Audio / video sync drift

[johnnyv]

Hi all, I am mixing a one hour awards show type of program. The clients have supplied me with their locked video, and I'm lining up audio to their video. So far so good.

However, when I print an act for them to insert, there is some drift between the audio I'm sending them and their locked video.

What would likely be the source of the drift? Would it be something where their audio recording process at the show was not properly receiving the same clock as the cameras with which they are shooting the show?

I'm presuming that at such an event, where you have multiple cameras and an audio rig to capture all the mics (there are musical performances with full band...lots of mics), that it's imperative that everything be locked to the same clock source. Is that correct?

Anyone have ideas why my drift is happening? Or is it probably because of the answer I just suggested?

Thanks in advance...trying to understand more than I currently know!

John

[reichman]

There are a lot of variables here. Tell us some more technical details about frame rate, video codec, and sampling rate and maybe we can help.

Also, I always ask video editors to make me a reference video that includes audio, so you can hear what they're used to hearing in video editorial. If that's in sync, then the fault lies with your system.

[johnnyv]

Hey Nathaniel,

Thanks for your questions.

First, I'm running PTHD 12.7.x on a Mac Pro, lots of RAM, Sync I/O and a Horita Blackburst Generator (not sure whether I need to use that anymore these days, but I still do out of habit).

OK, frame rate is 29.97 drop frame.

The video to which I am editing the audio is H.264 AAC. 48k sampling rate. I'm trying to confirm from the camera crew that they were, indeed, working at straight 48k.

Yes, the locked video I've been provided has rough audio attached, and I believe it to be in sync fully (not sure that I've ever watched the entire video listening to their hideous audio track). So anyway, I was given 64 or so wav files which I have imported into my PT system. I'm trying to find out how the audio was recorded on location. A full blown PTHD rig with a Sync I/O being fed their SMPTE? A laptop running Logic or Cubase? Etc. Maybe tomorrow I'll find out these details.

The actual show was like 150 minutes, which they've cut down to 65 minutes or so. So, after they provided me with locked video, I chopped up the audio to line it up with their video. Everything was fine for about 30 minutes, then they started to notice some drift in the wav file I was sending them to drop into their video.

Which brings me to trying to figure out why. I just want to make sure that my Session Setup window is absolutely set up properly...I know there are pull-up and pull-down options, and perhaps they play a role? I've got the frame rate set to 29.97 drop. I'm clocking to Video Reference and all looks locked both in my Session setup window and on the Sync IO itself.

I hope to know more tomorrow about how it was shot and how sync was set up with the audio recording at the show itself.

Any thoughts would be GREATLY appreciated, and I will report more tomorrow once I have some answers...

Thank you again, Nate!

[Basstian]

Originally Posted by johnnyv

Hey Nathaniel,

Thanks for your questions.

First, I'm running PTHD 12.7.x on a Mac Pro, lots of RAM, Sync I/O and a Horita Blackburst Generator (not sure whether I need to use that anymore these days, but I still do out of habit).

OK, frame rate is 29.97 drop frame.

The video to which I am editing the audio is H.264 AAC. 48k sampling rate. I'm trying to confirm from the camera crew that they were, indeed, working at straight 48k.

Yes, the locked video I've been provided has rough audio attached, and I believe it to be in sync fully (not sure that I've ever watched the entire video listening to their hideous audio track). So anyway, I was given 64 or so wav files which I have imported into my PT system. I'm trying to find out how the audio was recorded on location. A full blown PTHD rig with a Sync I/O being fed their SMPTE? A laptop running Logic or Cubase? Etc. Maybe tomorrow I'll find out these details.

The actual show was like 150 minutes, which they've cut down to 65 minutes or so. So, after they provided me with locked video, I chopped up the audio to line it up with their video. Everything was fine for about 30 minutes, then they started to notice some drift in the wav file I was sending them to drop into their video.

Which brings me to trying to figure out why. I just want to make sure that my Session Setup window is absolutely set up properly...I know there are pull-up and pull-down options, and perhaps they play a role? I've got the frame rate set to 29.97 drop. I'm clocking to Video Reference and all looks locked both in my Session setup window and on the Sync IO itself.

