exiftool ability to retrieve encoded chars

[Archive]

[Originally posted by jdmls on 2009-04-03 08:26:33-07]

Code:

Hi,

First: sorry but I cannot post the non ascii chars, so I replaced them with <CHAR>...
I have a little questions about exiftool ability to retrieve encoded chars...
I have a jpg picture for which exiftool gives me the following tags:

  Keywords : Aufnahmeger<BADCHAR>t, ...
  Subject : Aufnahmeger<GOODCHAR>t, ...

  <GOODCHAR> is http://tinyurl.com/dzjse5
  <BADCHAR> is http://tinyurl.com/dlkn9b

I would expect both sentences to be identical (although I am not the one who
made the picture)

The Keywords one seems to come from an 8BIM section:

  ed 1f 4e 50 68 6f 74 6f  73 68 6f 70 20 33 2e 30  ..NPhotoshop 3.0
  00 38 42 49 4d 03 ed 00  00 00 00 00 10 00 48 00  .8BIM.........H.
  00 00 01 00 01 00 48 00  00 00 01 00 01 38 42 49  ......H......8BI
  4d 04 04 00 00 00 00 00  b1 1c 02 00 00 02 00 02  M...............
  1c 02 05 00 09 5a 77 69  6c 6c 69 6e 67 65 1c 02  .....Zwillinge..
  19 00 0d 41 75 66 6e 61  68 6d 65 67 65 72 8a 74  ...Aufnahmeger.t

The weird thing is that I cannot find how exiftool converts 0x8a to <BADCHAR>
And maybe it should even be <GOODCHAR>...

The Subject one should come from the XMP section and looks fine as it is <GOODCHAR>...

  72 64 66 3a 6c 69 3e 41  75 66 6e 61 68 6d 65 67  rdf:li>Aufnahmeg
  65 72 c3 a4 74 3c 2f 72  64 66 3a 6c 69 3e 0a 20  er..t</rdf:li>.

Any idea?

Thx
JD

[Archive]

[Originally posted by exiftool on 2009-04-03 11:16:12-07]

Hi JD,

Please read https://exiftool.org/faq.html#Q10" target="_blank">FAQ
number 10 and let me know if you have any questions after
reading this.

- Phil

[Archive]

[Originally posted by jdmls on 2009-04-03 11:53:27-07]

Code:

I already read the FAQ... but I must admit I don't understand very well all these encoding rules.
The picture was supposed to have the same sentence in both tags.
But I just had a look at an extended ascii table and 8a = BADCHAR...
So, I guess exiftool read it correctly and the error must be in the original (photoshop error)

My bad,
JD

[Archive]

[Originally posted by exiftool on 2009-04-03 14:16:14-07]

The problem is that Photoshop uses the
local system encoding when writing IPTC information,
and there is no way to determine what encoding was
used.  As a result, display of special characters written
by Photoshop in IPTC is problematic.

This is idiotic,  really, and would have been fixed long ago
if Adobe wasn't pushing XMP.  The solution, of course, is
to write IPTC in UTF8 and set the CodedCharacterSet
accordingly so that other applications know which encoding
to use.  Simple.  But I just checked Photoshop CS4 and it
still doesn't handle IPTC encodings properly.

- Phil