[PDF]

Extracting information from music to create a playable rhythm game


Aidan Davey

08/05/2025

Supervised by Crispin Cooper; Moderated by Theodoros Spyridopoulos

This project presents the development of a rhythm game that generates levels from musical information extracted from user imported audio. Track tempo is estimated via the use of autocorrelation and resonance, and the two methods are compared for their accuracy. Key challenges faced include accurate onset detection for enemy generation and modification of tempo detection algorithms to improve accuracy. The final implementation presents a playable and responsive game experience, with semi-accurate tempo estimation.


Initial Plan (03/02/2025) [Zip Archive]

Final Report (08/05/2025) [Zip Archive]

Publication Form