Writing accented characters?

[Archive]

[Originally posted by jaypng on 2007-07-03 20:17:17-07]

While writing "accented characters" (cannot be rendered correctly here) results in "CRAP" (cannot be rendered correctly here).

How to do write them correctly?

I'm using Windows XP SP2 and exiftool 6.92 perl (not exe)

[Archive]

[Originally posted by exiftool on 2007-07-03 22:15:29-07]

By default, ExifTool assumes that input character strings are in UTF8 encoding.
But since you're using Windows, the characters you are entering are likely
encoded in Windows Latin1.  If this is true, adding a -L option to the
command line may solve the problems.

However, things are slightly different for different information formats,
so you may want to read
FAQ #10
for more information about character encoding.

- Phil

[Archive]

[Originally posted by jaypng on 2007-07-04 07:02:22-07]

I have more information.

When writing IPTC without the "-L" exiftool complains "Warning: Malformed UTF-8 character(s)" and writes nothing.

If I write to IPTC the "-L" it doesn't give any errors; an output to a txt file shows the characters written correctly, but an output to the screen shows them incorrectly.

If I write to XMP it doesn't give any errors; output to txt file is fine; output to screen is wrong.

But if I write first to IPTC with -L and then write to XMP and output to file, the IPTC field is wrong but XMP is fine.

But it gets better. Open the txt file in WordPad: output is wrong. Open it with Notepad: output is fine.

So the problem is with applications that are not Unicode aware. This means that command line output on Windows XP doesn't support unicode but Notepad does!?

I finally found some applications that read xmp metadata from png files on Windows: iView MediaPro and Extensis Portfolio.

Both can read but not write xmp data to pngs. They don't support IPTC on pngs.

Both can read the metadata if it was written "-L". But will not read it if it was written without. So although exiftool writes accented characters correctly to xmp without "-L" no other software can read it (on Windows at least).

How is the support in Linux?

[Archive]

[Originally posted by exiftool on 2007-07-04 10:35:26-07]

Did you use the -L option when you obtained your output?

If you want the output in Latin1, use -L.  Otherwise the output is
in UTF8.

I assume that WordPad uses Latin1 but that Notepad understands UTF8.

In Linux, things are simple because mostly everything understands UTF8.

- Phil

[Archive]

[Originally posted by eric00000007 on 2007-08-02 11:09:20-07]

Hi all,
I install the last version of exiftool. I was previously adding UTF-8 accented caracters without trouble. Now I can not add accented characters anymore. Is there a new setting ?

Nothing change on my machine. I tested that on different configuration.

Somebody has an idea ?

[Archive]

[Originally posted by eric00000007 on 2007-08-02 13:53:37-07]

I tested with my previous version 6.28 and it doesn't work
I tested also 6.90 and 6.93 and same thing

maybe it has never worked. Somebody has an idea ?

[Archive]

[Originally posted by eric00000007 on 2007-08-02 14:30:40-07]

I tested with my previous version 6.28 and it doesn't work
I tested also 6.90 and 6.93 and same thing

maybe it has never worked. Somebody has an idea ?

[Archive]

[Originally posted by eric00000007 on 2007-08-05 14:53:16-07]

I forget to say that I was testing it with the cwindows command line. It seems that it don't support utf-8. I am checking this information. So I tested 6.94 windows executable from an web application (Coldfusion) and it works.

So take care when using the windows command line. I will continue my tests this week.

Thanks again Phill for this great tool !

[Archive]

[Originally posted by exiftool on 2007-08-16 11:10:22-07]

Typically, the Windows command line uses the Latin1 character set,
so you should use the -L option with exiftool when writing accented
characters from the Windows command line.

- Phil