A Music Information Retrieval Approach to Classify Sub-Genres in Role Playing Games
Daeun Hwang, Xuyuan Cai, Edward F. Melcer, Elin Carstensdottir
TL;DR
The study tackles the lack of systematic, quantitative analysis of video game music (VGM) across RPG sub-genres and its relation to storytelling and gameplay. It adopts an MIR-inspired approach, assembling 27 instrumental tracks from 9 games across Adventure, Action, and Strategy RPGs, standardizing audio, and extracting features such as BPM, Zero-Crossing Rate, spectral centroid, chroma features, and MFCCs with Librosa, followed by a nearest-neighbors evaluation to assess sub-genre predictability. The results show that MFCCs do not sharply differentiate sub-genres, while Adventure RPGs exhibit broader MFCC ranges and more variable ZCR and chroma; tempo differences are limited and classifier accuracy is 33.3%, indicating limited predictive power with the current data. This work demonstrates a concrete, data-driven MIR workflow for VGM and highlights the need for larger datasets and potentially more sophisticated models to exploit feature-based cues for genre classification and storytelling analysis in game music research.
Abstract
Video game music (VGM) is often studied under the same lens as film music, which largely focuses on its theoretical functionality with relation to the identified genres of the media. However, till date, we are unaware of any systematic approach that analyzes the quantifiable musical features in VGM across several identified game genres. Therefore, we extracted musical features from VGM in games from three sub-genres of Role-Playing Games (RPG), and then hypothesized how different musical features are correlated to the perceptions and portrayals of each genre. This observed correlation may be used to further suggest such features are relevant to the expected storytelling elements or play mechanics associated with the sub-genre.