Temporal streaming based on fine structure cues

Abstract

Alternating sequence of two A and B bursts of sound tends to be perceptually organized in different auditory streams when introducing either a spectral or a temporal difference between A and B. In general, the channeling theory of streaming predicts that any salient difference between the excitation patterns evoked by A and B sounds would lead to a segregated percept. Some previous studies have, however, evidenced that a sequence of sounds with similar spectral properties but with different temporal properties can be heard as segregated. In particular, temporal cues are probably responsible for the segregated percept when hearing a sequence of bursts of noises that are amplitude-modulated at different rates. All previous streaming experiments involved stimuli that had frequency components below 5000 Hz. As a consequence, the individual contribution of temporal fine structure cues and envelope cues could not be dissociated and remains largely undetermined. The current experiment is dedicated to test further the relative importance of both envelope and fine structure cues to segregate sequences of unresolved complex tones with different fundamental frequencies. Two high-frequency regions above and below 5000 Hz and several phase relationships leading to several temporal peak factors have been used in a subjective streaming task.