Re: [rtcweb] On the topic of MTI video codecs

["Cullen Jennings (fluffy)" <[e-mail]>]

On Oct 31, 2013, at 6:11 AM, Silvia Pfeiffer <[e-mail]>
 wrote:

> On Thu, Oct 31, 2013 at 8:40 AM, Cullen Jennings (fluffy)
> <[e-mail]> wrote:
>> 
>> On Oct 30, 2013, at 12:20 PM, Leon Geyser <[e-mail]> wrote:
>> 
>>> Unfortunately like Jonathan pointed out H.264 will only be able to be used royalty free on certain(most popular) platforms.
>>> To be able to avoid negotiation failure we need a MTI codec that every potential now/future browser would be able to implement freely.
>>> I like what Cisco did, but the solution seems a bit half-baked.
>>> 
>> 
>> I think that Mozilla put it pretty nicely in their blog. What this annoucement gives us is not a perfect world. Mozilla is working towards a better future but in the mean time, this is the best thing we could possibly figure out on how to make the internet today be the best it can be for users.
> 
> 
> This is a very grand gesture of CISCO.
> 
> I'm just a bit cautious, since it looks too good to be true: have you
> considered that MPEG LA could change the license conditions on H.264
> next year and make this impossible?
> 
> Regards,
> Silvia.


Yes, I have deeply considered that including discussion with relevant lawyers. 

I do not see any path where significant changes could be be made that could stop this from continuing. Keep in mind that many of the companies in the MPEG LA 264 pool are in favor of just making H.264 RF for use on the internet. Theses same companies, and they include people like, cough, cisco, and apple, have to approve changes that happen in the future. I would not be surprised to see the discussion on future changes to the MEPG LA license be around just making the license pool RF. Or perhaps just moving the price of the cap up a bit. 

Now you can say, ok, that is way too good to be true, no big company would ever give away valuable IPR. Lets clear up one thing, the amount of money that any large company gets in royalties from 264 MPEG LA pool is so small it is a joke - and a quick back of the envelope estimate will show you the limits. But ask yourself this instead, Why does Cisco care about webrtc at all? what would a company like Cisco pay to MPEG LA to improve the ability for lots of companies, particularly small companies or open source projects, to be able to produce, consume, and distribute video on the internet? What would it take to make it so video on the internet was far more important to people's lives than watching cats on skateboards. 

This whole discussion reminds me of a real blast from the past. In 2000 Cisco acquired Vovida (where I worked) and open sourced 100% of the technology. People said said Cisco was nuts to acquire that and give it away for fee - the numbers were much larger than anything being discussed here. People like vonage starting using the code. People said "It's a trap" and included the appropriate star wars images - Cisco has patents on SIP and is going to get you for using the open source. Cisco did not go after people for using the software and companies like vonage effectively helped drive lots of voice traffic that resulted in good business for Cisco. Open source helped enable that. You might say that was then this is now but you might want to look at what we did this month - 2.7 billion for the Sourcefire, a company that's most interesting product is open source. 

Snort is not a trap, it's exactly what it seem. This is the same - Cisco wants there to be a video codec that can be widely used on the internet as it exists today - Cisco will continue to want that and will continue to pay MPEG LA to make that happen. 

Which leads of of course to why this codecs and not some other one. We don't believe VP8 is the one that works with the internet today and given the reasons have been on this before I don't want to reiterate here. I'd be doing a little dance of joy at the IETF microphone if Daala can meet their goals of better than existing codecs, royalty free, and wide adoption and I really hope that happens but it clear that is not here today and this is project still at early stage. I'm glad they are working on something better for the future. Older codecs like, just for example, H.263 might be easier / cheaper to get to be RF but, sadly they sort of blow chunks. I don't believe a business model like apple iTunes, Netflix streaming video, or even Skype could be successful in today's market if it had to use a codec like 263, or for that matter any of the older codecs - thus they don't really meet Cisco's goals of widespread viable codec people can use today.