Hi Michael,
I’m not sure how the timecode was sent to the cameras, I was provided the data from multiple P2 cards on a hard drive.
When you say reference do you mean reference audio? Two of the cameras did have native camera sound turned on, the third (which was actually the master shot) did not have any audio. Audio was recorded as timecoded .WAV files to a CF card that I was also given. I had gone through and Merged these .WAV files with the silent master footage before attempting to make the multiclip. I left the other two clips alone with their crummy (but built-in) audio.
It sounds like this might account for two of the cameras being in synch some of the time when the third one isn’t. All three cameras were shot 24 NDF, so I don’t think it’s a framerate thing, and setting the inpoint at the exact same timecode (I’m not using a visual reference) on each clip produces a perfect multiclip… is it maybe because each clip isn’t starting at the exact same timecode? They’re each off by a few seconds, but I would think using the timecode to align them would completely ignore when they actually started and just fill in black for the clips that started later.
-Jim