Music Structure Analysis

Music structure analysis (MSA) aims to automatically segment musical pieces into meaningful sections like verses and choruses, providing a simplified representation of a song's organization. Current research focuses on developing both supervised and unsupervised methods, employing techniques like dynamic programming algorithms operating on self-similarity matrices, deep learning models such as LSTMs and Transformers, and graph-based approaches analyzing symbolic representations. These advancements improve the accuracy of music segmentation and facilitate applications in music information retrieval, automatic music composition, and musicological analysis by providing more robust and efficient tools for understanding musical structure.