PDMX

dataset·public domain musicxml·active

Large MusicXML collection of scores labeled public domain, with score and user-interaction metadata.

Recorded facts

Official site	https://pnlong.github.io/PDMX.demo/ ↗
Geography	United States
doi	10.5281/zenodo.14648209
arxiv	2409.10831
github repository	pnlong/PDMX
size	scores: 254077 · deposit gb: 14.1 · estimated hours: 6250
scope	Copyright-conscious symbolic music processing and generation.
genres	multiple; strong classical/public-domain representation
license	CC0 / public-domain claim
steward	UC San Diego research authors
creators	Phillip Long; Zachary Novack; Taylor Berg-Kirkpatrick; Julian McAuley
languages	multilingual score metadata
modalities	MusicXML; PDF; CSV metadata; tags; user interactions
geographies	Global
access method	Zenodo archives, GitHub code and Hugging Face links.
intended uses	symbolic music generation; score analysis; data-quality research
related papers	PDMX: A Large-Scale Public Domain MusicXML Dataset for Symbolic Music Processing
annotation method	Platform metadata and interaction signals retained; subsets created for quality analysis.
collection method	Scores collected from MuseScore and filtered for public-domain status.
provenance claims	Authors describe the selected scores as public domain and the release as copyright-free.
current availability	Zenodo record and repository active.
commercial use limits	No dataset-level commercial restriction claimed.
us market scope basis	Publicly available or materially used in US-facing music/audio-AI research.
limitations biases disputes	Public-domain classification depends on metadata and jurisdiction.; MuseScore user metadata may be incomplete or incorrect.
redistribution restrictions	No restriction claimed for selected public-domain scores; source-platform terms require review.