A survey of generative music applications in video games, both academic and from the games industry
Abstract: Music is a common element in almost all video games. Most music in games is written by a human composer, and played as a linear piece behind gameplay. Adaptive and/or Generative music systems can be used to extend the musical content or create new musical content using algorithms and AI. While there is research into these systems, there has yet to be an organized examination of their architecture and use. We present a taxonomy of generative music in games, to allow for examination and discussion of generative music systems. In doing so, we also present a survey of the current state of the art of generative music systems in games, and discuss challenges and prospects of generative music for games.
Our survey of generative music in video games organizes, typologizes, and analyzes trends in past uses of generative music in video games. Overall, we find that academic approaches generally focus more on applying novel algorithms to the real-time composition of music, while approaches from the game industry generally focus on re-arranging and expanding composed and pre-recorded audio stems.
We typologize these systems and uses of generative music by musical features (the way that the music is constructed and adapts), gameplay dimensions (the way that the music exists in relation to the game world), and architecture dimensions (the way that the system generates or arranges the music).