Melodic and rhythmic pattern discovery in music: towards explicit modelling of implicit musical knowledge

In my talk I will address Anders Elowsson’s question on the usefulness of music concepts as mid-level presentations or intermediate presentations (as discussed in his PhD thesis) in a different context. I am interested in automatically finding repeated patterns in different musical dimensions, as they are often argued to influence many important musical concepts such as musical similarity, style recognition or characterizing the “most salient part” of a musical composition. Music is known to be extremely repetitive, and automatic pattern discovery algorithms find usually huge numbers of repetitive patterns in compositions. But which of these repetitions are musically meaningful to listeners in which context? In this talk I address the question as to when we perceive two folk melodies to be variants of each other, and when we consider a piece to be a ragtime, by employing automatic pattern finding. I will discuss how the computational modeling of concepts such as “repeated patterns” may help shedding light on implicit human musical knowledge.