How to convert utf8 to cp1251 to write ID3_V1 tag of mp3 file?

Question

ID3_V1 supports only latin1 encoding. In order to write V1 tags with Russian characters, cp1251 encoding is used. I would like to copy data from V2 tags (unicode) to V1 tags. I get V2 tags with the following code with eyeD3 usage:

tag.link(mp3path, v=eyeD3.ID3_V2)
mp3album_v2 = tag.getAlbum()
...
tag.link(mp3path, v=eyeD3.ID3_V1)
tag.setTextEncoding(eyeD3.LATIN1_ENCODING)
tag.setAlbum(mp3album_v2.encode('cp1251')) # ???
tag.update()

The following is returned:

>>> print mp3album_v2
Жить в твоей голове

>>> print type(mp3album_v2)
<type 'unicode'>

>>> print repr(mp3album_v2)
u'\u0416\u0438\u0442\u044c \u0432 \u0442\u0432\u043e\u0435\u0439 \u0433\u043e\u043b\u043e\u0432\u0435'

Looks like setAlbum expects utf-8 string (?):

def setAlbum(self, a):
    self.setTextFrame(ALBUM_FID, self.strToUnicode(a));

def strToUnicode(self, s):
    t = type(s);
    if t != unicode and t == str:
        s = unicode(s, eyeD3.LOCAL_ENCODING);
    elif t != unicode and t != str:
        raise TagException("Wrong type passed to strToUnicode: %s" % str(t));
    return s;

But if I try to do tag.setAlbum(mp3album_v2.encode('cp1251').encode('utf-8')), then I am getting an error UnicodeDecodeError: 'utf8' codec can't decode byte 0xc6 in position 0: invalid continuation byte

Answer 1

ID3v1 cannot reliably include any non-ASCII character. You can write cp1251-encoded bytes into ID3v1 tags but they will only render as Cyrillic on Russian-locale OS installs and even then not on all applications.

EyeD3 deals with Unicode strings internally and arbitrarily chooses to use latin1 (aka ISO-8859-1) as the encoding for ID3v1 tags. This probably isn't a good choice because latin1 is never the default locale-specific encoding on a Windows box (for Western Europe it's actually cp1252 which is similar but not the same).

However a property of this choice of encoding is that each byte in it maps to a Unicode character with the same code point number. You can take advantage of this by making a Unicode string that contains characters that, when encoded as latin1, will end up being the byte encoding of a chosen string in an encoding other than latin1.

album_name = u'Жить в твоей голове'
mangled_name = album_name.encode('cp1251').decode('latin1')
tag.setAlbum(mangled_name) # will encode as latin1, resulting in cp1251 bytes

This is a horrible hack, of doubtful benefit, and one of the reasons you should avoid ID3v1.