Chord Recognition in Symbolic Music: A Segmental CRF Model, Segment-Level Features, and Comparative Evaluations on Classical and Popular Music

TL;DR

This paper introduces a semi-Markov CRF model for chord recognition in symbolic music, effectively segmenting and labeling music into chords using rich segment-level features, and demonstrates superior performance on classical and rock datasets.

Contribution

The paper presents a novel semi-CRF based approach for joint segmentation and chord labeling, utilizing segment-level features for improved harmonic analysis.

Findings

Semi-CRF model outperforms previous methods with ample training data.

Model remains competitive with limited training data.

Extensive evaluation on classical and rock music corpora.

Abstract

We present a new approach to harmonic analysis that is trained to segment music into a sequence of chord spans tagged with chord labels. Formulated as a semi-Markov Conditional Random Field (semi-CRF), this joint segmentation and labeling approach enables the use of a rich set of segment-level features, such as segment purity and chord coverage, that capture the extent to which the events in an entire segment of music are compatible with a candidate chord label. The new chord recognition model is evaluated extensively on three corpora of classical music and a newly created corpus of rock music. Experimental results show that the semi-CRF model performs substantially better than previous approaches when trained on a sufficient number of labeled examples and remains competitive when the amount of training data is limited.