[whatwg] <video> element proposal

[Maik Merten]

Geoffrey Sneddon schrieb:
> 
> On 4 Mar 2007, at 14:08, Maik Merten wrote:
> 
>> - MPEG4: This is most common in forms of DivX and XviD. Predecessor of
>> H.264. As usual there's patent pool licensing involved. This means that
>> albeit XviD is open sourced it's not really free due to patent licensing
>> issues.
> 
> That's wrong – H.264 is MPEG4 Part 11 – it's part of the MPEG4 spec.

You're right (Part 10, I've read your correction mail). Thanks for the
correction. H.264 became part of MPEG4 long after "MPEG4" codecs were
already out in the wild, so many people say "MPEG4" to the old thing and
"H.264" to make clear they mean the "new" thing.


> I think we need to look at why the MPEG standards see near universal
> support and use: as you say, parts of MPEG4 are highly efficient (such
> as H.264 and AAC), whereas alternatives of things like Theora aren't
> anywhere near efficient. Also note that patents haven't stopped the web
> in the past (see: GIF).

Well, I'm pretty sure many people would have liked to avoid GIF if it
would have been an option. Sadly it wasn't.

I think patented formats should be avoided whenever possible, they're
nothing but trouble depending on what environment you're in. In case of
media formats there are alternatives available. Vorbis e.g. is a
top-performer _and_ free. MP3 performs far worse (and isn't free) and
AAC isn't performing better with its base profile either. Theora is
already outperforming the most widespread video streaming solution
(Flash Video with H.263) and is only hindered by its current encoder
implementation.

The fact that there will certainly be more powerful codecs in the future
(both free and non-free) is a strong argument for not disallowing other
media formats, though. It's not like e.g. <img> would disallow anything
beside JPEG and GIf to my knowledge, so why should a <video> element do so?


> I really believe that this is too political, as history has shown people
> will use whatever formats can be created easily, and are well supported.
> It could be perfectly possible that anything wanting to implement the
> spec is put off by needing to support a single format that (almost)
> nobody uses.

Well, Flash Video is everwhere on the web, but it's basically never used
outside of the web. I don't think the fact certain formats are popular
outside of the web is a strong reason to use them in web browsers.
Content used for usage out of the web (e.g. "DVD backups") is
inappropriate thanks to web streaming having narrow requirements on e.g.
bitrate - it would have to be converted anyway.

"Nobody uses SVG" - but browsers support it because somebody has to make
a start.

"Nobody uses PNG" - same here.

It all boils down to following questions:

* Do we a set of base formats to make sure <video> is actually usable?
(So it won't end up like "Safari uses QuickTime, Opera uses MPEG4,
Firefox uses FLV to be compatible with existing content - and Microsoft
uses Windows Media". That'd make sure people stick to Flash "because it
works 'everywhere'")

* Should at least one base format be freely available on all platforms?
Personally I'd say "of course!", but it's necessary to realize that even
if decoders for a special format are available it may still not be
acceptable in certain scenerios. I use Ubuntu. If I wanted to play back
(or generate!) certain formats I'd have to enter a legal gray area and
download codec packs from somewhere. As an end-user I have no way to
directly pay for the patents I'd use. And albeit Ubuntu may implement a
scheme to allow buying licensed codecs: What with Debian or Fedora? What
with users that want to live in a completely free userland - would they
have to strip <video> out of their browsers?

So if you agree on base formats: Better make sure there's at least one
format in there so content providers, users and distributors of software
can use even if they happen to not be able to meet certain licensing needs


Maik Merten