Enregistré dans:
Détails bibliographiques
Auteurs principaux: Ahn, Jaehoon, Hwang, Tae Gum, Jung, Moon-Ryul
Format: Preprint
Publié: 2026
Sujets:
Accès en ligne:https://arxiv.org/abs/2605.12287
Tags: Ajouter un tag
Pas de tags, Soyez le premier à ajouter un tag!
_version_ 1866913118751293440
author Ahn, Jaehoon
Hwang, Tae Gum
Jung, Moon-Ryul
author_facet Ahn, Jaehoon
Hwang, Tae Gum
Jung, Moon-Ryul
contents Over the past two decades, the task of musical beat tracking has transitioned from heuristic onset detection algorithms to highly capable deep neural networks (DNN). Although DNN-based beat tracking models achieve near-perfect performance on mainstream, percussive datasets, the SMC dataset has stubbornly yielded low F-measure scores. By testing how well state-of-the-art models detect beats on individual tracks in the SMC dataset, we identify three distinct failure modes: octave errors, continuity errors, and complete tracking failure where all metrics fall below 0.3. We reveal that state-of-the-art models tend to generate "confident-but-wrong" activations. Furthermore, we show that the standard DBN's default minimum tempo of 55 BPM prevents it from inferring the correct tempo for 21\% of SMC tracks, forcing double-tempo predictions on slow music. By exposing such fundamental oversights, we provide concrete directions for improving beat and downbeat detection, specifically emphasizing training data diversification and multi-hypothesis tempo estimation.
format Preprint
id arxiv_https___arxiv_org_abs_2605_12287
institution arXiv
publishDate 2026
record_format arxiv
spellingShingle The SMC Blind Spot: A Failure Mode Analysis of State-of-the-Art Beat Tracking
Ahn, Jaehoon
Hwang, Tae Gum
Jung, Moon-Ryul
Audio and Speech Processing
Sound
Over the past two decades, the task of musical beat tracking has transitioned from heuristic onset detection algorithms to highly capable deep neural networks (DNN). Although DNN-based beat tracking models achieve near-perfect performance on mainstream, percussive datasets, the SMC dataset has stubbornly yielded low F-measure scores. By testing how well state-of-the-art models detect beats on individual tracks in the SMC dataset, we identify three distinct failure modes: octave errors, continuity errors, and complete tracking failure where all metrics fall below 0.3. We reveal that state-of-the-art models tend to generate "confident-but-wrong" activations. Furthermore, we show that the standard DBN's default minimum tempo of 55 BPM prevents it from inferring the correct tempo for 21\% of SMC tracks, forcing double-tempo predictions on slow music. By exposing such fundamental oversights, we provide concrete directions for improving beat and downbeat detection, specifically emphasizing training data diversification and multi-hypothesis tempo estimation.
title The SMC Blind Spot: A Failure Mode Analysis of State-of-the-Art Beat Tracking
topic Audio and Speech Processing
Sound
url https://arxiv.org/abs/2605.12287