Hello all,
Have tried to some extent to separate them (M/S, Instruments, Layers) but the result is not optimum because of the interleaved frequencies.
Just wanted to know if the programmers are looking into integrating this function for the future releases.
OTOH, if they are not , to let us know how we can go around using the already existing tools.
BTW, I ALWAYS use the HIGH quality option when Unmixing Songs, takes a long time but it’s definitely worth it! 
Saludos!
It’s not a feature to be introduced….it’s a capability not yet invented by any underlying code on the planet 
I give it another 3-5 years for the demix-instruments-consisting-of-like-frequencies to get somewhat nailed by the thousands of github guys out there working for solutions.
That being said…..life is often full of surprises.
+1 or maybe +10
Again, I expect machine learning to learn and then to regenerate the parts….I still struggle to comprehend how unmixing layered/ dbl-tracked guitars and keys or same instruments playing harmony parts will ever get “unmixed”
but using multitudes of oscillators to regurgitate what has been scanned seems entirely possible…just not you in existence other than humans learning parts and regurgitating manually
I have not tried those models myself, but lots of others from this site with very good results.
You can try them yourself for free…
well, acoustic guitar harmonics and basic tones are quite different from electric guitars, obviously, CLEAN el gtrs are very similar to steel string ones, and nylon string gtr are a totally different beast, so I guess that the wizards/programmers can work with that to begin with…just saying…
ha …well….I suppose you can figure you’re telling all the thousands of GitHub intense-nerd-code-programming guys something they don’t already know
