A Large-Scale Study of Language Models for Chord Prediction
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.
Code (0)
등록된 구현이 없습니다.
Tasks
Chord RecognitionSimilar Papers 제목 키워드 기반
Chord Embeddings: Analyzing What They Capture and Their Role for Next Chord Prediction and Artist Attribute Prediction
Natural language processing methods have been applied in a variety of music studies, drawing the connection between music and language. In this paper, we expand those approaches by investigating \textit{chord embeddings}…
AttributePredictionMulti-Step Chord Sequence Prediction Based on Aggregated Multi-Scale Encoder-Decoder Network
This paper studies the prediction of chord progressions for jazz music by relying on machine learning models. The motivation of our study comes from the recent success of neural networks for performing automatic music co…
DecoderImproved Chord Recognition by Combining Duration and Harmonic Language Models
Chord recognition systems typically comprise an acoustic model that predicts chords for each audio frame, and a temporal model that casts these predictions into labelled chord segments. However, temporal models have been…
Chord RecognitionLanguage ModelingLanguage ModellingEmpirical Study of Pop and Jazz Mix Ratios for Genre-Adaptive Chord Generation
This revision updates a pop-to-jazz chord-generation rehearsal study. Best-epoch metrics still show that modest pop rehearsal preserves pop accuracy while improving jazz prediction, but v2 corrects released-checkpoint se…
Reward prediction errors arising from switches between major and minor modes in music: An fMRI study
Evidence has accumulated that prediction error processing plays a role in the enjoyment of music listening. The present study examined listeners' neural responses to the signed reward prediction errors (RPEs) arising fro…