Structure-Aware Audio-to-Score Alignment using Progressively Dilated Convolutional Neural Networks

TL;DR

This paper introduces a novel structure-aware audio-to-score alignment method using progressively dilated convolutional neural networks, effectively capturing structural differences in music performances even with limited annotated data.

Contribution

The paper proposes a new neural network architecture with varying dilation rates to improve detection of structural differences in audio-to-score alignment.

Findings

Models outperform standard methods in real performance recordings

Effective in limited annotated data scenarios

Captures both short-term and long-term musical context

Abstract

The identification of structural differences between a music performance and the score is a challenging yet integral step of audio-to-score alignment, an important subtask of music information retrieval. We present a novel method to detect such differences between the score and performance for a given piece of music using progressively dilated convolutional neural networks. Our method incorporates varying dilation rates at different layers to capture both short-term and long-term context, and can be employed successfully in the presence of limited annotated data. We conduct experiments on audio recordings of real performances that differ structurally from the score, and our results demonstrate that our models outperform standard methods for structure-aware audio-to-score alignment.