Re: Wide Color Gamut and High Dynamic Range displays

On Wed, Jan 28, 2015 at 3:42 PM, Tab Atkins Jr. <jackalmage@gmail.com>
wrote:

> On Wed, Jan 28, 2015 at 1:30 PM, Rik Cabanier <cabanier@gmail.com> wrote:
> > On Wed, Jan 28, 2015 at 1:13 PM, Tab Atkins Jr. <jackalmage@gmail.com>
> > wrote:
> >> On Wed, Jan 28, 2015 at 12:26 PM, Rik Cabanier <cabanier@gmail.com>
> wrote:
> >> > On Wed, Jan 28, 2015 at 9:02 AM, Mark Watson <watsonm@netflix.com>
> >> > wrote:
> >> >> It would be nice to be able to detect whether the display has the
> >> >> capability of rendering Wide Color Gamut and High Dynamic Range
> video.
> >> >>
> >> >> This is independent of codec support: in fact the video codec itself
> >> >> may
> >> >> be unaware of the colorspace and dynamic range of the encoded video.
> It
> >> >> may
> >> >> also be the case that the media pipeline in a device supports these
> >> >> things
> >> >> but the presently connected display does not.
> >> >>
> >> >> For WGC, the basic question is whether the display can interpret data
> >> >> coded in the BT.2020 or DCI P3 colorspaces (I say "interpret"
> >> >> deliberately,
> >> >> because I'm unaware of any displays that can render the full BT.2020
> >> >> space
> >> >> yet.)
> >> >>
> >> >> Would it make sense to add attributes for these properties to the CSS
> >> >> OM
> >> >> View Module ? Other suggestions ? Questions ?
> >> >
> >> > What are you planning on doing with that information?
> >> > AFAIK it is defined that pages are composited in sRGB and then mapped
> to
> >> > the
> >> > monitor profile.
> >>
> >> sRGB supports wide gamuts (at least theoretically). It's just outside
> >> the standard gamut, but still representable.
> >
> > Is there any browser that supports this and uses those values?
> > This is also not compatible with the BT.2020 or DCI P3 colorspaces that
> > Mark's requesting.
>
> Browsers *do* handle wide-gamut images, when they're appropriately
> tagged and going straight to the screen.


Can you give an example where that is happening?
It would be odd to have part of a document displayed with a different
profile.

I only know of the MS ICC sample page [1] but that just shows how the image
is converted to sRGB using its attached profile.


> And we (Chrome) are working
> on making it composite properly, too, though that's difficult.


Yes, the only way to do this would be to composite the whole page in a
difference colorspace which requires a lot of transformations of the sRGB
elements.
I believe only Apple has the framework in place to support this.


> >> > Would you use this to change color handling of a full-screen video?
> >>
> >> You can send different sources to <video> based on a media query.
> >
> > Since a video element is composited just like other elements, I don't see
> > how that would make a difference.
> > Pushing wide gamut pixels into an sRGB back buffer would just make the
> video
> > darker. (unless you map then to sRGB in which case there was no need for
> the
> > wide gamut video stream)
>
> We'd be in the same situation as wide-gamut images - if they go
> straight to the screen, we can send the better info; if they're
> composited, they get smushed into the sRGB gamut first, until we solve
> the technical challenges and can composite in a wider space.


1: http://ie.microsoft.com/testdrive/graphics/colorprofiles/default.html