- From: Peter Thatcher <pthatcher@google.com>
- Date: Tue, 17 Jun 2014 10:11:37 -0700
- To: "public-ortc@w3.org" <public-ortc@w3.org>
- Message-ID: <CAJrXDUFC6SfkU9_ed4rXkuvYogCp+FEKwKu0e6TqxrRKN6zkZg@mail.gmail.com>
I was going to write 3 emails covering 3 topics, but Bernard already covered all three topics in his recent flurry of big emails covering lots of stuff. But just to be complete, I'll reiterate my support for the changes he made that covers these things: 1. RTCP parameters. We need parameters for saying "don't use compound RTCP" and "don't use RTCP mux" and "use this ssrc for RTCP". We already have the last 2 and need to add the last one. Now that we have three, we might as well make a separate "RTCP parameters" object. And that's exactly what Bernard proposed, so I like that. 2. We need to control which header extensions are used on a per-codec basis. This does *not* mean control over the ID/URI mapping per codec, but just controlling if the header extension is used, with the mapping determined already. Thus, codecs need a way to say "use these header extensions" and refer to the. I think doing so by URI is a good choice, which is what Bernard proposed. I think the default if the codec doesn't specify which header extensions to use, would be to use all of the header extensions that are defined and make sense (ie no audio level header extension with a video codec). 3. We need a way for the browser to tell the JS the maximum number of spatial layers it can support, if there is a limit. Bernard also covered that, and included quality and temporal layers as well. Since we don't have control in parameters for the number of quality and temporal layers yet, I'm not sure what the JS would do with that info, but it seems reasonable to include that now and put in knobs for those things in parameters later. That's it. Thanks for doing so much work, Bernard!
Received on Tuesday, 17 June 2014 17:12:43 UTC