All you physiologists ... ("Albert Bregman, Tel: 514-398-6103" )


Subject: All you physiologists ...
From:    "Albert Bregman, Tel: 514-398-6103"  <IN09(at)MUSICB.MCGILL.CA>
Date:    Wed, 24 Jun 1992 22:55:24 EDT

Dear Colleagues: This query is directed to those of you who are knowledgable concerning the physiology of audition. We recently completed a series of studies. Excerpts from our write-up are quoted below. We end up by saying that we don't know of any physiological studies that show how the phasic response of auditory neurons is affected by the RATE of change in intensity of the signal (either as an increase or a decrease). Can anybody help us? We also welcome any other types of comments. - Al Bregman ================================================================= Excerpts follow: The experiments that we shall report were an attempt to clarify a phenomenon that we observed in the laboratory. We were experimenting with some software that could reverse the order of samples in a digital sound file, a process analogous to playing a tape backwards in a tape recorder. When a tone with a rapid rise and a slow decay was played forward, it sounded like a clear plucked-instrument tone. When played backward, however, it sounded less definite. Even more interesting was the observation that when a cluster of sudden-onset tones, highly overlapped in time, were played in the forward order, they sounded distinct and their order was clearly heard. However when the file was played from back to front, the tones blended together and the temporal order was unclear. Apparently this was due to the slow onsets of the tones in the reverse direction. We decided to design formal experiments to examine the role of abruptness of onset and offset in these phenomena. The findings concerning onset abruptness can be summarized as follows: When clusters of pure tones of different frequencies were highly overlapped in time, the ability to judge the order of onset of the components was strongly affected by the suddenness of their onsets (Experiment 1). As the onsets became more abrupt, the ability to discriminate the order became more acute up to and including the second fastest onset velocity used. At the fastest one, there was a small decline in performance. However, this decline could have been due to some disruptive effect caused by the most abrupt onsets, and tended to disappear with practice. Alternatively, it could have been a statistical artifact. When we ourselves listened to the clusters, the pitches of the more abrupt onset tones seemed more distinct than those of tones with slower onsets. When the onsets were synchronous but the offsets asynchronous (Experiment 2), abruptness of change was again important. More abrupt offsets made it easier to discriminate the order in which the different pitches dropped out. However, the ability to detect the order of offset was much worse than for the order of onset. When we listened to the asynchronous offset signals, the sudden offsets sounded like "plops" colored by the pitch of the component that was dropping out. Asking subjects to judge the quality and locations of pitches in the stimuli gave results that tended to confirm our own experiences. The rated "number of distinct sounds" declined as the onsets or offsets became less abrupt. More distinct pitches were heard at the onsets than at the offsets. The rated "clarity" of pitches also declined smoothly as the onsets became more gradual. The clarity of onsets was greater than those of offsets. The mean differences between conditions were less than for the ratings of "number of distinct sounds." This may mean that there was some perceptual process other than the one that asseses the pitches that was facilitated. This process, revealing itself in the judgement of "number of distinct pitches" may have played a role in the discrimination of the order of onsets and offsets in the earlier experiments. The judgements of "position of clear pitches" also showed effects of the abruptness of change. Judgements of "clear pitches at the beginning" or "clear pitches at the end" were strongly associated with sudden onset conditions while the response, "no clear pitches", was given primarily for the gradual onsets or offsets. It is interesting to recall that all the components were within a critical band. When they were not asynchronous, it was difficult to hear three different pitches in the mixture. Only when a component changed in intensity did its pitch become evident. This suggests that sudden amplitude changes cause a new pitch analysis to be undertaken. The components that have suddenly changed will dominate this analysis. Presumably the other pure tones do not interfere with the computation of the new pitch, despite being in the same critical band, because the auditory system does not respond to them as strongly as to the rapidly changing frequencies. The fact that offsets also benefit from the suddenness of the change suggests that Kubovy's name "onset-segregation effect" may be too specific. Perhaps it is a "sudden-change" segregation effect. In natural sounds, sudden onsets are more usual, being caused by sudden impacts, as in the piano, or by sudden releases, as in the guitar. Sudden drops in intensity would have to be caused by a sudden stopping of a vibration, as at the end of the word "hip". Although we distinguish sudden-onset sounds by a name, "percussive", we have no comparable name for sudden-offset sounds. The fact that sudden offsets do not give as clear pitches as sudden onsets may be due to the fact that by the time the auditory system is alerted to the sudden change, the energy has greatly diminished and there is less left to analyze. The well-known fact that many neurons in the auditory system have a phasic response, responding more strongly to onsets than to continuations of a sound could be the basis of the onset effect that we observed. Presumably, the phasic response is tied to the rate of onset of the sound. We are not, however, aware of any physiological observations that show this. The offset effect is also mysterious physiologically.


This message came from the mail archive
http://www.auditory.org/postings/1992/
maintained by:
DAn Ellis <dpwe@ee.columbia.edu>
Electrical Engineering Dept., Columbia University