|
Dear List, Please find below the announcement for a PhD position that could be available in my lab. Thanks for circulating it among your interested students. cheers, mat ------------------------------------------------------------------------ Ph.D. position in Psychoacoustics, Lyon, France “Individual predictions of speech intelligibility in noisy rooms” Summary In noisy environments, the ability to spatially separate a target voice from competing sources can greatly improve intelligibility. This spatial release from masking is generally
reliant on having two normally functioning ears — hearing loss and current hearing-aid processing can greatly reduce this benefit. There is currently no model able to predict the intelligibility of speech masked by competing speech, nor the associated effects
of hearing loss and hearing aid processing. Models have only been validated for different types of noise maskers, partly because in speech-on-speech situations the sources of masking can be either “energetic” or “informational”. Energetic masking (EM) refers
to a decrease in intelligibility when the target and masker signals overlap in time and frequency reducing target audibility. Informational masking (IM) refers to more central factors limiting intelligibility (e.g., inability to segregate the signals or to
attend the target speech), when the maskers are very similar to the target or otherwise highly distracting. The aim of this PhD is to develop a model able to predict the effects of EM on individual intelligibility for both normal-hearing (NH) and hearing-impaired (HI) listeners
in realistic environments. The contribution of IM will then be quantified as the remaining variations in the data once predicted EM effects have been factored out. This research will be based on a binaural model recently* proposed to predict individual intelligibility
differences associated with EM among HI listeners for speech intelligibility measured in noise, modulated noise, and competing speech. Compare to this earlier model, the aim will be to predict full psychometric functions (percent correct as a function of signal-to-noise
ratio, SNR) rather than being limited to differences in speech reception threshold (SNR for 50% intelligibility). The model will be tested on data measured for NH and HI listeners in recorded real-world environments varying in complexity.
This PhD is expected to support the development of hearing aids by allowing a relevant evaluation of their effects on perception in complex, realistic scenarios. The model will be
developed using MatLab (and made available to the community once validated), and this development could require the design and analysis of controlled listening tests. This PhD could involve an
international mobility, thanks to the Europen project SAV (Marie Curie Research and
Innovation Staff Exchange, 2023-2027). This project can fund up to a year of mobility for the PhD candidate in the lab of one of the international partners of the project: Jörg Buchholz (Macquarie University, Australia) and Virginia Best (Boston University,
USA). *
Lavandier et al. (2021) “Individual differences in speech intelligibility at a cocktail party: A modelling perspective”
J. Acoust. Soc. Am., 150 (2), pp. 1076-1087 Funding:
it is not secured yet. Application for funding has to be made with the selected candidate.
Starting date of Ph.D.: October 1st 2024 Deadline for application:
May 5th 2024 (send application, CV and the coordinates of 2 referees). Contact :
Mathieu Lavandier (mathieu.lavandier@xxxxxxxx,
+33 (0)4 72 04 70 30) Université de Lyon, ENTPE, LTDS, Rue M. Audin, 69518 Vaulx-en-Velin Cedex, FRANCE https://mathieulavandier.wordpress.com/ ----------------------------------------------------------------------------------------------- Mathieu Lavandier https://mathieulavandier.wordpress.com/ https://celya.universite-lyon.fr ENTPE, Laboratoire de Tribologie et Dynamique des Systèmes (LTDS, UMR 5513) Rue M. Audin, 69518 Vaulx-en-Velin Cedex, FRANCE |