Corrupt FFF files - shooting to card on H3D39II

Previous topic - Next topic

0 Members and 1 Guest are viewing this topic.

Eoin

I'm just processing some images from a recent shoot and appear to have 6 corrupt files towards the end of the batch. I was shooting to card on a H3D39-II. The card was a brand new Sandisk ultra which had been formatted (on the camera) just before I started shooting.

The raw 3FR files were copied from the card to computer as normal using a card reader, then imported via the latest version of Phocus (2.6.1) - turning them into FFF files.

The interesting thing is that the thumbnail image on the 3FR file shows the image intact, with no corruption at all. This leads me to believe these shots may be salvageable.

See the attached examples for what I mean when I say corruption. This occurs in the FFF file and subsequently anything I export to - such as DNG, JPG, TIFF. As far as I'm aware 3FR is not a format which can be edited as is - so needs to be imported to Phocus first, before saving as another workable format such as those just mentioned - but perhaps I'm wrong on this?

Has anyone experienced similar issues with files when shooting to card on this system? I usually shoot tethered so have not encountered it before.

The only thing I can think of, which may have caused this in the first place, is the fact that I deleted some images on the camera as the shoot progressed (using the approval system - yellow dots) - which in turn may have lead to some file system error - but on a brand new, freshly formatted card? Seems like a major flaw if it is the case.

Any ideas greatly appreciated - I really don't want to lose a couple of these shots.

Here are two examples...

http://img695.imageshack.us/i/a0000547.jpg/
http://img860.imageshack.us/i/a0000570.jpg/


Thanks

HLArt

#1
It seems like a writemistake from cam to card. Is on your cam the newest firmware installed? Have you tested to open the 3fr-files with adobe cameraraw? Is there the same error when shown the 3fr-files in adobe, as in the fff-files in Phocus?  the fff-files can edit only in Phocus, the 3fr-files from the H3DII-39 can edit in Adobe Camera Raw.

Greg

I had similar looking corrupt files on my CFV16.  The back had to go on holiday to Europe for a few weeks.

Greg

NickT

The reason the thumbnail looked oK is exactly that, it is only a preview file. The corruption could be a number of things up to and including a faulty DSP in the back itself. I don't think it's a file system error caused by deleting in camera. I would suggest trying to copy the files across again to ule out a read/write error. Try copying direct from the camera and also with a card reader, also try a couple of different destinations to rule out write errors.
Let us know how you get on.
Nick-T
Nick-T typing at you from Flexframe's secret location under a Volcano

Eoin

Quote from: HLArt on May 02, 2011, 03:27:17 AM
It seems like a writemistake from cam to card. Is on your cam the newest firmware installed? Have you tested to open the 3fr-files with adobe cameraraw? Is there the same error when shown the 3fr-files in adobe, as in the fff-files in Phocus?  the fff-files can edit only in Phocus, the 3fr-files from the H3DII-39 can edit in Adobe Camera Raw.

Yes the newest firmware is installed on the camera. Opening the 3FR file in Adobe Camera Raw shows the same corruption - same as the FFF file in Phocus.

Eoin

Thanks for all your input. It does now seem that this was an error made between camera and card when the file was originally being written.
Still it doesn't make sense that the thumbnail preview was generated without error. Well, maybe it does to someone....!


Eoin

Quote from: NickT on May 02, 2011, 08:34:48 AM
The reason the thumbnail looked oK is exactly that, it is only a preview file. The corruption could be a number of things up to and including a faulty DSP in the back itself. I don't think it's a file system error caused by deleting in camera. I would suggest trying to copy the files across again to ule out a read/write error. Try copying direct from the camera and also with a card reader, also try a couple of different destinations to rule out write errors.
Let us know how you get on.
Nick-T

Hi Nick,

Unfortunately I've tried all of those things and am still coming up with the same result. Looks like the shots are lost.

Considering the corrupt images are all towards the end of the batch, I have a feeling it was caused by me deleting the images as I went, in-camera, to make space for a few more shots. Perhaps the file system didn't manage the deletion and re-writing of data as it should have, leading to the corruption.

I would usually never delete anything on the camera like that under normal circumstances, but of course I was caught out here without a spare CF card (2 full ones in the bag, no laptop to transfer!).

Lesson learned, painful as it is.

DavidH

Hi Eoin,

Sorry to hear that you lost images.

So the problem only happened on a new CF card and not when firewire connected to the computer.

Has it happened since that time on your older CF cards?

After saving the images you might want to place the CF card on a computer and delete everything below the DCIM directory then format the card in the camera then test again (and also with your other cards). That is something that you don`t want to see again.

I`ve had a DSP chip go bad. I would take the photo and the display would look good for about a second then go bad. The fix was a vacation in Denmark. Make sure that you save the problem images as it can help Hasselblad to diagnose the issue.


David


Andy Johnson-Laird

It may also be worth noting that the in-camera format, at least based on my experiences with the H3D and H4D, only reformats the CF card's directory structure and file allocation tables, thus making the cards look "empty," rather than a low-level format of the actual data blocks on the card (thus obliterating the actual image data).

So the in-camera format is is not a "deep format" that writes data to every data block.

The corruption of the images appears to be an error in the internal structure of the 3FR file rather than anything to do with the relationship between the file directory/allocation table.

As NickT suggests this is possibly an error writing to the CF card, but, and I'm speculating, it could also suggest that the CF card in question has problems.

What I would consider doing is to create a really large data file (there are various utility programs that do this--If you can tell me whether you are Mac or PC I'll try to find one for you), just slightly smaller than the CF card's capacity, and then copy this file from the computer to the CF card, thus forcing a data write to the entire CF card.

I'd then run a compare between the large file on the computer and the large file on the CF card to see if they were the same. If they are, then you have some confidence that the CF card is the culprit. You have not eliminated the connector on the CF card and the socket in the camera body--but if the body connector was bad the odds are you would have seen other CF cards with mangled images.

Let me know if you're Mac or PC and need more details.

Regards
Andy
Forensic Software Analyst : H4D-50, HTS and beaucoup HC lenses
Portland, Oregon, USA

jonathan.lipkin

I've been advised to format the card as DOS-FAT in the compter every once in a while, in addition to formatting in-camera.

Eoin

Apologies for the slack response here - many thanks David and Andy for the detailled replies. Definitely some food for thought. I'd happily devote more time to figuring out this anomaly if I were shooting to card more often, but as it is I reckon 95% of the time the camera is tethered.

Not sure what I think of that extra DOS-FAT format, Jonathan, it seems like overkill but then again...! Who knows.

Let's hope this was a one off. I lost one particularly special shot on this batch unfortunately.