Chord Recognition in Symbolic Music: A Segmental CRF Model, Segment-Level Features, and Comparative Evaluations on Classical and Popular Music
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.
Keywords
Cite
@article{arxiv.1810.10002,
title = {Chord Recognition in Symbolic Music: A Segmental CRF Model, Segment-Level Features, and Comparative Evaluations on Classical and Popular Music},
author = {Kristen Masada and Razvan Bunescu},
journal= {arXiv preprint arXiv:1810.10002},
year = {2018}
}