[AUDITORY] Cadenza Challenge pre-announcement: Signal processing challenge for music and listeners with a hearing loss

[Trevor Cox <0000017832c3f089-dmarc-request@xxxxxxxxxxxxxxx>]

Background

http://cadenzachallenge.org/ is organising signal processing challenges for music and listeners with a hearing loss. Interested? Please join our Google Group https://groups.google.com/g/cadenza-challenge

 

Although hearing loss is an almost inevitable part of aging, the majority of adults who would benefit from hearing aids don’t use them. The purpose of the Cadenza challenges is to catalyze new work to radically improve the processing of music for those with a hearing loss. Even if you’ve not worked on hearing loss before, we’ll provide you with the tools to enable you to apply your machine learning and music processing and demixing/mixing algorithms to get going.

 

The first round with a full launch in March 2023 will focus on two scenarios:

• Track 1, listening over headphones.
• Track 2, listening in a car in the presence of noise.

Your task is to improve the perceived audio quality of the reproduction considering the listeners' hearing loss. This is about intelligent demixing/remixing of the tracks or processing the mixed music to compensate for the hearing loss of the listener. You might be doing this to make the lyrics clearer, correct the frequency balance, ensure the music has the intended emotional impact, etc.

 

The 2023 Challenge

 

Track 1: Headphones and demixing

You will be tasked with improving the audio quality of music samples for listeners with defined hearing losses. The listeners are not using their normal hearing aids, they are just listening to the signals you provide via the headphones. One machine learning challenge here is to demix stereo music using an evaluation metric that allows for hearing loss, to then allow an intelligent remix for the listener.

 

Track 2: Car

You need to enhance music samples played by the car stereo in the presence of noise from the engine/road. This will be for listeners with a defined hearing loss, who have a fixed hearing aid that we will provide.

 

Evaluation

Entries will be objectively evaluated using Hearing Aid Audio Quality Index (HAAQI). The best systems will go forward to be scored by our listening panel of people with a hearing loss.

 

What you will get

• Databases of music and scenes for training and evaluation.
• Listener characteristics, including audiograms.
• A complete end-to-end software baseline to build upon from sample selection to evaluation.
• Tutorials to learn about hearing loss, hearing aids, and our software.

Draft key dates

• 1st Feb 2023: Beta launch of challenge with software and datasets.
• 1st March 2023: Full launch of challenge with software and datasets.
• June 2023: Release of evaluation data.
• July 2023: Competition closed. All entrants submit (i) audio for evaluation and (ii) a draft of their technical report.
• Aug 2023: Entrants informed which systems are going forward to the listening test evaluation stage.
• Sept 2023: Submit two page technical reports to Cadenza-2023 workshop.
• Autumn 2023: Cadenza-2023 workshop.

The Team

• Trevor Cox, Professor of Acoustic Engineering, University of Salford
• Alinka Greasley, Associate Professor in Music Psychology, Leeds University
• Michael Akeroyd, Professor of Hearing Sciences, University of Nottingham
• Jon Barker, Professor in Computer Science, University of Sheffield
• William Whitmer, Senior Investigator Scientist, University of Nottingham
• Bruno Fazenda, Reader in Acoustics, University of Salford
• Simone Graetzer, Research Fellow, University of Salford
• Rebecca Vos, Research Fellow, University of Salford
• Jennifer Firth, Research Assistant in Hearing Sciences, University of Nottingham

Funders

• Cadenza is funded by EPSRC. Project partners are RNID; BBC R&D; Carl von Ossietzky University Oldenburg; Google; Logitech UK Ltd and Sonova AG.

 

Prof @trevor_cox

Acoustical Engineering, University of Salford

+44 161 295 5474; +44 7986 557419