-
Sync Drift-How It Was Fixed and Revelations!
I’ll try not to make this long-winded – but I thought I’d share the process I went through and the two solutions.
- Camera footage from two Canon C100s (23.98 fps AVCHD) rewrapped to MOV using EditReady, metadata maintained.
- Audio from a Sound Devices recorder (perhaps a 633 or 644)
- Camera guide track audio – 48kHz
- Audio Specs: Polyphonic WAV, sampling rate 48kHz*
- Problem: Sync drift when syncing footage in FCP X; in sync at beginning of clips, progressively and increasingly out of sync towards end of clips.
- Confirmed that storyline settings were matching.
- Confirm FCP X was reading all format/metadata correctly.
- Verified audio format in WaveAgent app.
- Tried changing timeline to solid 24FPS – still out of sync.
- Tried syncing in DaVinci 14.1.1 – still no luck.
PluralEyes 4
- Used PluralEye 4s for sync-drift correction feature
- Using FCP X sorting and show in Finder feature, I located each grouping of syncable video and audio clips and put those into PluralEyes 4 – one grouping at a time.
- PluralEyes corrected the sync drift about 85% of the time.
- Sent the multicam clips from PluralEyes 4 to FCP X
- Roadblock:Finished all of the syncing and then realized a major problem: PluralEyes mixed all of the channels of my polyphonic file and threw out the metadata. For those of you who have read some of my posting, you know I am not a fan of PluralEyes for many tested reasons, but I used it for this one sync-drift correcting feature.
Next Steps: I need those channels back. Retime tool to the rescue
- I had this idea of lining up the original audio with the PluralEyes corrected audio.
- Added an angle to multicam clips in FCP X or attached new audio to synchronized clips in FCP X.
- Lined the original audio (which has the channels and metadata) against the corrected audio from the head edge.
- You have to zoom way in on the time line to do this: Went to the end of the clip and used the retime tool on the original audio (which was typically longer than the sync-drift corrected audio) and dragged the retime tool so that the original audio tail matched the sync-drift corrected audio tail edge. Sometimes I still had to drag a little more looking at the waveform of both to match. The changes were so slight that the retime tool didn’t even register a percent change (from 100% to 99%)
- Repeat for each syncing group – yeah a lot of work.
A Friend to the Rescue.
Contacted a friend who’s more adept with all things sound – shout out to Jim Schaefer. We discussed the issues and the process.- Jim’s thoughts: A)Recordist thought we were shooting 24 fps and finishing in 23.98 so he did a trick to prevent us having to do a pull-down on the audio to match the telecine’d footage. B) He forgot to reset the sample rate.
- Jim’s revelation: There’s another setting Sound Devices has for changing the sampling to 48.048kHz. Try resetting the sample rate of the original audio to 48.048kHz in Wave Agent and see if that works.
- Tried Jim’s advice and synced the changed clip in FCP X. VOILA! IT WORKED.
Conclusion and Revelations
- All I would have had to do is change all of the audio to 48.048kHz and all of the audio (whether by FCP X automatically or me lining things up by slate) would have synced. And they did as I went back through this process; changing [a copy of] all of the original audio to 48.048kHz in Wave Agent and adding them to their appropriate multicam or synchronized clips in FCP X
- Why the recordist used this sampling rate, we don’t know.
- Why this information wasn’t communicated to us, we don’t know.
- FCPX doesn’t care about frame rate metadata on audio clips and try to adjust audio files based on FPS information.
- Time to get some Tentacle Sync devices or something similar 😉
Hope this helps.