adjusting recorded audio by system latency?

[Tweakster]

Hi guys. May sound like a total newb questions but it's nagging in my mind. I made a post yesterday about plug-in induced dealay and while it seems similar this question is still different so I didn't want to mix the two.

Say, I am recording a guitar track "dry" and monitor "wet" in PT, my amp sim (Pod Farm, Sans Amp, etc.) is in inserted in the audio track I am recording onto. I have 512 samples setting in my H/W buffer in Playback Engine - which is equivalent to 11,61 ms ( 512 / 44100 ) * 1000 with 44,1kHz setting, and I am recording with a click (and a drum loop - no effects!!!) on.

Now the questions that pop up in my mind:
1. I should here the click and the loop without any delay - correct?
2. As I play along (let's presume I am very tight and accurate) the dry signal should get recorded exactly alligned with the beat - correct?
3. But now I hear myself play with 11,61 ms delay and while that might not be noticeable - am I guessing it rigth from purely technical point: the recorded sample would be better alligned with the drum loop/click if I nudge the recorded region back by 512 samples backward (or by 256 to make it 50/50)?

I know this all may seem crazy - but I was using Cubase before getting into PT LE and I remember reading that I didn't have to worry about that and plug-in delay compensation as it was all taken care of by Cubase and since PT LE does not compensate for plugin delay automatically maybe there is also no compensation for the system (audio engine) latency when recording and monitoring via PT LE.

I'd very much appreciate your expert advise on this

[TOM@METRO]

You should hear audio exactly as you recorded it. If play behind for any reason, monitoring or whatever, it will play back that way and you should nudge accordingly. Why don’t you record at a lower buffer setting?

[Tweakster]

Thank you for your fast response !!!

It seems my Intel MacBook (2Ghz core duo) is pushed a bit hard at lower settings. So I am a bit reluctant to switch it from 512 to 256 samples. And 11 ms is not too much.

I asked that as a purely technical question - not that I noticed this sort of delay - just part of my getting to know PT . Say, if I use a really powerful plugin that would add to the monitoring latency. This could become an issue. I cannot use input direct monitoring as I play with distortion (as plugin) and and I don't have a hardware mixer to route for direct monitoring (and using a real pedal purely for monitoring). I should buy a small mixer though.

May I ask you to confirm whether I understand your response correctly - you are saying PT 8 LE does compensate for the system latency when recording - i.e the audio is "pushed back" automatically when it is placed on the track in the edit view after a take is recorded - correct?

Or you're saying that if I play it tight with the click/drum loop then the recorded take would not need to be nudged backward (neither automatically nor manually) and it doesn't matter that I get to hear what I am playing while recording with 11 ms delay (or lower)?

Thanks for your help !!!

[TOM@METRO]

The latter.

[AndyJCP]

Tweakster,

you don't have to compensate for the buffer delay (512 samples in your case). PT (even LE) does this automatically.

You may have to compensate for the delay induced by your amp sim or other plugins (if there is any).

Regards,

Andreas

[RiF]

So just to make sure (because I am coming from the civilized world where plugin compensation is quite natural for years):
Scenario A:

• I am using zero-sample-latency plugins ONLY.
• My hardware-buffer setting is fairly high (like 1024 samples)
• I monitor the material that I am recording in real time through my external mixing console (no PT or audio-interface monitoring, my rec-armed tracks in PT are muted).

Assuming that my recording performance's timing is sample-accurate tight, which is played to the playback given to me by Pro Tools, I do not have to nudge or compensate for anything and it is played back exactly where I played it during recording, right?

Scenario B (just a negative test):

• Same as A
• I have a plugin on my master bus that introduces 500 samples of latency.

This would delay the playback by 500 samples without PT "knowing" of the delay (OK, we all know it "knows", but it just doesn't care) , so I play/record to this delayed playback and when I play it back, my recording will be 500 samples behind, right?

If the above is right, I can derive these rules for me:

• Do not record with latency-introducing plugins activated (muting alone does not remove the latency, btw).
• It does not matter at which hardware-buffer setting I record, as long as I monitor externally (hardware-buffer independent), the recording will be aligned perfectly.
• If I do not monitor externally but run the signal through my audio interface, the monitored signal will be delayed by the hardware-buffer setting's amount of samples. A low setting (64-128 samples) will probably be less of a problem, I higher setting (256+) might (or will) be.

Did I get something wrong?

[AndyJCP]

Hi RiF,

you're perfectly right.
I made the experiment (your scenarios A and B, see this thread http://duc.digidesign.com/showthread.php?t=235056) and got exactly the results you mention.

One comment to your "rule 3": It's usually ok with 128 samples. Most of the musicians I recorded (not so many though) did not complain about the slightly delayed monitoring signal.

Regards,

Andy

[mikecpiazza]

If you are still in disbelief regarding the auto delay compensation, here is a test you can try. Create a click track and second track for recording. Now, record the click track to the second track from your monitors with a mic. Playback and listen for delay, try at different buffer settings. You can also try with a third track and record the audio from track two, this way you can see the delay, if any, in the edit window... MikeP

[RiF]

Thanx AndyJCP and mikecpiazza, I haven't been 100% sure about the delay thingy. Still waiting for ADC to happen in PTLE...

[TOM@METRO]

Who isn't?