[Vorbis-dev] Re: [ogg-dev] Peer review draft for the new media types/file extensions

Silvia Pfeiffer silviapfeiffer1 at gmail.com
Thu Oct 4 00:52:41 PDT 2007

On 10/4/07, Ian Malone <ibmalone at gmail.com> wrote:
> Ivo Emanuel Gonçalves wrote:
> > On 10/3/07, Ian Malone <ibmalone at gmail.com> wrote:
> >> This would mean that the inclusion of lyrics in a pop song
> >> would make it 'video'.
> >
> > If they are added in such way as to make it visual, then we are out of
> > the domain of audio/ogg.
> >
> I look at it like this: if you have video/ogg and you only play back
> the audio it's likely less meaningful.  If you have lyrics embedded in
> audio/ogg and only play back the audio you haven't lost very much.
> My reference to searching lyrics was intended to suggest that lyrics
> in a file could help identify its contents.

I agree with Ian here. audio/ogg suggests that it is a file that has
predominantly audio information. Annotations on the audio are great
additional information for an audio player, but it's still an audio
file and a video application will look in vain for a video track. It's
plain wrong to classify that as video.

> > I have a question, though; while Skeleton won't be pushed for Vorbis,
> > do Vorbis files with Skeleton tracks break existing players/refuse to
> > play?  I've never tested it, and I'm curious.
> >
> Certainly they will refuse to play on some existing players.  The
> Skeleton packet must start the stream; so anything that was written
> for Vorbis I and is going to get first packet to check it's Vorbis
> is going to find it isn't.  Some players do survive.

For backwards compatibility, we retain .ogg as VorbisI files, and
existing players will work with those. For moving forward, we can
define anything. It is expected that .oga will break players that
interpret them as .ogg files - but that's not a problem because they
will not even want to open a .oga file. As soon as .oga support is
implemented in the player, it will parse the skeleton track and know
what's inside and then use the appropriate audio decoder.

> >> This is application/ogg we are talking about. Ogg is a container
> >> format in principle ready for any generic time-continuously sampled
> >> data, not just audio and video. I would not want to restrict us for
> >> the future to just audio and video.
> >
> > We are not restricting to audio and video.  That's the whole point of
> > separating the media types.  BUT most people only deal with audio and
> > video, with basic Theora, Vorbis, and Speex files.  That's the
> > situation right now.  We want those people to use audio/ogg &
> > video/ogg.  We do not want them to use application/ogg.  That's for
> > "applications", stuff that is more complex than the basic media file.
> > You, yourself, explained me this.
> >
> > As such, it makes no sense to say application/ogg is for generic use.
> > Scientific applications and the like are not generic.

You misunderstand me. Scientific applications are an example. We don't
at this point in time know what applications will use application/ogg
and we don't want to restrict it. If we call it audio/video, we
restrict it. If we call it "generic", we don't.


More information about the ogg-dev mailing list