Different latencies when printing

[SimonITA]

Hi Darryl, thanks for the reply.
Apologies if the video is too long. I thought I'd show all the possibilities I tried and I was doing something useful.

When I wrote "I found that there is no difference between when I insert ProQ3 plugin before the hardware and when I don't" I meant that the offset found between the original track and the printed one is identical both when I insert the plugin before 'hardware, that when I don't.
The main problem remains, but regardless of the use of the plugin.
I use the ProQ3 often and have never had any problems, in any mode (low latency/linear phase/etc.).
This problem only occurs when I use (the same) hardware on different tracks. And, as I tried to show, in a "random" way.
Thanks so much for your suggestions

[Darryl Ramm]

Simon

I do appreciate you put a lot of time into the video. I just could not deal with going though it all and working out what you are doing/what looks broken in the time I have this week. I'll try to look at it more later, maybe over the weekend.

I am fascinated and frustrated by all the latency and ADC issues in Pro Tools.

I'd love to see what is going on in that session. I'll try to reproduce this myself. But even better can you please share a stripped down session that shows this problem for you, ideally just using core Pro Tools plugins, and no need for any audio content, it's better to be working with clicks anyhow, and I'll run that up where with hardware inserts on my RME interfaces.

So to be clear to make sure I understand this, with no plugins in the session at all you do get an offset for the hardware insert but it is the same on all tracks (and all insert pairs?). But when you add a plugin before or after the hardware insert it is not not consistent on all tracks. Correct?

This problem happens with all plugins you have tested? whether they are in front or behind the insert, and the offset error is different on different tracks for the same plugin, but is the same on an individual track regardless of if that plugin is in front of or behind the insert (what happens when it is both?... but needs careful very accurate measurements).

What about plugins with zero latency? Do they show an offset? (that's a potentially interesting test). And what about a zero latency plugin ahead and behind the insert?

Please double check for any hidden tracks, and actually remove all software plugins from all inserts.

When nothing is making sense it's great to validate/confirm what you know. So again it would be great if you could measure the H/W insert latency very accurately with a click test and set that up in the +/- field when a H/W insert is in use and confirm you get sample accurate correction for the hardware inserts, that you see this on the same insert on different tracks and different inserts pairs on the same track (use the slot bridge/loopback feature if needed) and different tracks.

IIRC you have some mono and stereo tracks going through the H/W inserts, do mono vs stereo show differences? I'm just looking for any things that possibly be different and might trigger a bug. And I'd have suspicion with any aux and sends in the signal path, and/or differences in signal paths that are pure high/latency disk playback vs. some signal path that are low-latency (i.e. have active inputs). Worth seeing if things like that are triggering misbehavior.

Thanks!

[SimonITA]

Hi Darryl, thanks for the reply.

Originally Posted by Darryl Ramm

Simon

I do appreciate you put a lot of time into the video. I just could not deal with going though it all and working out what you are doing/what looks broken in the time I have this week. I'll try to look at it more later, maybe over the weekend.

I am fascinated and frustrated by all the latency and ADC issues in Pro Tools.

I'd love to see what is going on in that session. I'll try to reproduce this myself. But even better can you please share a stripped down session that shows this problem for you, ideally just using core Pro Tools plugins, and no need for any audio content, it's better to be working with clicks anyhow, and I'll run that up where with hardware inserts on my RME interfaces.

Thanks bro.

Originally Posted by Darryl Ramm

So to be clear to make sure I understand this, with no plugins in the session at all you do get an offset for the hardware insert but it is the same on all tracks (and all insert pairs?). But when you add a plugin before or after the hardware insert it is not not consistent on all tracks. Correct?

Not exactly: when I use the hardware insert (same module that I pass through each track), the offset is different for each track.
I used the same module (in the same slot) just to exclude any differences between the various slots (but I will also try this option).

