Automatic Real-Time Music Generation for Games

Authors

  • Steve Engels University of Toronto
  • Tiffany Tong University of Toronto
  • Fabian Chan University of Toronto

DOI:

https://doi.org/10.1609/aiide.v11i1.12775

Keywords:

music generation, video game design, machine learning

Abstract

Music composition can be a challenge for many small- to medium-sized game companies, largely due to the expense and difficulty in creating original music for each level of a game. To address this, we developed a tool that automatically generates original music, by training a music generator on pieces whose style the game designer wishes to imitate. The generator then creates original music in that style in real-time, and switches between styles when signaled by the game. This software has been refined to produce music that is coherent and imitates a composer’s larger music structure.

Downloads

Published

2021-06-24

How to Cite

Engels, S., Tong, T., & Chan, F. (2021). Automatic Real-Time Music Generation for Games. Proceedings of the AAAI Conference on Artificial Intelligence and Interactive Digital Entertainment, 11(1), 220-222. https://doi.org/10.1609/aiide.v11i1.12775