google-code-export / beets

Automatically exported from code.google.com/p/beets
MIT License
0 stars 0 forks source link

Tag corrections not applied #454

Closed GoogleCodeExporter closed 9 years ago

GoogleCodeExporter commented 9 years ago
I seem to have a problem similar to the one here: 
http://code.google.com/p/beets/issues/detail?id=423

When using the command "import -L", I apply the corrections. When I run the 
command again the same corrections appear.

Here is the verbose log: http://pastebin.com/5RyzdjY2
Here is the mediainfo of one of the files: http://pastebin.com/AcwVNa6F
If you want the files, just contact me on irc (nickname: Axio)

I also have some files for which beets asks me to modify their length, but 
applying the corrections has no effects either. I don't know if it is normal or 
not:
Log: http://pastebin.com/2STqXmre
Mediainfo: http://pastebin.com/s0tcxYw2

Original issue reported on code.google.com by trash.xa...@gmail.com on 3 Nov 2012 at 10:42

GoogleCodeExporter commented 9 years ago
This actually isn't the results of tags going unmodified -- it's something a 
bit more subtle. When importing an album, beets tries to guess the album's 
artist and title by looking at the files' current metadata. To get the artist, 
it looks at each track's artist and chooses the most popular one (or 
arbitrarily if there's a tie). In this case, the track artists don't match the 
album artist. Since beets previews matches using the album artist, there 
appears to be a difference where there is, in fact, no change to be made.

The solution to this is to improve the algorithm that guesses the current 
metadata to look at the album artist as well as the track artists.

Also: beets doesn't change files' lengths; it just shows you differences so you 
know if something fishy is going on.

Original comment by adrian.sampson on 3 Nov 2012 at 7:31

GoogleCodeExporter commented 9 years ago
Similar bug again, although there isn't any apparent modification: 
http://pastebin.com/FfAZrT71

I have tried to apply the change many times, but the 93.5% similarity is always 
back.

Original comment by trash.xa...@gmail.com on 9 Dec 2012 at 10:15

GoogleCodeExporter commented 9 years ago
That's because the track lengths don't match the lengths in the MusicBrainz 
database. That won't change even when the choice is applied.

Original comment by adrian.sampson on 10 Dec 2012 at 1:14