On Sun, Jul 09, 2000 at 03:40:29PM -0700, Ralph Giles wrote:
> On Sun, 9 Jul 2000, Don Melton wrote:
>
> > Thanks. This fix is also in but in a slightly different way. I modifed
> > the actual tagging routines in lame.c to check for Ogg Vorbis output
> > rather than making the client code responsible, so we only have to do
> > the test in one place rather than in all clients.
>
> Huh. I thought it made more sense to make the client code responsible as
> long as the routines had 'id3' in the name.
No, it's more work for each client that way. Consider your original
patch: while you made the appropriate modification to "main.c", you
didn't do the same for "mp3rtp.c" which is also a client of the LAME API
defined in "lame.h".
Moving the responsibility for the check inside the API fixes the bug in
ALL clients without the client code having to change.
> If you are reworking things, please be aware that the vorbis comment
> header is more flexible, basically allowing arbitrary token-stringvalue
> pairs (though there is a standard set that players/encoders should
> recognize) and there can be multiple instances of the same tag. So it
> would be nice to be able to say:
>
> lame --ogg --ta "Peter Gabriel" --ta "Kate Bush" \
> --tag "CDDBID=foo" song.wav song.ogg
>
> and have it handle things properly. But obviously this complicates things
> quite a bit.
Agreed. And I don't know if we want to make LAME into a tagging utility
yet. :-)
But while the vorbis comment header and its arbitrary data model is
flexible, that doesn't necessarily make it useful outside a single
client of that data unless there are agreed upon conventions for those
tokens and the format of the string values. And I think the vorbis
folks have a ways to go on the standardization of that sort of thing,
especially since they still haven't decided on a "track" convention. :-)
This is what makes the ID3 version 2 standard so useful. It has
specific tags for damn near everything! And all the tagging utilities
agree on the formats. It's easy because they're all (well, at least
most of them) defined in the standard. Granted, ID3v2 is perhaps the
most over-engineered specification outside of ISO, but it IS thorough.
Personally, I never use anything except the "--tt" and "--ta" options in
LAME. :-)
> Thanks for the commit,
> -ralph
No problem. It was a good idea!
--
Don Melton
mailto:[EMAIL PROTECTED]
--
MP3 ENCODER mailing list ( http://geek.rcc.se/mp3encoder/ )
