[vorbis] Vorbis over RTP

[Roland Parviainen]

Monty wrote:
> 
> > So the codebook is sent out once per second, as 3 or 4 fragments ? If there is
> > 10% packet, then the chances of losing the codebook at ~ 30-40% - i.e., quite good.
> 
> (note) This is why the design I did uses an out of band TCP connect to
> quickly snarf headers before listening to multicast.  Of course, the
> TCP solution has its own drawbacks-- I just find them more manageable
> and they don't add a background bitrate overhead.
> 

For now it works good if: the users are few (multicast isn't that
widespread yet) and 
we only need the headers once. If we wan't to send a stream with
different headers we don't won't thousands of TCP connections in the
same instant... 

I currently use mp3 for a jukebox/radio-like application where the users
vote for songs they want to hear, I don't won't the server to have to
handle large amount of TCP connections too....

> > > The payload header also contains a 8 bit id that changes when client
> > > have to reset the decoder with a new codebook, so there is no problem
> > > with sending e.g. a 128 kbit/s file followed by a 160 kbit/s file,
> > > except
> > > for a small delay.
> > >
> >
> > So once you get it, you don't need it again, so if you loose it, you
> > can just use the last one ?
> 
> I don't fully understand.  Codebooks/encoder setup/state can't be
> counted in a small number of bits.  Whenever the ID changes, you need
> to make sure you have current headers.  The ID mechanism makes sure
> that if you have header set # 84, you're using it on audio packets
> also stamped 'set # 84'.  The streaming server will have to increment
> the ID when a new logical stream begins.
> 
> Monty
> 
Yes, that is what I do. Whenever the clients need new headers I add one
to the id. If the streams have a minimum length one bit would be enough.
(When the bit changes we through out any state and start all over again,
and hope that a late packet doesn't arrive)

//Roland

--- >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.