Working with diffuser models based on simulated annealing and other simple techniques the correct cone and reverberations of a sonic space is computed in real time. The representation of the sound sources and the computation of collisions is done generatively, by considering phase difference carried by binaural audio. This produces audio clues which can then be manipulated by a higher level framework.

We are continuing to explore other models with the advent of highly parallelized computing at affordable rates.