Re: [rtcweb] 5 tuples and rmcat-cc-requirements-01

On 01/01/14 02:47, Martin Thomson wrote:
> On 29 December 2013 16:02, Dave Taht <dave.taht@gmail.com> wrote:
>> In current implementations, is it possible to force flows onto
>> different tuples? What are the other flaws in suggesting or requiring
>> a different tuple?
>
> There are quite a few ways to achieve this, depending on your
> configuration.  All will require some fair amount of SDP cracking.
>
> If you have browsers at both ends, try creating two RTCPeerConnection
> objects on each, and treat one pair as audio and the other video.  You
> can do this by telling the answerer that audio (or video) is not
> wanted in the offer (set port zero for the unwanted component).
> offer.sdp.replace(/m=audio \d+ /, 'm=audio 0 ') should be done before
> sending the offer out (I don't know whether you can pass this to
> setLocalDescription safely, so I wouldn't).

One way of doing it without any SDP cracking would be to create two 
different MediaStream's out of the one obtained from getUserMedia; one 
with the audio MediaStreamTrack only, and one with the video one.

Then use one of them with one PeerConnection, and the other with another 
PeerConnection, and assemble into a MediaStream with audio and video at 
the receiving end.

(And this becomes more straightforward if we move to adding tracks and 
not streams to PeerConnection)

If you want to use one PeerConnection only, I think the simplest way 
would be to remove bundle from the SDP offer - either before setLocal or 
before sending the offer to the remote endpoint.

> _______________________________________________
> rtcweb mailing list
> rtcweb@ietf.org
> https://www.ietf.org/mailman/listinfo/rtcweb
>