Auditory Isochrony: Time Shrinking and Temporal Patterns

Abstract

It has previously been reported that the duration of short time intervals is conspicuously underestimated if they are preceded by shorter neighbouring time intervals. This illusion was called ‘time shrinking’ and it was argued that it strongly affects the perception of auditory rhythms. In the present study this supposition has been pursued in three experiments. In the first, temporal patterns consisting of two, three, and four intervals had to be judged for anisochrony, which was invoked by offsetting the last sound from its isochronous position. By a constant method, it was determined that the last sound of fast sequences (50 ms base interval) had to be delayed by about 30 ms in order for isochronous rhythms to be perceived. Another interesting finding was that for sound sequences with base intervals up to 200 ms it was the difference limen, rather than Weber's ratio, that was constant for anisochrony detection. In the second experiment, the temporal patterns comprised two intervals, presented serially or separately. The deviation of isochrony could be on either the first or the second interval. The data, gathered by an adaptive method, showed time shrinking to be effective even up to a base interval of 200 ms. The third experiment involved a constant method and anisochrony was implemented on the first interval of two interval patterns. Time shrinking affected perceived isochrony in sequences with base intervals of 50, 100, and 200 ms. It is argued that the paradoxical results of anisochrony detection can be explained in terms of time shrinking. Some anomalies of rhythm perception and production that are the result of time shrinking are discussed. Finally, a tentative setup for a model of anisochrony detection that defies Weber's law is offered.