The influence of different timbre attributes on the perceptual segregation of complex-tone sequences

Spectral factors such as differences in harmonic content are powerful cues in the perceptual organization of tone sequences. Temporal features such as rise time, however, have been shown to be poor cues [W. M. Hartmann and D. Johnson, Mus. Perc. 9, 155-184 (1991)]. The relative influence of these timbral features on perceptual segregation was investigated. Complex tones were sequenced in a repeating ABA- "gallop" format, under four conditions in which tones A and B had the same or different timbres as defined by differences in numbers of harmonics and temporal-envelope features. A sequence started with A and B tones at the same F0. The F0 difference between A and B then increased over the course of a trial, until a listener terminated the trial indicating perceptual segregation into sub-sequences comprising A and B tones, respectively. The F0 difference required to reach this crossover point of segregation provided a measure of the efficacy of stimulus features of A and B as cues for perceptual organization. Sequences combining differences in harmonic structure and temporal envelope required the smallest F0 change for segregation. Sequences of tones with the same harmonic structure and temporal envelope required larger changes in F0, while the other conditions fell in the middle of this range. The F0-tracking method used in this study facilitates measurement of the relative contribution of different stimulus features to stream segregation. It also holds potential as a tool using the point of segregation as a measure of the magnitude of timbre differences brought about by different physical features of sounds.