Sound predictability as a higher-order cue in auditory scene analysis

Bendixen, Alexandra and Denham, Susan L. and Winkler, István (2013) Sound predictability as a higher-order cue in auditory scene analysis. In: AIA-DAGA 2013 Conference on Acoustics, 18-21 March 2013, Merano, Italy.

Preview

Text Bendixen_Denham_Winkler_2013_DAGA.pdf Download (416kB) | Preview

Abstract

A major challenge for the auditory system is to disentangle signals emitted by two or more sound sources that are active in a temporally interleaved manner (sequential stream segregation). Besides distinct characteristics of the individual signals (e.g., their timbre, location, and pitch), one important cue for distinguishing the sound sources is how their emitted signals unfold over time. It seems intuitively plausible that signals that unfold predictably with respect to their acoustic features and time-points of occurrence, such as the repetitive signature of a train moving on the rails, can be more readily identified as originating from one sound source. Based on this rationale, predictive elements have successfully been incorporated into computational models of auditory scene analysis for many years.

Actions (login required)

Edit Item