Anyone else doing much with this?
I'm on a very limited budget, so when I found DMUCS GUI, which is open source, I got pretty excited about it, and have been having some good results, able to resurrect some very old demos, managing to extract some pretty usable vocals from some very bad recordings.
I'm also messing around with splitting out some of my favourite old commercial tracks and have been having some pretty mixed results.
Just wondering how you guys with the new Logic are finding it?, and I'm not really interested in spending money on stuff like Izotope for similar results as I am already seeing.
Obviously, the source material has a massive impact, and my expectations aren't that high, but I am doing a few null tests against the originals and am not sure what to make of the results.
An A-B listen against the original gives pretty good results, but on my last experiment, I used a few FLACs of really loud Primal Scream tracks, and DMUCS spits out WAV, which both get resampled in Reaper, and the null is pretty bad when I flip the phase on the FLAC original with FF pro Q3.
Not quite what I was expecting anyway, results are still pretty impressive - being able to get the drums and bass out of these very dense recordings.
I'll keep experimenting.
Comments
https://www.stemroller.com/
I've also got Logic for iPad
My Windows laptop doesn't have a compatible graphics card which means it can take 2-3 hours to split a track.
Logic takes seconds as I'm using an M2 iPad
I have found Stem Roller to be slightly better quality output, but it's not night and day.
Definitely would work to 'remix' the performance, it probably has more to do with the mic setup, and less to do with the recording method, obviously results may vary depending on source, but I'm very happy with Demucs Gui, and don't feel the need to spend money on the technology at this point.