Spectralayers 12 is a downgrade

SpectraLayers
21

Maybe you ought to read the original post again, clearly its about unmixing a song!
But never mind, you being a fan boy, Steinberg can do no wrong in your eyes!

22

For me it’s a downgrade too. Ara doesn’t work and I tried many solutions from users here and it stills doesn’t work.

1 Like
23

So true.

Steinberg should fix all (or almost all) existing known bugs and make a complete code review before implementing more and more features, that are at best beta.

And no! I am not willing to figure out some workarounds to get things done properly. Every function, every module should simply work as one can expect by common sense without any weird quirks or obscure side effects.

1 Like
24

nah, you just don’t provide detail and slam

I’m no SB fanboy! I guess you ain’t read any of my posts. Looks like you’ve been into SB a lot longer than I.

I know what I need SLP for and use it for that (mostly NR I cannot achieve any other way). We’ve been talking about unmixing music here quite a lot for the past 2-ish years I’ve been using SLP…I’ve made my experiences and feelings about unmixing music with SLP known here on this forum.

25

Well, its working for you, thats great, so nothing for you to see here, jog on!
Others also experiencing problems, so I know I am not alone!

1 Like
26

except you aren’t being clear

you said the “new models are worse”…I understand you to mean all the “new models”

what’s working for me? Unmix Song? Unmix Song works similar for me as it does for others…essentially not what I expected…so I do not unmix music very much

as I said : “I use SLP for particular NR work”…daily

you’ve come into this forum complaining about what all of us have complained about in the past couple years…in our various scenarios and then generally bashed all SLP “new models”…they aren’t all a downgrade

I can confirm, the SL12 unmix noisy speech module’s model is an improvement from Sl11 to SL12…Unmix Noisy Speech is a fantastic starting point in my daily work that is working for me

but yeah, sure, there are all kinds of annoyances and unexpected behaviour with SL; but for me, SL12 is waaaaaaaaaay more stable than SLP11 was…but then I do very little unmixing….mostly manual editing at the moment

27

I can only say, SLP 12 has been a great update for me and the type of work I do.

Unmix Noisy Speech has been fantastic and separation quality is much better than in SLP 11.

Unmix Soundtrack is a life saver for working on film projects. That feature alone has saved me weeks.

I don’t use Unmix Song that much, so can’t really tell if SLP 11 did a better job.

SLP 12.0.30 has been rock solid for me, but I have a RTX A6000 GPU with 48 GB of VRAM - that helps for sure. And I don’t use ARA, my DAWs are Reaper and Fairlight.

Several tools have been improved. My biggest gripe is still the user interface - there is still room for a lot of improvements.

1 Like
28

Here an example for a bug, that is really suckz when needed:

  1. Use a project of lets say 10 seconds and add a sweep for exactly 10 seconds.
  2. Now move cursor to the right (i. e. 10 second mark) and try to add another 10 seconds via Project → Insert Time (10 sec)
    → NOTHING happens!!!
  3. Now, move the cursor slighty to the left (i. e. 09:959 seconds for instance) and then add 10 seconds → this works.

So, this is just one of those bugs, that are really annoying and a sign of bad to none quality checking.
Those kind of bugs make the programm seem like a careless developer version.

Please fix this and all the other known bugs and glitches.
SL in its current state is a plethora of bugs.

Solution?

Use <= instead of <
or > instead of >=

Those kind of bugs cost me hours/days of working around.
(To be clear, this is just one example of tons of bugs to find!)
But year, maybe, I am not smart enough to appreciate all this…

29

I know I won’t get an answer to this, but how is it possible that Logic is much faster and better at stem separation than SL12? I’ve tried it with several songs, and Logic has always, ALWAYS been faster and better.

I really don’t know what to say except that it’s a shame.

30

Spectralayers is something more than just stem separation, tbh I didn’t spend money for that but for audio editing and repair. It might be not the best software for stem separation but this won’t me regret the purchase. I’m sure future versions will improve the features and will make you happy. I imagine you didn’t spend that load of money just for stem separation, did you?

3 Likes
31

As much as I disagree with you on this forum, I did recently need to comp several different takes of cicada “ambience” together in one job and ran into this. I don’t often work on sound files of different duration and I felt it had to be done this way to see the timing in the Spectral Display. I did find adding time in SLP was not as expected and user is on their own as the documentation provides no real guidance; the documentation just repeats the the UI labelling and not much more.

I can’t recall if my UIX was exactly as you describe, but certainly I had similar issues.

32

yes, you’re absolutely right, SL can do more than just stem separation. I expected it to be better in stem separation, since it’s specialized software designed specifically for this kind of thing, unlike the “free” integrated Logic tool.

2 Likes
33

Robin answered that, some weeks ago. It’s because of the multiplatform AI framework that is used in SpectraLayers. It doesn’t support the Apple hardware the same way as a native Apple software. It would require a rewrite without the framework and a separate codebase for Mac and Windows to get a speed improvement.

2 Likes