Re: [rtcweb] revisiting MTI

[Randell Jesup <[e-mail]>]

On 12/16/2014 3:42 PM, Roman Shpount wrote:
> On Tue, Dec 16, 2014 at 2:56 PM, DRAGE, Keith (Keith) 
> <[e-mail] 
> <mailto:[e-mail]>> wrote:
>
>     As far as I am concerned an IPR free codec is irrelevant if I
>     cannot talk to the endpoints I need to, and just concentrating on
>     the webrtc community as the available endpoint may meet some
>     deployers use cases, but not others.
>
>
> There is no such thing as IPR free codec, unless you are talking about 
> H.261 or MJPEG where IPR has expired. What we are looking for is a 
> video codec with an acceptable quality (both VP8 and H.264 qualify) 
> and reasonable licensing (both VP8 and H.264 have serious issues 
> here). If it is confirmed the VP8 licensing issues are resolved (i.e. 
> it went through the standardization process, all IPR declaration 
> against it which prevent licensing are confirmed to not apply in court 
> or due to some sort of licencing agreement), then VP8 is clearly 
> superior. If H.264 fixes its licensing (i.e. present a clear, 
> published, no royalty licensing terms with no use restrictions), it 
> can be a superior codec. Legacy interop support is important, but to 
> me at least, it is secondary to unrestricted use. Please see Opus vs 
> AMR-WB+ for clear similarities.

+1.  "IPR-free" barely would even apply to H.261 - you'd also have avoid 
any of the (patented) tricks on the encoding side that have shown up 
during later codec evolution from that  - but you always *can* avoid 
them by basing on ancient implementations.

> Also, VP9 or H.265 would clearly  improve video quality, but both VP8 
> and H.264 with the current connection speeds allow for the highly 
> usable video communications. I think that quality wise, either one of 
> VP8 and H.264 can serve as usable MTI for a lot longer then the next 
> two years or whatever time is required for the next great video codec 
> to be developed. Especially in another 5-6 years when all the IPR on 
> these codecs will expire.

+1 (Well, most of the patents on the decoder side and basic 
non-performant encoders may expire.  Patents are weird.)

We have better things to do...

-- 
Randell Jesup -- rjesup a t mozilla d o t com