Way back at UNM, I worked on the problem of audio positioning via convolutions and HRTFs, using wavelets. The results weren’t as hoped (details here, the filters didn’t compress well at all), but I keep an eye out for the state of the art. I just found SweetSpotter, an unfortunately-Windows-only thesis project of amazing promise.

It ups the ante of audio positioning by using your webcam to see where you are, and adjusting the filtering in realtime to correct.

That is, to use the proper technical adjective, fucking amazing.

Sebastian Merchel did the work. I am impressed. As soon as the borked webcam on my MacBook gets fixed, I’ll see if I can run this in VMWare or VirtualBox.