I need Amarok to correctly display the Hebrew information in ID3 tags that were created when the computer was running Windows XP Home. I have tried setting the encoding in Settings -> Amarok Settings -> General -> Encoding to CP 1255, UTF-8, and ISO-8859-8I, but none of them has made any difference. I did restart Amarok after each try.

A screenshot of the wrongly-encoded result is here:
Amarok

What else can I do? Thanks in advance.

Hi,
as best as i can understand it the problem is with the tags and not with amarok
Qt/KDE uses utf-8 internally so nothing utf8 can confuse it.
however only id3v2.4 uses utf8 properly
you can use utf16 and it should work.

the problem with the windows tags must be the first byte still says it's latin1.

there must be some way to set it right
you could try running this on them

http://www.cs.berkeley.edu/~zf/id3iconv/
java -jar id3conv.jar song.mp3
java also uses utf8 internally just like KDE

if you have id3v2.4 you can try
eyeD3
http://eyed3.nicfit.net/
to convert them
i think you would use
eyeD3 --force-update --set-encoding=utf8 song.mp3
might need more command switches -- not sure

Thanks, I will give eyeD3 a try. This at least gives me a lead where to continue if it does not work- I now know that the files must be fixed and not Amarok.

Is there a way to do this recursivly? I have about 20 gigs of mp3 files in recursive directories...

you could do it with bash scripting since the program doesn't support -R

i am so bad at shell scripting perhaps someone else will come along and show better.

basically something like this
once you figure out the right command substitute for what i put
(caution danger test first) i am a bad scripter
you always have to give shell scripts a
chmod 755 scriptname.sh
to get them running

foo_bar_foo,
Thanks for the trying to write the script. It failed on two accounts, though:
1) It breaks on file names with spaces. I seem unable to correct this. This is, however, the first time that I have ever tried to work with a shell script (I know nothing about them).
2) The tags seem to be v2.3. This was the output of every file that did not contain spaces:
What can I do from here? Is there a way I can convert from v2.3 to v2.4?

hmmmmm windows filenames with spaces
i can't figure out how to do that

for the other thing you can try the java app or upgrade libid3tag

i am sure thee other guys here can fix the script to do windows filenames if you can get there attention perhaps with a new thread "whitespace in filenames" or something
give them the script and they can fix it.

Thanks, foo. Now that I've proceeded to two seperate problems, I'll start two new threads on each. I really appreciate the assistance.

Well, I've got those two problems solved, but I still can't read the Hebrew ID3 tags in Amarok. Now that I've converted a test batch of files to v2.4 and then to UTF-8, the problem persists- I still see Latin extended characters instead of Hebrew. The simple conversion might work for Latin characters, but not for the Hebrew characters in my ID3 tags.

Any ideas?

Problem solved. For future problem solving:
I had to change my locale to UTF-8. I had to mount the FAT32 partition with charset UTF-8. I had to override the latin encoding in the Amarok settings and use UTF-8.

There was no need to modify the mp3 files themselves (all my modifications were on a test batch- proved unnessaccary).

Thank you!