AI-Based Stem Separation

Enhancement Summary

I use a stem separator quite often when exploring the makeup of other tracks, and equally when remixing or looking for accapellas.
Anyway, it's a bit of a shame to have to open up demucs every time, when LMMS could contain this feature. I know FL recently got a similar feature.

Implementation Details / Mockup

Likely a sample track based feature, where one would be able to right-click, then select an option to open a stem separation menu. [ BADLY PHOTOSHOPPED EXAMPLE BELOW ]

This would then call demucs using a Python API ( from what it seems, pybind11 would probably be best ).
After separation, new sample tracks would be created, each containing a stem of the targeted sample.
This may require some resolution of #735 for ideal functioning.
I'd be more than happy to try and make a first draft ( I know my way around C++ pretty well, but not so much the LMMS codebase ) - just wanted to check that this was an acceptable suggestion.

NOTE : I'm aware of #4587 and the Python discussion. Therefore, I'm wondering whether people think a problem would arise from such a dependency addition ?

Please search the issue tracker for existing feature requests before submitting your own.

• I have searched all existing issues and confirmed that this is not a duplicate.

Start by reading issue #735 and the Python discussion in #4587, then inspect LMMS’s existing sample-track and right-click entry points. Clarify the proposed Demucs integration and dependency implications before defining the implementation scope. Done means an agreed approach and a feature that creates new sample tracks for the separated stems.