W3C home > Mailing lists > Public > public-html@w3.org > April 2007

Re: Mandated Video Format

From: Henri Sivonen <hsivonen@iki.fi>
Date: Fri, 6 Apr 2007 11:19:43 +0300
Message-Id: <C49EB3D7-613F-4E57-96FC-68C27C45974A@iki.fi>
Cc: <Chris.Wilson@microsoft.com>, <doug.schepers@vectoreal.com>, <public-html@w3.org>
To: Robert Brodrecht <w3c@robertdot.org>

On Apr 6, 2007, at 04:09, Robert Brodrecht wrote:

> Henri Sivonen said:
>> <img> de facto mandates GIF87a, GIF89a, PNG and Huffman-coded
>> Baseline JFIF, which leads to interop.
>>
>> If Microsoft supports WMV9/VC-1, Apple supports a subset of H.264,
>> Opera supports Theora and Mozilla supports Dirac (just illustrating
>> the point; no actual Mozilla commitment to Dirac implied), interop
>> won't be achieved.
>
> Would a few baseline de facto video formats not eventually appear?   
> I've
> said it before, and I still may be naive: in this age of moving toward
> interoperability between browsers, I don't see why the four major  
> vendors
> (IE, Moz, Opera, and Apple) can't come to some diplomatic decision  
> about
> what formats could be supported.  There has to be some overlap  
> somewhere.
> If we can get IE and Apple on one codec

Microsoft and Apple have taken different sides in the codec battle  
before <video>. So far, the market hasn't sorted this out.

Apple has put it weight behind H.264 (standardized at ISO, subject to  
an MPEG LA patent portfolio). Microsoft has put its weight behind  
WMV9 (standardized after the fact as VC-1 at SMPTE, subject to an  
MPEG LA patent portfolio).

> and Moz and Opera on another,

That's a reasonable expectation.

> This is better than 4 different codecs.

Well, you'd get three codecs unless you can get Apple and Microsoft  
to solve their pre-existing bigger codec battle.

> Further, I'm still of the opinion that the video does not have to  
> be HDTV
> quality.  As a web designer, I'm more concerned about the file  
> being easy
> to create and implement.  If anything, Flash has provided that  
> ability.
> Their videos aren't amazing quality.  But they are relatively small  
> and
> they work without much fuss.  That is what <video> is competing  
> against.
> Maybe there is some other "more open" codec that is older and less
> restrictive that is RF and more distributable?

That would be Theora.

If you want older than that and less submarine risk, you get Motion  
JPEG (no interframe compression) and uncompressed PCM audio in an AVI  
container, which is totally unsuited for Web used due to the size of  
files.

The reality of the matter is that with the current patent regime in  
place, every video compression method that is technically suitable  
for Web use carries a submarine patent risk. Due to the way the  
nature of patents has been twisted, it is impossible to do an  
exhaustive patent search. Theora is RF for all known patents. H.264  
and VC-1 are not RF and licensing the known patents from MPEG LA does  
not remove the submarine risk. Moreover, the MPEG LA patent license  
is incompatible with copyleft licensing.

> At this point, the lines have clearly been drawn (ignoring  
> Mozilla's lack
> of opinion, AFAIK):

I cannot give an official Mozilla opinion, so I just point out that  
the patent licensing for H.264 and WMV9/VC-1 is incompatible with the  
licensing model of Firefox and the licensing of Theora is compatible  
with the licensing model of Firefox.

-- 
Henri Sivonen
hsivonen@iki.fi
http://hsivonen.iki.fi/
Received on Friday, 6 April 2007 08:35:45 UTC

This archive was generated by hypermail 2.3.1 : Monday, 29 September 2014 09:38:42 UTC