Re: FEC ssrc

On 20 Oct 2014, at 12:45, Sergio Garcia Murillo <sergio.garcia.murillo@gmail.com<mailto:sergio.garcia.murillo@gmail.com>> wrote:

Making the FEC stream has it's own SSRC and avoiding the current RED/FEC non-sense has always seemed the correct approach to me. So this are great news!


There is also rmcat-adaptive-fec, which uses FEC for adapting media stream, For example when the media streams are data-limited or the sender doesn’t want to increase the encoding rate but still wants to probe the network.

Do you know if this has been already presented/discussed at rtcweb?

This should be discussed at the upcoming IETF meeting, we are trying to figure out if it will be AVTCore/Ext, the appropriate chairs will respond. We’ll update the SDP section, it is not yet bundle ready.

But reviews on the document is appreciated :)

In regards ORTC, this opens another question, how to associate several SSRCs in one RTPSender/RTPReceiver or be able to group different objects together. This seems to be already an issue with RTX packets (and I was already working on an email in this regards).


Repeating from another email (that I sent today) about FECing packets across media streams, if that is out-of-scope?

Best regards
Sergio
On 20/10/2014 11:31, Singh Varun wrote:
There are several issues with RFC5109, we have attempted to fix those in https://tools.ietf.org/html/draft-singh-payload-rtp-1d2d-parity-scheme-00. I hope WebRTC folks will review and give feedback. Discussion on the PAYLOAD WG, and Github if you have pull requests https://github.com/vr000m/payload-1d2d-parity-fec/


      Synchronization Source (SSRC): The SSRC value SHALL be randomly

      assigned as suggested by [RFC3550<https://tools.ietf.org/html/rfc3550>].  This allows the sender to
      multiplex the source and repair flows on the same port, or
      multiplex multiple repair flows on a single port.  The repair
      flows SHOULD use the RTCP CNAME field to associate themselves with
      the source flow.


On 17 Oct 2014, at 23:19, Sergio Garcia Murillo <sergio.garcia.murillo@gmail.com<mailto:sergio.garcia.murillo@gmail.com>> wrote:

Yes sorry, dyslexia, quite late here..  :)

On 17/10/2014 22:16, Peter Thatcher wrote:
I think you mean 5109

http://tools.ietf.org/html/rfc5109

On Fri, Oct 17, 2014 at 1:11 PM, Sergio Garcia Murillo <sergio.garcia.murillo@gmail.com<mailto:sergio.garcia.murillo@gmail.com>> wrote:
Hi all,

One question, in the current draft it is stated:

9.10.1 Dictionary RTCRtpFecParameters<http://ortc.org/wp-content/uploads/2014/08/ortc.html#idl-def-RTCRtpFecParameters> Members

mechanism of type DOMString

The Forward Error Correction (FEC) mechanism to use.

ssrc of type unsigned long

The SSRC to use for FEC.


But according to RFC 5190:

   Synchronization Source (SSRC): The SSRC value SHALL be the same as
   the SSRC value of the media stream it protects.

So it doesn't seem right to allow configuring the ssrc of the FEC stream

Best regards
Sergio

Received on Monday, 20 October 2014 13:44:35 UTC