Rich,
I looked at the two files you sent and they appear to be OK until the end of the image. If you view these in Photoshop and zoom down to the bottom right corner of the image, you will see some image corruption (a block about 8 pixels tall which is how JPEG is formatted). I first noticed some corruption in the file itself in a hex editor, so then I looked at the image in Photoshop and saw the problem visually too.
About 2000 bytes at the end of the image looks like it was overwritten with RAW Bayer CCD data (e.g. from the NEF). The JPEG has no 0xFFD9 terminating marker. So somehow this image is corrupted near the EOF. Although Photo Mechanic does append its image preferences to the EOF, it only appends and doesn't grab Bayer data from somewhere else when it writes.
Photoshop must be ignoring this JPEG corruption (it often does), but PM is getting an error and refuses to show the full resolution image. I'll see if we can somehow ignore the error we are getting from JPEGLib.
I think you may have an issue with your machine (e.g. memory or disk) especially since this is happening randomly to your photos and we haven't seen this type of problem before. I would run the diagnostics from your boot CD (if you still have it).
--dennis
Dennis et al -
In addition to my previous reply (of 1/18/07... see below), I have observed that the files on my back-up DVD's show no corruption. The JPG and NEF files display properly in both Photoshop CS2 and PM4433. I do not recall the chronology of my back-ups in this case. This is what I usually do: I download my card(s) in the field via my iBook and PM4433 onto a portable LaCie Firewire external drive. I then usually burn a couple DVD's in the field. Upon return to my studio, I usually then copy the respective image folders from the portable Lacie drive onto a desktop LaCie HD.
I need to underscore that I am not sure of the chronology in this case. It might seem that the file corruption was introduced when I copied the portable drive's files to the desktop HD. However, I would really like to understand the problem and when/how it is caused.
What do you think the leading suspects might be?
I have run Disk Utilities and DiskWarrior, but nothing seemed amiss.
Not to digress, but just in case this adds a clue:
I have intermittently experienced some unusual image corruptionwhen editing in CS2. This corruption takes the form of a tiny series of dots or lines only involving a small number of pixels, creating what looks like a tiny bar code somewhere in the image. This is never there when the image is first opened in Photoshop; it is created during editing. Usually (perhaps always) it is only created in the green channel.
Best, Rich Frishman
Dennis -
Thanks for your detailed insights into the underlying causative issues. Aside from the apparent file corruption issue I appear to be experiencing in those last few bytes, I have one other problem: my thick head. I cannot fully appreciate the elegance and wisdom of your comments because "0xFFD9 terminating marker" and "RAW Bayer CCD data" are not in my lexicon. No comprendé. Sorry.
So, here are my top questions at this juncture:
1 - Are you suggesting that the NEFs are also corrupted? Having compromised RAW data would be a lot worse than damaged derivative jpgs.
2 - Are you suggesting the problem is likeliest to begin with the CF card, my download computer's RAM or disks, or my archiving computer's RAM or disk?
3 - When you suggest running a diagnostic, are you referring to something like Disk Utilities, or Norton Disk Doctor or TechToolsPro?
Feel free to contact me offline if you prefer.
Best, Rich Frishman
