Greater Manchester, March 3
Author: Simone Graetzer, Research Fellow in Speech Signal Processing for Hearing Devices at Acoustics Research Centre, University of Salford.
The Clarity project is running a series of machine learning challenges to revolutionise signal processing in hearing aids. The first enhancement challenge has just been launched. For more information go to claritychallenge.org.
In this first round, we’re challenging entrants to improve speech heard in living rooms against background noise. Entrants are given the task of producing signal processing that enhances the intelligibility of speech.
We are providing open-access datasets, models and code to help entrants take part. These include
- Tools for generating realistic audio examples for different listening scenarios.
- A model of hearing loss.
- A hearing aid model for entrants to improve on.
- A model for estimating speech intelligibility.
- Databases of speech and speech perception in noise.
These will allow entrants to develop their own algorithms for speech and hearing aid processing, even if they haven’t considered hearing impairment in previous research.
This is a first of a series of enhancement challenges. Later we will also run prediction challenges to develop and improve methods for predicting speech intelligibility and quality.
This EPSRC-funded project involves researchers from the Universities of Sheffield, Salford, Nottingham and Cardiff in conjunction with the Hearing Industry Research Consortium, RNID, Amazon, and Honda.
1st February 2021
- Release of documentation and description of baseline system.
- Example dataset published.
15th February 2021
- Full dataset
- Scene generation code
- Baseline system.
1st May 2021
Evaluation data released.
1st June 2021
Download the data via GitHub