[Fwd: Re: [JDEV] Videoconferencing with jabber / Re: [speex-dev] Videoconferencing with speex and jabber]

[Jean-Marc Valin]

> to speex: *now here comes the more important part, can we build a c++
> component which does what avrelay does? is it practicable to de/encode
> 100 streams with a c/c++ speex de/encoder in realtime?* COMMENTS WELCOME

At low bit-rate (6-8 kbps) and lowest complexity, it's probably possible
to encode 100 streams on a 3 GHz machine (and decoding is cheap), but
that's all you're going to do on that machine. However, I tend to favor
the approach of sending multiple streams instead of doing a
decode-mix-reencode. For instance, if you're encoding at 8 kbps, you can
transmit up to two streams, which would make the maximum bit-rate 16
kbps (max 16 kbps), with no loss of quality. If you decide to go with
mixing, you'll probably have to go with 18.2 kbps because Speex is not
tuned for encoding multiple voices. Even then, the quality will likely
be lower because of the tandeming (reencoding). The only case where
mixing might be good is when working with high bit-rates (e.g. 15 kbps
narrowband or 28 kbps wideband).

        Jean-Marc


-- 
Jean-Marc Valin, M.Sc.A., ing. jr.
LABORIUS (http://www.gel.usherb.ca/laborius)
Université de Sherbrooke, Québec, Canada


-------------- next part --------------
A non-text attachment was scrubbed...
Name: signature.asc
Type: application/pgp-signature
Size: 190 bytes
Desc: Ceci est une partie de message numériquement signée
Url : http://lists.xiph.org/pipermail/speex-dev/attachments/20031126/81230dde/signature.pgp