Tempo perception: The multiple-look strategy is limited by a short temporal window

Abstract

Previous studies [Drake and Botte (1993)] indicated that sensitivity to tempo changes in isochronous sequences improves as the number of intervals in the sequence increases. Further studies suggested that this multiple-look process only works within a limited temporal window of 0.5 or 2 s depending on the tempo of the sequence [Panissal-Vieu et al. (1995)]. This temporal window represents the time required to extract all the necessary temporal information (mean interval duration and dispersion). Two experiments investigated how this mental representation is used to make temporal judgements. In a 2 AFC paradigm, trained listeners identified the slower of two isochronous sequences. The number of intervals was varied in either the first or second sequence, with the temporal window saturated in the other sequence. The number of intervals required to saturate the temporal window of the first sequence varied with tempo, as predicted. However, only three intervals were required in the second sequence. These results suggest a buildup of a stable and accurate memory trace during the first sequence which is subsequently compared with the first two intervals of the second sequence. There is no need to recalculate the memory trace for the second sequence.