VST3 plugins are buggy and very slow in Spectral Layers

I was checking to see if that hadn’t tripped you up…it did me when I started with SL. It’s nice to confirm that a more powerful computer is delivering similar results to my 6 yr old machine with SL11 FFT size settings.

I tried to use Pro-Q3 in SL11 today and didn’t get anywhere, not that that mattered…that’s not how I am using SL…I’m running modules then further manual NR and voice untangling and mixing/ levelling…and SL11 is a game changer for me…I think things can only improve with SL over time

@Sunnyman can you post a video recording (with the whole screen visible) ? I remember @Joey_Kapish having a much older CPU. As shown in my video it runs pretty smooth with a CPU similar to yours - even slightly slower.

1 Like

Engaging with us positively and constructively here on our forum is undoubtedly the best way of providing your feedback. There are many Steinbergers reading and contributing to these forums and your experiences with our software are actively discussed and considered when deciding how to prioritise our development resources.

I can assure you that YouTube videos aren’t nearly as effective.

4 Likes

Noooooooooooo! Not true!

I was on the AMD 6900HX then updated to the 7940HS.

@Robin_Lobel

I believe the problem mostly revolves around selections. Try to make the selection (as I did in my video) and then preview (exactly as I did) and post your video here.

Well, year. But if those glitches would not be there, we would not have all this discussion, right?
It seems to me, that Spectral Layers is not high priority, when it comes to quality of code. Wavelab, for instance, runs much much smoother than Spectral Layers. Why is that so? Maybe a matter of priority… ?
I can also imagine, that Spectral Layers will some time in the future be completely given away or maybe merged together with wavelab.
Wavelab and Spectral Layers are the “complete solution” for audio editing and they belong together.
But that is just speculation. I love both programs, but SL… damn, so good.
Spectral Layers relates to Wavelab as the pseudospectrum relates to mere eigenvalues of a matrix.
I just wish, that SL becomes so smooth as Wavelab is today.
And pls, keep the numerical issues in sight.

@Joey_Kapish Same FFT settings, similar selection : vst3 preview2.mov - Google Drive

1 Like

@Sunnyman please open a thread about numerical issues, so this can be identified and fixed, I don’t know what issue you’re referring to here.

Can you also provide a video recording of the issue you see on your computer ? As shown in my 2 videos posted here I definitely don’t experience any issue like that on a configuration very similar to yours.

1 Like

I just tried the example shown in the videos

thanks for the KHs plugs :smiley:

Learned something there with those vids

I get similar performance to @Joey_Kapish at 8K+ FFT size except the preview falls behind RT playback

at 2048 smp I get smooth playback on a 48/16 audio file

also @Robin_Lobel in Joey’s vid he sweeps from 200Hz to 5KHz
in yours you sweep from 0Hz to 300Hz

does that no make a difference?

aren’t audio interface drivers an issue here? or is it GPU?

I’m running RME UFX+

Here we go…

Plugin

This happens with every plugin.
And exclusively in Spectral Layers, not in Wavelab, not in Audacity.

The plugin does not audible work at all with FFT-Size greater or equal than 16384
The file itself is 44,1 kHz an 16 bit.
Anyway, I do not expect any trouble with different sampling rates or bit depth. It just has to work with every setting - I mean, this should not even be worth mentioning.
When pressing “Apply”, the change will take place.
So maybe it is an issue with the sound system of windows?
Settings
But as said above, it works smotthly with other programs.

1 Like

just tried again at 8129 and crashed SL

anyway, I realize my machine is a bit weaker than y’all :slight_smile:

@Sunnyman

Again! I’m trying to help you to get to a resolution to this issue. If you want a solution to this problem, I would do what I suggested above. I’m advising you not to argue here because I’ve done it already (all you need to do is go back and search/read all the topics about optimization) with no results. There’s numerous topics (not only from me but from other users) about Spectralayers sluggish laggy performance… All you(@Sunnyman ) need to do is address the issue once and then when what you addressed is ignored then take drastic measures publicly (it works, trust me! I’ve gotten similar results when I privately addressed my concerns with certain companies and publicly spoke about my issues openly. Once it goes public then companies will try to scramble to address your issues, because they don’t want the bad publicity). If I were you(@Sunnyman) I would not entertain this topic any further and take matters elsewhere.

@Robin_Lobel

I did see the video and that’s strange that you’re not experiencing the same issues we are all here experiencing.

Hi all,
I tried many VST plugins and I don’t have the issue described.
I don’t have KHs installed though. I tested with IK Multimedia, Waves, SSL, Sonnox.

However, during the process I had a “crash” when switching the VST plug in the first slot to an IK Multimedia vst while playing the preview (it does not happen if the preview is not playing). I didn’t manage to reproduce the crash with other VST brands.

@Robin_Lobel : when SL crashes with no message, is there a dump file or log file that we can post here ?
EDIT: never mind I found the information in another thread, I will send you the dump file at divide frame.

Here is a screen shot just before the crash :

EQ-73 is in slot 1, I switched to EQ-PA and it crashed.

1 Like

@Sunnyman what I see in this video is very different from what I see in @Joey_Kapish video. Here it seems pretty clear it’s a sync issue between the playhead and the preview buffer, and not a slowness/optimization issue, as the playhead is moving smoothly and the spectrogram updated smoothly as well.
However the video lacks context for me to actually see what happened :
Can you do a video capture with the vst module also visible, and more importantly, also capturing the moment you click Preview?

Can you also provide a screenshot of Edit > Preferences > Devices? Are you using WASAPI or ASIO? If using ASIO, what happens if you use WASAPI?

1 Like

No, I can not make a video of SL with the VST-plugin, because the Alt-Window-G function seems to capture only the application-window.
The videos become quickly greater than 4mb. How to upload those?
image
When trying ASIO, than this happens
image
image
image
Since the machine has 64GB RAM, it should be enough…

1 Like

As an alternative you could film your screen with your smartphone, holding the phone with one hand and clicking Preview with your free hand ? (stay in WASAPI mode for audio)

1 Like

It looks like a syncproblem.
The machine is up to date with software-updates, though.

1 Like

Ok, I can’t see the playhead when the preview button is clicked on the video, but for some reason there seems to be a sync issue indeed. I have some hypothesis I’ll try to repro locally, and see how the sync can be enforced with the VST3 Effects module.
Do you have sync issue with other modules as well, or just this one ?

I’m using OBS Studio to screen record. I upload to Vimeo and embed here on the forum. Once you figure it out, it’s pretty easy :slight_smile:

It took me a little while to figure out how this forum tags the media embed- it is essentially automatic

I figured it out a week or so back and wrote about it in this thread:
https://forums.steinberg.net/t/groups-nested-in-groups-solo-again-after-mute-issues/926616/9?u=ctreitzell

1 Like

@Robin_Lobel
I was trying out in SL11 some VSTs I typically use

fab filter Pro-Q3 spectrum analyzer does not animate for me in when preview in SL11…

nothing animates in the UI for Pro-L either

I didn’t try DS or MB

is that expected behaviour?

many thanks

sample rate 96000
is that your typical setting?

I’m running RME drivers which I think dynamically follow output sample rate unless I set it to 32KHz manually…then can get stuck in 32KHz