Indirect Acquisition of Percussion Gestures Using Timbre Recognition
Adam Tindale, Ajay Kapur, W. Andrew Schloss, George Tzanetakis · 2005
There are many techniques available to capture the gestures of a performer. By utilizing digital signal processing and machine learning techniques we are able to capture, process and classify signals in real-time in order to provide data for use by musicians in a performance context. A common technique that achieves similar results is the use of sensors. The system that we describe uses data that is already available in miked environments, therefore not impeding or modifying the performer. The system is accurate and works in real-time and could be used in the future for a live performance. This system has been implemented with a snare drum and tabla for as demonstrations. The specific algorithms used, experiments and advantages and disadvantages of this technology will be discussed.