[vorbis] bitrate peeling

[Jack Moffitt]

> Yes, this would be a killer solution, coupled with a proper transfer 
> protocol for streaming, in which the client sends quality information 
> back to the server, so it knows 'how much' to send.

It can be done on TCP, and ____ (I forgot his name, unfortunately) wrote
test code that did this for genereated test streams.  The basic idea is
to use the same logic that TCP itself uses to determine how fat the
pipes are.  It seemed to work quite well.

> This is all done by RealAudio for serveral years now though.

RealAudio has certainly had SureStream for several years, but I'd make
the case that this isn't the same thing.  You take a large quality hit
with each jump and I don't know if it works on live streams.  It's
certainly not as practical as just encoding one stream.

Also, with SureStream, I believe it's also still quite obvious when it
switches, usually causing a rebuffer, etc, but it's been a while since
I've used it.  I think a well implemented peeling streamer would blow
the doors off of it, but of course, RealAudio works now, and I'm talking
about code that has yet to be written.

jack.
--- >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.