Hi,
we provide not way of doing this at the moment, but I can see a way this could work if you assume that the you can put the two processing routines in series or parallel. For instance, you would have your C++(or pyo) code pre-process the audio or sensor input to SuperCollider. Or you could have the C++(or pyo) post-process the output of SuperCollider, or the two could process in parallel and be combined before being sent to the outputs.
- the most hackish approach to this would be recompile supercollider including the pre-processing / post-processing/parallel processing stage.
- better would be to add hooks in the supercollider core to call external functions pre/post processing. This would be a one-off modification to the SC source code.
- I think the best approach would really be to write your own supercollider ugens, which could easily be made to wrap pyo or C++ or libpd. I have never done this, but it is something that could turn out easier than the rest and would be the most appropriate, flexible and powerful way of doing it in my opinion, as then your additional processing stages would be handled as regular ugens within SuperCollider.