ubuntu 11.10 rhythmbox Japanese display error

jack_green · 01-20-2012, 09:39 PM

Hello,guys. I love Ubuntu Desktop OS and I use it from ubuntu 10.04. But what the thing bother me is that I cannot make rhythmbox working perfectly. I have downloads many Japanese songs but rhythmbox cannot display Japanese fonts. Any solutions ? Thanks so much.

David the H. · 01-21-2012, 08:32 AM

Your system most likely uses UTF-8 as the character set, and vorbis and flac comments are also in UTF-8 by default. But id3 comments are often found in local encodings. I'd bet most of those file's tags are in shift-JIS encoding.

Some player programs have the ability to force the read encoding for id3 comments. I don't know about rhythmbox though, so check your options.

A better choice would be to convert the comments of the files themselves. I recommend using easytag for this, as it has clear options for both the reading and writing encoding to use.

jack_green · 01-21-2012, 08:43 AM

Quote:

Originally Posted by David the H.

Your system most likely uses UTF-8 as the character set, and vorbis and flac comments are also in UTF-8 by default. But id3 comments are often found in local encodings. I'd bet most of those file's tags are in shift-JIS encoding.

Some player programs have the ability to force the read encoding for id3 comments. I don't know about rhythmbox though, so check your options.

A better choice would be to convert the comments of the files themselves. I recommend using easytag for this, as it has clear options for both the reading and writing encoding to use.

sorry, it cannot help.

David the H. · 01-21-2012, 08:59 AM

Umm, what "cannot help"? What part of the advice I gave doesn't work for you?

It's quite obvious to me that the problem is with the encoding of the tags, so either you have to configure your player to use the correct one, or convert the tags themselves.

jack_green · 01-21-2012, 09:46 AM

Quote:

Originally Posted by David the H.

Umm, what "cannot help"? What part of the advice I gave doesn't work for you?

It's quite obvious to me that the problem is with the encoding of the tags, so either you have to configure your player to use the correct one, or convert the tags themselves.

Well, I followed your advice and installed easytag on my ubuntu. But when I run that software, 30s later the PID of easytag automatically died.

David the H. · 01-22-2012, 03:16 AM

So something about your system or the files is making easytag crash. Then it's a problem with that particular program, not with my advice, and it would've been polite to mention that.

If you can post the output you get when you run it in a shell it may be possible to to diagnose the problem. Don't forget to do a google search too to see if anyone else is having the same problem, and whether there are any fixes already available.

Otherwise, there are other tag editors out there that may be able to do the job. I often use kid3 myself, for example (use the kid3-qt version if you don't need the full kde integration).

I just did a quick test, however, using an mp3 file I converted to shift-JIS in easytag, and while it can write tags properly in both id3v1 and id3v2 tags, it can only be set to read arbitrary encodings from id3v1 tags. id3v2 tags set to anything other than unicode* or iso8859-1 are unsupported and you only get mojibake with them.

(*It appears that id3v2.3 only supports UTF-16, but that id3v2.4 also supports UTF-8. I haven't tested this personally, and it's probably not an issue, as I'm sure all tag readers know how to interpret them properly. id3v1 tags aren't designed for storing international encodings, BTW, but they can store the raw strings, which is why so many programs let you interpret them according to different encodings).

That could still be enough to correct your files though; if they contain both v1 and v2 tags, then you can copy the readable v1 tags to the v2 fields, and at least get a partial correction.

In other news, I've also just discovered mid3iconv, supplied with the python-mutagen package, which is a cli program designed for converting encodings. I just ran a test like this (using -d and -p so that it only outputs the changes that would be made), and everything worked except the comment, for some reason. There seems to be something about the comment tag format that's confusing it.

Code:

mid3iconv -p -d -e SHIFT-JIS file.mp3

So if you try it, be sure to operate on copies of the files, at least until you're sure it's working correctly.

And here's one more option for you, although it's a bit more work. If you get a command-line program like id3v2 that outputs tag data in a useful way, then you can run it through the regular iconv to convert it into a clean utf-8 encoding, and use whatever tagging program you like to manually update the files accordingly. This command worked cleanly on my test file:

Code:

id3v2 -l file.mp3 | iconv -f SHIFT-JIS -t UTF-8 >> tagile.txt

The last part is there to redirect (appending) the output to a text file, of course.

jack_green · 01-22-2012, 04:12 AM

Quote:

Originally Posted by David the H.

So something about your system or the files is making easytag crash. Then it's a problem with that particular program, not with my advice, and it would've been polite to mention that.

If you can post the output you get when you run it in a shell it may be possible to to diagnose the problem. Don't forget to do a google search too to see if anyone else is having the same problem, and whether there are any fixes already available.

Otherwise, there are other tag editors out there that may be able to do the job. I often use kid3 myself, for example (use the kid3-qt version if you don't need the full kde integration).

I just did a quick test, however, using an mp3 file I converted to shift-JIS in easytag, and while it can write tags properly in both id3v1 and id3v2 tags, it can only be set to read arbitrary encodings from id3v1 tags. id3v2 tags set to anything other than unicode* or iso8859-1 are unsupported and you only get mojibake with them.

(*It appears that id3v2.3 only supports UTF-16, but that id3v2.4 also supports UTF-8. I haven't tested this personally, and it's probably not an issue, as I'm sure all tag readers know how to interpret them properly. id3v1 tags aren't designed for storing international encodings, BTW, but they can store the raw strings, which is why so many programs let you interpret them according to different encodings).

That could still be enough to correct your files though; if they contain both v1 and v2 tags, then you can copy the readable v1 tags to the v2 fields, and at least get a partial correction.

In other news, I've also just discovered mid3iconv, supplied with the python-mutagen package, which is a cli program designed for converting encodings. I just ran a test like this (using -d and -p so that it only outputs the changes that would be made), and everything worked except the comment, for some reason. There seems to be something about the comment tag format that's confusing it.

Code:

mid3iconv -p -d -e SHIFT-JIS file.mp3

So if you try it, be sure to operate on copies of the files, at least until you're sure it's working correctly.

And here's one more option for you, although it's a bit more work. If you get a command-line program like id3v2 that outputs tag data in a useful way, then you can run it through the regular iconv to convert it into a clean utf-8 encoding, and use whatever tagging program you like to manually update the files accordingly. This command worked cleanly on my test file:

Code:

id3v2 -l file.mp3 | iconv -f SHIFT-JIS -t UTF-8 >> tagile.txt

The last part is there to redirect (appending) the output to a text file, of course.

I am so sorry for my last answer.

Thanks so much

David the H. · 01-22-2012, 05:07 AM

No problem. Just let us know if and how you get a final fix to work.

BTW, it's not generally necessary to quote the entire previous message when responding either. It's clear who you're responding to, so it just clutters up the thread. Save the quote tags for when you really need to direct a comment towards a particular poster or a specific part of a post.

jack_green · 01-22-2012, 09:30 AM

Hello, guys. This is my easytag bug, any suggestion ?