[asterisk-dev] SIP: handling multiple m=video or m= audio lines

[Olle E. Johansson]

10 sep 2009 kl. 11.33 skrev Vadim Lebedev:

> Olle E. Johansson wrote:
>>
>> 9 sep 2009 kl. 23.02 skrev Mark Michelson:
>>
>>

>>>
>>>
>> And, to repeat myself, in order to make sure that we answer  
>> everything
>> properly, including totally to us unknown media streams, we need to
>> save it in text format.
>>
>> /O
>>
>
> I think it would be good occasion to implement better handling of  
> fmtp attributes for streams
> handled in transparent way.  I mean that if endpoint A  calling  
> endpoint B has a fmtp attribute in video stream descriptor it should  
> be fowarded to B by asterisk.
> Today it is scrapped.

And yes, there we open the flood gates for a new media architecture.  
If we start working with these, there will be a lot of work. Check the  
videocaps branch where we try to map all of these combinations and  
make i twork better. The big problem is not really A calling Asterisk  
calling B. It's B's answer that get translated into a binary YES/NO  
and is returned that way, without any FMTPs in the ANSWER to the first  
SDP offer. Videocaps is one solution, that has been input for a huge  
discussion and outlines for a new architecture. I haven't seen any  
work begin though and videocaps is now multiple years old, but still  
used in production in many places. Looking back, we propably should  
have merged it at the point when it was contributed to the project by  
Aupix in the UK. It's easy to say that afterwards though.

In short, just forwarding FMTPs sound easy but is not. I don't think  
it's part of this short term project focused on fixing the SDP stream  
handling, but part of the new media architecture/codec negotiation  
project that we all wish we had proper  resources to be able to start.

Cheers,
/O