This sounds.... interesting: I know it's just a demo but in the end, what happened? Was it actually used in some games?
I doubt it, it's not hard to code audio panning. Haha and I've got a further draw distance on my NY libGS splitscreen game
Remember is says it combines music channels and simulates surround, that's more than basic panning. A helicopter is a terrible demo...
AFAIK surround sound should be trivial to implement as long as you use two voices per sample and apply left/right pan on both to simulate four pseudo channels. Of course, you can't do that with libsnd, which only works with a limited sound environment that you can barely override. There was an official sample with custom handling of spu samples that mentioned this very issue in a 3D sound function. The main issue is when you have tracker music, as that usually reserves at least 16 voices for BGM, so you're left with only 8 voices = 4 surround samples at the same time. You could work around that by having streamed music -- with XA you would have all 24 voices available, while SPU streams leave you with 22 voices. While that is super convenient for most genres, there would be no mean to cache new data while the CD unit is busy. SPU economy is a real bitch in most cases, but it all comes down to what your game requires.
