[rtcweb] Interlace (Re: Video resolution SHOULDs (Re: resolutions in draft-cbran-rtcweb-codec-01))

[Harald Alvestrand <[e-mail]>]

On 11/30/2011 06:18 AM, James M. Polk wrote:
> At 10:56 PM 11/29/2011, Ralph Giles wrote:
>> On 29 November 2011 11:49, Harald Alvestrand <[e-mail]> 
>> wrote:
>>
>> > Interlace? Just Say No!
>>
>> Works for me. I'm an optimist.
>
> not meaning to open a can of worms, but what still requires interlaced 
> video? Most new things convert I to P, but less of these convert P to I.
Just hit the same conversation here in Geneva at MPEG meeting. Funny 
what you learn!

Interlaced material requires quite extensive resources on the decoding 
end to do properly, and if you just ignore the interlace marker and 
display the half-frames on top of each other, the resulting display is 
clearly substandard.
But converting is a small-but-significant cost to the broadcasters with 
stored interlaced material, and the hardware guys who already have to 
implement it for some profile see no extra cost to having that support, 
so they don't push back much against it.

In contrast, WebRTC is much more greenfield, and has far less 
established interlace support than TVs.

If we can push interlaced material up the food chain into those devices 
that have to deal with it because it's in the archives, and get them to 
do the I-to-P conversion, the Web will be better off.
</rant>

                       Harald