Well, that's my theory, anyway.
I'm importing RED footage for a feature using the Red Rocket and ROCKETcineX1720. Everything is hunky dory and real-time fast. However, I have a bunch of failed imports.
-No error messages, just unreadable Quicktimes.
-All and I mean every single one is 2GB in size.
-A quick scan of all my successful imports reveal sizes under 2GB.
Anyone have any idea how to solve this? I'm logging and transferring as many files in FCP but it will waste at least a day with how slow this goes.
The same issue exists with the RA! app. When they are rendering files that end up larger than 2GB over AFP, those files end up corrupted. What "really" happens is that the quicktime api they use seems to get confused and stop making write calls. No data is being written, but the app keeps going as if everything is working. (this points to the API since it has to tell the app something is wrong).
I have a bug going with apple and I'm waiting on some debug output from Red to help with that. I suspect these issues are related. I know they also have an open thread on this with apple and are perhaps working it that way.
In the meantime, those clips can be captured locally and copied over. (I know that stinks, but it works). This does not seem to be a realtime issue. It's related to 32 bit block addresses vs 64 bit.