I am starting to dislike SL, because of the bad implementation of the VST-interface.
And no, I am not going to give any more details on this or that plugin. SL crashes so often and reacts so terrible to changes in plugins. The visuals of Fabfilter, for example, don’t work at all.
Other programs, like Wavelab or Cubase work much much better on this part.
As it is, it is trash!
Please make the VST-interface a proper interface for the year 2025 and leave the beta-phase.
It is plain unusable! Very bad code-quality!
Another simple question: Why can VST-plugins only be used in preview and not in realtime with all other gui-movements, like jumping from one place to another?
Year, it’s in the spectrum and there must be some calculations be done, sure. But I guess, the modern PC today is fast enough for this. It feels so unfinished as it is now.
It think you have made your dissatisfaction with SL’s offline VST handling quite clear to us, it seems to have been a favourite subject of yours in many posts this last year. 
I use SL for several hours almost everyday and I have hundreds of VST3 plugins that I use; many of them in the VST3 module in SL. Daily.
I’m very pleased to have the ability to add plugin functionality in my work with SL. That module is working offline like all the other modules in SL.
I have not managed to make any VST3 plugin crash SL for the last years.
Except for Zynaptiq’s Punch and Balance (their two latest plugins, released this year).
So I mailed Denis at Zynaptiq on the matter, they got the crash reports and pretty soon they released fixed updates.
So now everything is in order again, no more crashes due to prakt plugins. 
2 Likes
Well, just compare the usability of VST in SL with Wavelab or Cubase.
SL has the great advantage of seeing the result of a plugin almost immediately (400 ms stutter step delay, which completely sucks with EQ, delay or other stuff ).
And why not use plugins for every layer? Year, because SL is an editor, not a DAW…
Does it make any sense to not have this option? Nope!
Also, SL has so many bugs and flaws , that it just is not fun tu use.
Why is the selection process so creepy and slow, when having some more datapoints?
Why does selecting and adding make SL slower and slower?
For instance, take 1 sec, copy and add. Than we have 2 sec, copy and add, makes 4 s, 8s, 16s,… and so on.
In Audacity or Wavelab this goes smoothly without slowing down that fast. Seems the internal datahandling of SL needs some improvement on the codebasis. (I mentioned this in an earlier thread).
Why does normalization take so long?
Lots and lots of other stuffcan be said about SL.
Nope! I do not like the state SL is in, right now.
The concept is ingenious, but the (algorithmic) code quality is really bad.
Yes, SL uses the GPU for some parts, fine. But it’s not all about bling bling. The basics, the fundamentals have to be flawless at first. This is something Steinberg seems to neglect.
Imho, SL is one of those great programs, that change the way people work in their subject. It deservs better quality assuarance, maintance and support, that it might enjoy now.
When it comes to “stutter-step”, this picture should be enough to clarify the issue!
Moving the pitch combined with a delay gives a nice sounding effect, but 400ms of latency are just…plain suckz → And you know it!
This is not exeptable!
Much better in Wavelab (though the display in Supervision does not seem to look that much “professional” at all, but this is another topic)
