BUG: Match alternates out of Sync to AAF

[dieffe]

Originally Posted by Frank Kruse

Thanks for the update guys. Luckily these machines are rather rare. At least in my hemisphere. YMMV.

Wouldn't hurt to notify Nagra and Sonosax about this if you are an owner of their gear.

Frank.

Unfortunately no, I only received the pd for an upcoming project, and it's the first time that I got audio files from the Nagra VI.
And btw, is there any way to fix this issue (maybe inside mediacomposer, maybe with a batch import of the corrected files?) other than manually fix the sync for every single region in pro tools?

Davide

[Frank Kruse]

Originally Posted by dieffe

Unfortunately no, I only received the pd for an upcoming project, and it's the first time that I got audio files from the Nagra VI.
And btw, is there any way to fix this issue (maybe inside mediacomposer, maybe with a batch import of the corrected files?) other than manually fix the sync for every single region in pro tools?

Davide

You could batch-truncate all files to full frames BEFORE they load them into the MC. But I have no idea if there's a programm that does this while being fully bext and iXML-transparent.
Other than that there is no way. Thing is: Once it's in the AVID the sync is already off if they use tc-slates or auto-sync.
So the AAF won't be a sync reference anymore from the start. It's a huge dilemma. Chicken vs. Egg situation.


Frank.

[dieffe]

Originally Posted by Frank Kruse

You could batch-truncate all files to full frames BEFORE they load them into the MC. But I have no idea if there's a programm that does this while being fully bext and iXML-transparent.
Other than that there is no way. Thing is: Once it's in the AVID the sync is already off if they use tc-slates or auto-sync.
So the AAF won't be a sync reference anymore from the start. It's a huge dilemma. Chicken vs. Egg situation.


Frank.

Thanks Frank. I checked with the assistant editor, and she's syncing the dailies manually, with the slate, so no autosync error inside mediacomposer.
Do you know a software that can batch truncate sound files to full frame?

Anyway, I fear that's it's too late and we'll need to check the sync of the expanded Pro Tools audio tracks against the aaf and the picture.

Davide

[octagondd]

Originally Posted by Frank Kruse

Thing is: Once it's in the AVID the sync is already off if they use tc-slates or auto-sync.
So the AAF won't be a sync reference anymore from the start. It's a huge dilemma. Chicken vs. Egg situation.


Frank.

So, what you are telling me is, all the AAF/OMFs we have received for the last few years that use a file based workflow, from AVIDS and did not use SD or Cantar recorders, have been out of sync from the picture edit bay. Since there is so little time and budget, Television Post sound has almost always considered the AAF/OMF or guide created from those same tracks as sync reference, so basically every clip aired since has been slightly off by up to 2001 samples. Obviously sync is subjective and there is the whole clapper issue and monitor/electronic processing delays, but this just blows my mind.

Question, since the sync is off in the AVID, when we use the Field Recorder workflow to assemble the poly files, are they actually correct for sync since they do not sync with the AAF tracks?

[ErikGuldager]

I may be wrong but below are my current conclusions on the subject.

As long as the Avid is running a full MXF project the audio files themselves will be in sync with the avid. That's what the "fillers" are for.
And as long as the AAF is exported as a MXF AAF it will still be in sync when it gets into the daw. Now the only way to get the match alternates to stay in sync is to convert the AAF session to bwav before matching. But that will in turn remove some metadata that is stored in the MXF container.

It's quite a mess.
As stated earlier only some machines do not start recording on a frame edge, but if you don't sync up often enough it will still be off even with a cantar. But then it will get even worse in a way, because everything will look and work ok all through the chain, but in reality any drift between the cantar and the camera frame edge is likely to be ignored. Thus the auto or clapper synced audio even in the avid will be up to a frame out of sync.

[ErikGuldager]

But I doubt the cantar will drift a frame over a day, not so sure about all the different cameras though.

