Improving speech intelligibility

Speech playback becomes less intelligible in noisy environments: just think of how hard it usually is to understand a public announcement in a big and crowded space, such as a train station or an airport.
Speech pre-enhancement, also known as NELE (Near End Listening Enhancement) is a technology aimed at modifying the speech signal before playback - i.e. before it becomes mixed with noise - in order to improve intelligibility for the listener.



Some pre-enhancement algorithms have proven to work very well in additive noise for normal hearing listeners. However, there is still lack of combined approaches that tackle both additive noise and reverberation, and there is still little research on how these methods can potentially benefit the users of hearing prosthetics.
The goal of the ENRICH network is to modify or augment speech with additional information to make it easier to process. The University of Edinburgh's CSTR (Centre for Speech Technology Research) is taking part in this initiative, hosting 2 PhD projects.
Read on to the website if you'd like to know more!


This project has received funding from the EU's H2020 research and innovation programme under the MSCA GA 675324