Finally got to do this test.
Setup
Shot with UM4.6K camera A connected to the BMVA12G7 ‘camera B’ via SDI for both clips. I recorded the audio for one clip with my Sennheiser 416 MKH connected via XLR to the UM4.6K camera feed and the other clip had the Sennheiser connected via mini-XLR directly to the BMVA12G7. Timecode was supplied to the UM4.6K via Tentacle Sync and that’s passed through to the BMVA12G7.
In DaVinci Resolve I loaded both videos and ensured my A and B videos were in sync by selecting Edit / Show Timecode Overlays. It’s important to use this tool for multiple ‘cameras’ as it quickly shows if there are any video and audio frame count differences on your Timeline.
The subject for both clips was an edited music video promo (Gordon Lightfoot Tribute) playing back on my Mac Pro with the 32” XDR display. This was a good choice as both clips show the abrupt transition from the end of one song to the beginning of the next song. Effectively this is like having the clapper board close but it’s even more precise.
With the videos aligned (while linked to their respective audio) things were decent but not perfect. I looked at the highest display resolution of the audio and found differences in the audio waveforms from the different video tracks.
With the video and audio no longer linked, I used Trim to shift audio on the UM4.6K audio track. On the first clip of Camera A, the UM4.6K audio represented the Sennheiser mic and on the second clip of Camera A, the UM4.6K audio represented the camera mics with the Camera B audio provided by the Sennheiser mic directly feeding the BMVA12G7.
For the first take: Camera A Trim / Nudge audio one FRAME left.
This might have been less than a SUBFRAME too far. Since audio is recorded at 48,000 Hz, it’s inevitable that there are tiny variations when different cameras or external recorders are compared.
For the second take: Camera A Trim / Slip Audio one SUBFRAME Reverse.
This was actually a tiny bit more than I needed but I don’t know how to Slip Audio less than a SUBFRAME.
Takeaways:
Using the external microphone plugged into the BMVA12G7 did produce a more accurate result and best if you are handing off a ProRes video to a client (but using the SDI feed from the UM4.6K was good).
In Resolve if your playback sounds like you’re recording in an auditorium, you can correct it easily in post via the Trim / Slip Audio or Nudge menu items.
What did I learn?
I had previously thought audio in the BMVA12G7 was about two frames off, but I was wrong.
It’s very close after you match the video Timecode with the respective audio Link selected so the audio moves with video.
Use the Resolve / Edit / Show Timecode Overlays to quickly show if video clips on the Timeline are not the same Timecode.
https://vimeo.com/851785873/55a6d1fabdEdit
Apologies if this link to the promo isn’t working for you as this video content is still ‘unlisted’ on Vimeo, you may need to paste the URL into your web browser or Vimeo app. I’ve asked the client to release the video to the public, but awaiting approval of the band producer.