Generative Structures in Improvisation: Computational Segmentation of Keyboard Performances

AbstractWe assessed the hypothesis elaborated by Pressing and by the first author that improvisation is normally based on the generation of segmental contrasts in musical features. Nine experienced improvisers performed a series of free- and 3-section referent-based pieces, each of about three minutes. Each player undertook eight improvisations with 'referent' instructions, preceded and followed by a free improvisation. A MIDI-equipped grand piano was used, and audio and MIDI were recorded. Computational analyses of the MIDI data assessed whether performers realized the referents successfully, and then determined the large-scale segmentation of the free improvisations. The performers almost always fulfilled the referents (68/72 improvisations, p < 0.0001), and their free improvisations were also susceptible to large-scale segmentation (p < 0.005). Since performers were only told of the referent structures as the session proceeded, the similar segmentation of both the first and second free improvisation suggests that such segmentation is common in solo free improvisation, in accord with the hypothesis.Keywords: performancemusic analysisrepresentationtonality AcknowledgementsThe participation of our improvisers is greatly appreciated. Naturally, their data are not individualized above, but most were happy to be listed, and many of their recordings and performances, and other aspects of their work, are accessible internationally. They included: Diana Blom, Leonie Cohen, the late Robin Howat, Kevin Hunt, Mark Isaacs, Dave MacRae, Matt McMahon, and Alister Spence. We also thank Dr. Mary Broughton for assistance in developing the methods for the experimental study, and Dr Rebecca Killick (University of Lancaster) for advice on Changepoint and other helpful comments. Dr David Temperley kindly suggested the use of pitch class entropy in the context of gradients of tonalness.