Artificial Neural Networks Jamming on the Beat

TL;DR

This paper proposes a simple method for generating symbolic music by creating drum patterns and using neural networks to produce melodies, addressing long-scale correlation challenges in music generation.

Contribution

It introduces a novel approach combining drum pattern generation with neural networks to improve long-scale structure in symbolic music synthesis.

Findings

Neural network can generate melodies aligned with drum patterns.

Latent space exploration enables style-specific drum pattern creation.

System produces music with song-like structure and enhanced long-scale correlations.

Abstract

This paper addresses the issue of long-scale correlations that is characteristic for symbolic music and is a challenge for modern generative algorithms. It suggests a very simple workaround for this challenge, namely, generation of a drum pattern that could be further used as a foundation for melody generation. The paper presents a large dataset of drum patterns alongside with corresponding melodies. It explores two possible methods for drum pattern generation. Exploring a latent space of drum patterns one could generate new drum patterns with a given music style. Finally, the paper demonstrates that a simple artificial neural network could be trained to generate melodies corresponding with these drum patters used as inputs. Resulting system could be used for end-to-end generation of symbolic music with song-like structure and higher long-scale correlations between the notes.