Affective Music Generation and its effect on Player Experience
Research output: Book / Anthology / Report / Ph.D. thesis › Ph.D. thesis
This interest has been mainly motivated by how complex games have become, requiring a huge amount of assets to be created to support them.
While most focus has been put on generating levels, textures, and 3D models, there are also examples of games that generate music dynamically, which is the focus of this thesis (e.g. Spore).
In games, unlike in traditional linear storytelling media such as novels or films, narrative events unfold in response to player input. Therefore, the music composer in an interactive environment needs to create music that is dynamic and non-repetitive.
This thesis investigates how to express emotions and moods in music and how to apply this research to improve player experience in games.
This focus on the emotional expression that procedurally generated music should express has also been identified by Collins as one of the missing features that currently prevent procedurally generated music being more widely used in the game industry.
The research therefore focuses on investigating the expression of moods, and the effect affective music can have on the listener during game play.
In this thesis three systems are described: the MetaCompose affective music generator, its prototype, and a system for co-evolution of improvisational modules (Primal-Improv).
The characteristics of MetaCompose are: (i) expressing different affective states using a variety of AI-techniques, (ii) generating such music in real-time, and (iii) reacting in real-time to external stimuli.
Its architecture is comprised of three main components: the composition generator, the real-time affective music composer and an archive of compositions.
A novel feature of our approach is the separation of composition and affective interpretation: the system creates abstractions of music pieces (called “compositions”) and interprets these in real-time to achieve the desired affective expression, while simultaneously introducing stochastic variations.
Notably, the abstraction generation component includes a graph traversal-based chord sequence generator, a search-based melody generator and a pattern-based accompaniment generator.
The melody generation uses a novel constrained multi-objective evolutionary technique combining FI-2POP and Non-dominated Sorting Genetic Algorithm (NSGA-II).
The thesis presents the results of several evaluation studies, evaluating both the design of the systems, their affective expression, and exploring the effects on player experience when integrating MetaCompose with the game of Checkers.
|Publisher||IT-Universitetet i København|
|Number of pages||183|
|Publication status||Published - 2017|