Convolutive Block-Matching Segmentation Algorithm with Application to Music Structure Analysis

TL;DR

This paper introduces the Convolutive Block-Matching algorithm for Music Structure Analysis, which segments songs into sections using autosimilarity matrices and dynamic programming, achieving competitive unsupervised performance.

Contribution

The paper presents a novel unsupervised CBM algorithm for MSA that operates on autosimilarity matrices derived from audio features, improving segmentation performance.

Findings

Competitive performance on 3 out of 4 metrics compared to supervised methods

Effective use of autosimilarity matrices from audio features

Unsupervised approach suitable for music segmentation

Abstract

Music Structure Analysis (MSA) consists of representing a song in sections (such as ``chorus'', ``verse'', ``solo'' etc), and can be seen as the retrieval of a simplified organization of the song. This work presents a new algorithm, called Convolutive Block-Matching (CBM) algorithm, devoted to MSA. In particular, the CBM algorithm is a dynamic programming algorithm, applying on autosimilarity matrices, a standard tool in MSA. In this work, autosimilarity matrices are computed from the feature representation of an audio signal, and time is sampled on the barscale. We study three different similarity functions for the computation of autosimilarity matrices. We report that the proposed algorithm achieves a level of performance competitive to that of supervised State-of-the-Art methods on 3 among 4 metrics, while being unsupervised.