M4-ESR: Modelling auditory stream segregation based on a temporal coherence analysis

There is no satisfactory quantitatively model yet to describe the ability of human listeners to perceptually segregate acoustic sources. Inspired by recent physiological findings and a conceptual modelling analysis, we aim in this project to develop a framework that consists of a model of human auditory (pre)processing as a front end and a temporal coherence analysis as the back end which evaluates the coherence across tonotopic channels to group activity across frequency. An analysis of the essential components will help understand the effects of individual hearing impairment and hearing-instrument processing on perceptual grouping and sound source segregation.

Host institution: Danmarks Tekniske Universitet (DTU)
Supervisor names: Torsten Dau and Bastian Epp
Start date: 1.6.2013
PhD enrolment: Yes
Employment contract: Full-time early-stage research fellow, 36 months