[Marten Berkman]

Rafael and Michael, thank you for your feedback!

First I must share more of the situation that i discovered:

When XDCAM transfer in FCP hit a corrupted clip, it balked at all subsequent clips, giving me an error response. Thus i assumed all subsequent clips were affected. When I tried a random selection of those clips in camera, they played fine.

turns out that I was playing good clips that remained on the card. the first clip of the group that XDCAM transfer hung on would not play.

I opened XDCAM transfer independent of fcp, and found this more stable, in that i could preview a clip until it reached the corrupted point, though I still had to restart XDCAM transfer after it hit a corrupted point. I also noticed that the card registered “in use” even after XDCAM transfer was closed, was impossible to eject properly, and I had to restart the camera and restore media before executing USB transfer again.

So i thought, having a sense of where the corrupted point was, why not export only a selection of the clip before that point in XDCAM transfer. That worked. unfortunately one of the corrupted clips failed about three seconds in.

I then tried Log and transfer plug in in fcp, and was delighted to discover that at least even im portant corrupted clips were transferred successfully, though only to the corrupted point, without hanging anything. However, i was still missing important media.

(some clips would not transfer at all in log and transfer, and by situational deduction, i think these were clips where there was rapid stop/start of recording…i was still able to salvage them…read on…)

My friend Cory Trepanier had encountered a corrupted clip and his solution was to select short sections of the clip one at a time for export, which would work until a corrupted point was found. at least media was being successfully exported, albeit in pieces.

So i tried this in XDCAM transfer and was delighted to discover I could export media at the end of the clip. I did successive exports, working backwards, ie increasing the clip size until the corrupted point was hit and was no longer able to export.

So Rafael, same idea as Craig’s link that you sent me, though i am happy to say i was able to do the clip dividing in XDCAM transfer.

And thanks for the SDI suggestion, Michael, alas I do not have facilities to record SDI, though i was thinking i would be calling on a friend who does. I was also thinking doing at least an iLink out in HDV, but the material has to be recorded in SD, not HQ to do that.

Now the question remains, is this a corrupted card problem? or a problem of too rapid powering down, or recommencement of shooting? when i have time I will try to recreate this to discover if that isolates the problem.

It seems odd that a point midclip or near the beginning of the clip should be affected by powering down or recording again before the card indicator lights are green. However i am not familiar with the architecture of a clip’s data. Could it be that something recorded near the end is critical info for frames so much earlier in? or is the data recorded chronologically (except for key frames of course). perhaps this is neccessary for VBR?

A note of warning to those using log and transfer for xdcam: seeing as it can download a corrupted clip successfully up to the problem point, this may go unnoticed in a batch process until actually viewing the clip. However the downloaded clip length in your browser will indicate what was successfully captured, while the media length in the log and transfer window will indicate original shot length. scan and compare these two after log and transfer and any discrepancies will be clear.

whew what a long post for my initial problem! I hope sharing this will help anyone else finding themselves in this situation. Thanks Michael and Rafael for your help. If any light gets shed on whether this was a card issue, or a power down/stop/start issue, i’ll post it.

cheers,

marten