B ird's‐ E ye V iews of the M usical S urface : M ethods for S ystematic P itch ‐C lass S et A nalysis
Erkki Huovinen, Atte Tenkanen · Music Analysis · 2007
ABSTRACT This study presents a series of methods for the analysis of the average characteristics of pitch‐class material on the surface level of musical works. The methods rely on so‐called tail segmentation, the partitioning of a musical work into a large number of overlapping pitch‐class sets of equal cardinality. The resulting data can be used as a means of scanning a piece of music in order to detect changes in the local prevalence of some chosen pitch class–related feature. For example, scanning the musical surface for each individual interval class in turn results in a ‘stretched interval‐class vector’ represented by six curves. In another application termed ‘comparison set analysis’, similarity measures for set classes are used to detect similarities with respect to some chosen set class. One may thus measure local changes in, for example, the relative degree of ‘diatonicism’ within a piece of music. Alternatively, one may compare whole corpuses of compositions in terms of their overall surface similarity with respect to a chosen comparison set. In such methods of systematic pitch‐class set analysis, the focus is shifted from identifying structurally important pitch sets or pitch‐class sets to measuring selected properties within the set of all locally available pitch‐class sets.