Charpentier Sacred Music Dataset (CSMD): An Annotated Corpus for AI Music Research
收藏资源简介:
The Charpentier Sacred Music Dataset (CSMD) is a curated collection of 52 symbolic scores in MusicXML format covering a broad range of Marc-Antoine Charpentier's (1643–1704) sacred output: Masses, Magnificats, Grands Motets,Petits Motets, Noëls, Hymns, Psalm settings, and Dialogues. Each score is accompanied by structured JSON annotations at four musicological levels: (1)phrase structure with cadence-type labels (PAC, IAC, Phrygian, Plagal, HC); (2) harmonic progressions as Roman numeral sequences per beat; (3) melodicfeatures of the soprano/dessus voice (scale degree, interval direction); (4)voice ranges per part. The dataset includes multi-format exports (MIDI, ABCnotation via xml2abc) and voice-reduction augmentations (1-, 2-, 3-, and4-voice variants). A Jupyter notebook provides complete analysis with datavisualizations: key distribution, voice ranges, interval histograms, chordtransition heatmaps, and phrase-length statistics.



