Output garbled characters when ID3V1 contains Chinese characters. about tinytag HOT 7 CLOSED

chardet return latin1 when title contains chars Chineses.
I think chardet needs more characters to determine the encoding, not enough characters in the title

from tinytag.

Added: ID3V2 has same bug.

from tinytag.

I try with locale module, it works.
locale.getpreferredencoding(False)
or
decode using locale.getpreferredencoding(False), if failed, try 'latin1'?

from tinytag.

Hey cdhigh,

I have another idea to make this possible that is similar to your PR, which is to use chardet instead of the system locale. Maybe you can make a test using that library and we could add it as optional dependency to tinytag.

Chardet uses the "letter-frequency" of the bytes to determine which encoding might be successfully used.

https://pypi.org/project/chardet/

from tinytag.

Hey @cdhigh ,

Could you please provide a sample file so I can reproduce the bug?

from tinytag.

https://we.tl/t-cs2f6K6j66

from tinytag.

Hey @cdhigh

Thanks for providing a sample file, I was able to reproduce the bug. Unfortunately, there is no way to automatically detect the correct encoding for gbk encoded files. (As you also saw, chardet misdetects it as another encoding).

I don't want to fill the library with code that guesses encodings based on system settings and whatnot, so instead I opted to allow for overriding the default encoding.

You can now use TinyTag.get(filename, encoding='gbk') and it will then use gbk encoding whenever no encoding has been specified inside of the mp3 file. Since gbk also contains the ASCII character set, this should work pretty well for you, even when you also have english files in between.

You can use the version in the master branch until I release a new version containing the feature.

from tinytag.