DeepJ: Style-Specific Music Generation

TL;DR

DeepJ is a neural network model that generates music conditioned on specific composer styles, allowing style control and demonstrating improved performance over previous methods through human evaluation.

Contribution

We introduce DeepJ, a novel end-to-end generative model capable of style-specific music composition with controllable parameters.

Findings

DeepJ outperforms Biaxial LSTM in human evaluations.

The model effectively learns musical style and dynamics.

Style control is demonstrated as a proof of concept.

Abstract

Recent advances in deep neural networks have enabled algorithms to compose music that is comparable to music composed by humans. However, few algorithms allow the user to generate music with tunable parameters. The ability to tune properties of generated music will yield more practical benefits for aiding artists, filmmakers, and composers in their creative tasks. In this paper, we introduce DeepJ - an end-to-end generative model that is capable of composing music conditioned on a specific mixture of composer styles. Our innovations include methods to learn musical style and music dynamics. We use our model to demonstrate a simple technique for controlling the style of generated music as a proof of concept. Evaluation of our model using human raters shows that we have improved over the Biaxial LSTM approach.