senders vs receivers codec selection

My understanding was this:
A RTCRtpSender picks a single codec for sending but a receiver can 
receive on any of the defined RTCRtpCodecs.

Here's a few questions:
a) If a sender wants to change sending codecs, do they setup a new 
sender object?

b) If a receiver latches onto a stream with a particular payload type, 
is that latch permanent (i.e. if sender changes sending codecs, does a 
new receiver need to be setup to receive a payload type even on the same 
SSRC)?


As things are defined now, the sender must pick a sending codec. The 
receiver does not have enough information to receive upon any codec 
because the layering information only works for one codec's scenario 
(despite having a list of codecs available) thus the receiver can only 
receive upon one payload type.

-or-

The receiver can only work on one codec if using SVC, but can switch 
between codecs if the other codecs are not configured using SVC (since 
only one SVC scenario can be setup).


Therefor, it's doesn't appear to be currently possible for the receiver 
to configure a bunch of SVC codecs and then the sender pick which one to 
use, even though the receiver can configure a bunch of non-SVC codecs 
and can have the sender pick which one to use.


I want to confirm my understanding is correct (and if others agree that 
this should be correct).

-Robin

Received on Sunday, 11 May 2014 14:41:47 UTC