A framework for interactive generation of music for games
Abstract
Tonal music has had a rich history in Video Games and movies, and in fact, music generation has played a minor role in the history of game music as well. Recent developments in Music Theory has derived representions of chord progressions using geometric topologies. Unlike prior generative music, the framework proposed in this paper attempts to approach tonal music generation by building networks of chords using the geometric topologies. This geometric network of chords can then be used inside of reinforcement learning models for learning the best motions in the progression. The proposed method uses Q Learning models, by rewarding acceptable chords, such as Major and minor chords. Rewards are also given to chords within a specified scale. The proposed framework approaches tonal chord progressions by keeping a tonal center in the progression. Methods for creating interactive and unique music for video games are also discussed. © 2012 IEEE.
