Year: 2007
Type: conference-paper
Abstract: This paper describes software to facilitate research on the automatic derivation of hierarchical (Schenkerian) musical structures from a musical surface.Many MIR tasks require information about musical structure, or would perform better if such information were available.Automatic derivation of musical structure faces two significant obstacles.Firstly, the solution space of possible structural analyses of a piece is very large.Secondly, pieces can have more than one valid structural analysis, and there is little firm agreement among music theorists about how to distinguish a good analysis.To circumvent the first of these obstacles, software has been developed which derives a tractable 'matrix' of possibilities from a musical surface (i.e., MIDI-like note-time information).The matrix is somewhat like the intermediate results of a dynamic-programming algorithm, and in a similar way it is possible to extract a particular structural analysis from the matrix by following the appropriate path from the top level to the surface.It therefore provides a tool to facilitate research on the second obstacle by allowing candidate 'goodness' metrics to be incorporated into the software and tested on actual music.
Author: Marsden, Alan
Language: English
Subfield: Music
Field: Arts and Humanities
Domain: Social Sciences
Open Access status: green
OpenAlex ID: W31442318