A Large-Scale Study of Language Models for Chord Prediction

TL;DR

This study compares N-gram and recurrent neural network models for chord prediction across extensive datasets, revealing that RNNs can adapt to individual songs, enhancing local musical context understanding.

Contribution

It provides a comprehensive comparison of language models for chord prediction and demonstrates RNNs' ability to adapt to specific songs, advancing context-aware chord recognition.

Findings

RNNs outperform N-gram models in chord prediction.

Certain RNN configurations adapt to individual songs at test time.

The study offers insights into hyper-parameter tuning for RNNs.

Abstract

We conduct a large-scale study of language models for chord prediction. Specifically, we compare N-gram models to various flavours of recurrent neural networks on a comprehensive dataset comprising all publicly available datasets of annotated chords known to us. This large amount of data allows us to systematically explore hyper-parameter settings for the recurrent neural networks---a crucial step in achieving good results with this model class. Our results show not only a quantitative difference between the models, but also a qualitative one: in contrast to static N-gram models, certain RNN configurations adapt to the songs at test time. This finding constitutes a further step towards the development of chord recognition systems that are more aware of local musical context than what was previously possible.