Re: [rtcweb] On video codec for rtcweb

[Basil Mohamed Gohar <[e-mail]>]

On 03/23/2012 11:24 AM, Matthew Kaufman wrote:
> Timothy B. Terriberry:
>
>
>> Matthew Kaufman wrote:
>  > > or rather, in order to achieve interoperable communication without the cost  > (and security concerns) of transcoding sites will in fact start to deny  > service to particular users based on their choice of browser.
>
>> I agree this is an excellent argument for having _a_ mandatory to implement codec. But it's an argument against having that codec be an encumbered codec. Much better to have it be a free one that everyone is able to deploy. Even Microsoft (via Skype) now deploys VP8.
> I wasn't commenting on that part, just that if Mozilla is going to soften up on H.264 you'd think it'd happen to RTCWEB first, VOD second... not the reverse.
>
> As for VP8, we don't actually know if it falls into the set "a free one that everyone is able to deploy", do we?
>
> Matthew Kaufman
With all due respect, this is FUD that is baseless.  A blog post[1], an
abandoned patent pool[2], and vague comments by Steve Jobs[3] are all
that really stand as doubt as to the royalty-free basis on VP8.  What we
do know is that On2 had some patents on VP8, and these have been
licensed for royalty-free usage after their acquisition by Google. 
Google themselves have deployed VP8 widely on YouTube via the WebM
format and implemented and distributed it on Android, which is quite a
large surface area to expose if they had doubts about its safety. 
Patent licensing targets don't get much more juicy than Google.

As for implementation, any choice of format will come with challenges,
and the licensing on H.264 vs. VP8 make it clear that VP8 can receive a
much wider adoption in the long run, especially one that does not
unnecessarily restrict this forthcoming standard only to existing
players in the telecommunications market.  There are already plenty of
hardware chips that natively support VP8 decoding and encoding[4], and
that will grow over time, as well, because many parties are strongly
interested in a royalty-free alternative to H.264.  Let alone there are
ARM implementations of VP8 as well[5].

Where is the legitimate doubt, then, that VP8 is not an acceptable,
royalty-free format that can be used for the purposes of WebRTC?  We're
talking initial deployment.  Technology is always changed, and WebRTC 2
is free to adopt Daala[6] when it is released.  Believe me, the
developers would love to ensure it is suitable for WebRTC, amongst other
things.

[1] http://x264dev.multimedia.cx/archives/377#more-377
[2] http://www.mpegla.org/main/pages/CurrentPools_LightBox.aspx (note,
it's not listed there!)
[3] http://www.theregister.co.uk/2010/05/20/jobs_on_vp8/
[4] http://wiki.webmproject.org/hardware/arm-socs
[5] http://www.webmproject.org/tools/vp8-sdk/changelog.html
[6] https://www.xiph.org/daala/