# Supervised Metric Learning for Music Structure Features

arXiv: https://arxiv.org/abs/2110.09000 · HTML: https://mlsanity.com/paper/2110.09000/

## Key facts

- Cited by 6 papers in the map (data through 14 Oct 2026).
- First posted on arXiv 18 Oct 2021.

## Built on by (most cited first, 6 of 6)

- [Pitchclass2vec: Symbolic Music Structure Segmentation with Chord Embeddings](https://mlsanity.com/paper/2303.15306/) (2023)
- [Self-Similarity-Based and Novelty-based loss for music structure analysis](https://mlsanity.com/paper/2309.02243/) (2023)
- [SSM-Net: feature learning for Music Structure Analysis using a Self-Similarity-Matrix based loss](https://mlsanity.com/paper/2211.08141/) (2022)
- [Do Foundational Audio Encoders Understand Music Structure?](https://mlsanity.com/paper/2512.17209/) (2025)
- [Barwise Compression Schemes for Audio-Based Music Structure Analysis](https://mlsanity.com/paper/2202.04981/) (2022)
- [Of All StrIPEs: Investigating Structure-informed Positional Encoding for Efficient Music Generation](https://mlsanity.com/paper/2504.05364/) (2025)

## Abstract (arXiv)

Music structure analysis (MSA) methods traditionally search for musically meaningful patterns in audio: homogeneity, repetition, novelty, and segment-length regularity. Hand-crafted audio features such as MFCCs or chromagrams are often used to elicit these patterns. However, with more annotations of section labels (e.g., verse, chorus, and bridge) becoming available, one can use supervised feature learning to make these patterns even clearer and improve MSA performance. To this end, we take a supervised metric learning approach: we train a deep neural network to output embeddings that are near each other for two spectrogram inputs if both have the same section type (according to an annotation), and otherwise far apart. We propose a batch sampling scheme to ensure the labels in a training pair are interpreted meaningfully. The trained model extracts features that can be used in existing MSA algorithms. In evaluations with three datasets (HarmonixSet, SALAMI, and RWC), we demonstrate that using the proposed features can improve a traditional MSA algorithm significantly in both intra- and cross-dataset scenarios.