Integrated Composition, Improvisation and Technology (ICIT)

Music often conveys a sense of “gesture”, anevocation of motion and energy, which makes it dramatic, exciting, and expressive. One common challenge in the production of algorithmically-generated computer music is the question of how to imbue the sound with the excitement and vitality of liveperformance. In the case of interactive computer music, one has the additional challenge of programming the computer to interpret the expressive qualities of musicbeing performed in real time. This lecture presents an approach to automaticallyanalyzing and characterizing gesture in musical sound, as a way of improving a computer’s interaction with a human performer in a live improvisation. By describing music as patterns of changing parametric data, the computer can store and categorize descriptors of musical gestures. As an extension of that research, we can then consider how derivatives of that analysis data—the ways in which the data changes over time—characterize thegestural quality of a performance. In the algorithmic generation of music, control of those derivatives of change in musical parameters can improve the expressivepotential of computerized improvisation.