Adaptive Music Scores for Interactive Media

This thesis introduces a framework for adaptively scoring interactive media based on principles of cognitive science, media studies, musicology and artificial intelligence. The framework consists of three core, interconnected foundations: a novel approach to modelling user emotion for interactive media and music, cognitive models of knowledge organisation for representing virtual worlds and approaches to music composition based on processes of adaptation and improvisation. The framework is used to produce a working adaptive music system for interactive media. A user study of game players shows that the system increased feelings of immersion, correlation of music with events and consistency of mood compared to the official soundtrack in two tested games, supporting the effectiveness of the framework in scoring interactive media products.