Automation Timing issues, and Lowend

Hii to all Cubase users out here,
I would love to address an issue i hope some are aware and know a fix around the issue of timing of automation (at most when there is a 1 bar loop), as a electronic producer i always feel Cubase fall behind some features such as sidechain with volume automation, of course there are many tricks to achieve that and each has its sound, When i try to do volume automation as a ducking effect from a reverb for a Kick Rumble the timing of the automation is not precise(sounds even quite and random) not like in FL studio and Ableton which is effortlessly achieved and on point each time especially at 150bpm +, also unfortunately there is no Utility like plugin in cubase so i always attempted this with many plugins like, Pro q3, Frequency, free blue Cat-gain etc… yet result unfortunately sounds unprofessional, any one experienced this specially hard techno producers?



Best,
Sam

Hello and welcome,

I am not sure what your screenshots are supposed to show. At least the Cubase one does not show volume automation (of the volume fader) , it shows automation of a plugin. That is a big difference. Whereas real volume automation in C12 is sample accurate (well, you can configure it how sample accurate it is), the automation precision of a plugin depends on the buffer size and the plugin itself, how it handles automation data.

Also, the screen shots all show very different automation curves, and it seems like you are auromating different plugins, which just might behave differently (e.g. different scaling of gain), so from that I cannot understand how you come to the conclusion that Cubase sounds “unprofessional” or “random”? How can you compare if the settings are different?

There might be one thing which could possibly affect automation precision. Unlike Ableton (not sure about FL), Cubase does have a hybrid engine with ASIOGuard, meaning it does have extra bigger buffers for all tracks not in the realtime path (i.e.: record enabled). But I am speculating here, I have no idea whether the additional ASIOGuard buffer actually has an influence on automation precision, and if so, it should only be on playback, not on render.

There actually were some reproducible automation issues when the buffer size of your audio interface is set to 2048 samples or more, so I don’t recommend doing that.

TLDR: if you require maximum precision for volume automation, automate the channel fader, not some third party plugins. Set your audio interface buffer to 512 or lower.

Hey Fese,
thank you for taking the time to answer,
The screenshots show the automation of the volume of the plugins(just to get a sense) the curves are slightly different because i was trying to replicate FL curve sound in both Live and Cubase , since Cubase don’t have a Utility like plugin which have a gain knob or Fruity Balance I used a third party plugin volume to achieve the effect in the insert chain as it has a specific order, so unfortunately using the channels volume isn’t something i want to have as it deliver different sound, regarding Cubase sounding unprofessional for the automation reading result not in General :slight_smile:

So you suggest working and bouncing at a lower buffers to achieve a precise automation reading so the reverb don’t sound chopped or random as it follows the automation curve setting in the plugin?, because if i bounce at high buffer its going to sound chopped in wave file as well :frowning:
and then i assume bounce them because at some point must increase the buffer rate ?

what your thoughts regarding ASIOGuard and Multi processing should the kept off?

ASIOguard and multiprocessing should be on, generally. Especially if you have higher core count on your CPUs (i think without ASIOguard, Cubase uses max 12 cores or so, or 14, don’t remember the number).
The audio interface buffer size can be kept lower, because of ASIOguard. I have mine at 128 permanently, which works for me when playing in instruments or monitoring through Cubase.

BTW, I just did some tests, and there can be quite some differences in plugins with automation:

Test: DMG TrackControl, Acon Digital Equalize (out gain) and Cubase volume automation, each going from their lowest to their 0dB value. As you see, there are quite some differences, in the scaling as well as in the rendered result: TC does the most smoothing, Equalize is quite good, channel volume automation (as to be expected) is the best. All are perfectly fine, though, and no artifacts can be seen.
So I guess whatever it is that you have a problem with, it is most likely a setup problem or a problem of the plugin…

(If anyone has already seen this - I deleted it because I made a stupid mistake…)

I got you XD, still i wish the curves where more pierces hope it get implemented soon with a Utility plugin like in live and Fl

(sorry delete the previous by mistake)

I create a volume automation and rendered the result.

The signal went from -3dB to -inf in 9 samples (48kHz). That’s 0.1875ms. To me this looks reasonably fast. Faster than anything, the human brain could perceive.
grafik

I kinda agree, there is something fishy going on, at least with automation of VST3 plugins. I actually got different renders when I started Cubase in safe mode with user prefs disabled, like this:

That is somewhat uncool imho… (unless I made further mistakes, which I would not rule out)

And then, if I draw the automation in precise 90° angles, with a VST3 plugins (Acon Equalize), it render consistenly even, but if I draw a slope, the results are inconsistent:

I cannot say whether that is a problem of the VST3 plugin or Cubase, though…

Volume fader automation is absolutely precise, though.

