Genres where does Gmusicbrowser read them from?

[blackest]

Hi
Over the last few weeks i've been playing around with genres using a bit of python to read and write the genretag

The genre tags are separated with a semicolon so Rock;Pop ;80's would be  read as 3 genre's
I started out with a script i found which read the top 20 tags from lastfm for a particular track this put these tags in the comment field of the mp3.
I  then copied these tags to the genre field that was interesting I had around 80,000 genres for around 16,000 tracks !
It was also totally unmanagable for gmusicbrowser. Especially for editing.

So my next task was to analyze the tags I had gathered so i wrote another program which sorted the genres in order of popularity in my Mp3 files
One thing became obvious around 38,000 tags were unique existing in only 1 file and many others were almost as rare.

So my next program took the top 3550 tags as a control and I rewrote the genretags so each genre had to be one of the 3550.

This was much improved Gmusicbrowser could pretty much handle the Genre data but there was a problem. Although I had specified 3550 genres there were in fact  around 7000. so why the extra?

Well for one there were ape tags in my mp3 files using puddletag i was able to remove the ape tags but that only reduced the genres to around 5000.  So it seems they must be stored elsewhere.

I think it may just be that some mp3's have multiple frames of genre information. instead of the single frame i was writing too.

so really i'm looking for confirmation that this is the case and i can write something to find and delete these extra frames.

i'm hoping to generate some interesting playlists already i can select say female vocalists  and not have them classed separately to say rock or pop. I kind of decided it is too difficult a task to use a single genre for a song I often disagreed with the tag that had been selected also some were insane what is christian gangsta rap. I also have multiple languages which can be an important filter.   

Anyway the real question is where does gmusicbrowser look for genres and is it reading multiple frames to find these genres?
     

[Quentin Sculo]

It does read genres in multiple frames, though it is kind of rare so I'm not sure that is what cause your problem.
Older versions used to read genres from all available tags, but I changed it so that it reads the genres from only one tag. So for example if the id3v2 tag has genres, it won't read the genre from id3v1, but if the id3v2 tag doesn't have a genre, the genre from the id3v1 tag will be used.
So I don't know where these genres are coming from, what you should do is try to find one that is not in your list, apparently 1/2 are not, so it shouldn't be too hard, and look at the file using the advanced tag editing dialog (button at bottom of song properties dialog)