W3C home > Mailing lists > Public > public-webrtc@w3.org > June 2013

Re: SDP wrapper? Object-oriented API?

From: Iñaki Baz Castillo <ibc@aliax.net>
Date: Mon, 17 Jun 2013 12:54:24 +0200
Message-ID: <CALiegf=XJaMpOBPWA8Z3dwh8pGTjnS16YRipeTm6MWofatkueg@mail.gmail.com>
To: tim panton <thp@westhawk.co.uk>
Cc: public-webrtc <public-webrtc@w3.org>
2013/6/17 tim panton <thp@westhawk.co.uk>:
> Currently as a javascript coder I have to :
> 0) intercept the offer
> 1) look for an m=audio line
> 2) look for a subsequent a=rtpmap lines
> 3) remove all a=rtpmap lines that don't contain OPUS or telephone-event
> 4) rewrite the m=audio line so that it only contains the types for opus and telephone-event
> 5) make sure 2->4 only apply to lines before a subsequent m= line.
> 6) set the re-written offer as a  local description and send it.

It can be worse:

- A offers OPUS and ALAW (in that order or preference).
- B replies with OPUS and ULAW (in that order or preference).

At this point A could send audio in ULAW codec, and B could send audio
in ALAW codec. A would detect the exact codec it is receiving from B
by inspecting the RTP payload type (same for B). Asymmetric and
anti-expectable.

Just another "cool" feature of SDP, the monolithic, unmanageable
anti-API-able multimedia description format imposed in WebRTC for
"legacy" interop (but at the same time those "legacy" decives need to
implement SRTP-DTLS, ICE, rtcp-mux, SSRCs, MSIDs, etc).

Fail.


--
Iñaki Baz Castillo
<ibc@aliax.net>
Received on Monday, 17 June 2013 10:55:12 UTC

This archive was generated by hypermail 2.3.1 : Monday, 23 October 2017 15:19:33 UTC