Êtes-vous un étudiant de l'EPFL à la recherche d'un projet de semestre?
Travaillez avec nous sur des projets en science des données et en visualisation, et déployez votre projet sous forme d'application sur Graph Search.
Enhancing speech intelligibility for hearing-impaired subjects in complex acoustic conditions is still a challenging topic of research. To mitigate the detrimental effects of background noise and reverberation, current hearing instruments incorporate various hardware and software strategies, among which speech enhancement algorithms are of primary importance. In this paper, two algorithms based on the multichannel Wiener filter (previously reported in the literature) and one proprietary algorithm are experimentally assessed and compared. All of them make use of a remote microphone worn by the speaker of interest. The objective of these algorithms is to improve the speech contribution within the hearing aid microphone signals. The algorithms are assessed in terms of interference reduction performance, speech quality, spatial hearing preservation, and technical requirements. Using a recorded database of audio signals, the effects of the signal-to-noise ratio and of the delay between the remote and hearing aid microphone signals are studied. The results show that the proprietary algorithm provides a good performance and yields the lowest distortion of the binaural localization cues, while being the most efficient in terms of computational cost and wireless usage. The main drawback is the degradation of the output sound quality that is observed when the remote and hearing aid microphone signals are not temporally aligned.
Vincent Pierre Olivier Grimaldi
Hervé Lissek, Gilles André Courtois, Vincent Pierre Olivier Grimaldi
David Atienza Alonso, Giulio Masinelli, Adriana Arza Valdes, Fabio Isidoro Tiberio Dell'Agnola