[rtcweb] Video codecs and the staw poll

Hervé W. <H.O.W.aka.V+ietf@gmail.com> Thu, 30 January 2014 16:05 UTC

MIME-Version: 1.0
Sender: h.o.w.aka.v@gmail.com
In-Reply-To: <9E34D50A21D1D1489134B4D770CE0397680AAFC5@SZXEMA504-MBX.china.huawei.com>
References: <BFDBDCA9-937E-4B90-97B1-A23EEB65CF9A@iii.ca> <9E34D50A21D1D1489134B4D770CE0397680AAFC5@SZXEMA504-MBX.china.huawei.com>
Date: Thu, 30 Jan 2014 17:05:37 +0100
Message-ID: <CAKBXTOecGK3raXXPN5Wt65GS23NZPXci7JtAyCyJ5uaRXvsSag@mail.gmail.com>
From: "Hervé W." <H.O.W.aka.V+ietf@gmail.com>
To: "Chenxin (Xin)" <hangzhou.chenxin@huawei.com>, "rtcweb@ietf.org" <rtcweb@ietf.org>
Content-Type: multipart/alternative; boundary="001a11c37082679ff204f1323a5c"
Subject: [rtcweb] Video codecs and the staw poll
Precedence: list

On Tuesday, 28 January 2014, Chenxin (Xin) <hangzhou.chenxin@huawei.com>
wrote:
> +1
>
> One thing could be decided by the poll is that the MTI video codec
candidates are only { H.264, VP8}. Will eliminate many discussions....
>
> Best Regards,
>      Xin
>
>
>>-----Original Message-----
>>From: rtcweb [mailto:rtcweb-bounces@ietf.org] On Behalf Of Cullen Jennings
>>Sent: Tuesday, January 28, 2014 1:13 PM
>>To: rtcweb@ietf.org
>>Subject: [rtcweb] Video codecs and the staw poll
>>
>>
>>Dear WG,
>>
>>After reviewing the poll results found here:
>>http://www.ietf.org/mail-archive/web/rtcweb/current/pdfWd2PIhOY9y.pdf the
>>chairs concludes that the working group still believes that an MTI is
required for
>>the WebRTC ecology to develop.    There are a number of options which did
>>not garner significant support; essentially only options 1, 2, 3, 4 seem
to have
>>enough support that they might be the eventual basis of working group
>>consensus.  The chairs do not view the other options as having sufficient
>>support to warrant further working group activity or discussion.
>>
>>There is no obvious leader between VP8 and H.264, however, nor obvious
>>support for selecting both.  Each has similar numbers of supporting
positions
>>and objections, and both have the support of well over half the
participants in
>>the straw poll.  Given that, we are no closer to being able to choose
between
>>them at this time.


Unless something comes out of MPEG or MPEG LA or the VP8 Nokia
claims/lawsuits
are resolved, I think it's unlikely the IPR related objections will be
lifted.

H.261 did not get the numbrrs, but also got some objections that seemed of
a technical nature,
but were either relying on other peoples opinions or were expressions of
doubt on performance.
(Not all objections were like that, but some were.)
If the stalemate continues into September, I suggest those objectors test
encoder
implementations themselves and/or view the already produced samples and
form their own opinions.

An option that didn't make it to the straw poll: Reconsider Mpeg-1 video
(first release 1993)
 related message:
http://www.ietf.org/mail-archive/web/rtcweb/current/msg09794.html
+ demonstrably better quality/bitrate than H.261 and MJPEG
+ arbitrary resolutions unlike H.261
+ lawsuits and patents concerning H.262/MPEG-2 Video (first release
1995/1996) don't necessarily relate to MPEG-1 implementations;
* 3 years in between releases
* patents on 20xx tech are unlikely to apply to 1993 implementations

As far as I know, the H.261 samples were largely produced without any data
on the target devices.
When the stalemate continues, can we get some data on those? Available
bitrate, resolution, CPU


>>The chairs therefore propose tabling the discussion of a mandatory to
>>implement video codec until about 6 week before the start of the IETF 91
>>meeting in November 2014. This is so that the working group can focus its
>>energy on completing other work.  We do expect to begin work on the video
>>document (draft-ietf-rtcweb-video) to meet its milestone of December, but
>>initially without specifying which of the two codecs is the WG consensus
for MTI.


+1


>>When we return to the discussion, the working group chairs currently
expect to
>>run a consensus call on support for each codec to be mandatory to
implement.
>>This expectation may change, however, based on new information or working
>>group experience.
>>
>>If anyone has concerns about tabling this discussion until September 29,
2014
>>please let us know by February 4.
>>
>>Thank you,
>>
>>Cullen, Magnus, Ted <the chairs>


- Hervé

Re: [rtcweb] Video codecs and the staw poll Peter Saint-Andre
Re: [rtcweb] Video codecs and the staw poll Eric Rescorla
Re: [rtcweb] Video codecs and the staw poll Chenxin (Xin)
Re: [rtcweb] Video codecs and the staw poll Stefan Håkansson LK
[rtcweb] Video codecs and the staw poll Cullen Jennings
Re: [rtcweb] Video codecs and the staw poll Christer Holmberg
Re: [rtcweb] Video codecs and the staw poll Simon Pietro Romano
Re: [rtcweb] Video codecs and the staw poll Victor Pascual Avila
Re: [rtcweb] Video codecs and the staw poll Roman Shpount
Re: [rtcweb] Video codecs and the staw poll Mary Barnes
Re: [rtcweb] Video codecs and the staw poll Harald Alvestrand
Re: [rtcweb] Video codecs and the staw poll Ted Hardie
Re: [rtcweb] Video codecs and the staw poll Martin Thomson
Re: [rtcweb] Video codecs and the staw poll Peter Saint-Andre
Re: [rtcweb] Video codecs and the staw pol Bernard Aboba
Re: [rtcweb] Video codecs and the staw poll OSCAR DIVORRA ESCODA
Re: [rtcweb] Video codecs and the staw poll Alexandre GOUAILLARD
[rtcweb] Video codecs and the staw poll Hervé W.
Re: [rtcweb] Video codecs and the staw poll Jeremy Laurenson (jlaurens)