|
Decoding the Human Conducting GestureDannenberg, RogerCarnegie Mellon University Music conducting is a sophisticated form of gestural communication. Despite centuries of practice, there is no consensus regarding even what is communicated, much less how the information is encoded and decoded. This project develops simultaneous advances in the nature of conducting through instrumentation and measurement; and in the techniques for sensing, detection, and interpretation of human gesture. The project records conducting gestures using wearable, wireless sensors and synchronizes gestural data with video and audio. Working closely with musicians and conductors, the project will develop new hypotheses about conducting gestures, develop new sensors to capture relevant information, and then use machine learning to create a conducting gesture recognition system. The results are a characterization of the information conveyed by conductors and an improved generation of wearable sensors and techniques for sensing human motion. The broader impact of this work includes expanding the role of experimental science in the arts, especially music. Through work with professional and student orchestras, the methods and benefits of research will be clearly communicated to non-scientists. This work will also benefit society through a better understanding of conducting and music. This new understanding may also lead to new techniques for music education. Another application of this work is the creation of interactive exhibits at museums where people can conduct a virtual orchestra, and perhaps people will someday enjoy conducting a personal electronic orchestra at home. Advances in gestural sensing have enormous potential in the field of health care. Using human motion to detect levels and types of activity as well as emotional state could offer great benefits, allowing computers to monitor recovery from injury, at-risk patients, and the elderly. Last modified 16 August 2007 at 6:43 am by haleden |