About
430 million people worldwide require rehabilitation to address hearing loss (World Health Organization). Yet even in developed countries, only 40% of people who could benefit from hearing aids have them and use them often enough. People believe that hearing aids perform poorly, with the difficulties of hearing speech-in-noise being a key issue.
Speech enhancement is a major research area with thousands of papers each year, yet only a tiny percentage of these explicitly consider improvements for listeners who have a hearing loss. Consequently, this signal processing challenge is designed to get the latest advancements in speech enhancement applied to hearing aids.
The Challenge
Entrants will be tasked to enhance speech-in-noise for input into our baseline hearing aid processor. The hearing aid will be tuned to the hearing characteristics of particular people. Thus you can enter without in-depth knowledge of hearing aids, and just concentrate on the task of de-noising. The scenario is listening to speech in the presence of typical domestic noise. We will provide the signals captured by the microphones on a pair of behind-the-ear hearing aids and those captured at the eardrum. The target speech will be a short sentence. The interfering noises will be a mix of speech, domestic appliance noise and music. The audio includes the simulation of the acoustic of typical small living rooms.
The challenge is to improve the speech intelligibility without excessive loss of quality. To this end, entries will be evaluated using an objective metric that is an average of the Hearing Aid Speech Perception Index (HASPI) and Hearing Aid Speech Quality Index (HASQI).We will provide
- Premixed speech + interferer scenes for training and evaluation.
- Listener characteristics, including audiograms and speech-in-noise testing.
- Software tools for evaluating enhanced signals including a hearing aid amplifier and HASPI/HASQI hearing aid performance metrics.
- Databases of target sentences, along with speech, noise and music interferers, plus tools for training data augmentation
For full details of how to register and take part visit the Clarity Challenge site.