External Effects delay compensation in Cubase 9

I’m using Cubase 6 and one of the things that really bothers me is the fact that the delay compensation for external effects doesn’t work (with my RME Fireface UFX)!
Can somebody confirm that this issue has been solved since then?

Thanks!

what is it that doesn’t work?
respecting a set delay, or detecting a delay?

I’ve got a RME MadiFX and both work, but you can also easily detect manually.

There hasn’t been any changes in how external fx are handled.
But works for me, RME Raydat and a mix of old ad/da

Thanks for the fast replies!

Actually I would need a negative delay for the external effect, which is not possible in Cubase 6.
For example: if I have two identical audio tracks and add an external effect on one of them (output connected to input via patchbay), I have to set the track delay to 0,39ms for the two tracks to be in phase. This means I would have to set the delay for the external effect to -0,39!

No, that’s not how it works
External fx is about round trips
Take a sample with a strong transient
Place it on the grid and record it.
Without ext fx it should record on the grid
With ext fx it has an extra rondtrip and will record BEHIND the grid
The offset it creates is a positive value you measure in samples, when you divide that by the sample rate, you get the offset in ms. This value you can enter into the connections dialogue

Remember to Not direct monitor the signal when pinging. I had totalmix (by mistake)monitoring an ext. FX. And the ping did not work until I turned it off.

It is a bit confusing how it’s done on tracks - you have to set negative there - but external FX or instruments, you only set positive since it’s used as a delay and create negative, kind of, by delaying all other tracks to it lines up.

Raphie had good example how to calibrate system.
I’d like to extend a little bit.


#1. Calibrate without any FX at all, plugins or external, first - to see if reported ASIO latency is enough.
When having pure digital interfaces, meaning it does not have AD/DA built in - it cannot report latency for anything but ASIO buffer.
So to calibrate your basic setup - you can do like this - so you also get AD/DA in there:

a) make analog loopback between analog out, any way you listen to it, and to each analog input you use - one at a time and record click/metronome.
If having some preamps/line input over ADAT to interface - use these also.
So repeat and take notes of result of this test.

b) check how it lines up with grid as recorded - since click/metronome does as it’s generated.
Take notes how that input line up. Repeat on every device - one input at a time and take notes.

Any deviation you get here is what ASIO could not report to host - and host cannot compensate for this.
So add this extra delay to ASIO settings in Cubase - then repeat record and check.

In this process you may find that certain external gear produce difference delays. Let’s say you have a couple of different ADAT preamp AD/DA equipped gear - maybe different brand or model - test each of these interfaces, one input and see if it differ lining up in Cubase.

I even discovered that my addon board Ai4S had different delay on AD compared to the two channels built in by default. And it made a difference making a lined+amped signal from guitar sound boxy when adding them while mixing. So I discovered this delay and compensated on track itself - to make it line up - and then it sounded like expected and I could mix lined and amped signal.

Anyway - if you note that some interfaces you use and those AD converters create more delay than others, you have to individually compensate on each track for those delays - since ASIO setting allow one manual entered value. ASIO itself cannot report individuall delay on each channel - I checked with RME on this - ASIO only can report to host two values, built in AD/DA and buffers. Some host I use show both - but Cubase only total.

Then we must trust Cubase to compensate for individual plugins added once system is calibrated.

I do understand what negative and positive delay mean in this regard!
If I do as you described, I end up with the second sample (recorded through external FX) being AHEAD of the grid, hence the need for a NEGATIVE delay value in the external FX dialog.

Correct larioso.
I’ve measured interfaces, as well as instruments and outboard, it’s Interesting to see how they differ between components, though most hover between the 3.5 and 5ms in my setup. (44.1kHz at 256k buffers)

If it ends up AHEAD of the grid you’re not measuring in an empty project without direct monitoring, but you have already got a bigger offset somewhere else in the project.

I AM measuring in an empty project without direct monitoring! :confused:

I do this:

  • Open a new and empty project (44,1kHz)
  • Add a sample from Media Bay on a new audio track
  • Add a group channel and route the audio track with the sample to the group channel
  • Add two audio channels and set their input as “group output” (so I can directly record the output of the group channel in Cubase)
  • Now on the first audio track I record the signal with just the described routing (“DIRECT”)
  • Then I add the external effect (output to input on patchbay) on the group channel and record its output on the second audio track (“EXTERNAL FX”)

As you can see in the attached picture the “EXTERNAL EFFECT” recording is ahead of the direct sample! When I measure the time difference I get 17 samples which correspond to 0,39ms at 44,1kHz.

So there is no way for me to compensate this delay with the external effect settings.

Edit: Larioso, you’re always writing faster than me… :laughing:

Maybe it is like you suggest - it is a bug in your version of Cubase then. I only use for external instruments and there delays are hundreds of samples and positive values generate correct result. So it’s delay for midi out to gear(nothing ASIO bothers with) - and then audio input. So it’s possible something changed since your Cubase version, I cannot tell.

But also check - make basic calibration of system first - because I think your basic system shóuld have a value added to reported ASIO delay.
So make loopback without your fx involved first - is what I would do.
And do it for one the channels that external fx use, disconnect fx and just do loopback.
And check the same for the normal mike input route that might be what ASIO is reporting for.

It may be there is a basic calibration needed - and the route your fx is connected to something that differs.
If so, you may have to set basic settings for system to shortest delay(fastest interface route) - so anything you add for external is longer and then your FX can be made lining up.

But agree in general - would not hurt if able to set negative for external gear. If that route is shorter delay than rest of system it creates a lot of extra work calibrating.

I can’t believe there is nobody else who has (or had) this problem in Cubase 6… :confused:

Larioso, could you explain what you mean with “basic calibration”?

btw.: sorry for the late reply… :blush:

This problem has been around for ages. Here we go again:

https://www.steinberg.net/forums/viewtopic.php?f=226&t=71278

I have had the same issue for many years, and I can report tha as of Cubase 11, it is still not fixed. As I understand it, the RME driver reports to cubase a basic ASIO delay of 32 samples in and 32 samples out (64 round trip). This is based on the internal AD/DA of my RME card. If I hook up and external AD/DA via ADAT connectors, the signal comes back exactly 64 samples EARLIER than the origional. This makes since because the RME driver has no idea what interface is connected to it. My solution has been to add a plugin that delays the signal 64 samples before the external FX plugin.

You do not need to use this unless you are using digital outboard of which will have its own AD/DA and processing latency which of coarse, your computer nor ADDA, nor Cubase is going to see.

It is for latency, outside of your system. If the drivers for your ADDA are done right, all your internal system latency should be accounted for automatically.

test it your self with an out and back in roundtrip DA->AD test using a sound blip.

At least, this what I have always understood, and what my tests have shown.

Like I said, I’ve been batteling this for years.
PingDelay

the image is too small, why is it so small? just upload regular size