Satellite and Delay compensation

[Norad155]

With 2 identical PTHD rigs running with satellite...wouldn't it be nice if satellite linked the delay compensate on all the systems so audio that passing between them is in sync?

The only work around I found was to manually enter compensation numbers on the track that feeds audio from one system to another. And to come up with that number it's a system of trail and error recording beeps from one system to another. There doesn't seem to be any mathematical way to come up with that number. Subtracting total system delay comps doesn't yield a correct result.

Any thoughts?

Thanks!

[Bob Brown]

This should work.

How are you measuring the delay comp offset?

Do you have delay comp enabled on both systems?

If you are recording from one system to another, do you have the prefs checked for compensating for i/o when recording (check I/O Setups Output pane for that pref).

[Norad155]

Hey Rob! thanks for the reply.

I have checked to make sure delay comp is on "short" for both machines. Also compensate for input/output delay after record pass is checked for both machines as well.

Here is the test I am running.

System #2(total delay comp 39 samples) has a stereo track with a 1 frame beep at 01:00:00:00 that track feeds an Aux (with it's delay comp turned off!) that aux goes to a digital output that normals into system #1.

System #1(total delay comp 39 samples) has a stereo aux track that receives the input of system #2 and routes it to a stereo audio track that records it.

The result I get is that the beep recorded onto System #1 is 24 samples late. (it is at 01:00:00:00 + 24 samples)

I have also tried turning the delay comp on system#2 aux on which results in a recorded beep that is 42 samples late... (which makes sence because the delay comp on that aux is 18 samples.

Now... if eliminate the auxes(not a useful workflow at all, but hey it's a test)... I get a beep that is 1 sample early....

Also I am not using pre-roll or post-roll as I know there is some sort of sync issue with that.

Thanks for your help!

[Bob Brown]

Just to check...

Are you using SYNC (I/O or HD) on both systems?
Do both SYNC boxes have video reference going to them?

I'm pretty certain you are setup properly, but I want to double check.

The second test that you ran where they beeps were 1 sample out is what I would expect. We support/claim "near sample accuracy". Having the systems be within 1 to 2 samples is within our tolerance. That's actually pretty tight if you think about what is going on. There are various limitations in the hardware/firmware/software in how the SYNC reads the video reference that cause this 1 to 2 sample discrepancy. Additionally, if your systems are properly loop sync'd or driven by WordClock all of the samples are being clocked out exactly right such that you will have no pops/clicks between systems when routing digital audio, etc.

With respect to the aux tracks, it looks like you have run into a bug with delay comp and aux tracks. I'll check without our delay comp expert and see what he thinks. I seem to remember there being some issues with aux tracks or something similar.

[Norad155]

Bob,

Thanks for your response again!
Both are running sync i/o boxes and are synced to video ref.

1 sample is completely acceptable, I am going to run some more tests tomorrow AM with a little bit more complex setup(different delay comp totals on each system) to see if I can still maintain accuracy. I will let you know the results.

Thanks!

[SBP]

Bob -just a note we are seeing the same issue here.

Delay compensation on three machines all turned on all sync works fine with a very simple test session ie just tracks to outputs.

Once the sessions get bigger with AUX inputs and into more real world set ups and the delay values on the three machine gets bigger (and different) -one machine around 3000 samples one around 1000 and one around 200 samples we see audio from our playback machines being recorded early looks like by the number of samples that the record machine is showing as its sample delay number. I have a work around for current project however you might like to look into it. It is an area that could work better.

We are enjoying satellite -thank you.

All the best

Bruno

[Norad155]

Hey Bob,

Another round of tests:

same setup as above:
System 1 large delay comp (around 3000)
System 2 small delay comp (around 200)

both systems have long delay comp engine
synced to aes ref

With audio going through an aux on the output and aux on the input, consistently 24 samples late. I can compensate for this by putting in -24 samples on the input aux of system 1 and consistently it will be right on. I can add and remove plugins on both systems to change the total comp and it will still be right on. Not sure what's up with the 24 samples I have to enter manually.

If I put it right to a track on system 1 without going through an aux, the beep is way early(approx 3000 samples)

Now if system 2 has a longer total delay comp, I can't even get it close. Always about 400 samples late.

Anywho... Hope this helps track down the problems... Let me know if you have any suggestions...

Thanks!

[Bob Brown]

Originally Posted by Norad155

Another round of tests:

same setup as above:
System 1 large delay comp (around 3000)
System 2 small delay comp (around 200)

both systems have long delay comp engine
synced to aes ref

In this configuration, do you have video reference going into each system?

Is your positional reference set to Video Reference in the Session Setup Window?

[Norad155]

Originally Posted by Norad155 http://duc.digidesign.com/images/buttons/viewpost.gif

Another round of tests:

same setup as above:
System 1 large delay comp (around 3000)
System 2 small delay comp (around 200)

both systems have long delay comp engine
synced to aes ref

Above test was performed with AES as clock source reference. With positional reference set to LTC (there is no video ref option under positional reference)

I re-tested with clock source set to our house Video Reference with same results and behaviors for all tests.

Again thank you so much for your continued responses!!

[Bob Brown]

Originally Posted by Norad155

Above test was performed with AES as clock source reference. With positional reference set to LTC (there is no video ref option under positional reference)

I re-tested with clock source set to our house Video Reference with same results and behaviors for all tests.

Again thank you so much for your continued responses!!

Thanks for the clarification. As long as you have video ref coming into the system (and the Video Ref Present LED is lit), you will be aligned to the video reference signal.