Re: Discuss HDR at CSS WG next week ?

On 2016-09-20 13:53, Mark Watson wrote:
>
> ​I think the two things which came up specifically with the color 
> specification were:
> (i) the predefined "dci-p3" profile is mis-named. There should be some 
> changes to clarify that this is not DCI P3, but rather a profile 
> defined in this specification which makes used of the DCI P3 color 
> primaries​. You could call it "web-p3" or "display-p3" or whetever.
Yes, agreed. i discussed that with Dean from Apple today, he is sending 
me an ICC profile that defines it. And certainly it needs to be renamed.

> (ii) when the update to iccMax is done, it would be good to have at 
> least one predefined profile for HDR (e.g. BT.2100 / PQ), especially 
> since I heard at least one implementor say they did not intend to 
> implement arbitrary ICC profiles, but rather just the predefined ones

Yes, also agreed (but that will likely be in a later version of the 
spec, because IccMax is a bit early to rely on while ICC v.4 is mature).

It is certainly possible to start a css color 5 as a placeholder for 
that sort of thing, in parallel with css color 4 development.
>
> The question of where to map the SDR / sRGB luminance range when 
> compositing with HDR is unsolved. It may be that for consistent 
> results across platforms this needs to be specified.
Yes. That would likely be in a new level of the compositing spec.

> But there needs to be a period of experimentation on that before we 
> could conclude either that specification is needed or that CSS color 
> is a good place to specify it.

Yes.

(3000 lines of quoted thread deleted for space)

-- 
Chris Lilley
@svgeesus
Technical Director, W3C Interaction Domain