ICASSP Grand Challenge: Lyric Intelligibility Prediction
Trevor Cox
This ICASSP challenge is now live and the data ready to download at https://cadenzachallenge.org/
The Cadenza project is working to improve music for those with hearing loss. Hearing loss causes problems when listening to music including making picking out lyrics more difficult.
In speech technology, having metrics to automatically evaluate intelligibility has driven improvements in speech enhancement. We want to do the same for music. There has been little research into Lyric Intelligibility Prediction, so there is lots of scope to learn from Speech Intelligibility Prediction and do novel research. There are differences between speech and sung lyrics, which we hope will make this an interesting problem to tackle.
The top 5 ranked teams will be invited to submit papers to ICASSP-2026, Barcelona, Spain.
We will challenge entrants to develop systems that predict lyric intelligibility from song excerpts of popular Western music. The systems will take stereo audio as input and estimate the word correct rate a listener is likely to achieve in a perceptual test.
What will be provided:
- A new dataset with thousands of paired song excerpts with intelligibility scores for training and evaluation.
- Song excerpts will be presented either unprocessed or with mild or moderate hearing loss simulated.
- Software baseline.
- A leaderboard
Important Dates:
- 1st Sept 2025: Challenge launch.
- 3rd Nov 2025: Release of evaluation data.
- 17th Nov 2025: Teams submit intelligibility predictions and technical reports.
- 1st Dec 2025: Results released. Top 5 ranked teams invited to submit papers to ICASSP-2026.
- 7th Dec 2025: Invited papers submitted to ICASSP-2026.
- 4-8th May 2026: Overview paper and invited papers presented at ICASSP-2026 special session.