Skip That Beat: Augmenting Meter Tracking Models for Underrepresented Time Signatures

2025-02-18

AI Summary

Existing beat and downbeat tracking models heavily rely on 4/4 time-signature data, resulting in poor generalization to non-dominant meters such as 2/4 and 3/4 (e.g., Brazilian samba). To address this, we propose a beat-interval–controlled data augmentation method: by systematically pruning specific beat positions from 4/4-annotated sequences, we synthesize high-fidelity 2/4 and 3/4 training samples without additional manual annotation, effectively mitigating the problem of sparse beat representation. We fine-tune and evaluate multiple state-of-the-art beat/downbeat tracking models under cross-meter transfer settings. Our approach yields substantial improvements in downbeat detection for non-4/4 meters—achieving significant average F1-score gains on 2/4 and 3/4 test sets, and a 12.3% absolute F1 improvement on an unseen samba dataset—while preserving performance on 4/4 meter.

Technology Category

• Machine Learning: Transfer, Domain Adaptation, Multi-Task Learning
• Computer Vision: Motion & Tracking
• Search and Optimization: Metareasoning and Metaheuristics

Application Category

• Economics, Online Markets and Human Computation: Data quality aspects of human-annotated datasets
• Web Mining and Content Analysis: Robustness and generalizability of Web mining methods
• User Modeling, Personalization and Recommendation: Metrics for user behavior and evaluating success

Abstract

Beat and downbeat tracking models are predominantly developed using datasets with music in 4/4 meter, which decreases their generalization to repertories in other time signatures, such as Brazilian samba which is in 2/4. In this work, we propose a simple augmentation technique to increase the representation of time signatures beyond 4/4, namely 2/4 and 3/4. Our augmentation procedure works by removing beat intervals from 4/4 annotated tracks. We show that the augmented data helps to improve downbeat tracking for underrepresented meters while preserving the overall performance of beat tracking in two different models. We also show that this technique helps improve downbeat tracking in an unseen samba dataset.

Problem

• Improving generalization of meter tracking models
• Augmenting underrepresented time signatures like 2/4
• Enhancing downbeat tracking in Brazilian samba

Innovation

• Augments meter tracking models
• Removes beat intervals
• Improves underrepresented time signatures