Introduction to Acoustics

13.4 Temporal Processing in the Auditory System
This section is concerned mainly with temporal resolution
(or acuity), which refers to the ability to detect
changes in stimuli over time, for example, to detect
a brief gap between two stimuli or to detect that a sound
is modulated in some way. As pointed out by Viemeister
and Plack [13.71], it is also important to distinguish
the rapid pressure variations in a sound (the fine structure)
from the slower overall changes in the amplitude
of those fluctuations (the envelope). Temporal resolution
normally refers to the resolution of changes in the
envelope, not in the fine structure. In characterizing temporal
resolution in the auditory system, it is important
to take account of the filtering that takes place in the peripheral
auditory system. Temporal resolution depends
on two main processes: analysis of the time pattern occurring
within each frequency channel and comparison
of the time patterns across channels.
A major difficulty in measuring the temporal resolution
of the auditory system is that changes in the
time pattern of a sound are generally associated with
changes in its magnitude spectrum, for example its
power spectrum (see Chap. 14). Thus, the detection of
a change in time pattern can sometimes depend not
on temporal resolution per se, but on the detection of
the change in magnitude spectrum. There have been
two general approaches to getting around this problem.
One is to use signals whose magnitude spectrum is not
changed when the time pattern is altered. For example,
the magnitude spectrum of white noise remains
flat if the noise is interrupted, i. e., if a gap is introduced
into the noise. The second approach uses stimuli
whose magnitude spectra are altered by the change in
time pattern, but extra background sounds are added
to mask the spectral changes. Both approaches will be
described.
13.4.1 Temporal Resolution Based
on Within-Channel Processes
The threshold for detecting a gap in a broadband noise
provides a simple and convenient measure of temporal
resolution. Usually a two-alternative forced-choice
(2AFC) procedure is used: the subject is presented with
two successive bursts of noise and either the first or
the second burst (at random) is interrupted to produce
the gap. The task of the subject is to indicate which
burst contained the gap. The gap threshold is typically
2–3 ms [13.72, 73]. The threshold increases at very low
sound levels, when the level of the noise approaches the
Psychoacoustics 13.4 Temporal Processing in the Auditory System 473
absolute threshold, but is relatively invariant with level
for moderate to high levels.
The long-term magnitude spectrum of a sound is
not changed when that sound is time reversed (played
backward in time). Thus, if a time-reversed sound can be
discriminated from the original, this must reflect a sensitivity
to the difference in time pattern of the two sounds.
This was exploited by Ronken [13.74],whousedas
stimuli pairs of clicks differing in amplitude. One click,
labeled A, had an amplitude greater than that of the other
click, labeled B. Typically the amplitude of A was twice
that of B. Subjects were required to distinguish click
pairs differing in the order of A and B: either AB or BA.
The ability to do this was measured as a function of the
time interval or gap between A and B. Ronken found that
subjects could distinguish the click pairs for gaps down
to 2–3 ms. Thus, the limit to temporal resolution found
in this task is similar to that found for the detection of
a gap in broadband noise. It should be noted that, in this
task, subjects do not hear the individual clicks within
a click pair. Rather, each click pair is heard as a single
sound with its own characteristic quality.
20 log (m)
0
–5
–10
–15
–20
–25
–30
2
10
100
1000
Modulation frequency (Hz)
Fig. 13.15 A temporal modulation transfer function
(TMTF). A broadband white noise was sinusoidally amplitude
modulated, and the threshold amount of modulation
required for detection is plotted as a function of modulation
rate. The amount of modulation is specified as 20 log(m),
where m is the modulation index (see Chap. 14). The higher
the sensitivity to modulation, the more negative is this
quantity. (After Bacon and Viemeister [13.70])
Part D 13.4