Re: [rtcweb] Codec extensibility (Re: Interoperability between browsers (MTI Video))

[Hadriel Kaplan <[e-mail]>]

On Mar 17, 2013, at 3:17 AM, Harald Alvestrand <[e-mail]> wrote:

> On 03/16/2013 10:17 PM, Bernard Aboba wrote:
>> 
>> On Mar 16, 2013, at 15:18, "Leon Geyser" <[e-mail]> wrote:
>> 
>>> In my opinion there really is a need for a MTI video codec for interoperability between browsers.
>>> 
>>> Here are some scenarios on MTI decisions:
>>> 
>>> -- Decide on no MTI codec:
>>> This would split the browers in two sides.
>> [BA] No. It just removes the stick that each side can use to beat the other side.  The need for interoperability still remains, and most likely will either be addressed by the marketplace choosing one over the other decisively, or if deployment is split, by support for (limited) codec extensibility.
> An interesting angle on codec extensibility is what license agreement one needs in order to deploy a codec extension; it would seem natural to assume that one would need an MPEG-LA license in order to deploy a H.264 codec extension.

Sure, or an extension that uses an API embedded in the local hardware - i.e., if you're on a vendor-X phone and vendor-X provides the extension to work in your browser(s).
Or perhaps the patent holders of H.264 will convince the holdouts to accept an RF license for a browser extension, restricted to use only in browsers.


> An API for access to raw incoming/outgoing data, and for registering a codec in the negotiation machinery, would need to be defined, but doesn't seem extremely complicated to do (although the details will take some care). But the license issue might mean that the first plug-in codecs developed would be the ASCII codec, the H.261 codec and VP8.

But that's a good thing, no?  VP8 everywhere!
Besides, it might even foster innovation in new codecs - like 3D video, or *color* ASCII/ANSI-code video. :)

-hadriel