Re: [rtcweb] On video codec for rtcweb
"Schleef, David" <ds@ti.com> Fri, 23 March 2012 22:38 UTC
Return-Path: <ds@ti.com>
X-Original-To: rtcweb@ietfa.amsl.com
Delivered-To: rtcweb@ietfa.amsl.com
Received: from localhost (localhost [127.0.0.1]) by ietfa.amsl.com (Postfix) with ESMTP id 5F4C021F85D3 for <rtcweb@ietfa.amsl.com>; Fri, 23 Mar 2012 15:38:50 -0700 (PDT)
X-Virus-Scanned: amavisd-new at amsl.com
X-Spam-Flag: NO
X-Spam-Score: -10.599
X-Spam-Level:
X-Spam-Status: No, score=-10.599 tagged_above=-999 required=5 tests=[BAYES_00=-2.599, RCVD_IN_DNSWL_HI=-8]
Received: from mail.ietf.org ([12.22.58.30]) by localhost (ietfa.amsl.com [127.0.0.1]) (amavisd-new, port 10024) with ESMTP id ki1QpqYaZkDe for <rtcweb@ietfa.amsl.com>; Fri, 23 Mar 2012 15:38:49 -0700 (PDT)
Received: from devils.ext.ti.com (devils.ext.ti.com [198.47.26.153]) by ietfa.amsl.com (Postfix) with ESMTP id 74F7C21F856C for <rtcweb@ietf.org>; Fri, 23 Mar 2012 15:38:48 -0700 (PDT)
Received: from dlep26.itg.ti.com ([157.170.170.121]) by devils.ext.ti.com (8.13.7/8.13.7) with ESMTP id q2NMcmMx013147; Fri, 23 Mar 2012 17:38:48 -0500
Received: from DFLE70.ent.ti.com (localhost [127.0.0.1]) by dlep26.itg.ti.com (8.13.8/8.13.8) with ESMTP id q2NMcmH4014599; Fri, 23 Mar 2012 17:38:48 -0500 (CDT)
Received: from DLEE08.ent.ti.com ([fe80::8145:89a6:94e5:8cad]) by DFLE70.ent.ti.com ([fe80::db3:609a:aa62:e1ec%22]) with mapi id 14.01.0323.003; Fri, 23 Mar 2012 17:38:47 -0500
From: "Schleef, David" <ds@ti.com>
To: "chris.cavigioli@intel.com" <chris.cavigioli@intel.com>
Thread-Topic: [rtcweb] On video codec for rtcweb
Thread-Index: Ac0JHgA8Dirv+pElTfm4dOmisEvpPwAJ7Ryc
Date: Fri, 23 Mar 2012 22:38:46 +0000
Message-ID: <1332542323.27047.37.camel@ula0273821>
References: <4F6C5A5E.6050100@ericsson.com> <4F6C6138.6010908@mozilla.com> <E44893DD4E290745BB608EB23FDDB76219E813@008-AM1MPN1-041.mgdnok.nokia.com> <E36D1A4AE0B6AA4091F1728D584A6AD215EC3936@ORSMSX104.amr.corp.intel.com>
In-Reply-To: <E36D1A4AE0B6AA4091F1728D584A6AD215EC3936@ORSMSX104.amr.corp.intel.com>
Accept-Language: en-US
Content-Language: en-US
X-MS-Has-Attach:
X-MS-TNEF-Correlator:
x-originating-ip: [157.170.170.90]
Content-Type: text/plain; charset="utf-8"
Content-Transfer-Encoding: base64
MIME-Version: 1.0
Cc: "rtcweb@ietf.org" <rtcweb@ietf.org>
Subject: Re: [rtcweb] On video codec for rtcweb
X-BeenThere: rtcweb@ietf.org
X-Mailman-Version: 2.1.12
Precedence: list
List-Id: Real-Time Communication in WEB-browsers working group list <rtcweb.ietf.org>
List-Unsubscribe: <https://www.ietf.org/mailman/options/rtcweb>, <mailto:rtcweb-request@ietf.org?subject=unsubscribe>
List-Archive: <http://www.ietf.org/mail-archive/web/rtcweb>
List-Post: <mailto:rtcweb@ietf.org>
List-Help: <mailto:rtcweb-request@ietf.org?subject=help>
List-Subscribe: <https://www.ietf.org/mailman/listinfo/rtcweb>, <mailto:rtcweb-request@ietf.org?subject=subscribe>
X-List-Received-Date: Fri, 23 Mar 2012 23:01:21 -0000
There are several key factors about mobile devices that make your conclusion on H.264 Baseline a lot less strong than one might think. Power consumption for a VC application in current generation mobile devices is mostly due to running the camera, the display, and the 4G or wifi radio. Encoding/decoding, while obviously making *some* contribution, is not really significant at VC sizes and bitrates. And encoding/decoding in software doesn't contribute much more. Secondly, "hardware codec" can mean a lot of different things. It's almost always some form of DSP or microcontroller doing the sequencing of tightly-coupled hardware blocks that offload some of the heavy lifting. Software codecs on CPUs are just one end of this. Some vendors on the DSP end already support VP8 encoding/decoding in existing products. As it has for H.264, VP8 will inevitably progress toward more hardware/less processor in upcoming generations, and thus lower power. Also, not related to mobile, the primary (only?) goal of a baseline codec is to provide 100% coverage. To get to 100%, _someone_ will have to use a software codec. There are no barriers to an implementor using VP8 in software; that's sort of the entire point of free codecs. There are business barriers that will never change for companies like Red Hat to use H.264 in software. Thus, you will never get 100% coverage from H.264. Ever. Adding VP8 is, as you say, a "very trivial software burden". From my vantage point, your suggestion of H.264 as a baseline codec comes down to "want to save a few percent of power usage only for the current generation (which will barely see WebRTC anyway), willing to sacrifice actually attaining the goal of a baseline codec". David On Fri, 2012-03-23 at 12:34 +0000, Cavigioli, Chris wrote: > Let's be clear on profiles. If we seek widest mobile hardware footprint: > - ITU-T H.264 Constrained Baseline Profile is probably the most ubiquitous hardware video