XMP Update destroys ACR settings in TIFF files

[Ethan Welty]

The programming team has been aware of this problem since the beginning of the year, but I wanted to post it here as well as a warning to other users.

Applying an XMP update on a TIF file in PM will destroy any Adobe Camera Raw settings embedded in that file.
Mac OS X 10.8.2
PM 4.9 (final) and 5 (build 13764)
Adobe Camera Raw 7.2.0.46

To recreate the problem:
Open a JPG file into Adobe Camera Raw. Change the settings around. Click 'Done'. Return to Photo Mechanic to set the star rating, etc. Reopen in Camera Raw. On my system, the settings are unchanged. Now, from Camera Raw, export to a TIF. Open that TIF into Camera Raw and make changes to the settings. Click 'Done'. Adjust metadata in PM. Reopen in Camera Raw. On my system, all the settings have disappeared.

This is a big issue for my (scanning) workflow, and hope a solution is released soon.

[Kirk Baker]

Ethan,

The programming team has been aware of this problem since the beginning of the year, but I wanted to post it here as well as a warning to other users.

Applying an XMP update on a TIF file in PM will destroy any Adobe Camera Raw settings embedded in that file.
Mac OS X 10.8.2
PM 4.9 and 5 (latest)

Please state your PM 5 build number.  You can view the build number in the About box.

-Kirk

[Ethan Welty]

Hi Kirk, I'm using PM 5 Build 13764.

[dennis]

Ethan,

Well I stepped through the debugger and PM is not deleting the ACR settings anymore for TIF files since I made the fix a while back.  But for some reason ACR is not showing the adjustments (after PM edits the XMP) even though they are indeed still inside the XMP for the TIFF.  So I don't know what to say.

However, if you were to have a TIFF file saved by PS like I tested with (and not saved out by ACR by Exporting), then load that TIFF file into ACR and make adjustments, then modify the TIFF file with PM (e.g. tag), then do you still see the ACR adjustments?  I do.  Something to do with the path of creating this TIFF file (which seems a bit convoluted).  Perhaps it is an ACR bug.

--dennis

[Ethan Welty]

Hi Dennis,

I wish... No, on my machine, it doesn't seem to matter what software exported or last saved the TIFF file (PM, ACR, Photoshop), the result is always the same: lost ACR adjustments following modification in PM. Should this bug report be sent to Adobe?

[dennis]

Should this bug report be sent to Adobe?

Ethan,

Not quite yet (and I'm not sure it would help anyway).  I will do some more investigation on why ACR isn't responding to its own settings.  It may be an issue with its edit history and/or some mismatched internal document ID that is triggering it.  Unfortunately Photoshop and ACR write the XMP in a condensed format and it isn't straightforward to compare XMP before/after PM has touched it.  But in stepping through the code I can see that PM is not deleting the ACR settings for TIFF files as it used to (this was before ACR added the option to edit TIFF files).  I may need to write some extra test code to diagnose what is happening.

I'll keep you posted...

--dennis

[dennis]

Ethan,

OK I think I finally figured out what is happening.  In the XMP there is an ACR setting called "AlreadyApplied".  This is normally "False".  I had thought this was only related to the crop but apparently it applies to all ACR settings.  So even though all the adjustments were still present in the XMP, if you remove "AlreadyApplied=False" then ACR assumes that they HAVE been applied and ignores its own settings.  So now PM will no longer write to this "AlreadyApplied" field and this seems to do the trick.  The differences in behavior may have been related to whether or not an image had a crop.  If the crop existed then PM would still write "AlreadyApplied=False" and then ACR settings should stick.  If PM edited a photo without a crop then it would clear this "AlreadyApplied" and the settings would "disappear".  Fun.  This will be in the next update.  I'll try to get you a build to verify (Mac according to your earlier post).

Thanks for pointing this out and sticking with us on this tricky issue!

--dennis

[Ethan Welty]

That is exciting progress. I look forward to testing out the bug fix.