[vorbis-dev] Ogg VEO Codec, video in an Ogg stream

Thomas Marshall Eubanks tme at 21rst-century.com
Fri Sep 1 13:36:48 PDT 2000



Ralph Giles wrote:

> On Fri, 1 Sep 2000, Thomas Marshall Eubanks wrote:
>
> > Hello;
> >
> >    If you are planning on doing this, and streaming it, then you should read
> >
> > http://www.stardust.com/content-engineering/whitepapers/fgs_01.htm
> >
> > and build a layered protection system in from the start, so that it doesn't have
> > to be implemented after the protocol is designed as in the MPEG 4 case.
>
> Thanks for the link. And references! :-)
>
> The paper doesn't really address the issue of packet loss, though it is
> mentioned in the introduction. All the figures assume deliberate
> filtering for bandwidth scalability, not random drops. Could you comment
> on this?
>

Well, the idea is you could protect the base layer with FEC and make sure
you still have a picture! You could just send the base layer twice, for
example, and then have a factor of two protection on your picture being there,
at a cost of a bandwidth increase of only about 20%.

>
> One can imagine a scheme where no packet is essential to a frame; has
> anyone tried an approach like that?
>

Sure - Reed Soloman FEC would make this easy. I think it would be better
to have a system where no FRAME is essential. As it stands now, with MPEG-4 the
difference (P) frame can be lost, but not the I frames - if they are lost, the
following P frames are useless. A scheme where ALL frames were difference type, but
with sufficient redundancy so that no packet was "more equal than another", would be much
better.

The basic thing is that MPEG 2 and 4 was not designed for streaming - so, for
video especially, streaming is a kludge. Designing streaming video from scratch
would make a lot of sense.

>
>  -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-dev-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.


--
                                 Regards
                                 Marshall Eubanks

T.M. Eubanks
Multicast Technologies, Inc
10301 Democracy Lane, Suite 410
Fairfax, Virginia 22030
Phone : 703-293-9624
Fax     : 703-293-9609
e-mail : tme at on-the-i.com     tme at multicasttech.com

http://www.on-the-i.com http://www.buzzwaves.com

--- >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-dev-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.




More information about the Vorbis-dev mailing list