Timing in auditory perception
Ira J. Hirsh · The Journal of the Acoustical Society of America · 1987
Recognition of sequential patterns of sound is basic to speech recognition, music perception, and identification of environmental sounds from birdsong to thunderclaps. While we define these sound patterns as sequences of brief sounds, the usual acoustical description has principally been the spectrum. For some 30 years, work has been reported on fusion and successivity, duration discrimination, perceived order, and the influence of temporal variables on more traditional psychoacoustic dimensions. At Central Institute, with colleagues Divenyi, Lauter, Grant, Singh, and Monahan, our emphasis has been on listeners' sensitivity to changes in order and timing, and influence of nontemporal features on such perception. Here, some experiments are described that address interval perception in a series. In an isochronous series of three, six, or ten tones with intertone intervals of 200 ms or more, listeners can detect a temporal offset or “jitter” of between 5% and 10%, irrespective of the position of offset. For shorter intervals, performance is worse and there is a strong effect of position in the series, with offsets made late in the pattern being more easily discriminated than those made earlier. When a single pitch change is made in one tone in a series of six, detection of an offset is impaired when it coincides with the different pitch. Phrased patterns (do-re-mi-do-re-mi) on the other hand, have not shown such a strong position effect. It is recognized that temporal patterns influence, and are influenced by stimulus dimensions as well as by dimensions learned from formal systems like the rules of language and music. Our search continues to focus on an in-between perceptual territory, which lies beyond sensory processes, but yet is less dependent on rule learning, [Work supported by NINCDS and AFOSR.]