I hope to know more tomorrow about how it was shot and how sync was set up with the audio recording at the show itself.

Any thoughts would be GREATLY appreciated, and I will report more tomorrow once I have some answers...

Thank you again, Nate!

So the sync problem only occurs on their side? In your system it is in sync until you playout?

[johnnyv]

Originally Posted by Basstian

So the sync problem only occurs on their side? In your system it is in sync until you playout?

Yes, when I load their locked video (H.264 AAC) into ProTools and then align the multitrack audio to it, it all plays fine. Everything looks great. Good lip lock.

I mix the multitrack audio.

Then I export my audio mix as a .wav file and then send it to them to import into Adobe Premier, in which they are editing the show.

When they watch the video in Premier, there are areas where my audio appears 3-4 frames ahead or behind their picture. Typically, these areas occur not in the early part of the edited show, but later (say, after 30 minutes into the show).

I wish my knowledge in this area were deeper...

Thanks for your post!

[Basstian]

Originally Posted by johnnyv

Yes, when I load their locked video (H.264 AAC) into ProTools and then align the multitrack audio to it, it all plays fine. Everything looks great. Good lip lock.

I mix the multitrack audio.

Then I export my audio mix as a .wav file and then send it to them to import into Adobe Premier, in which they are editing the show.

When they watch the video in Premier, there are areas where my audio appears 3-4 frames ahead or behind their picture. Typically, these areas occur not in the early part of the edited show, but later (say, after 30 minutes into the show).

I wish my knowledge in this area were deeper...

Thanks for your post!

Then they problem ist on their side. Sounds like they are using frame rate conversion on import.

[dr sound]

Originally Posted by johnnyv

First, I'm running PTHD 12.7.x on a Mac Pro, lots of RAM, Sync I/O and a Horita Blackburst Generator (not sure whether I need to use that anymore these days, but I still do out of habit).

OK, frame rate is 29.97 drop frame.

The video to which I am editing the audio is H.264 AAC. 48k sampling rate. I'm trying to confirm from the camera crew that they were, indeed, working at straight 48k.

Yes, the locked video I've been provided has rough audio attached, and I believe it to be in sync fully (not sure that I've ever watched the entire video listening to their hideous audio track).

Any thoughts would be GREATLY appreciated, and I will report more tomorrow once I have some answers...

29.97 ... Really? You sure it's not 23.98 ? Are you locked to Black as a reference instead of Tri Level sync?

[johnnyv]

Originally Posted by Basstian

Then they problem ist on their side. Sounds like they are using frame rate conversion on import.

I will check on that...thanks for the suggestion.

[johnnyv]

Originally Posted by dr sound

29.97 ... Really? You sure it's not 23.98 ? Are you locked to Black as a reference instead of Tri Level sync?

Marti, if I choose 23.98 in the ProTools session setup window, I get the RED lettering display in the track info area of the video track in the ProTools edit window. It's only when I change it to the info they've given me (29.97) that the display doesn't show in red.

As far as my PT setup, in the Session Setup window, I'm choosing Video Reference as my sync, and my Sync IO has a Horita black burst gen plugged into Video Reference on the back of the Sync IO.

What I'm trying to find out from the camera crew is whether there is any pull-up and pull-down that I should be choosing in the PT Session Setup window.

But maybe that's not the problem at all...

[minister]

Don't use H264 in PT, use ProRes LT or DNxHD 36 instead.

I just noticed you said this:

Originally Posted by johnnyv

The actual show was like 150 minutes, which they've cut down to 65 minutes or so. So, after they provided me with locked video, I chopped up the audio to line it up with their video. Everything was fine for about 30 minutes, then they started to notice some drift in the wav file I was sending them to drop into their video.

If it is a a multi-camera show, they should make sure that the settings for the camera used after 30 minutes were crystal locked.

If they are using Final Cut 7, then there is a bug that you should be aware of: if their capture settings when they created the project were not the same as the project, Final Cut likes to "help" you by pulling the audio to the capture settings, not the project settings. But that doesn't account for : in sync for 30, then drift.