ICASSP 2026 Cadenza Challenge

dataset·lyric intelligibility prediction·closed

A completed challenge to predict lyric intelligibility for popular music, including listeners with hearing loss.

Recorded facts

Official site	https://cadenzachallenge.org/docs/clip1/results ↗
Geography	United Kingdom
size	source music signals: 11072
scope	Prediction of lyric intelligibility from song excerpts under listener and hearing-condition variation.
genres	popular Western music
license	dataset/challenge-specific
steward	Cadenza project
creators	Cadenza Challenge organizers
languages	English singing
modalities	music audio; lyrics; listener scores; challenge submissions
geographies	Global
access method	Challenge site and dataset paper.
intended uses	lyric intelligibility prediction; music accessibility
consent claims	Inherited from CLIP listener study and source tracks.
related papers	The Cadenza lyric intelligibility prediction dataset
annotation method	Human transcription and listening tests provide target intelligibility scores.
collection method	Uses the Cadenza Lyric Intelligibility Prediction dataset.
provenance claims	Official results page states the ICASSP 2026 challenge is complete.
current availability	Challenge complete; results and documentation online.
commercial use limits	Per-track FMA rights and challenge terms apply.
us market scope basis	Publicly available or materially used in US-facing music/audio-AI research.
limitations biases disputes	Restricted to English singing and popular Western music.