Predicting individual differences in speech-on-speech intelligibility for listeners with hearing loss - Forum Acusticum 2020 Accéder directement au contenu
Communication Dans Un Congrès Année : 2020

Predicting individual differences in speech-on-speech intelligibility for listeners with hearing loss

Résumé

The model of Vicente et al. (2019) [Proceedings of the 23rd International Congress on Acoustics, 9-13 September 2019, Aachen, Germany] was revised to describe individual differences in hearing-impaired (HI) listeners for the intelligibility of speech masked by competing speech. In order to account for severe hearing loss, a floor-SNR value was added to the model and estimated using speech reception thresholds (SRTs) measured in the presence of stationary and modulated noises. The revised model also relies on the average SRT across listeners as input, so it cannot make a priori predictions. On the other hand, it can be used to make a posteriori estimations of the amount of energetic masking that was present and to account for the differences across listeners in a given condition. This was demonstrated using the SRTs from Best et al. (2017) [J. Acoust. Soc. Am., 141 (1), pp. 81-91]. In that study, the target speech was presented (using HRTFs) in front of the listener in the presence of two speech maskers at +/- 90 degrees or four speech maskers at +/- 90/45 degrees. SRTs were measured for 10 normal-hearing (NH) and 7 HI listeners. In each masker configuration, the model was able to account for the variation in HI SRTs and the best NH SRT based solely on differences in audibility. The model could not describe the observed variability among NH listeners, which is unrelated to audibility and may reflect informational masking.
Fichier non déposé

Dates et versions

hal-03234179 , version 1 (25-05-2021)

Identifiants

Citer

Mathieu Lavandier, Virginia Best, Christine Mason, Gerald Kidd. Predicting individual differences in speech-on-speech intelligibility for listeners with hearing loss. Forum Acusticum, Dec 2020, Lyon, France. pp.3491-3491, ⟨10.48465/fa.2020.0101⟩. ⟨hal-03234179⟩

Collections

ENTPE FA2020 UDL
30 Consultations
0 Téléchargements

Altmetric

Partager

Gmail Facebook X LinkedIn More