Datasets

Popular-music signals with lyrics and listener-derived intelligibility scores for the ICASSP 2026 Cadenza Challenge.

A test-only benchmark converting diverse MIR tasks into standardized music instruction-following evaluation.

Preference data and evaluation for reward models scoring musicality, text alignment and compositional instruction following.

Synthetic four-part chamber-ensemble music with mixtures, stems, MIDI and fine-grained performance annotations.

Pozalabs symbolic dataset of 11,144 professional-composer MIDI samples with 12 metadata fields.

Large predominantly Brazilian music database for cover-song and live-song identification, with public metadata and features.

Controlled singing-voice deepfake corpus generated from public singing data with synthesis and voice-conversion systems.

Full-song lyric, vocal-note and audio alignments at note, word, line and paragraph granularity.

Collection of harmonically and formally annotated symbolic music corpora.

Music excerpts and full songs with continuous and song-level valence and arousal annotations.

Synchronized egocentric audiovisual dataset for music understanding and hearing-enhancement research.

Pop-piano audio and MIDI clips labeled for perceived emotion, supporting recognition and emotion-conditioned generation.

Curated and annotated corpus of traditional Georgian vocal music for computational musicology and MIR tasks.

Expanded human drum-performance corpus with MIDI-aligned audio rendered across 43 drum kits.

MusicCaps prompts regenerated with multiple text-to-music systems for synthetic-music detection and generator attribution.

Full-length Creative Commons music with audio, features, metadata and hierarchical genre labels for MIR.