Music Score Arrangements with Audio Alignment Metadata

Operation Legato contains 22,060 tokenized music score arrangements sourced from MuseScore. The dataset was created by user hidude562 and was last updated on the platform in April 2026. Each record includes arrangement metadata and alignment information with original audio references.

Use Cases

• Train sequence models for music score generation using the tokenized music score text format.
• Develop audio-to-score alignment algorithms using the provided score and original audio alignment information.
• Classify music by genre or instrument using the arrangement metadata fields like instrument(s) and genre.
• Analyze arrangement complexity levels for educational or musicological research.

Strengths

• 22,060 distinct music score arrangements provide a substantial collection.
• Includes multimodal data with text-based scores and references to original audio (YouTube).
• Built from the PDMX source of 254,000 public domain MusicXML scores.

Limitations

• The specific row count for audio-aligned entries is unknown, as YouTube references are only provided 'when available'.
• Unknown sample size and file formats limit immediate assessment of data structure.
• Potential geographic or stylistic bias inherent to the source of public domain MuseScore arrangements.

Provenance

Source

Built from the PDMX dataset, which contains public domain MusicXML scores from MuseScore.

Collection Method

Derived and processed from an existing collection of 254,000 scores.

Freshness

Last updated on the platform in April 2026.