Truncated IPTC tags

Started by hendrich, May 25, 2012, 04:05:12 AM

Previous topic - Next topic

hendrich

Hi all,

I use exiftool to read and manage EXIF and IPTC tags from my own (Java-based) image viewer.  This used to work fine for my needs, until I switched to a recent version of exiftool after upgrading my Linux and Windows installation.

Problem is, the new versions of exiftool enforce strict IPTC compatibility, truncating tags that I spent hours to create them in the first place. Of course I never looked at the IPTC specs, but just went for appropriate key names and let exiftool do the work. Now those same tags end up crippled. The -m option sometimes works, but for some unknown reason not reliably - some tags still end up truncated.

Q1:  I have trouble following the change-logs. What is the latest version of exiftool that writes IPTC tags without TRUNCATING them? Is there any way to still download that version of exiftool somewhere?

Q2: If I try to write multiple "invalid" (too many characters) tags, -m doesn't seem to work reliably with my version (exiftool-8.80). Is there any way to write those tags at all with current exiftool?  Perhaps multiple -m tags on the command line? Or multiple invocations of exiftool, one tag at a time??

Q3: If I need to compile an old version of exiftool from source, I might as well compile the latest version and undo the IPTC "compatibility" fix... any tips on this?

Best,
  Norman

Phil Harvey

Hi Norman,

Give me an example where the -m option doesn't work and I'll look into this problem.

- Phil
...where DIR is the name of a directory/folder containing the images.  On Mac/Linux/PowerShell, use single quotes (') instead of double quotes (") around arguments containing a dollar sign ($).

Ken Rahaim

Hi Phil,

While searching for similar issues I ran across this thread in your forum. I think we have an instance of the -m option not working (that or we're making a mistake in its use).

For background info, I work for the Smithsonian Institution in the Digital Asset Management Systems branch of our IT division. We use exiftool to read embedded metadata which is then written into our DAMS database. Our problem occurs in the "IPTC:CopyrightNotice" field. There we have the boilerplate statement (or some variation of it):

QuoteThis image was obtained from the Smithsonian Institution. The image or its contents may be protected by international copyright laws.

Earlier this summer the string began getting truncated at the 128 character point. From what I understand, this is consistent with Exiftool's recent enforcement of IPTC limits.

Looking for a way to turn off the enforcement of IPTC limits, we discovered the -m option. Unfortunately, it doesn't appear to be working in our case. With -m set in the command line we're finding the string is still being truncated. Here is the command line we're executing:

C:\Users\phand>exiftool --IFD0:ProfileHueSatMapData1 --IFD0:ProfileHueSatMapData2 --XMP-crs:all --ExifIFD:ImageWidth --ExifIFD:ImageHeight -m -L -X -c "%.6f Degrees" "h:\Test Data\Images\copyright\natm cast castle_0012-18c.jpg" > "h:\Test Data\Images\copyright\natm cast castle_0012-18c.jpg.txt"

We are using Exiftool version 9.02. The command was executed on a Windows 7 machine.

I've attached the text output results.

Any help or pointers you have would be greatly appreciated.
--
Ken Rahaim
DAMS Branch
Smithsonian Institution

Phil Harvey

Hi Ken,

ExifTool doesn't truncate IPTC when reading.  It only does this when writing.  So the -m option only affects this behaviour when writing IPTC tags.

- Phil
...where DIR is the name of a directory/folder containing the images.  On Mac/Linux/PowerShell, use single quotes (') instead of double quotes (") around arguments containing a dollar sign ($).

Ken Rahaim

Hi Phil,

Understood. Thanks for the explanation.

It still makes it very strange that exiftool's output for the "IPTC:CopyrightNotice" field has a truncated version of our copyright boilerplate info in it.

Looking at the exiftool ouput further, the correct copyright info appears in the "IFD0:Copyright" field.

Looking at the file in Adobe Bridge or File Info, its "IPTC Copyright Notice" field shows the full copyright info.

I can't understand where exiftool is getting the truncated version of the copyright info from.

I've made a copy of the file in question available on my dropbox here:

https://www.dropbox.com/s/mey44gvh8simrmj/natm%20cast%20castle_0012-18c.jpg

Any insight you can provide would be greatly appreciated.

P.S. I'm not sure my original exiftool output file was attached to my first post so here's the link to the output file on dropbox:

https://www.dropbox.com/s/45ghdmiz93cim7r/natm%20cast%20castle_0012-18c.jpg.txt

Phil Harvey

Only the IPTC is truncated.  The other application is probably reading the duplicate information from EXIF which isn't truncated.   ExifTool will read everything and show you where it came from with -a -G1.  Also see the MWG tags documentation for a feature that may be useful.

- Phil
...where DIR is the name of a directory/folder containing the images.  On Mac/Linux/PowerShell, use single quotes (') instead of double quotes (") around arguments containing a dollar sign ($).