The main focus of this project is now the Kinect 360 ambisonic interface that allows you, the audience to interact with the performance.
Since my first update on this project I’ve changed and refined many things, the most interesting change I reckon is the added speech to visuals that allows you to become part of the visuals as well as the performance. It works by using the python library FasterWhisper a faster version of Whisper by Open AI (hence the name) to transcribe voices picked up by the omni-directional mic, the text then gets sent to the DALL-E 3 API by Open AI.