Consider the OT has 8 tracks of recorder trigs, per step sampling of different lengths and source/sources, playback of all recorders freely assignable to all tracks per step, different track length/scale, ability of the play tracks to playback the sample(s) instantly warped in many ways(slices, pitch shift, rate, fx, etc…) and all of the recorder buffer placeholders update while sampling…
There’s a different style of super crazy sampling power in all of that…
It’s conceivable that you could program an automated sample/warp/resample process to play an extremely complex soundscape or even a whole song just by sampling a note, or a noise even, and the process you programmed would make the soundscape/song out of the one sample and it’s automated resamples… Initiated just by pressing play and feeding it that note or noise, no samples loaded.
That could be just from one sample with resampling, much more extremely diverse and complex audio could come from beginning with several sources of more complex audio…(4 harmonizing passages sent from midi tracks and returned to the inputs for example.) And these processes can dynamicly change by feeding in different initial sample audio or better yet constant sampling…
How hard is it?
I have no idea…