Loading available alignment metrics…
Part-aware choral audio–MIDI dataset
PawChorale
Mastered choir recordings, isolated vocal parts, and aligned per-part MIDI labels in one research-ready collection.
- Works
- 200
- Master audio
- —
- Isolated parts
- —
- Vocal MIDI notes
- —
Download
Get PawChorale.
Download all 200 curated works at once from the permanent Zenodo record. Four smaller GitHub archives remain available as mirrors.
Complete dataset · one ZIP
4.45 GiB · 200 works · MP3 · part + mixture MIDI · MusicXML · manifests
Checking release status…
GitHub mirror archives
4 partsInteractive demo
Hear and inspect the alignment.
Choose any of the 200 works, play its master and available isolated parts in sync, and inspect the aligned per-part MIDI piano roll.
Loading…
Preparing catalog
- Duration
- —
- Voicing
- —
- MIDI notes
- —
- Matched notes
- —
Audible tracks
Toggle the mixture and parts during synchronized playback.Source metadata
- Public work ID
- —
- Source score
- —
- CPDL edition
- —
- Source license
- —
- Conversion
- MusicXML → per-part MIDI
Rendering configuration
- Audio
- —
- Layout
- —
- Time reference
- Seconds from recording start
Selected-work downloads
Work files
Dataset contents
One work, one folder.
Every work has a stable public ID. Available voicing is preserved as recorded; complete SATB coverage is not required.
Master mixtures
One mastered choir recording for each retained work.
Isolated voices
Available Soprano, Alto, Tenor, and Bass recordings.
MIDI + MusicXML
Per-part and mixture notes, original scores, source records, and rendering manifests.
001/ · 001_master.mp3 · 001_Soprano.mp3 · 01_Soprano.mid · choral_mixture.midi · choral_music_score.xml
Work catalog
Browse all 200 works.
| ID | Work | Audio parts | MIDI files | Archive |
|---|---|---|---|---|
| Loading catalog… | ||||
Citation
Using PawChorale?
The author list can be updated with the paper release. The current dataset citation is:
Meng, H. (2026). PawChorale: A Part-Aware Choral Audio–MIDI Dataset (Version 1.0.0). Zenodo. https://doi.org/10.5281/zenodo.21777039