[FFmpeg-devel] [PATCH] "Mojibake" in Japanese

Vladimir Mosgalin mosgalin at VM10124.spb.edu
Mon Feb 6 14:39:07 CET 2012

Hi tetu!

 On 2012.02.06 at 09:49:49 +0900, tetu wrote next:

> I made patch for "Mojibake" in Japanese.
> It is for corresponding to Japanese strange culture. FFMpeg have no problem.
> In Japan, ID3 Tags in a lot of mp3 file is written in Shift JIS as
> ISO-8859-1.
> So, when libavformat read ISO-8859-1 ID3 Tags as UTF-8, unable to read
> correct string.
> (called "Mojibake" in Japanese)
> The patch has no meaning to people using ASCII,
> But it will be so useful to many Japanese.

The same problem exist for russian .mp3 files, where CP1251 encoding is
used de-facto instead of ISO-8859-1 in ID3 tags. Because of that, all
music players that are popular around here (and tag managing software
like easytag, too) support that and have options to treat non-unicode
ID3 tags as CP1251-encoded.

While SJIS is compatible with ascii, treating any other encoding as SJIS
would cause disaster..

I believe it would be unwise to automatically force treating ID3 tags as
SJIS-encoded.. (not to mention that it's violation of ID3 standard).
At very least it should be some option to select encoding, where you can
pick CP1251, SJIS or some other encoding (I can imagine that this
problem exists not only in Russia and Japan).



More information about the ffmpeg-devel mailing list