JultikaUniversity of Oulu repository

Abstract

The video game industry has grown quickly from its humble beginnings to one of the largest entertainment industries in the world. Fuelled by the continuous advancements in technology, the quality and quantity of content in AAA video games continues to rise along with customer expectations. But with the ever-higher ambitions, the development budgets and durations rise with them, making the cycle unsustainable on the long run.

Procedural content generation is a technique that has the potential of helping break the cycle. The automatic generation of game content, such as levels, could help game developers reach the desired quantity of content with a fraction of the time and money required. However, commercial applications of procedural content generation so far have been largely limited in scope and lacking in quality, with the more successful cases being found in smaller budget indie games.

In this study, the possibility to use the idea of rhythm in guiding procedural level generation towards better quality was studied. Using a design science research approach, the gameplay rhythm of original Super Mario Bros. levels was extracted and used to build a rhythm-based procedural 2D platformer level generator. The nature of the generated levels was investigated by computational metrics, and the quality of them was evaluated by a series of playtests.

It was found that the existing platformer levels included an extractable rhythm. The rhythm-based level generator that was built upon the found rhythm data produced levels that were closely on par with the original levels, indicating that rhythm has potential applications in informing how a procedural content generator could create more meaningful and higher quality content. Finally, this experimental approach in incorporating music theory to procedural content generation opens up many interesting new avenues for future research.