Re: [MMUSIC] [rtcweb] About unified Plan and multiple m lines

[Eric Rescorla <[e-mail]>]

On Wed, Jul 24, 2013 at 9:09 AM, Emil Ivov <[e-mail]> wrote:

> (redirecting to mmusic)
>
>
> On 24.07.13, 16:29, Eric Rescorla wrote:
>
>>
>>
>>
>> On Wed, Jul 24, 2013 at 7:08 AM, Iñaki Baz Castillo <[e-mail]
>> <mailto:[e-mail]>> wrote:
>>
>>     Hi, let's suppose WebRTC 1.0 borns with "Unified Plan" built-in, and
>>     my browser receives a session invitation from a conference server, and
>>     there are 8 audio participants in the conference.
>>
>>     So the SDP I receive has 8 "m=audio" lines, maybe with different audio
>>     codecs in each line. And following SDP O/A rules my browser will
>>     generate SDP answer that will also contain 8 "m=audio" lines.
>>
>>     Now, where it is supposed my browser will send my single audio track
>>     to the conference server? What does it mean that I receive 8 "m=audio"
>>     lines while I'm just able to send a single audio track? which m=audio
>>     line should I use for sending RTP?
>>
>>
>> One of them will be a=sendrecv. The others will be a=sendonly.
>> The sendrecv line is the one you use to send media.
>>
>
> I have a few questions here but before that, could you please clarify the
> following:
>
> Are you saying that the above is going to be the case on the offer that
> the callee receives? In other words, the caller "knew" that the callee
> intended to only send one stream so it preset all directions accordingly?


It depends on the context obviously. But in the case of (for instance) a
Hangouts
like system if the conference bridge were the offerer, it would send a pile
of
m-lines with only one being sendrecv (or maybe even recvonly) for the
user's media and the rest being sendonly.

-Ekr