Modelling music similarity in MUSIVA - New Trends in eHumanities

Anja Volk, Utrecht University


September 13, 2012


Modelling music similarity in MUSIVA: addressing interdisciplinary challenges of computational humanities

In this talk I present my interdisciplinary VIDI-project MUSIVA at Utrecht University on the topic of music similarity and will discuss interdisciplinary challenges we are facing in computational humanities with examples from computational musicology. The VIDI-project “Modelling musical similarity over time through the variation principle” (MUSIVA) investigates musical similarity, a key concept within music information retrieval (MIR) for developing search engines for digital music collections. While similarity is considered fundamental in Cognitive Science, no comprehensive model exists for similarity in the domain of music. Musicology offers a general approach to similarity through the universal principle of variation.  MUSIVA aims to build cognition-based computational approaches to music similarity grounding in the variation principle employed in classical, folk and popular music. This requires substantial contributions from Musicology, Music Information Retrieval and Cognitive Science. Discussing these contributions, I address issues of interdisciplinary collaboration within computational humanities, such as the use of models and abstractions for humanities research and the evaluation of computational models with respect to both standardized methods in computer science and regarding their contribution to humanities research. I will reflect on how differences between the disciplines, rather than being just stumbling blocks of mutual understanding, might become a stimulating source for interdisciplinary research and discuss what we might envision as important contributions of humanities to e-Humanities beyond the mere provision of “data”.


Anja Volk’s research is dedicated to integrate humanities research and computational methods into a comprehensive approach to the study of music. She holds master degrees in musicology and mathematics and received her PhD in computational musicology from Humboldt University of Berlin. The results of her research have contributed to areas such as computational musicology, music information retrieval, digital cultural heritage, music cognition, and mathematical music theory. After two post-doc periods at the University of Southern California and Utrecht University, she has been awarded a prestigious VIDI grant from the Netherlands Organisation for Scientific Research in 2010, which allowed her to start her own research group at Utrecht University. Her current project ‘Modelling musical similarity over time through the variation principle’  (MUSIVA) investigates music similarity in an interdisciplinary manner comprising Music Information Retrieval, Musicology and Cognitive Science. Anja is a board member of the international Society for Mathematics and Computation in Music.

(Presentation slides)