[vorbis] Video codec

[Ralph Giles]

On Fri, 8 Sep 2000, Sean Wieland wrote:

> > Squish or uncompressed as an option as well. MNG and tarkin (the lossy
> > video codec-to-be) as video formats, and mng for annotative overlays like
> > DVD-style subtitles. XML substreams for closed captioning, real subtitles,
> > metadata, etc. And of course we can have alternate tracks for commentary
> > or translations.
> > 
> > My aim is to properly reimplement all the features of DVD, except
> > for the non-linear and interactive bits. I've not seen a clear way to
> > implemen this last within a multimedia framework, and so leave it to the
> > next layer up.
> > 
> > Hope that helps,
> >  -ralph
> 
> While I think this is an excellent and noble goal.  (I know I'd love to
> have an on-demand archive of any movie I'd ever want to see, rather than
> having to actually find the DVD and deal with all of the hoops of CSS
> and macrovision; the same way I love having all of my CDs as a huge
> on-demand Vorbis archive.)  What are your thoughts on streaming video? 
> MPEG-1 doesn't cut it except for broadband users, and the
> Real/QuickTime/ASF battle is rediculous.

Well, at this point, I think you're better off buying a net-accessible
jukebox; while the DVD format is ugly, it doesn't make any more sense to
transcode DVD->Ogg than it does to transcode mp3s, unless you're also
seriously downsampling (e.g. for internet distribution).

As far as streaming goes, we can use ogg for http streaming just as we can
use it for local files. To do things "properly" (RTP), one doesn't 
want ogg, but I imagine we can work out an equivalent "application
profile" with the same substreams and features.

Streaming video is never going to work without (at least!) a broadband
connection. Chad's estimates are that the current design can be pushed to
a factor of 2 or 3 better than MPEG-1, which will allow the VHS quality
movie on a CD (loosely defined) claimed for MPEG-4 (around 1.5 Mbps).
I have no hope that we're going to get DVD-quality video under the limits
of broadband connections. As far as that goes, I expect we'll just
be another contender in the Real/QuickTime/ASF battle.

> Also, where is the MNG/tarkin source code?  Where exactly is the
> development status?

The mng reference implementation (beta) is available at libmng.com.
There's currently no code for ogg integration. This would be pretty
simple, if you're looking for something to do. :)

It has been alleged that the tarkin test source will appear in the
'tarkin' cvs module, but it's not there yet. AFAIK, there's never been a
public release, though Jack recently described the algorithm. See the
ensuing discussion for immediate followup ideas:

  http://www.xiph.org/archives/vorbis-dev/1329.html

Cheers,
 -ralph

--- >8 ----
List archives:  http://www.xiph.org/archives/
Ogg project homepage: http://www.xiph.org/ogg/
To unsubscribe from this list, send a message to 'vorbis-request at xiph.org'
containing only the word 'unsubscribe' in the body.  No subject is needed.
Unsubscribe messages sent to the list will be ignored/filtered.