Sidechain PDC bug and has carried over to 9.5

Hi there…

There is a bug in Cubase’s PDC engine when it comes to Sidechains. I have verified this now with every single plugin i possess that is vst 3 with sidechain inputs, that also has latency.

Basically, if the plugin you are using to receive the sidechain, has latency itself, it can’t compensate. For example, you have an audio track with a pad and insert a latent gate on it. You feed this gate a drum loop as it’s sidechain. It will be rhythmically out of time. If you use a zero latency gate however, there is no issue.

Also, you can put even plugins with massive delays before or after the sidechain plugin, and there is no issue. Only the sidechain plugin itself can not have latency.
I also checked that said plugins are reporting their latency correctly to cubase, and in all cases they were. I do this very easily by duplicating a track, putting the plugin on one but not on the other, and flipping phase on one. Null means plugin is reporting latency perfectly:)

If you want a really easy way to see the issue, just try for example, plug in alliance G8, and set the look ahead to 50 or 100 milliseconds, then sidechain something into it. presto.

Now, because Cubase has an exhaustive PDC engine that is usually top of the heap, i really do not think this is by design, but rather a bug.
I mean even logic, whose PDC is very flawed, does this particular thing perfectly.

If any further explanation is needed, please ask me and i’ll try clarify more. Cheers!

Hmm, surprised no one else interested in this…

would you guys like a quick video to show what I mean?

Ok it seems no one else cares about this…

So rather than start a new topic for a workaround, i will ask here…

Can i nudge a clip by milliseconds or preferably, samples in Cubase?

My aim it to say, move the audio clip feeding the sidechain back by 4096 samples, cause the plugin receiving the sidechain has 4096 samples of delay which can’t be compensated by Cubase PDC engine.

I care. I’d like a video if your offer still stands- please.

No problem… I will get it done this weekend latest :slight_smile:

Please note, on my macbook pro where C9 is installed i don’t have any audio interfaces, i just use the built in output at the moment to my iloud via 3.5mm jack…

it’s absolutely fine for composing with Vi’s which i can then transfer to my studio computer… Right now i do not want to install C9 on that one as i am still not sure if i am buying it.

Unfortunately, anything plugged into the mac output disabled the internal microphone, as of 2014 models, which sucks.

So i will be recording audio for the screencast from the macbook’s own built in speakers to the internal microphone… it won’t be the best quality, but the issue will still be clearly audible.

I’ll post a link when done.

This is HUGE! Anybody else? I believe I’m seeing this in Cubase 10, when I use Neutron 2 Dynamic EQ to carve space for other sounds, sidechained. Does no one else here get that technical to notice. It seems off slightly. I’ll report back with more details.

I am having the same issue in C10, had it in 8 and 9.5 also.
Anyone experiencing the same issue ?

I still don’t know if I’m imagining it or if it’s an actual thing.

I thought the very same thing, until I started using the offset on LFO Tool.

I’d love for some hardcore evidence with a nice step through explanation. It would ease my mind.


Has this issue since been resolved? I am unable to make it happen on Cubase 10.0.60 (but thought I had encountered it with an earlier version).

I tested with a project containing just two audio tracks, one sidechaining the other. I tested with Cubase Gate, FabFilter Pro-G, and Unfiltered Audio G8. Results were as expected when enabling lookahead: audio stayed in phase, and the gate window/envelope moved earlier in time.