11th International Society for Music Information Retrieval Conference (ISMIR 2010) COMPUTATIONAL ANALYSIS OF MUSICAL INFLUENCE: A MUSICOLOGICAL CASE STUDY USING MIR TOOLS Nick Collins Department of Informatics, University of Sussex, Falmer, Brighton, BN1 9QJ, UK
[email protected] ABSTRACT of genres is implicit in much discussion of the philosophy of stylistic categories in music [1, 10], and related to sim- Are there new insights through computational methods to ilar questions in biology concerning speciation events and the thorny problem of plotting the flow of musical influ- memetics [4, 6]. ence? This project, motivated by a musicological study of Automated methods for the analysis of musical similar- early synth pop, applies MIR tools as an aid to the inves- ity provide a new angle on relationships between works, tigator. Web scraping and web services provide one an- whether comparing individual pieces or within larger cor- gle, sourcing data from allmusic.com, and utilising python pora. For example, the data-driven analyses explored by APIs for last.fm, EchoNest, and MusicBrainz. Charts of David Cope across MIDI files [3] are primarily used for influence are constructed in GraphViz combining artist sim- synthesis, but can also help to explore the links between ilarity and dates. Content based music similarity is the sec- composers. Symbolic analysis tools in MIR parallel such ond approach, based around a core collection of synth pop movements in algorithmic composition: McKay and Fu- albums. The prospect for new musical analyses are dis- jinaga [11] discuss the application of their jSymbolic fea- cussed with respect to these techniques.