This isn't actually a duplicate, as it applies to more than just MIDI.
The issue with this is, is it feasible? The algorithms I've seen to do this sort of thing are generally not fast enough to run in realtime.
As for drawing a distinction between synth and waveform audio, that's an architectural change I'm not very comfortable with, as it would require totally redoing our sound/audio system.
Well, I didn't request this. I am confused how it is almost identical to my post. I asked for MIDI in mine because I knew that had the potential to be super simple.
Former userAccount Deleted
For PCM audo one could just skip so many or compute an average of n samples. This will increase the speed and the pitch. A method that leaves pitch as is but speeds up the audio is more complicated though.