[rtcweb] VP8 "profile" text - proposed replacement

Harald Alvestrand <harald@alvestrand.no> Fri, 28 November 2014 15:36 UTC

Message-ID: <54789695.2010406@alvestrand.no>
Date: Fri, 28 Nov 2014 07:36:53 -0800
From: Harald Alvestrand <harald@alvestrand.no>
User-Agent: Mozilla/5.0 (X11; Linux x86_64; rv:31.0) Gecko/20100101 Thunderbird/31.2.0
MIME-Version: 1.0
To: "rtcweb@ietf.org" <rtcweb@ietf.org>
Content-Type: text/plain; charset="utf-8"
Content-Transfer-Encoding: quoted-printable
Archived-At: http://mailarchive.ietf.org/arch/msg/rtcweb/qFWE8hdXC7L72ZYIii-msPk-ZCo
Subject: [rtcweb] VP8 "profile" text - proposed replacement
Precedence: list

The current text in draft-ietf-rtcweb-video-03 text about VP8 that I
commented on in Hawaii, and my proposed replacement:

** OLD **

   For the VP8 codec, defined in [RFC6386], endpoints MUST support the
   payload formats defined in [I-D.ietf-payload-vp8].  In addition they
   MUST support the 'bilinear' and 'none' reconstruction filters.

** NEW **

   For the VP8 codec, defined in [RFC6386], endpoints MUST support the
   payload formats defined in [I-D.ietf-payload-vp8].
   All VP8 decoders have to support all features of the VP8 bitstream as
   defined in [RFC6386], including all 4 defined values (0-3) of the
"version"
   number defined in section 9.1 of that document.

** END CHANGE **

RFC 6386 isn't very clear about the fact that all features must be
implemented by all decoders, but that has been the reality since VP8 was
first released. So the new sentence should be equivalent to saying
nothing, but it is OK to be super-clear about this.

I believe this is already clarified in the MPEG version of the VP8
specification too.

-- 
Surveillance is pervasive. Go Dark.

[rtcweb] VP8 "profile" text - proposed replacement Harald Alvestrand