C. J. Darwin
Nicholas I. Hill
Lab. of Exptl. Psychol., Univ. of Sussex, Brighton BN1 9QG, UK
This paper reviews the effect of two powerful cues to sound segregation, onset asynchrony and mistuning, in a number of different experimental paradigms. Although both cues substantially influence grouping in all three paradigms, there are quantitative differences across the paradigms. An onset asynchrony of a few tens of milliseconds is sufficient to allow listeners to hear out a single low-numbered harmonic from a complex and for it to be lateralized largely independently of the other harmonics. It is also enough to remove that harmonic from the calculation of vowel quality. But a substantially greater asynchrony (over 100 ms) is needed in order to remove a single (mistuned) harmonic from a complex for the purpose of calculating its pitch. When a single low-numbered harmonic is mistuned, a mistuning of less than 1% is generally sufficient to allow it to be heard out. At about 3% mistuning, the mistuned harmonic can be lateralized largely independently of the other harmonics, but will still make a full contribution to the pitch and the vowel quality of the complex. A substantially reduced contribution both to the vowel quality and to the pitch of the complex requires a larger mistuning---around 8%. [Work supported by UK SERC & MRC.]