Yin-Yang: Developing Motifs With Long-Term Structure And Controllability

TL;DR

Yin-Yang is a novel framework for generating long-term structured motifs in music, combining a phrase generator, refiner, and selector to improve controllability and structural coherence over existing transformer models.

Contribution

It introduces a new motif development framework with a corruption-refinement training strategy and a novel evaluation metric for musical structure.

Findings

Outperforms state-of-the-art transformer models in structural coherence

Enables controllable and semi-interpretable musical motif development

Improves long-term musical structure generation in datasets with shorter durations

Abstract

Transformer models have made great strides in generating symbolically represented music with local coherence. However, controlling the development of motifs in a structured way with global form remains an open research area. One of the reasons for this challenge is due to the note-by-note autoregressive generation of such models, which lack the ability to correct themselves after deviations from the motif. In addition, their structural performance on datasets with shorter durations has not been studied in the literature. In this study, we propose Yin-Yang, a framework consisting of a phrase generator, phrase refiner, and phrase selector models for the development of motifs into melodies with long-term structure and controllability. The phrase refiner is trained on a novel corruption-refinement strategy which allows it to produce melodic and rhythmic variations of an original motif at generation time, thereby rectifying deviations of the phrase generator. We also introduce a new objective evaluation metric for quantifying how smoothly the motif manifests itself within the piece. Evaluation results show that our model achieves better performance compared to state-of-the-art transformer models while having the advantage of being controllable and making the generated musical structure semi-interpretable, paving the way for musical analysis. Our code and demo page can be found at https://github.com/keshavbhandari/yinyang.