Pattern matching in music and its use for automated composition [PhD Probation Report]

Collins, Tom (2010). Pattern matching in music and its use for automated composition [PhD Probation Report]. Technical Report 2010/09; Department of Computing, The Open University.

DOI: https://doi.org/10.21954/ou.ro.000160a3

Abstract

It is something of a truism to say that self-reference abounds in classical music. That is, given a single piece of music, it will most likely contain instances of verbatim repetition as well as more subtle variation. Why then, when programming a computer to generate music, is little attention paid to ensuring the program's output contains self-reference? If the program is intended to test a model of musical style then its inattention to self- reference may result in failure of the test. This report gives further exposition of the above issue and the research question thus motivated. Current computational methods for pattern matching in music are exemplified and criticised. An introduction to Markov models for music is also provided, which is intended to be suggestive of more intricate computational models of musical style. The research proposal emphasises the need to integrate pattern matchers and music generators, and makes feasible plans to pursue this objective.