Characterizing consequences of hearing impairment and hearing-aid processing on speech perception in competing-talker scenarios

Paolo Attilio Mesiano

Speech comprehension is crucial in daily life as it is the means for exchanging information and the basis of social interactions. The difficulty of recognizing a target speech signal depends on the acoustical condition. A very challenging every-day situation is a competing-talker scenario, where normal-hearing (NH) listeners exhibit exceptional abilities in understanding speech, while hearing-impaired (HI) listeners often experience great difficulties. Currently, the perceptual mechanisms that facilitate these abilities in NH listeners and limit the performance in HI listeners are not entirely understood.

The goal of this project is to extend our understanding of the processing mechanisms responsible for speech perception in complex acoustical scenarios and to evaluate the benefit that different hearing-aid compensation strategies provide for this task. In a first phase, speech perception in NH listeners will be characterized with psychoacoustic experiments and state-of-the-art computational speech intelligibility models will be extended to simulate the experimental results. In a second phase, HI listeners’ performance will be experimentally characterized both with and without hearing aids and the computational models will be adapted to account for their results. The outcome is expected to provide insights regarding the hearing-aid signal processing approach that can best help HI listeners in competing-talker scenarios.

The project will be carried out in collaboration with Eriksholm Research Center / Oticon A/S.

Supervisors:
Torsten Dau (DTU), Johannes Zaar (DTU), Niels Henrik Pontoppidan (Eriksholm Research Centre)

To be completed: June 2021
DTU Orbit

Contact

Paolo Attilio Mesiano
PhD student
DTU Health Tech