It’s important to continue calling back into the plugin while it’s bypassed, even though the plugin is not in the signal path. That’s the key difference between bypass and on/off. Failure to do so leads to the very problem this thread is about 
. It used to work in earlier versions of Cubase. At this point I’m sure Steinberg understands the problem and there’s no need to belabor the point, especially since there are workarounds.
I tried several plugins
they all work the same… bypass disables the passing of audio
even these plugins with an own implementation of the bypass switch do not process audio or it is intended to do so it is not a real bypass,
like in the UAD Pultec emulation
I think this is intended behaviour and I can only guess on the reason, I think they did it to save processing power…
what you refer to is the old implementation from analog consoles where the insert swich toggles between unprocessed and processed signal
the send works always…
but after all it is not a bug, it is just different then before
some call it good, some call it bad and many even don’t care
I never missed it really…
if we talk about the feature I think it should work like:
plugin window bypass should toggle between input and output of the plugin
the insert slot in the MixConsole windows should take the plugin completly out of the chain
but nevertheless the idea to automate an inserted delay FX with the bypass is a bad idea since the delay could introduce always clicks and pops to the signal in the moment of switching, the better way would be to use a send FX instead
Just to be clear, I didn’t mean to imply that your problem is necessarily related to the bug reported in this thread. It might be, but I don’t know that.
Thanks, I understood that. I don’t know the answers to questions in either thread. I think some of the posts have good thoughts about this general issue of what is happening with Bypass. The actual under-the-hood reasons for this are way out of my league. I’ve found that discussions by users are meaningful to the coding side at times. So, perhaps these conversations will be useful. I’m glad you made me aware of the other thread and cross posted about it.
Hi,
What Cubase version do you refer to, please? Do you mean Cubase 6 as you mentioned? Are you sure, it was correct and now it’s not? Isn’t it the vice versa?
Hi,
I don’t agree with this statement (sorry). The plug-in cannot be taken out of the signal chain, otherwise the delay compensation wouldn’t be taken in account. And you are right, even if I Bypass the plug-in, the Total Channel Latency field still shows the delay of the (bypassed) plug-in.
If you bypass the plug-in, the plug-in is still feet by the signal. You can try it with Delay plug-in.
- Insert Delay plug-in.
- Bypass it.
- Playback very short signal (for example any drum hit), which the plug-in is bypassed.
→ The Delay has been feet by the signal. - Disable Bypass.
=> You can hear the Delay reflections.
if your statement would be true… the whole thread would be useless
because the “bug” is that it not feeds the plugin
and it does not
insert the Fabfilter ProQ3 or any similiar plugin with pre processing RTA
bypass plugin… no RTA
It was working correctly in Cubase 6, but it is not working correctly in Cubase 10. It says right in the user guide how it should work, so there’s no need to guess what the intended behavior is . I’ll quote the user guide below. Cubase should be processing the plugin in the background, but it’s not. That is leading to the undesirable audio artifact the OP reported.
As a plugin developer, I suppose I’m more sensitive to this issue than most people. However bypass only makes sense if Cubase continues to process the plugin while bypassed. In other words, there is a good reason why bypass should work like it is described in the user guide.
Bypassing Insert Effects
If you want to listen to the track without having it processed by a particular effect, but do not want to remove this effect completely from the insert slot, you can bypass it. A bypassed effect is still processing in the background. This allows for crackle-free comparison of the original and the processed signal.
PREREQUISITE
You have added an insert effect to an audio channel.
PROCEDURE
- In the track list, select the audio track with the insert effect that you want to bypass.
- In the Inspector, open the Inserts section, and click Bypass Insert.
RESULT
The effect is bypassed, but still processing in the background.
Hi,
Sorry tester fail.
I forgot, when I Bypass the plug-in the buffer is somehow frozen and it’s played back once I disable the Bypass.
So my steps where:
- Insert Delay plug-in.
- Play the Audio file when the plug-in is not bypassed.
- Bypass the plug-in before all reflections come.
- Hit Stop and start playback from the same position as before again (while the plug-in is bypassed).
- After the Drum hit, disable the Bypass.
=> You can hear the delay reflections.
But the reason is not because the plug-in has been feed by the playback while the plug-in was bypassed. You can still hear the reflections of the very first playback, when the plug-in was not bypassed, they are playing back now. 
Sorry…
You are right, the plug-in is not feed.
ok you’re right… it’s a bug
and it is confirmed for Nuendo 10.3 as well
Reading this thread and would like to hear a “bug report” type of post about all this if possible? It sometimes takes a few tries to describe a complex problem uncounted during normal use of the DAWs.
Random thought:
If a delay (or other), insert or send, is bypassed, whatever signal is hitting that track is still being processed by the delay, bypass state does not matter. The delay is “listening” in bypass. If you “Disable” the plug-in you’ve pulled the plug. I’m assuming it is not processing anything, but honestly don’t know for sure.
Hi,
I’m sorry, I still hesitate to report it as a bug, because I expect it’s specified like this.
Sorry for my ignorance, how does it work on the analog consoles? I can imagine it works the very same way, as Cubase does work.
See my post above that describes the specified behavior. It is a bug. Write your own test plugin to verify this if you don’t believe what everyone is telling you .
However, whether this is important enough to fix, is another matter .
Hi,
How do you know Cubase specification? Where is this specification written, please?
This is not on plug-in side, this is on Cubase side. It doesn’t make sense to write a new plug-in. All plug-ins behave the same in Cubase.
Please read the user guide for the details of the specified behavior.
Please write your own test plugin to verify a bypassed plugin is not having its process() callback called.
After you have done these two things, the problem should be obvious to you.
Hi,
Do you refer to this page? And the sentence:
The effect is bypassed, but still processing in the background.
Or do you refer to something different?
You said it’s specified somewhere, so I would like to see, where it is specified (I’m not aware of it). Then I could write the report. Thank you.
Sorry, I don’t have knowledge to be able to do so.
It’s not about my/yours feeling expectation. It’s about the specification.
Btw: Cubase specification could be different than VST plug-in specification.
I already answered that. Maybe you didn’t read my reply to you above?
Again, I don’t consider this a serious problem since there are workarounds. I’m sure that’s why this problem has gone undetected until now.
exactly, that is the way it is working in analogue consoles with switch able insert points
the insert send is always working only the return is swiched
and the page in the manual implies exactly this
so we can call it a bug
In my first porta–studio, if you un-muted the send, you’d get a nice, fat snap, about -12 down, printed onto the record track. I’m still waiting for Tascam tech support to get back to me on that. 
.
Hello
I have new findings…
with my Voxengo LFMax Punch and some Waves plugins it is working as intended
Bypass enabled still shows signal inside the plugin.
So could it be that it matters wich plugin is inserted?