SD error, frozen cam and differences between images on the two cards

[afx]

Went to shoot a concert last night. The D750 was loaded with two 64GB Sandisk Extreme Pro cards.
When I tried to check an image I got a message that the file could not be displayed.
Wen back a few images and they showed up fine.
Took out the first card, still the same error.
Could not turn the cam off.
Took out the battery, check the images and found images that look like parts where shifted.

I went on shooting, everything fine.

Copied the images verbatim to the computer.
If it is a card error, then the two cards should show a difference on the offending images.
But all images show differences between card one and two??!!!
Of the 1010 files,
993 show a difference at offset 32641, 17 at offset 32481 which adds up to 1010.
So there is always a difference between the cards. Where does that come from?

Why are the the images on two two cards always different?
 
Sill me did not record the number of the offending images, so it will take a while before I find them.
I'll keep a set of the verbatim copies for analysis.

cheers
afx

[Bjørn Rørslett]

A bit error during transmission of the files to the cards? That usually makes parts of the image shifted or striped, or filled with garbage.

I also wonder whether Slot 1 and Slot 2 stored the exact same file version. Such as one only does RAW, the other RAW+JPG. Or their file create time differed at the second level.

Guess I really don't know other than virtually all the time, DSLR writing to their storage cards) seem to be successful in my experience. Early in the DSLR era, I had the occasional error issue when batteries got low during a transmission. That could lock up the camera and/or corrupt the card. But haven't seen that problem at least for a decade.

[Frank Fremerey]

First thought: Are these cards on the Nikon compatability list? I had similar effects with my D3 when I used Transcend Cards recommended by a friend in his D3. I then exchanged cards to the fasted biggest and most expensive on Nikons compatability list and did not have any trouble since.

Second thought: There is counterfeit Sandisk product in the market very difficult to distinguish even for experts. Are the cards genuine?

[afx]

I also wonder whether Slot 1 and Slot 2 stored the exact same file version. Such as one only does RAW, the other RAW+JPG. Or their file create time differed at the second level.

Me thinks a byte compare would be nonsense of the second card had anything but a straight backup...
And file create time could be the culprit, but then this is offset constant, and file create times are stored in seconds, not subseconds vs image create time which does have subseconds but should be identical for the backup image.

First thought: Are these cards on the Nikon compatability list?

Yes.

Second thought: There is counterfeit Sandisk product in the market very difficult to distinguish even for experts. Are the cards genuine?

Did all the usual checks when I got them in December 2014.

What puzzles me more than the error is the constant difference between master and backup...

cheers
afx

[afx]

Is there any RAW file analyzer out there that shows the file sections?
FastRawviewer and RawDigger seem to only care about image data analysis, but do not show dissected files.

cheers
afx

[golunvolo]

Out of my league problem. I had one problem with the d300 -sandisk rescue to the rescue- once and one recently. I clean up the connectors in the card and in the computer reader, no problems since.

[Andrea B.]

The Raw Digger guys would be the ones to ask about this even if their current software does not show file sections. They seem to know all there is to know about raw files.

[afx]

The Raw Digger guys would be the ones to ask about this even if their current software does not show file sections. They seem to know all there is to know about raw files.

Yes, I already asked on D4scussion and got some pointer from Iiliah.
Exiftool to the rescue;-)

> exiftool in one of "-v" modes should help.

And then the -htmlDump option with the offset makes it even more clear.

It is in the fileinfo section:

This is the diff of two files with exiftool -v4

2232c2232
< | | | 7f7c: 30 31 30 30 00 00 65 00 1e 1e 00 00 00 00 00 00 [0100..e.........]
---
> | | | 7f7c: 30 31 30 30 01 00 65 00 1e 1e 00 00 00 00 00 00 [0100..e.........]

Still no idea why ;-)

cheers
afx

[Andy]

Andreas,
I am currently travelling, so can't check by myself.

i would not be surprised, if Nikon used the liberty of makernotes in the exif section of a file, to indicate which file is the primary copy of the image and which is the second copy. I.e. using code "00" for the first copy and "01" for the second copy - and all other things are identical.

Could someone please check with a camera with dual write mode, if the two files of this other camera have also one bit difference?

Doesn't explain the frozen status, but would provide some insights in the 993 photos with differences.

As said, just a guess - but easy to find out,
Andy

[Akira]

Andreas, just a thought: didn't you shoot continuously and the data bus was overloaded, which might cause the error in writing data onto two cards simultaneously and the freezing of the camera?

[afx]

Buffer issues? Not really at 3FPS, but then, you never know....

Meanwhile Mike Braca on D4scussion pointed out that the bit is right in front of folder and file number in the fileinfo section, probably denoting the slot number.
I looked that the exiftool fileinfo description which documents only three tags:
http://www.sno.phy.queensu.ca/~phil/exiftool/TagNames/Nikon.html#FileInfo
Maybe a fourth is needed. I've posted the info in the exiftool forum.

cheers
afx