Rhythm, Chord and Melody Generation for Lead Sheets using Recurrent Neural Networks
This work addresses the challenge of generating structured and coherent music for lead sheets, which is incremental as it builds on existing recurrent neural network methods.
The authors tackled the problem of generating coherent lead sheets by proposing a two-stage LSTM model that first creates harmonic and rhythmic templates, then generates melody notes conditioned on them, resulting in improved perceived musical coherence over baselines in subjective listening tests.
Music that is generated by recurrent neural networks often lacks a sense of direction and coherence. We therefore propose a two-stage LSTM-based model for lead sheet generation, in which the harmonic and rhythmic templates of the song are produced first, after which, in a second stage, a sequence of melody notes is generated conditioned on these templates. A subjective listening test shows that our approach outperforms the baselines and increases perceived musical coherence.