I’ve seen a similar issue on the show I’m currently working on. Our workflow is all XDCAM based, and our plan was to ingest on the cheap using all U1 and U2 drives. Mount XDCAM disc, AMA, transcode for offline. The plan was then to assign tape names to each of XDCAM disc, and then use an F1600 deck to batch capture for online. The theory being, save money on ingest buy renting U1 drives for offline, but batch capture the old fashioned reliable way. What we found in testing this workflow was that our timecode was consistently off by 12 frames. Every single clip had to be slipped 12 frames. Weird things is, sometimes 12 frames forward, sometimes 12 frames backward, but always 12 frames.
In doing some investigation, we looked at the starting timecode of the transcoded clips versus the clips when linked in AMA, and indeed the timecode was off. We never modified the timecode so, this was very strange behavior.
I then tested a workflow where I decomposed my sequence for offline. I found that I couldn’t point the AMA linked clips to link to the new offline master clips, but if I made subclips out of all of those master clips, that the AMA linked clip would link to those subclips. I then transcoded those subclips to my High Res workspace, one tape at a time. With all that media now on the high res workspace, I relinked the offline master clips to that it came online perfectly accurate. to expedite making subclips out of exist master clips, I exported an ALE of the bin with my master clips, and then imported that ALE to the same bin, with the shotlog import option “Merge Events with known sources and automatically create subclips”
It’s not the most straightforward process, but I tested the workflow several times and it was perfectly accurate each time despite the straight timecode variations I was experiencing.
Los Angeles, CA
“We are the facilitators of our own creative evolution.” – Bill Hicks