[Frank Kruse]

Originally Posted by octagondd

So, what you are telling me is, all the AAF/OMFs we have received for the last few years that use a file based workflow, from AVIDS and did not use SD or Cantar recorders, have been out of sync from the picture edit bay. Since there is so little time and budget, Television Post sound has almost always considered the AAF/OMF or guide created from those same tracks as sync reference, so basically every clip aired since has been slightly off by up to 2001 samples. Obviously sync is subjective and there is the whole clapper issue and monitor/electronic processing delays, but this just blows my mind.

Question, since the sync is off in the AVID, when we use the Field Recorder workflow to assemble the poly files, are they actually correct for sync since they do not sync with the AAF tracks?

What I'm saying is that there is a flaw in the workflow. It's a bit of a chicken and egg situation.
This has nothing to do with subjective sync. The issue I mention is a technical thing that will always happen no matter how tight or lose your interlock between camera and sound recorder. It also has nothing to do with drift or any other sync-issues that might occur on top of what I've posted earlier. And even if there is drift the mentioned problem will be added on top.

Fact is that certain machines (Deva, Sonosax, Fostex, Nagra and possibly more) do not time-stamp their files on full frames (or ideally 00-frames) which leads to this problem because on one side they use TC as an interlocking grid on the other side they do not time-stamp the files on the grid they are resolving to. IMO, for a machine made to sync to picture it makes no sense whatsoever to start a file in between frames. Imagine a multi-machine setup and all of them stamping their files at random in-points "somewhere close to the next frame boundary". The AVID however assumes that everything coming in is frame-accurate. This is the dilemma.

What this leads to can be read in the earlier posting of this thread.

This issue could be solved by the recorder mfg but some of them say that it's an AVID issue. Problem is that when you expand to track that there's another point where your files are pulled out of sync relative to the AAF. This is where it gets really confusing especially on shows that used multiple different recorders.

I know that we can adjust sync manually and all of those arguments. The point is that within the whole chain there are now multiple stages where things change their sync reference randomly every time you swap out a leg of a recording that should be using the identical TC reference.

Frank.

[octagondd]

Originally Posted by Frank Kruse

What I'm saying is that there is a flaw in the workflow. It's a bit of a chicken and egg situation.
This has nothing to do with subjective sync. The issue I mention is a technical thing that will always happen no matter how tight or lose your interlock between camera and sound recorder. It also has nothing to do with drift or any other sync-issues that might occur on top of what I've posted earlier. And even if there is drift the mentioned problem will be added on top.

I hear what you are saying regarding the chicken and egg thing. I am not really concerned with who is to blame or who should make a fix, just concerned with the technical aspect you mentioned. Am I correct in saying that all the file based AAFs clips we have received from AVIDS using recorders other than Cantar and SD have been off by up to a frame? I just want to clarify that I am reading this thread and the other one on jwsoundgroup correctly.

Originally Posted by Frank Kruse

Fact is that certain machines (Deva, Sonosax, Fostex, Nagra and possibly more) do not time-stamp their files on full frames (or ideally 00-frames) which leads to this problem because on one side they use TC as an interlocking grid on the other side they do not time-stamp the files on the grid they are resolving to. IMO, for a machine made to sync to picture it makes no sense whatsoever to start a file in between frames. Imagine a multi-machine setup and all of them stamping their files at random in-points "somewhere close to the next frame boundary". The AVID however assumes that everything coming in is frame-accurate. This is the dilemma.

What this leads to can be read in the earlier posting of this thread.

As I mentioned above, I have read both threads, and am just curious if you think the FRW without using Titan actually puts the audio back into sync, since the AAF clips are out of sync due to the moving of the file to the frame edge.

Originally Posted by Frank Kruse

I know that we can adjust sync manually and all of those arguments. The point is that within the whole chain there are now multiple stages where things change their sync reference randomly every time you swap out a leg of a recording that should be using the identical TC reference.

