Is that possible? How is this done in theory? Would I simply make a very small buffer packet size and then delay the playback artificially? I assume that's the low-brain way? How could I do it in such a way that the sound is smooth and doesn't suck?
Delaying playback will cause stuttering in the audio. Perhaps this question might point the way?