Re: Some ideas on SVC support in WebRTC 1.0 (Take 2)

mediasoup implements RID and uses it for handling Firefox's simulcast:

https://github.com/versatica/mediasoup/blob/master/worker/src/RTC/RtpListener.cpp?ts=2#L251




El jue., 12 jul. 2018 21:25, Bernard Aboba <Bernard.Aboba@microsoft.com>
escribió:

> Inaki said:
>
>
> "SFUs may just implement RID"
>
>
> [BA] Which SFUs implement RID (and MID)?
> ------------------------------
> *From:* Iñaki Baz Castillo <ibc@aliax.net>
> *Sent:* Thursday, July 12, 2018 11:48:35 AM
> *To:* Bernard Aboba
> *Cc:* Harald Alvestrand; WebRTC WG
> *Subject:* Re: Some ideas on SVC support in WebRTC 1.0 (Take 2)
>
> IMHO, SFUs may just implement RID (as some of them already do) and we are
> done.
>
>
> El jue., 12 jul. 2018 20:12, Bernard Aboba <Bernard.Aboba@microsoft.com>
> escribió:
>
> Harald said:
>
>
> "Bernard, I believe you reported some pushback on the removal from SFU
> vendors in relation to the discussion about switching to Unified Plan
> (where explict SSRCs in the SDP is not required).
>
> But I can't find an issue on that now. Do we need to raise one?"
>
>
> [BA] Here is the issue: https://github.com/w3c/webrtc-pc/issues/1174
>
>
> Since SFUs currently do not support RID header extensions and SSRC
> multiplexing is widely used, SFU vendors need a way to obtain the SSRCs so
> as to allow them to be signaled.
>
>
> At this point, it appears that "a=ssrc" lines will be included in Unified
> Plan implementations, so that SFU vendor needs can be met that way.  Given
> this, there did not appear to be a compelling reason to revert the removal
> of SSRCs from the object model.
>
>
>
> ------------------------------
> *From:* Harald Alvestrand <harald@alvestrand.no>
> *Sent:* Wednesday, July 11, 2018 11:57:11 PM
> *To:* public-webrtc@w3.org
> *Subject:* Re: Some ideas on SVC support in WebRTC 1.0 (Take 2)
>
> Bernard, I believe you reported some pushback on the removal from SFU
> vendors in relation to the discussion about switching to Unified Plan
> (where explict SSRCs in the SDP is not required).
>
> But I can't find an issue on that now. Do we need to raise one?
>
> Den 11. juli 2018 18:59, skrev Bernard Aboba:
> > Harald said:
> >
> > "> The problem of having each SVC layer defined in a
> >> RTCRtpEncodingParameters entry (within encodings array) is that all
> >> those entries belonging to the same stream must share LOT of fields
> >> (such as pt, ssrc, etc). Super error prune IMHO.
> >>
> >
> > RTCRtpEncodingParameters doesn't contain either pt or ssrc. Perhaps it
> > should have for the simulcast case (didn't we have an issue on that?),
> > but at present it doesn't.
> >
> > These would be read-only, I think, so the only source of error is the
> > browser, not the user's app."
> >
> > [BA] Right.
> >
> > We removed PT and SSRC from RTCRtpEncodingParameters as well as
> RTCRtpRtxParameters and RTCRtpFecParameters because we did
> > not expect WebRTC 1.0 browser applications (as opposed to SDP-consuming
> > servers) to use this (read-only) information.
> >
> > BTW, it does not appear necessary to reintroduce those parameters to
> support SVC, assuming we only support SRST codecs.
> >
> >
> > "
> >
>
>
>

Received on Thursday, 12 July 2018 19:49:37 UTC