Factors Influencing Stream Segregation Based on Interaural Phase Difference Cues.

Abstract

Interaural time differences are often considered a weak cue for stream segregation. We investigated this claim with headphone-presented pure tones differing in a related form of interaural configuration-interaural phase differences (ΔIPD)-or/and in frequency (ΔF). In experiment 1, sequences comprised 5 × ABA- repetitions (A and B = 80-ms tones, "-" = 160-ms silence), and listeners reported whether integration or segregation was heard. Envelope shape was varied but remained constant across all tones within a trial. Envelopes were either quasi-trapezoidal or had a fast attack and slow release (FA-SR) or vice versa (SA-FR). The FA-SR envelope caused more segregation than SA-FR in a task where only ΔIPD cues were present, but not in a corresponding ΔF-only task. In experiment 2, interstimulus interval (ISI) was varied (0-60 ms) between FA-SR tones. ΔF-based segregation decreased with increasing ISI, whereas ΔIPD-based segregation increased. This suggests that binaural temporal integration may limit segregation at short ISIs. In another task, ΔF and ΔIPD cues were presented alone or in combination. Here, ΔIPD-based segregation was greatly reduced, suggesting ΔIPD-based segregation is highly sensitive to experimental context. Experiments 1-2 demonstrate that ΔIPD can promote segregation in optimized stimuli/tasks. Experiment 3 employed a task requiring integration for good performance. Listeners detected a delay on the final four B tones of an 8 × ABA- sequence. Although performance worsened with increasing ΔF, increasing ΔIPD had only a marginal impact. This suggests that, even in stimuli optimized for ΔIPD-based segregation, listeners remained mostly able to disregard ΔIPD when segregation was detrimental to performance.