Round 1: 2021/2
To allow the development of better hearing aids, we need ways to evaluate the speech intelligibility of audio signals automatically. In other words, we need a prediction model that takes a listener’s characteristics and the audio transmitted by their hearing aids and estimates the intelligibility score that the real listener achieved in listening tests.
For the prediction challenge we will provide the following:
- Audio produced by a variety of (simulated) hearing aids for speech-in-noise.
- The corresponding “clean” reference signals.
- Characteristics of the listeners (pure tone audiogram, etc.).
- The WERs and measured speech intelligibility scores from the listening tests, where the listener was asked to say what they heard in the speech-in-noise.
- Challenge launch, 9th November 2021
- Final deadline, April 2022
- Workshop, May/June 2022
There are cash prizes for the systems that top our evaluation.