codec today with reasonable power/performance ratio > - ITU-T G.711, both u_law (North America) and A_law (Europe) formats, can be supported by everyone with very trivial software burden > -chris > > -----Original Message----- > From: rtcweb-bounces@ietf.org [mailto:rtcweb-bounces@ietf.org] On Behalf Of Markus.Isomaki@nokia.com > Sent: Friday, March 23, 2012 5:15 AM > To: tterriberry@mozilla.com; stefan.lk.hakansson@ericsson.com > Cc: rtcweb@ietf.org > Subject: Re: [rtcweb] On video codec for rtcweb > > Hi, > > Timothy B. Terriberry wrote: > > > >The primary reasoning behind that decision, i.e. the one > >factor that was not in our power to change, was the availability of content on > >websites. This is not a problem for WebRTC, as browsers control both ends of > >the communication. > > > > In WebRTC context what is somewhat analogous is the interconnect to existing non-web video systems. What codecs are used in them is not under browser control. I believe those mostly use H.264. It's probably possible to use transcoders but so it is between browsers too. I'd say that even if couldn't mandate a single video codec (as it has seemed to me since the beginning), in practice H.264 support would be highly RECOMMENDED (a SHOULD). There might be good reasons not to follow that SHOULD, as it seems to be the case for Mozilla. > > Markus > > > >-----Original Message----- > >From: rtcweb-bounces@ietf.org [mailto:rtcweb-bounces@ietf.org] On Behalf > >Of ext Timothy B. Terriberry > >Sent: 23 March, 2012 13:41 > >To: Stefan Hakansson LK > >Cc: rtcweb@ietf.org > >Subject: Re: [rtcweb] On video codec for rtcweb > > > >Stefan Hakansson LK wrote: > >> So, I propose that h.264 should be mandatory to implement. > > > >Mozilla strongly opposes such a proposal. Many will be familiar with our > >announcements regarding our use of platform codecs on B2G and possibly > >Android last week. The primary reasoning behind that decision, i.e. the one > >factor that was not in our power to change, was the availability of content on > >websites. This is not a problem for WebRTC, as browsers control both ends of > >the communication. > > > >What _is_ a problem is licensing and distributing encoders in an open-source > >project. This is quite a bit different than using system decoders, as outside of > >mobile, encoders are usually not available, and even if they were, they might > >not have particularly good quality, nor the features to control latency, loss > >robustness, and other aspects necessary for real-time communication. > > > >If you thought that a concession in one place would make Mozilla more likely > >to compromise in others, I assure you the opposite is true. > >Maintaining the use of RF codecs in WebRTC is now even more important than > >it was before. See the words directly from Mozilla's CTO: > >http://groups.google.com/group/mozilla.dev.platform/msg/53892ad3ea9eeb > >98 > >_______________________________________________ > >rtcweb mailing list > >rtcweb@ietf.org > >https://www.ietf.org/mailman/listinfo/rtcweb > _______________________________________________ > rtcweb mailing list > rtcweb@ietf.org > https://www.ietf.org/mailman/listinfo/rtcweb > _______________________________________________ > rtcweb mailing list > rtcweb@ietf.org > https://www.ietf.org/mailman/listinfo/rtcweb
- Re: [rtcweb] On video codec for rtcweb Bernard Aboba
- Re: [rtcweb] On video codec for rtcweb Matthew Kaufman
- [rtcweb] On video codec for rtcweb Stefan Hakansson LK
- Re: [rtcweb] On video codec for rtcweb Timothy B. Terriberry
- Re: [rtcweb] On video codec for rtcweb Lorenzo Miniero
- Re: [rtcweb] On video codec for rtcweb Markus.Isomaki
- Re: [rtcweb] On video codec for rtcweb Timothy B. Terriberry
- Re: [rtcweb] On video codec for rtcweb Cavigioli, Chris
- Re: [rtcweb] On video codec for rtcweb Markus.Isomaki
- Re: [rtcweb] On video codec for rtcweb Basil Mohamed Gohar
- Re: [rtcweb] On video codec for rtcweb Timothy B. Terriberry
- Re: [rtcweb] On video codec for rtcweb Timothy B. Terriberry
- Re: [rtcweb] On video codec for rtcweb Timothy B. Terriberry
- Re: [rtcweb] On video codec for rtcweb Timothy B. Terriberry
- Re: [rtcweb] On video codec for rtcweb Martin Thomson
- Re: [rtcweb] On video codec for rtcweb Matthew Kaufman
- Re: [rtcweb] On video codec for rtcweb Basil Mohamed Gohar
- Re: [rtcweb] On video codec for rtcweb Roman Shpount
- Re: [rtcweb] On video codec for rtcweb Tim Panton
- Re: [rtcweb] On video codec for rtcweb Serge Lachapelle
- Re: [rtcweb] On video codec for rtcweb Basil Mohamed Gohar
- Re: [rtcweb] On video codec for rtcweb Schleef, David
- Re: [rtcweb] On video codec for rtcweb Roman Shpount
- Re: [rtcweb] On video codec for rtcweb Gregory Maxwell
- Re: [rtcweb] On video codec for rtcweb Cameron Byrne
- Re: [rtcweb] On video codec for rtcweb Roman Shpount
- Re: [rtcweb] On video codec for rtcweb Jean-Marc Valin