I agree that in Post we can and should make sure things are in sync, and here is my endgame. I work in a post sound house that works on 20 or so different television shows a year. The budgets get smaller each year and we have to find ways to be more efficient. Since Post Conform disappeared, I had hoped Digi/AVID would come up with another assembly solution. Titan works great and we use it, but when I heard about the FRW back in PT 7 days, I thought, ok, this could be the new assembly method if all the metadata could be preserved in the chain. It would make assembly quicker and more accurate. Obviously there are always clips to chase down or metadata missing, etc., but from a loading time point of view, if FRW worked then I cut the loading time by like 75%. On 400-500 episodes a year, that is significant. Not only that, but sync would be perfect, or so I thought. We finally get some shows to work with the FRW but all the clips are off by varying amounts, so back to Titan to phase. No time saved.

If, as I suggest above, that the FRW without using Titan to phase is actually better sync, or possibly even actual sync, then I can now stop using Titan and move forward using the FRW and saving time and money.

[Frank Kruse]

Originally Posted by octagondd

I hear what you are saying regarding the chicken and egg thing. I am not really concerned with who is to blame or who should make a fix, just concerned with the technical aspect you mentioned. Am I correct in saying that all the file based AAFs clips we have received from AVIDS using recorders other than Cantar and SD have been off by up to a frame? I just want to clarify that I am reading this thread and the other one on jwsoundgroup correctly.



As I mentioned above, I have read both threads, and am just curious if you think the FRW without using Titan actually puts the audio back into sync, since the AAF clips are out of sync due to the moving of the file to the frame edge.



I agree that in Post we can and should make sure things are in sync, and here is my endgame. I work in a post sound house that works on 20 or so different television shows a year. The budgets get smaller each year and we have to find ways to be more efficient. Since Post Conform disappeared, I had hoped Digi/AVID would come up with another assembly solution. Titan works great and we use it, but when I heard about the FRW back in PT 7 days, I thought, ok, this could be the new assembly method if all the metadata could be preserved in the chain. It would make assembly quicker and more accurate. Obviously there are always clips to chase down or metadata missing, etc., but from a loading time point of view, if FRW worked then I cut the loading time by like 75%. On 400-500 episodes a year, that is significant. Not only that, but sync would be perfect, or so I thought. We finally get some shows to work with the FRW but all the clips are off by varying amounts, so back to Titan to phase. No time saved.

If, as I suggest above, that the FRW without using Titan to phase is actually better sync, or possibly even actual sync, then I can now stop using Titan and move forward using the FRW and saving time and money.

It depends on how they sync dailies. Over here many projects are synchronized in the AVID using clapperboards. By doing this the maintains initial sync the old-school way. If they rely on TC-sync using the TC meta-data in the AVID the files the sync will move when you use the FRW.

Which is more sync the AAF or the conform needs to be thoroughly tested further. Which I can not do at the moment. But one thing's for sure. Things randomly change sync during this workflow because of a relatively small problem at a very early stage in the food-chain. Thing is: Neither the recordists nor we can fix the cause. It has to be done by AVID and/or the recorder manufacturers.

I agree what you say about the FRW. It is a huge time-saver since it can conform an entire reel in just a few seconds without the downsides of EDLs based "Titaning".

Over on JW the consensus seems to be "stop whining and fix it in post just like you did the last 50 years." "Sync is subjective." "By the time it's on TV it will be 5 frames off anyway." which I don't find overly satisfying as a "solution" I must say. and I think mostly said because the issue isn't really understood.

Or like an editor I work with always says: "No solution, no problem". That seems to be the motto.

There's really no reason why the FRW shouldn't be bang on down to a sample these days.

Sorry, I don't have a recipie for you at the moment. No time for further tests at the moment but the cause is pretty clear. A combination of sound for picture recorders not time-stamping their files on the TC-grid and AVID MC shifting them because it assumes all incoming media is frame-accurate.

Frank.

[dieffe]

I had a discussion with Michael Maroussas and Douglas Murray a few weeks ago about this issue: Michael was kind enough to recap it in a post here.

Davide