Statistical Analyses of Encoded Music Performance - Jeff Smith Doctoral Dissertation Defense
Statistical Analyses of Encoded Music Performance
Statistical analyses of encoded music performance by a large and diverse international community of amateur performers reveals insights into fundamental questions of musical behavior. Specifically, by observing demographic data encapsulated in a corpus of performance recordings, we can conjecture about cultural, geographical, topographical, socio-political, economic and other potential influences, as well as explore possible ’universals’ in musical thought and practice.
A century ago, Bela Bartok visited what were then remote regions to identify and characterize folk music at its source. Subsequent investigators and collectors, particularly in the context of colonialism and with a variety of objectives (Agawu [1992]) sought to clarify and categorize music from a broad spectrum of cultures and regions. In some cases western music was introduced to an indigenous populace specifically in order to observe and record listener reaction [see Sachs, 1962, p17].
Although fraught with issues (1), there is a great deal to gain by examining how a particular populace interprets a foreign object. In terms of cultural objects such as a work of music, inferences can be drawn as to what (if anything) is ’universal’, as well as how to characterize cultural differences. As music delivery, through increasingly pervasive mobile devices, becomes more available as well as more interactive, new opportunities arise to study musical practices. Interactive applications merge audio playback with recorded performance, effectively providing users novel musical instrument interfaces that are amenable to mastery by amateurs.
These software instruments are limited in their acoustic richness and expressive control. They also are largely biased in their available sounds and tunings toward mainstream western musical practice. On the other hand, within a relatively short time and with minimal frustration, users both young and old can learn to perform and record music including a wide range of pre-composed works in diverse styles and genres and from numerous origins. More importantly, despite (perhaps, because of) these instruments’ limitations, particular aspects of performance — specifically rhythmic agogics — are recorded with accuracy and stored efficiently.
The underlying premise of this thesis is that, embedded in the encoded music performances by this diverse community of amateur performers is a wealth of information about musical performance practices. Statistical analysis of these performances can provide insight into musical behaviors as a whole, as well as comparative observations on the cultural, social, economic and geopolitical influences on music performance.
This study attempts to address questions of musical practices from an entirely novel perspective, specifically taking advantage of massively popular game-oriented music performance programs on mobile devices.
(1) Ultimately, the study of what (if anything) unites, as well as how to characterize the seemingly vast and diverse differences in musical practices amongst humans, is hampered by what Agawu labeled the ’us/them’ dichotomy.