Recording With Delay Compensation On 2.0

[ryan de topanga]

Just wanted to follow up on the informative (and passionate ;-) thread that is now closed.

In my workflow I output all my tracks to a stereo aux track (not a master fader), and that track is routed to a 192. Masterduk said that in order to compensate properly, tracks that are in record should be outputted directly to a 192. I understand this, but have a question: If I don't have any inserts on the aux input then the delay that the performer hears should be only the amount of samples that the aux track adds (usually little or none). Is this correct? If we're talking anything from 0 to 40 samples of delay then I can live with this. If it's something more than what is shown in the delay comp window for the aux track, I'd like to know why.

Thanks in advance...

[Mt.Everest]

Originally Posted by ryan de topanga

Just wanted to follow up on the informative (and passionate ;-) thread that is now closed.

In my workflow I output all my tracks to a stereo aux track (not a master fader), and that track is routed to a 192. Masterduk said that in order to compensate properly, tracks that are in record should be outputted directly to a 192. I understand this, but have a question: If I don't have any inserts on the aux input then the delay that the performer hears should be only the amount of samples that the aux track adds (usually little or none). Is this correct? If we're talking anything from 0 to 40 samples of delay then I can live with this. If it's something more than what is shown in the delay comp window for the aux track, I'd like to know why.

Thanks in advance...

First - are you using external converters or is this purely about ADC using Digi hardware for ALL your I/O, including your headphone feeds?

And also, I know ADC is ON, but are the " compensate input and output... " checkboxes checked in the I/O page?

For a quick answer, you're fine. Any delay you incur is going to be minimal. Under 20 samples. You can test this with your specific setup by sending an audio click that is right on the grid to the performers headphones, and putting those headphones right up to your mic and record that click back onto your performers track. Then you can measure the difference VISUALLY between the click track and the newly recorded click track. (In my setup it was like 22 samples or something - insignificant). You can also listen to the audio click and the newly recorded click and hear how they playback virtually in sync (give or take the few samples).

I believe the only way to have ZERO delay is if your A/D, D/A AND headphone feeds are all coming from the 192 directly.

But it really doesnt matter. Delay is inherent in sound, period. I mean, if a singer is 12 inches from the mic and does a take, then is 3 inches from the mic and does a take, there would be a few samples difference between the two takes based on the time it takes for the sound to hit the mic, but its so insignificant that its nothing to think about.

Id like to know your exact setup tho. What converters you are using (if not all 192s) and how you feed your headphones and what settings are turned on in PTs. Etc..

MT

[crizdee]

Originally Posted by ryan de topanga

Just wanted to follow up on the informative (and passionate ;-) thread that is now closed.

In my workflow I output all my tracks to a stereo aux track (not a master fader), and that track is routed to a 192. Masterduk said that in order to compensate properly, tracks that are in record should be outputted directly to a 192. I understand this, but have a question: If I don't have any inserts on the aux input then the delay that the performer hears should be only the amount of samples that the aux track adds (usually little or none). Is this correct? If we're talking anything from 0 to 40 samples of delay then I can live with this. If it's something more than what is shown in the delay comp window for the aux track, I'd like to know why.

Thanks in advance...

Hi,

As long as there are no plugins on your aux's then the delay would be almost negligible. you can manually bypass the compensation on an aux, but this would not work if there are other tacks routed to the same aux! so in your case above it would not work as your aux is your master bus and has all the other tracks routed to it. it goes without saying that you can't have any plugins on the master outputs.

I you want to be totally sure then route the audio track you are recording onto directly to your master fader for the left/right monitor path. as ONLY the record enabled track is bypassed from the delay comp when recording, so that track needs a direct route to the left right monitor path.

As always if recording with ADC on you need to select the compensate for input/output delays in the i/o setup.

Chris

[Sean Russell]

Originally Posted by crizdee

Hi,

I you want to be totally sure then route the audio track you are recording onto directly to your master fader for the left/right monitor path. as ONLY the record enabled track is bypassed from the delay comp when recording, so that track needs a direct route to the left right monitor path.

Chris

Does anyone know why we supposedly *have* to go out 1-2 and not bus 1-2 ? Is there a reason for it?

[ryan de topanga]

Thanks for the replies. First off, Mt. Everest, my i/o is is all 192’s. No third party i/io.

I did some testing at 24 bit 48K that you might be interested in. I recorded an MPC click (with a nice big spike) on an audio track (the “source” track). I then tried out several routing scenarios to see what happens and doesn’t happen. Here are the results:

Scenario #1: The source track’s output was sent through a PT buss to another audio track. With Delay Compensation on, as expected, the audio recorded on the new track was sample accurate. With delay compensation off the audio on the new track was 10 samples late.

Now, using a patch cord I connected the +4 output of channel 1 of my first 192 to the +4 input of channel 1 of my first 192. This created an analog out, analog in signal chain in order to be able to measure delay.

