We've run into a little bit of a conundrum here at work. We're still going to do more testing, and I'm sure we'll be able to figure it out, but I thought I'd ask here to see if anyone has experienced something similar.
The overall test was for a multi-camera shoot using a SONY EX3 XD camera and a Canon 5D DSLR. The EX3 was feeding timecode wirelessly to a Zaxcom Deva 5.8, which is jamming continuously in Auto-Load mode (automatically starts running as soon as time code starts running at the camera). We were also using a Denecke TS-C slate which was hard cabled to the time code output of the Deva. Of course, the Deva was time code stamping all of the audio.
One of the issues is that we only had one semi-useful shot from the EX3 with the slate (it wasn't facing the camera, but at least was in frame), but this made sense as we were trying to give the 5D a valid sync point (since it doesn't generate time code when it takes video).
The recorder was definitely at least 1 frame behind the EX3, maybe 2. Again, wasn't on axis, so I'm not 100%, but am fairly confident it's only 1. I can see the wireless introducing this, but it's not supposed to. Definitely more testing there.
The slate was definitely 2 frames behind the Deva. We had the close of the sticks from the 5D to match with the timestamped waveform.
So there's the initial problem. It's the first encounter. So, I already plan to do more and far more rigorous testing. We also need to test our Sound Devices 744T in the same workflow; so, hopefully that will offer up additional useful data in the experiments. In the meantime, I thought I would appeal to the collective wisdom here. I'll of course share any and all results I think might be useful to others as well.
First nugget - avoid workflows with a DSLR. They're a pain in the butt.