DECIBEL: Improving Audio Chord Estimation for Popular Music by Alignment and Integration of Crowd-Sourced Symbolic Representations

TL;DR

DECIBEL enhances automatic chord estimation in popular music by integrating audio, MIDI, and tab data through alignment and fusion, outperforming existing methods and demonstrating the value of combining heterogeneous symbolic representations.

Contribution

It introduces DECIBEL, a novel system that leverages MIDI and tab data to improve audio-based chord estimation in popular music, addressing limitations of current methods.

Findings

DECIBEL improves ACE accuracy by over 3% on average.

Integration of symbolic representations enhances chord estimation performance.

The approach demonstrates the effectiveness of combining heterogeneous musical data.

Abstract

Automatic Chord Estimation (ACE) is a fundamental task in Music Information Retrieval (MIR) and has applications in both music performance and MIR research. The task consists of segmenting a music recording or score and assigning a chord label to each segment. Although it has been a task in the annual benchmarking evaluation MIREX for over 10 years, ACE is not yet a solved problem, since performance has stagnated and modern systems have started to tune themselves to subjective training data. We propose DECIBEL, a new ACE system that exploits widely available MIDI and tab representations to improve ACE from audio only. From an audio file and a set of MIDI and tab files corresponding to the same popular music song, DECIBEL first estimates chord sequences. For audio, state-of-the-art audio ACE methods are used. MIDI files are aligned to the audio, followed by a MIDI chord estimation step. Tab files are transformed into untimed chord sequences and then aligned to the audio. Next, DECIBEL uses data fusion to integrate all estimated chord sequences into one final output sequence. DECIBEL improves all tested state-of-the-art ACE methods by over 3 percent on average. This result shows that the integration of musical knowledge from heterogeneous symbolic music representations is a suitable strategy for addressing challenging MIR tasks such as ACE.