| alsbergt@gmail.com 2006-09-12, 10:17 pm |
| (Shamelessly copied from my question in Adobe Forums)
I experience the following problem with Photoshop CS2 (version 9.0.1)
on Windows XP:
If I have JPEG files with IPTC tags in UTF-8 (Unicode) encoding,
containing a tag (e.g. Caption/Description) with non-English characters
(e.g. umlauts or European accents), when opening that file in
Photoshop, looking in the File Info dialogue, those tags are read as
ISO-8859-1 encoded (although the CharacterSet/1:090 IPTC attribute is
properly set to hexadecimal 1b 25 47 to indicate that the other tags
are in UTF-8 ).
This means, that all non-English characters in such files are decoded
as some arbitrary two characters. For example, the word "Caf=E9" is
turned to "Caf=C3=A9".
Saving the file from Photoshop will then remove the CharacterSet
attribute from the file, so it will further on be read by other
software also as ISO-8859-1, with the non-English characters ruined.
Furthermore, if a file is created in Photoshop CS2, and fill through
the File Info menu, tags with non-English characters, they are saved in
ISO-8859-1 encoding, even if they contain non Latin-1 characters (which
cannot be encoded in this encoding) - in this case they are replaced by
different characters. For example, entering a caption of "Caf=E9", will
result in Caf=E9 written as ISO-8859-1. Entering "Pawel", will result
in "Pawel" being written (note the 'l' turning into an 'l').
Admittedly, the full information is stored in the XMP tags when saving
such files (although it is not read correctly when opening files with
only IPTC tags in UTF-8, so in such a case information is lost), so if
I entered the tags through Photoshop, I will have them available next
time. However, when opening the files through other software which does
not support XMP, the IPTC information is used, which contains the "bad"
tags.
I need the IPTC information in the files - the XMP information is
useful for what supports it, but I depend on some software, which like
most programs today, still supports only IPTC and Exif, not XMP. The
current behvaiour I observe with Photoshop inhibits me from using
non-English characters with it.
The correct behaviour should be - when opening files with IPTC tags,
read them in the encoding indicated in the file. When saving files, use
any encoding that is capable of representing all the characters used.
If only English characters are used, ASCII or ISO-8859-1 is OK, but if
non-Latin-1 characters are used, the text should be written in Unicode.
Can anybody suggest some solutions to get Photoshop to interoperate
with other software with non-English IPTC tags?
Best regards, any help appreciated,
-- Tom
|