[Asterisk-Dev] IAX codec negociation (again...)

Goldenear goldenear at free.fr
Mon Jun 20 08:50:43 MST 2005

Kevin P. Fleming a écrit :

> Goldenear wrote:
>> When I see how codec négociation works with sip/ser, I just guess it 
>> could be as easy with * :)
> Asterisk is not a proxy.

I know it :) By the way, does any iax2 proxy exist ? (like ser but for IAX2)

>> I'm wondering, couldn't * ask both sides of a call their codecs 
>> capabilities before "bridging" them and so force them to use  the 
>> same codec if there is a common one ?
> Asterisk is not a proxy. The codec decision for the peer originating 
> the call is made long before we have any idea if the call is going to 
> be terminated to another peer (or an application, or anything else).
You're right, but as I wrote in an other post, it would be nice to have 
the choise between the actual way asterisk works and they way I'm 
talking about.

>> In this way, it works :) A switch from gsm to ulaw during the call.
>> So, I wonder why having asterisk in the path would break codecs 
>> nogociation.
> Asterisk is not a proxy :-) Putting it in between two peers will _not_ 
> be transparent, under any circumstances. In this scenario, A does not 
> 'switch from gsm to ulaw", the call is set up that way and never 
> changes. What happens if B transfers A to peer C, which only supports 
> GSM? How would you expect Asterisk to handle that?
wouldn't it be possible to have a codec renegociation after each tranfer 
(if configured) ?

> With all that said, it _is_ possible for this to be done, but it will 
> require Asterisk to be able to know when it needs to renegotiate a 
> codec with a peer, and to initiate that operation, and peers/protocols 
> that support it.

we may have a "codecrenegotiate=yes" for peers in iax.conf or sip.conf

> _______________________________________________
> Asterisk-Dev mailing list
> Asterisk-Dev at lists.digium.com
> http://lists.digium.com/mailman/listinfo/asterisk-dev
> To UNSUBSCRIBE or update options visit:
>   http://lists.digium.com/mailman/listinfo/asterisk-dev

More information about the asterisk-dev mailing list