Let me explain better (I hope I remember the delay values ​​precisely, but that's the concept):

- mono kick track passed through the Six CH module (inserted in slot 1 of the Cranborne) has a delay of 179 samples;

- mono bass track passed through the same Six CH module (inserted in slot 1 of the Cranborne) has a delay of 186 samples;

- stereo track of strings passed through two Six CH modules (inserted in slot 1 and in slot 2 of the Cranborne) has a delay of 600 samples.

If I check the DLY value indicated under the fader in the Pro Tools Mix window, these values ​​are always the same (I remember 2048 or something like that), at the precise moment in which I insert the hardware.

If I add the Fabfilter before the hardware (in the insert slot preceding the hardware one), only the DLY value indicated under the fader in the Pro Tools Mix window changes (if I use it in “natural phase” mode, it goes up and if I use it in “linear phase” mode rises even more), BUT the real offset that I detect after printing on a new track is the same (as if pro tools were able to perfectly compensate for the delay introduced by the plugin).
And, in fact, the offsets are the same (179 for the kick, 186 for the bass, 600 for the strings)…

I hope I have explained it better (that's why I created that long video...).
I'm sorry if I can't explain myself, but I have some difficulty with the language and often use the translator

The session that I created specifically for the video is actually very small: I only imported a mono kick track, a mono bass track, a stereo strings track and a stereo percussion track. No aux track, no send, nothing special. Very basic, precisely to exclude other unknowns.

However, I will try all the options you indicated (click print, plugin before and after the insert, etc.) and I will post the results here, this time without video

Yes, I confirm that all this is frustrating, also because other DAWs have overcome the problem centuries ago...
Anyway….


Post scriptum: I was hoping that Tope D (who, if I understood correctly, works at Avid) could also tell me something....

[SimonITA]

Ok, first of all I would like to thank Darryl who was a great help to me.

From in-depth checks, I understood that working at 44,100 with a buffer size of 1024 and using one of the modules installed into the Cranborne 500R as an insert, a latency not compensated by Pro Tools of 184 samples is obtained, i.e. 4.17 ms (also confirmed with RTL Utility).
By putting this value in H/W Insert Delay, everything aligns.
However, the latency value is not correctly indicated by PT (which simply indicates a dly of 2331 instead of 2515).
By also inserting plugins (ProQ3 linear phase for example), nothing changes: the offset is always the same.
I'll have to learn to live with this annoyance... Or I'll go back to Cubase which doesn't have this problem....

[Darryl Ramm]

And just a note to follow up, the problem here is the class compliant driver reporting incorrect latency for the Cranbourne which Pro Tools then uses to correct the H/W insert latency.

This would not happen if Cranbourne had a real properly done CoreAudio driver that reported the correct latency, or would be easier to work around if Pro Tools had a ping utility to measure H/W insert latency, like you know most other serious DAWs. But if you work out what is going on it's not hard to correct.

Attached is a screenshot from RTL Utility which Simon checked the latencies with.

Pro Tools ADC is correcting for the 2331 samples that CoreAudio is telling it the RTL latency is, but the actual latency is 2515 samples. So the signal thought the insert is still 2415-2331 = 84 samples behind where it should be. And that can be corrected in the track +/- ADC field or in this case since the delay is positive it can be corrected in the HW Insert Delay page.

Even with quality drivers from the interface manufacture you will see ADC failing to correct H/W inserts latency properly when using say ADAT or MADI connected converters where there is no way for those boxes to communicate the actual latency to the device driver and then Pro Tools. Pro Tools really just needs to have a ping capability, and has needed this for years.

The main point: When dealing with H/W inserts and latency/timing/phase problems in real sessions things can get confusing fast, start by checking this basic insert behavior with just a straight through cable/hardware loop-back. You can measure all this in Pro Tools, but RTL Utility is also a handy way to see what is going on https://oblique-audio.com/rtl-utility.php

[SimonITA]

Originally Posted by Darryl Ramm

Pro Tools really just needs to have a ping capability, and has needed this for years.

I agree. This is exactly the point