Research Seminars

On Onsets On-The-Fly: Real-Time Event Segmentation and Categorisation as a Compositional Effect
Nick Collins, Music Department, Cambridge University
29 September 2004

Abstract

Compositional applications for real-time event segmentation are discussed. A causal real-time onset detector which makes onset data available as fast as possible is introduced, based on work by Klapuri, Hainsworth and Jensen and Andersen. This analysis frontend informs algorithmic cutting procedures which respect the events of the incoming audio stream. A further refinement stores events to particular buffers based on a coarse categorisation between snare, kick or hihat classes. Algorithmic composers running playback of these buffers thereby respond to changing timbral events of a live feed from an instrumentalist or other audio source. The use of an onset detection threshold to create abstracted rhythms based on some existing source is further examined.

Further details of Nick's work on live audio processing can be found at:
http://www.cus.cam.ac.uk/~nc272/