W3C home > Mailing lists > Public > public-webrtc@w3.org > October 2014

Re: [Bug 20810] SDP inadequately defined

From: Stefan Håkansson LK <stefan.lk.hakansson@ericsson.com>
Date: Thu, 30 Oct 2014 13:41:10 +0000
To: tim panton <thp@westhawk.co.uk>
CC: "Matthew Kaufman (SKYPE)" <matthew.kaufman@skype.net>, Harald Alvestrand <harald@alvestrand.no>, "public-webrtc@w3.org" <public-webrtc@w3.org>
Message-ID: <1447FA0C20ED5147A1AA0EF02890A64B1D0AF188@ESESSMB209.ericsson.se>
On 30/10/14 14:34, tim panton wrote:
>
> On 30 Oct 2014, at 13:22, Stefan Håkansson LK
> <stefan.lk.hakansson@ericsson.com
> <mailto:stefan.lk.hakansson@ericsson.com>> wrote:
>
>> On 29/10/14 23:33, Matthew Kaufman (SKYPE) wrote:
>>>
>>>
>>> Matthew Kaufman
>>>
>>>> -----Original Message----- From: Harald Alvestrand
>>>> [mailto:harald@alvestrand.no] Sent: Tuesday, October 28, 2014 8:35
>>>> PM To: public-webrtc@w3.org <mailto:public-webrtc@w3.org> Subject:
>>>> Re: FW: [Bug 20810] SDP
>>>> inadequately defined
>>>>
>>
>>>> But the WEBRTC (W3C) bugtracker is about tracking issues that can
>>>> be solved by *modifying W3C specifications*. There is nothing here
>>>> that can be fixed by modifying a W3C specification.
>>>
>>> Sure it can. The W3C specification could specifically call out the
>>> existing specs explicitly in such a way that one would know what to
>>> implement.
>>>
>>> The W3C specification could more explicitly define the specific SDP
>>> that is to be generated and understood.
>>>
>>> The W3C specification could *not use SDP as an API* and then this
>>> problem goes away entirely.
>>
>> Is that really the case? In the W3C document, as far as I can see, the
>> actual SDP (or RTCSessionDescription) is never dealt with in any other
>> way than applying to the PeerConnection or sending off to a remote peer.
>>
>
> And yet every time I build something in webRTC, I almost always need to
> extract and parse and
> often modify the SDP.
> Recent examples include :
> Displaying the video bandwidth used - The stats object doesn’t currently
> specify the stream type - just the codec - so you need to
> dig in the sdp to deduce the video ssrc id, then use that to filter the
> streams.

Seems like the stats object should be updated.

> Reducing the audio bandwidth to work over 2.5g - need to set the average
> bitrate with code like this:
>                          if (sdpLines[i].search("a=rtpmap:") == 0) {
>                              var bits = sdpLines[i].split(" ");
>                              if (bits[1].search("opus") == 0) {
>                                  var num = bits[0].split(":")[1];
>                                  var line = "a=fmtp:" + num + "
> minptime=50; maxaveragebitrate=8000;";
>                                  sdpLines.splice(i, 0, [line]);
>                              }
>                          }

We'll talk about RTPSenders today, this kind of functionality should be 
offered on that API surface (if there is need for it).

> Tim.
>
Received on Thursday, 30 October 2014 13:41:36 UTC

This archive was generated by hypermail 2.3.1 : Monday, 23 October 2017 15:19:42 UTC