I've been fiddling with Linux, sound servers, Mathematica, and so on, and I got a fairly good solution worked out. I had previously been just cat-ting the audio stream to /dev/dsp, but then I switched distros and my /dev/dsp disappeared, though sound still works (weird). Anyway, I installed arts and set things up so Mathematica will play through the sound server, i.e. multiplexing sounds is possible. Copy the attached sound.m file to ~/.Mathematica/Kernel/ and add this line to ~/.Mathematica/Kernel/init.m
It should be pretty easy to modify this for esound, if you want. If I feel inclined, I'll try to give it the ability to understand different SampleRates, SampleDepths, and channels. If you do improve it, I ask that it be reposted here so that all may benefit.
Attachment: sound.m, URL: ,