[whatwg] Codecs (was Re: Apple Proposal for Timed Media Elements)

[Maik Merten]

Maciej Stachowiak schrieb:
> This is true of hardware audio decoders, but not hardware video
> decoders, which use dedicated circuit blocks. If Ogg suddenly became
> popular it would likely be a several year pipeline before there were any
> hardware decoders.

I'd say that any hardware player using hard-wired codec functionality is
a bad design. There are many embedded CPUs out there that also contain
DSP functionality. Those can easily support Theora.

> Most Flash video uses on the Sorenson Spark codec which is based on
> H.263. This is a much less processor-intensive codec than more modern
> options, but also gives worse compression. H.264 has been approved as
> one of the codecs for 3GPP so you can expect it to be supported by
> mobile devices in the future. Modern hardware decoders these days
> support H.263, MPEG-4 Part II, and H.264. These also happen to be the
> 3GPP codecs.

Those devices can also easily decode Theora on their general purpose
CPUs as has been demonstrated on the Nokia web tablet devices - without
even touching the DSP functionality!

Plus we're talking of web video here. That means we should choose a
codec that also decodes well on mobile platforms that don't have
hardware accelerated codecs on-board (like PocketPCs or gaming consoles).

Although I can see why Apple would want to support the MPEG codecs
(they're technically good and are popular and Apple is involved with
MPEG) I don't see why this rules out Theora being mentioned as base
format. Apple can safely ignore this and still be compliant, while
stripping out mentioning Ogg Theora may lead to no base format being in
place at all.