algorithm? I don’t understand what you mean
Yeah, I thought SL11 was going to be better based on the marketing demos from Steinberg. The ARA2 integration is great, but 9/10 times I get considerably better result from UVR5. Why not let us use the open source algos in SL11, at least we get ARA integration then. Ex, if I just need the bassline, it can be destructive. I don’t care if it is, what I care about is clean extraction with good transients, and UVR5 gives me that, but no ARA integration.
1 Like
do you mean you used your algorithm in another program and imported the stems into SL? did you need to do much editing after using your algo?
would you be willing to do a in depth video where you would share with us deplorables your vast knowledge ?? 
1 Like
interesting…anybody actually to a direct comparison video? I’ve never tried UVR…but if it’s free…seems like something to try out…and compare with SL
For my NR purposes I’d be surprised if there is better than SL…other than the next SL version
answering my own question
Yes! My own algorithm. I spent about 1 day (roughly about 8 hours) refining. However I chose that specific audio because if you run that specific 8-bit/chiptune(Pokémon National Park theme) through any other algorithm it fails completely. The bass from that audio is a midtone bass and is actually 2 separate tones(timbres) and I merged the 2 together, however it can be unmixed to 2 different tones (if Inwanted to). Go ahead and try to run the “Pokémon National Park Theme” through any algorithm that separates stems and please share your results.
1 Like
First of all, thanks for the earlier correction from another poster. There is no compression from SL. Of course, as you suggested, I’m already STARTING with compression usually.
I did try Unchirp, but I haven’t gotten the hang of it yet. I can see how making changes to the bass will get it closer to pulling it out.
Didn’t have as much luck with the piano and other mid-frequency instruments as I had hoped. I DEFINITELY changed what I got out. Just haven’t figured out how to best utilize it.
The same company has an unmix drums plugin which I tried. Of course, while it works well, it still winds up damping out the higher level frequencies over the course of the sing.
As I said, though, this is definitely SL11. I compared an unmix in 10 and 11 and found that 11 had much cleaner separations EXCEPT you wind up with drum bleed (usually the hats) on most of the other tracks. And it doesn’t seem that the transient remover does much about getting rid of it.
If anyone else figures something out, I’d be much obliged.
BTW, the my goal may be different than many of yours. I’m trying to unmix a song so I can transcribe it. And I’d love it if I could use some of the AI transcription tools to LISTEN to the audio and create the music. There are several that are actually getting quite good at it provided I can give clean enough audio to get there.
I’m a keyboardist. I find piano is much easier. It’s practically impossible to isolate a really good instance of Rhodes or Wurlitzer out there. And FORGET about an organ. LOL
1 Like
Algorithm definition:
“In mathematics and computer science, an algorithm is a finite sequence of mathematically rigorous instructions, typically used to solve a class of specific problems or to perform a computation.[1] Algorithms are used as specifications for performing calculations and data processing. More advanced algorithms can use conditionals to divert the code execution through various routes (referred to as automated decision-making) and deduce valid inferences (referred to as automated reasoning)”
so, you have written an algorithm and are running it in SL somehow? Or standalone in some other way?
Yes! Thanks for reminding me that the transient tool needs a revamp and a overhaul. Maybe a way to minimize/control the width of the entire selection (just like with the frequency selection) where you have the option to choose how thich you want the selections (pixel size).
Not true, I’ve done this plenty of times
1 Like
After you do the first unmix, it’s rare that the job is perfect. From there, you have to go into each layer, and further separate things, creating a new layer, then merging the new layer’s extracting audio into the layer where it should ideally live. This can often be time consuming, but the results are substantially better than running unmix and trying to use the first sepration of things. Hope this helps.
1 Like