😲The difference in quality is huge, VDJ clearly has the edge for now. I hope Serato improves quickly because it took them 2 years to launch Stems so they are late to the party
I left these two out because: - MDX had quite alot of bleeding in the INSTRUMENTS stem, so I preferred GSEP over MDX to include. - Ultimate Vocal Remover (its own models) doesn't have a way to separate into 4 stems directly. Also I was reacting to this other post (and comments) on reddit, where both weren't mentioned.
heh, look at Virtual DJ - the absolute, undisputed king shit of wild, random implementation of every potentially interesting feature the dev team comes across (they even supported Intel's RealSense depth cams at one point, probably still do). if you consider that they launched their stem separation almost simultaneously with DJay, and back then already delivered seriously good results even on a low-voltage i5 CPU, and now updated their algorithm as Serato announced they're ready to catch up after 2+ years, it's honestly impressive that they're leading the pack. it's a shame their GUI is almost irredeemably fugly - but I have to say, as a heavily disappointed longtime Traktor user, I'm seriously considering giving them their fair shot.
@@CWAent I know, their CEO is notoriously resistent to the idea of a redesign (which was necessary 20 years ago) and users have to fix it. thing is, I expect not to have to do free work for a paid piece of software. it's kind of like Bethesda actually assumes players will fix their broken shit (see: what Skyrim is today vs. in what state it was left after release). when everyone else has at the very least decent UI, this is a massive turn-off TBH.
What is the best algorithm to extract an acapella specifically concentrate on the acapella and lose all the instrumental, I have really used but not so good the subtractive equalization method called TDR NOVA is free, you should make a video with that method to extract an acapella and then pass it with artificial intelligence
We're having trouble with the standalone .app right now, so that takes a bit of time to doctor out what's causing this. Meanwhile, you can install it manually based on the description here: github.com/Anjok07/ultimatevocalremovergui/tree/uvr_5_5_MacOS If you need a video how to install manually, let me know and I'll consider.
macOS app out now! Tested on M1 + x86 (like Intel core based processor) using Catalina/Big Sur/Monterey so far. Download link in the description here: github.com/Anjok07/ultimatevocalremovergui#macos-installation
love mvsep, so much that i don't share it because it's getting flogged lately lol.. but yeah those few top algos it has are great, also found on UVR5.5, worth checking out for the new 6 stems, so glad to see guitar and piano now split from each other, but because they are also on their own tracks, they split guitar from anything synthy. can also get some pretty decent results using a UVR Karaoke model (banking the vocal result) and sending the instrumental result back through using the 'main' MDX model and keeping the vocal track there too. 2 vocal tracks, one backing. How are you finding Rx10 Bas, is it any improvement on 9 or 8? seems like RipX does better than them all right?
Tell ya what, i really expected to hear better drums come out of Lalal, shouldn't be hearing bass plucks on a drum track when they claim to have models for like 6 other things no one else has. Turns out they've gone for quantity over quality.
