Tsung-Ping / motif_discovery

An algorithm for discovering repeated pattern in symbolic music.

motif_discovery

This repository implements the algorithm for discovering repeated patterns in symbolic music as described in the paper: Yo-Wei Hsiao, Tzu-Yun Hung, Tsung-Ping Chen, Li Su, "BPS-Motif: A Dataset for Repeated Pattern Discovery of Polyphonic Symbolic Music," International Society of Music Information Retrieval Conference (ISMIR), November 2023.

Requirements

• numpy
• pretty_midi
• Beethoven Piano Sonata Motif Dataset
• Baseline algorithms

Usage

1. Clone all required datasets and repositories, and put all data in the same file.
2. Execute experiments.py.
3. For the algorithm described in the paper, see find_motives in SIA.py.