BTW, did you do the test in the other DAWs with VST2 or VST3 plugins? The automation behavior change quite a bit between those formats afaik, and most if not all third party plugins don not support the possible sample accurate automation that is theoretically possible with VST3 (though not always useful)
Also, just for a test, try to do your tests with Cubase started in safe mode with prefs disabled

Hey Johnny thanks for sharing your results, yes the volume Fader reacts pretty good but I would love to have similar results in a gain plugin inside a chain of plugins in a the track.

Cheers!

OK, I have a suspicion that this might be a problem specific to plugins. I did another test with TrackControl, and that one renders also very precisely and consistent (albeit with its own kind ow smoothing, which is fine):

So if you want consistent volume automation with a third party plugin, for the time being I recommend Track Control (and it’s free, too, yay :wink: )

100% true it can be the plugins issue of reacting like this, thats why i wish cubase had a gain dedicated plugin like the other DAWs, ill do my tests and ill update you.
Cheers

Nicee man good one :slight_smile: i appreciate your efforts.

The thing is its not good enough the smoothing wont work, i guess ill have to check for vsts alternative that will deliver exactly the result i need like in FL for example. and ill post if i find so people wont have to search.

Cheers!!

The smoothing here is like ~50ms. That should be absolutely fine for any situation.
You’d be hard pressed to find better third party plugins than those programmed by DMG. Dave knows what he is doing, the different slopes for fade in and out here are a prime example of that.

Ok, I used the gain plugin from kiloHearts (as it is free) and I see what you mean.
In this example (square wave) it takes 250 samples from -30dB (lowest value of plugin) to 0dB. At 48kHz that is 5.21ms.
Still almost impossible to perceive by the ear.

I would like to expand @fese 's opinion: The reason could be the plugin, it could maybe be the VST plugin standard or even the host. I guess we’d need to hear from a plugin developer whether sample accurate changes can be achieved in VST3.

Don’t need a developer for that, it is possible, it is one of those features that Steinberg always used to bring forward the argument that VST3 is better than VST2. The thing is that I don’t know of any third party plugin that makes use of that, mostly because last time I checked, JUCE doesn’t support it, and thus no plugins build with JUCE can do it.
Most of the time I’d argue that it doesn’t even matter, if the internal smoothing is done right (which VST2 plugins could already do), and you stick to normal buffer sizes.
Sample accurate automation can get CPU intensive…

Perfect thanks a lot man, I also found the Gain plugin from DMG working pretty good so i guess i have 2 options :wink:. Dave knows his craft for sure.
If you have another options even paid in the near future let me know!!
and still i hope steinberg makes a dedicate plugin for the upcoming event.

I was excited to test the Track Control plugin but the results where not as good :frowning:
any ideas?

Yes, that is the range and the scale of the fader in TC, it goes from -200dB or so to +18dB. Works as designed. Make sure your “lowest” point of the automation is at ~-30dB or so in TC.
Have you tried “FX Modulator” in Cubase? That seems more suited to what you’re trying to achieve…

Looking at the screenshot, it seems to me that Track Control (and likely other plugins) have a small, built in fade in and out. It also looks like that fade in/out time is constant which makes me think it’s intentional. If that “feature” is in the plugin or the host, I don’t know.

Did you run Acon Equalize in linear phase mode? I wonder if this plugin is a good test subject because of its different built in delays it uses to combat phase issues. I don’t know.
Do you get the same inconsistencies with a stock plugin, such as StudioEQ with a sloped automation curve?

No, I switched Equalize to minimum phase and deleted all EQ bands. I used it because it had a gain knob with a high range and in generals has a high quality… It just looks like there is no smoothing/averaging on the master gain - which is totally fine, it’s not the primary purpose of that parameter to serve for volume automation.
I also tested StudioEQ, but the result was really bad, also not smoothed and it produced some artifacts where it should have been quiet. I am pretty sure it doesn’t even support sample accurate automation, despite being a VST3 from Steinberg. At least not for the gain knob…

Track Control was actually the best I’ve tested, the fade in/out are intentionally differently averaged and smoothed. It works as designed.

Really interesting, never thought about how something that is so simple at first glance like volume automation can have such pitfalls :thinking: