Two positions are available as part of the FP7 Marie Curie Initial Training Network Investigating Speech Processing In Realistic Environments (INSPIRE). This network provides research opportunities for 13 PhD students and 3 postdocs. As part of this program, you will become a member of an international team of researchers whose aim is to gain a better understanding of how listeners recognize speech, even under non-ideal circumstances. You will contribute to urgently needed solutions that help alleviate the serious communication problems that arise, especially for older and hearing-impaired persons, when different combinations of ’adverse’ conditions affect the speech processing system.
Both projects with be supervised by Prof. Torsten Dau and carried out at the Centre for Applied Hearing Research (CAHR) at the Technical University of Denmark (DTU). At the centre, we conduct fundamental and applied research with a focus on human speech communication, auditory processing and perception, hearing impairment and hearing instruments. For more information about CAHR, please visit our website.
Position 1: Modelling speech intelligibility based on the signal-to-noise ratio in the modulation domain
Proposed starting date: August 1, 2012
Closing date for applications: April 15, 2012
Topic: As a PhD student you will investigate the modelling of speech intelligibility based on the signal-to-noise ratio in the modulation domain. A major challenge is to solve the “noise reduction paradox” that refers to the mismatch between predicted and actual speech intelligibility following noise reduction signal processing. It is postulated here that the ratio of the speech-to-noise energy in the modulation domain at the output of auditory signal processing is an important indicator of speech intelligibility, in contrast to the STI that considers only effects on the (speech) signal. Various distortions in transmission channels and effects of hearing-instrument processing are analyzed in this framework. The work is relevant for the evaluation of hearing-instrument algorithms, but also in context- and listener-specific speech modification to improve communication with automatic dialogue systems for groups such as the elderly or non-natives in realistic conditions.
This position will also be co-supervised by Dr. Mark Huckvale (Department of Speech, Hearing and Phonetic Sciences, University College London). As part of the project, you will also spend a few months at UCL.
Position 2: Characterization and prediction of perceptual consequences of individual hearing loss
Proposed starting date: September 1, 2012
Closing date for applications: May 15, 2012
Topic: There are enormous differences in hearing performance across individual listeners. To choose the right compensation strategy one needs to understand what the sources of this variability are. In this project, you will characterize individuals’ hearing loss based on estimates of sensitivity, amplitude compression, spectral and temporal resolution, cognitive capabilities, and their relation to speech intelligibility performance. A computational auditory model will be used to predict the consequences of individual deficits and to account for the variability in the data across listeners. The results are expected to have implications for advanced compensation strategies in hearing instrument technology.
This position will also be co-supervised by Dr Bert Cranen () , Centre for Language and Speech Technology (http://www.let.ru.nl/), Radboud University, The Netherlands. As part of the project, you will also spend a few months at Radboud University.
To be eligible for either position you must meet the FP7 Marie Curie requirements for ITNs:
- You must not have resided or performed your main research activity in Denmark for more than 12 months in the last three years
- You must be willing to work in at least one other country in the INSPIRE network
- You must have fewer than 4 years of research experience since you obtained your Masters degree, and not hold a PhD
For further details regarding the positions and how to apply, please see the attached pdf1 and pdf2 or email Prof. Torsten Dau ().