In collaboration with the National Ballet School of Canada
Shaun Amyot (choreography)
Don Sinclair (interactive visuals)
Doug Van Nort (interactive music composition)
Origin8 was an interactive dance/media piece which utilized MYO sensing technology. This was a collaboration with the National Ballet School of Canada, and my work focused on creating interactive music composition that was driven by the 21 dancers. The piece was created for the quadrennial Assembleé Internationale festival which brings together top Ballet schools and their select students internationally. In this work, the interactive music applied machine learning (continuous neural networks) to the task of learning mappings between MYO sensing and sonic output. In order to amplify the salience of particular phrases within the piece, we decided to specialize MYO placement for a given dancer to upper/lower positions on either arm or leg, as can be seen in the above figure. In this project, mappings would iterate over the same phrase multiple times. There was also a distributed aspect of the development in that dancer phrases were videoed in parallel with time-aligned sensor data being captured. This allowed me to repeatedly watch video of dancer movements and to develop a vocabulary of sonic gestural shapes. The out-of-real-time sensing data was then used to improvisationally train mappings using the Wekinator-based machine learning techniques through iterative in-studio “performance” in parallel with the video/data representations. The project thus followed a similar improvisational approach to development of machine mediated mappings as in Elemental Agency but in a much more unidirectional fashion given that the 21 dancers were flown in just days before the premiere. As part of the interactive music design, the overall musical structure emerged from the actions and movements of the dancers, so that each of the three performances had an overarching form but differed in their details.