Scenario #2: The source track’s output was sent directly to the 192 (output #1) and the record track’s input was set to the 192 (input #1). With Delay Compensation on and i/o input and output delay compensation enabled, the resulting track was one sample early (!) With Delay Compensation on and i/o input and output delay compensation disabled the resulting track was 86 samples late. With Delay Compensation off and i/o input and output delay compensation disabled the resulting track was 102 samples late.

Scenario #3: The source track’s output was bussed to an aux input, which was set to 192 (output #1). The record track’s input was set to the 192 (input #1). With Delay Compensation on and i/o input and output delay compensation enabled, the resulting track was again one sample early (!) With Delay Compensation on and i/o input and output delay compensation disabled the resulting track was 86 samples late (exactly the same as when the source track is sent directly to a 192). With Delay Compensation off and i/o input and output delay compensation disabled the resulting track was 108 samples late—6 samples more than using the direct 192 out on the source track.

So here’s what I can see from all this. Without Delay Compensation (at 48K) a buss will add 6-10 samples of latency. When Delay Compensation is on and Input / Output Delays are compensated for (in the i/o setup) there is no timing difference between scenario 2 and scenario 3. The only thing I can think of is that if I feed the musician the same stereo mix that I hear in the control room, s/he will hear an additional 6-10 samples of delay on their input (less than .001 second). So far no one has complained about this and that’s what I intend to keep doing. Unless, of course, someone thinks I’ve got it wrong. Luckily that rarely happens on the DUC ;-)

Cheers.

PS. I still can’t figure out why Delay Comp and i/o input delay comp puts things one sample early. Could it be that all 192’s are not created equal?

[Sean Russell]

Ryan -

Thanks for the testing. It would seem not to be as big of a deal (recording with tracks feeding auxes) as I had thought (if the IO boxes are checked). Excellent. I have no idea why you'd get a sample early on those tests. Maybe it's 'over-compensating'?

[Mt.Everest]

ryan - thanks for those test results. pretty much follows what I got with mine. Aside from that I am using a 3rd party converter so in your scenario 2, you ended up with a -1 sample early (weird i know, but still amazingly accurate) and in mine I ended up being around 10 samples late.

Which is expected since the i/o compensation doesn't work precisely with the Apogee, but it still HELPS. (without the checkboxes on, its 100+ samples late). And since its consistent, and such a small difference than using the 192 as A/D, Im fine with that minor discrepancy.

I just make sure that I dont mix converters on multi-mic recordings that need to be exactly sample accurate for proper phase alignment.

[crizdee]

Originally Posted by Sean Russell

Does anyone know why we supposedly *have* to go out 1-2 and not bus 1-2 ? Is there a reason for it?

Hi,

Because aux's are not disabled from the delay comp engine when recording!!
ONLY the recording track is disabled from delay comp.

If your record track is routed through an aux your audio will no be 100% correctly placed on the time line, although with no plugins on the aux that amount of delay caused is miniscule.

It all comes down to how sample accurate you want your audio recording to be! 100% or 99%


Chris

[Sean Russell]

Originally Posted by crizdee

Hi,

Because aux's are not disabled from the delay comp engine when recording!!
ONLY the recording track is disabled from delay comp.

If your record track is routed through an aux your audio will no be 100% correctly placed on the time line, although with no plugins on the aux that amount of delay caused is miniscule.

It all comes down to how sample accurate you want your audio recording to be! 100% or 99%


Chris

thanks, Chris. I feel like I knew that ... At some point.

I don't think a few samples through an aux (even with plugs - gasp!) will affect sales.

[Mt.Everest]

Originally Posted by crizdee

Hi,

Because aux's are not disabled from the delay comp engine when recording!!
ONLY the recording track is disabled from delay comp.

If your record track is routed through an aux your audio will no be 100% correctly placed on the time line, although with no plugins on the aux that amount of delay caused is miniscule.

It all comes down to how sample accurate you want your audio recording to be! 100% or 99%


Chris

Actually this is where I think I still have a bit of confusion-

Lets say you have a session with a total delay compensation of 1000 samples. The session has a bunch of Aux-subs with plugins on em, which all feed a Master Aux, which feeds a Master Fader. (There are NO plugins on the Master Aux or Master Fader however)
Turn ADC on, everything is peachy.

Now, you create 1 more audio track and feed it to a new Aux-sub. Right away, what ever audio passes thru that new aux will be delayed 1000 samples since that is the total system delay, correct?

Then, you begin recording audio on that new track. The track itself is suspended from ADC but the Aux it feeds obviously is not. So, the singer will be singing to an output that is 1000 samples delayed, thus hearing themselves 1000 samples late in the cans, correct?

If true, lets say they can deal with that delay while monitoring while they sing.

When you disarm the record track after a pass, what exactly happens to the pass you just recorded? Will it play back in time? Will it sound 1000 samples late? Or do the i/o checkboxes compensate for this?

I think my confusion still lies in how the checkboxes in the i/o page relate to ADC in the software.

And lets assume this is all with Digidesign hardware. The verdict?

Thanks
MT