Finding music to move to: Relevance in Music Information Retrieval

Relevance, a notion at the heart of information retrieval (IR), has received prolific attention in the textual IR domain. While the creation of rigorous and practicable theories concerning the nature of relevance has long been identified as a key priority for the field of Music Information Retrieval (MIR), relevance-related research has remained scarce. In this talk, I present the outcomes of a large-scale systematic analysis of the user-focused MIR literature to identify different conceptualizations of relevance in a musical context. The outcomes of the analysis establish a broad account of the state of knowledge in the field by triangulating convergent findings of disparate studies in order to identify areas of commonality, and outline several under-explored areas, pointing the way for future research.

Building on this foundation, I present an investigation of rhythmic information as a relevance criterion, focusing on beat salience, a measure of the perceptual prominence of the beat in the context of finding music to move to. Employing a convergent-methods approach investigating perceptual beat induction, sensorimotor synchronization, and beat salience judgement, I assess the validity and reliability of beat salience as a situational relevance criterion for use cases involving synchronized movement to music.

David M. Weigl is a postdoctoral research associate at the University of Oxford e-Research Centre. His work involves the application of Linked Data and semantic technologies in order to enrich digital music information and facilitate access to a variety of musical data sources. His research interests revolve around music perception and cognition, and music information retrieval.