Markerless pose estimation using deep neural networks performed in real-time and mapped to sounds using Bonsai. With just a webcam, moving one wrist up and down controls sound frequency (pitch) and the other wrist moving left and right to controls sound amplitude (volume).