invalid character encoding in tags - garbled accents

[arthurlutz]

First of, thanks for groovebasin, looks awesome and great features. And easily installable on debian jessie. Kudos to all contributors.

Am having problems with accents :

File names appear correctly in shell. They are garbled too when using eye3d, so the problem might be in the tags...

[arthurlutz]

OK, I can confirm that doing a cleanup with http://beets.radbox.org/ I get the correct accents. Any chance this could be fixed on the fly or it has to be fixed in the tags ?

[thejoshwolfe]

what format were the files in? can you send us one of the problem files so we can diagnose the issue?

I suspect that the tags are encoded using something other than utf8, and groovebasin is trying to decode them as utf8. depending on how obvious the misinformation is, groovebasin may or may not be able to detect that anything is wrong, let alone propose a solution.

a bigger picture solution, and the solution I suspect that the beets utility is using, is to ignore the present tags entirely, and use the audio data itself to figure out what the coocorrect tags should be by consulting a third-party database like MusiBrainz. this feature is planned for groovebasin, but it looks like there's no issue for it yet.

[thejoshwolfe]

got the file. thanks.

It looks like my original prediction was backwards. The tag really is encoded in utf8, but groovebasin appears to be decoding it with iso8859-1 (or similar) instead. Very curious.

I'll need to consult with @andrewrk on this one.

[andrewrk]

I notice that Ex Falso reads the tags correctly.

This is a bug in our underlying dependency, libav:

andy@andy-bx:~/dev/libav$ ./avprobe ~/tmp/02.\ Amitié_Et_Amertume.mp3 
avprobe version v12_dev0-686-g3a6dfec, Copyright (c) 2007-2015 the Libav developers
  built on Jan 26 2015 09:46:49 with gcc 4.9.1 (Ubuntu 4.9.1-16ubuntu6)
[mp3 @ 0x2412f60] max_analyze_duration 5000000 reached
Input #0, mp3, from '/home/andy/tmp/02. Amitié_Et_Amertume.mp3':
  Metadata:
    album           : Identité en Crescendo
    artist          : Rocé
    title           : Amitié Et Amertume
    track           : 02/13
    genre           : Rap
    date            : 2006
  Duration: 00:04:25.58, start: 0.000000, bitrate: 128 kb/s
    Stream #0.0: Audio: mp3, 44100 Hz, 2 channels, s16p, 128 kb/s
    Metadata:
      encoder         : LAME3.99r
    Side data:
      replaygain: track gain - -7.400000, track peak - unknown, album gain - unknown, album peak - unknown, 
# avprobe output

I will report the issue in the libav bug tracker and also look into the issue a little to see if I can solve it.