An Analysis Method for Metric-Level Switching in Beat Tracking

TL;DR

This paper introduces the annotation coverage ratio (ACR), a new analysis method for beat tracking that captures metric-level switching behaviors, providing deeper insights into model performance on expressive music with tempo changes.

Contribution

The paper proposes ACR, a novel evaluation metric that accounts for tempo variations and switching behaviors, enhancing the analysis of beat tracking models beyond existing metrics.

Findings

ACR effectively captures metric-level switching behaviors.

ACR provides new insights when used with existing metrics.

Experiments on diverse datasets demonstrate ACR's usefulness.

Abstract

For expressive music, the tempo may change over time, posing challenges to tracking the beats by an automatic model. The model may first tap to the correct tempo, but then may fail to adapt to a tempo change, or switch between several incorrect but perceptually plausible ones (e.g., half- or double-tempo). Existing evaluation metrics for beat tracking do not reflect such behaviors, as they typically assume a fixed relationship between the reference beats and estimated beats. In this paper, we propose a new performance analysis method, called annotation coverage ratio (ACR), that accounts for a variety of possible metric-level switching behaviors of beat trackers. The idea is to derive sequences of modified reference beats of all metrical levels for every two consecutive reference beats, and compare every sequence of modified reference beats to the subsequences of estimated beats. We show via experiments on three datasets of different genres the usefulness of ACR when utilized alongside existing metrics, and discuss the new insights to be gained.