Reverberation and Echoes
Most of the time we are unaware of how much of the sound that we hear comes
from reflections from environmental surfaces. Even out of doors, a significant
amount of energy is reflected by the ground and by surrounding structures
and vegetation. However, we only notice these reflections when the time
delay gets longer than about the 30- to 50-ms echo threshold,
in which case we become consciously aware of them and call them echoes.
Special rooms called anechoic chambers are built to absorb
sound energy, so that only the directly radiated energy reaches the ears.
Upon entering an anechoic chamber for the first time, most people are astonished
by how much softer and duller everything sounds.
If reflected sound is so common in ordinary acoustic environments, one might
wonder why these reflections do not interfere with our ability to localize
sources. The answer seems to be that we quickly adapt to a new environment,
and our auditory system uses only partially understood mechanisms to suppress
the effects of reflections and reverberation. The fact that we localize
on the basis of the signals that reach our ears first is known as the precedence effect or the Law of
the First Wavefront (see Blauert
). This is not to say that we are unaware of the reflections that follow.
Indeed, we subconsciously use this information to estimate range. However,
unless reverberation is severe, the reflections have relatively little effect
on our ability to localize sounds.
However, the precedence effect does force us to modify Rayleigh's Duplex
Theory. In a typical room, reflections begin to arrive a few milliseconds
after the initial sound. For a low-frequency sound whose period is longer
than the time for reflections to arrive (for example, below 250 Hz), the
reflections begin arriving before even one cycle is completed. By the time
several cycles have arrived and the auditory system can begin to estimate
pitch, the sound pattern in the room is a jumble of standing waves, and
it is now impossible for the auditory system to estimate interaural time
differences. Thus, in a reverberant room, low-frequency information is essentially
useless for localization.
However, that does not mean that interaural timing differences are unimportant.
The important timing information comes from the Interaural Envelope
Difference (IED), e.g., from the transients at the onset of a new
sound. This is vividly demonstrated by the Franssen
Effect. If a sine wave is suddenly turned on and a high-pass-filtered
version is sent to Loudspeaker A while a low-pass filtered version is sent
to Loudspeaker B, most listeners will localize the sound at Loudspeaker
A. This is true even if the frequency of the sine wave is sufficiently low
that in steady state most of the energy is coming from Loudspeaker B. Basically,
the starting transient provides unambiguous localization information, while
the steady-state signal is very difficult to localize, and in this circumstance
the auditory system simply ignores the ambiguous information. With some
risk of oversimplification, we can generalize and say that in reverberant
environments it is the high-frequency energy, not the low-frequency energy,
that is important for localization.
Back
to Range 
Up
to Psychoacoustics