Syncing Audio to Footage - HELP!

[Ravi Kiran]

We need a solution to our problem syncing the audio files to our video files. We recorded audio on a separate sound recorder, which was providing Time-of-Day timecode to the Red. Normally the camera would roll, then a few seconds later the audio would roll.

Our problem is, we're manually syncing every audio clip to every video clip! Because we used TOD we don't have unique timecode for every clip. Is there an easier way to sync all the audio files to the video files?

 

[Tom Visser]

http://www.robgalbraith.com/bins/content_page.asp?cid=7-10040-10109
http://www.singularsoftware.com/

I haven't tried it myself, but hear that it works pretty good.


Some devices TOD timecode can utilize the date information too, is this not the case with the material you did.

It won't help now, but some suggestions...
Use the slate mic feature of your recorder / mixer if it has it, if not, at least have the DIT, AC, or whoever is slating the scene to at least verbally call out the scene and take information while the recorder is rolling so post can quickly reference this info.
...but then it sounds like you might not have been slating at all?
Check out basic guides to film making and proper production methods with respect to slating and notations.

 

[Morten Eriksen]

We use Sync'N'Link for a feature film at the moment

http://www.assistedediting.com/Sync-N-Link/

Morten

 

[Ravi Kiran]

It won't help now, but some suggestions...
Use the slate mic feature of your recorder / mixer if it has it, if not, at least have the DIT, AC, or whoever is slating the scene to at least verbally call out the scene and take information while the recorder is rolling so post can quickly reference this info.
...but then it sounds like you might not have been slating at all?
Check out basic guides to film making and proper production methods with respect to slating and notations.

We did slate. The sound guys have to sync each audio clip with the corresponding video clip using the slate, though I don't think anyone verbally called out the scene and take numbers. Thanks for the suggestions, guys, I will look into them today.

 

[Chris Pickle]

We did slate. The sound guys have to sync each audio clip with the corresponding video clip using the slate, though I don't think anyone verbally called out the scene and take numbers. Thanks for the suggestions, guys, I will look into them today.

Welcome to the suck!

It's much harder than most people realize, especially on long-form. I swear that the folks that say it's easy have never done it.

We: Shot Red and recorded audio to either an Edirol or sometimes a protools rig on a cart. Up to 6 channels total.

We down converted Red to 720 pro-res. Gave the entire 720 project drive to audio guys.

They matched the audio to the video in a pro-tools project. They then re-exported the audio to exact frame run-time for it's corresponding video clip.

A fellow reduser wrote an automator to match the video and audio from 2 folders, open QT and re-save the video with the audio referenced. We could have exported the video to self-contained but that would have been another step and I couldn't get a straight answer anywhere if the self-contained would be lossless.

We put the folders on FCP project drive and imported just the video folder. When a clip gets dragged on the timeline, the QT file references the audio and brings the audio to the timeline as if it was a self-contained file.

We weren't sure if this reference system would cause any extra process power, but it seemed to work fine. Although when we put the drive on an old G5 tower, it chugged bad when there was lots of audio layers. Not sure if it was the reference issue or just because it was not intel.

Overall, this method blows. I wouldn't do it again, and if low budget I would patch the audio through a levels board and try to record staight into camera. (depending on how many audio tracks)

Hope that helps a little.

Chris

 

[Ravi Kiran]

Okay, so after looking at the files yesterday, here are some findings:

We recorded audio in BWF format at 24fps, which was what the Red was running at. We tried using the Sync-N-Link demo but the audio would always be off from the video by some random number of seconds. While scrubbing through the audio we noticed the timecode was being read at 30fps, and the logging info confirmed this. I guess Sync-N-Link was trying to fit a square peg into a round hole (24fps video with what it read as 30fps audio)

Accidentally, while importing a mono BWF file we saw the logging info was 24, but FCP would read polyphonic files at 30. A little research dug up that FCP has a problem reading polyphonic BWF files.

BWF2XML converts the BWF files into QTs (or .wavs), which FCP had no problem reading at 24. The program lets us choose the timebase. Initially we thought we had to re-export every poly file into multiple .wavs. Using BWF2XML with Sync-N-Link gave us the results we wanted. The audio files on top and camera audio on bottom, perfectly synced.

Now I have another question. After using the Merge Clips function in FCP, is there a way to replace the original clip with the merged clip? The new merged clip is a reference (with "merged" added to the file name) and not a newly rendered clip. If we can re-link these clips in FCP without using Sync-N-Link that would be great.

 

[Ravi Kiran]

Update: Sync-N-Link works great! We had some issues, but we talked to a very nice man named Greg from the company who responded to all our queries very quickly. It took our sound designer only a few hours to do all the syncing an now he's off and running. Thanks for the help!