exhibit a flutter echo
where the sound bounces back and forth between the
walls with little diffusion. A room with more archi-
tectural features or multiple surfaces will tend to
scatter the sound more, creating a denser and more
evenly distributed decay. - Envelopment. The sense of the reverberation coming
from all around you rather than a specific location is
controlled, making different patterns for the different
playback channels. This can be effectively achieved
using two-channel stereo as well as in systems with
multiple dedicated speakers. The most important
differences are in the pattern of the early reflections.
The decay tail portions should keep the same with the
direct to reverberant energy ratio and decay time but
can be randomized to produce a denser sound field.
Portions of the randomized signal can be altered in
their frequency response to mimic the ear’s nonuni-
form response to sounds from behind, further
increasing the sense of envelopment.
Figure 24-8. Energy Time Curve showing the delay of
sound in a room60
50
40
30
20
10
00
10
20
30
40
501024 samples from 1780 Hz to 2220 Hz in 4.0 s, HammingRT60 = 0.88 s EDir/ERev = -4.37 dB %ALcons = 4.93 dB down = 20.02Time–ms200 300 400 600 700 800 900