Page 44 - Spring2022

WILLIAM A. YOST
Figure 3. Left: Neural activation pattern (NAP) of a 200- Hz harmonic complex from model of Patterson et al. (1995). Right: NAP of an iterated rippled noise (IRN), three iterations with 5-ms delay. Adapted from Yost (2009). See text for detailed discussion.
produces a pitch corresponding to the inverse of the delay. Bill has investigated extensively how the pitch and pitch strength of IRN changes with its various parameters and has concluded that a temporal model that extracts periodicities in the fine structure of IRN best accounts for the data. Bill’s work on IRN is his most frequently cited and has contributed to making IRN a standard stimulus in many other areas of research on hearing.
Temporal Modulations in Sound
Most sounds of interest to us in nature change over time. The messages we convey through speech, the pleasure we take in music, and the actions we track in the sound events unfolding around us all derive from modulations
in the sound spectra over time. Without these dynamics, the sounds about us would combine to produce a single unpleasant drone.
Two parallel and largely independent lines of research have evaluated the influence of temporal modulations on sound source perception. The first has focused on the phenom- enon of auditory streaming (Bregman, 1990). This refers to the listener’s subjective impression of when a sequence of sounds, typically tones, is heard to split into separate perceptual objects or entities (streams). Striking exam- ples occur when the tones in the sequence are made to differ in frequency and rhythmic pattern (for demos, see http://auditoryneuroscience.com/index.php/scene-analysis).
The second line of research has focused on auditory mask- ing, an objective measure of the influence one sound (the masker) has on the listener’s ability to detect, discrimi- nate, or recognize another (the target). The early view of auditory masking, dating back to Fletcher (1940), was that it is caused by the overlap of neural excitation pro- duced by the target and masker in the auditory periphery. Bill would publish one of the early studies, indicating that the process is much more complex and possibly con- nected to auditory streaming (Yost et al., 1989).
Figure 4 shows three conditions of that study. The listener’s task was to detect an increase in the base modulation rate of a target tone (Figure 4, right). The target was either presented alone (Figure 4, top), presented with an unmod- ulated masking tone (Figure 4, center), or presented with the masking tone modulated at the same base frequency as the target (Figure 4, bottom). Little masking was expected in the two masking conditions because there was always a two-octave separation between target and masker; indeed, the unmodulated masker had little effect on threshold, consistent with that expectation. The modulated masker, on the other hand, produced a significant, unexpected increase in threshold, suggesting a perceptual interference created by the common modulation. The results are remi- niscent of those from the streaming experiments where common temporal modulations in the frequencies of tones cause those tones to fuse into a single auditory image (Bregman, 1990). Bill’s results on the effects of temporal modulations on masking and those of many other studies
Figure 4. Three conditions adapted from the study by Yost et al. (1989). See text for discussion.
44 Acoustics Today • Spring 2022