Re: Daughter of CODEC (was Re: Alternative decision process in RTCWeb)

[Eric Burger <eburger@xxxxxxxxxxxxxxxxxx>]

Funny, but by my math, MPEG2 patent expiry is roughly equal to the date a brand spanking new video codec RFC would be published. I see only two differences between a new codec and MPEG 2. In the brand spanking new video codec’s corner is I would expect it to run circles around MPEG2 from a performance perspective. and we would know there could be no royalties on MPEG2, but the brand spanking new video codec would have the identical cloud Opus has hanging over it. That is, we all know there must be IPR against it, but no one has raised their hand (yet) saying they have it.

[Is PHB & EWB agreeing on something the new version of Dave Crocker agreeing with jck?]

On Dec 2, 2013, at 1:32 PM, Phillip Hallam-Baker <hallam@xxxxxxxxx> wrote:

On Mon, Dec 2, 2013 at 12:15 PM, Mary Barnes <mary.h.barnes@xxxxxxxxx> wrote:

We had a videocodec bof  @IETF-85 and the WG just never got chartered:  http://trac.tools.ietf.org/bof/trac/wiki/BofIETF85.   Perhaps the ADs can fill in the gap as to why that didn't happen as I'm not sure whether a ball was dropped or there was a reason not to charter.

Regards,

Mary. 

How long is such a group likely to take and would it be developing a new CODEC or selecting from existing unencumbered choices?

Is there an expectation that such a group would produce a CODEC better than MPEG2 or report substantially earlier than the expiry of the last patents in the MPEG2 pool (expected to be ~2017)?

Half the MPEG2 patents have expired already and it is possible that a subset of the MPEG2 CODEC could be defined that avoided IPR encumbrances. But it is hard to see such a CODEC gaining significant ground in the next 4 years.

MPEG2 is not as efficient as H.264 and it does not support all the same modes. But it is almost certainly good enough for MTI.

The best approach to this problem looks to me to be to do what we did with the Diffie Hellman and RSA patents and wait for them to expire. We did start pushing DH and El Gamal based schemes between 1997 and 2000 when those were out of patent but RSA was still covered. Doing the same with H.264 makes sense.

The long tail patents on H.264 is very long, there is at least one in the pool that does not expire till 2027 

--
Website: http://hallambaker.com/

Attachment: signature.asc
Description: Message signed with OpenPGP using GPGMail