[argyllcms] Re: Towards understanding rendering intents
- From: Graeme Gill <graeme@xxxxxxxxxxxxx>
- To: argyllcms@xxxxxxxxxxxxx
- Date: Fri, 17 Jul 2020 12:41:32 +1000
johan@xxxxxx wrote:
Hi,
In some types of measurement (reflection, transmission) there is a natural
100% value, while in others (emission) there is no such value.
Yes, there is. Peak white of an emissive display is used as the 100% value.
I was thinking about measurement in general, not a specific device.
Yes, a display will have a maximum white value that may be used
as a reference.
However, there is a debate on the extent to which this applies to HDR where
diffuse white is defined by 75% of the hybrid log gamma (HLG) tone curve
Last time I looked, there didn't seem to be any define diffuse white
in the popular HDR encoding standards. There seemed to be an assumption amongst
many users that 100 cd/m^2 was diffuse white, but nothing defined in
a standard. This, and the use of an absolute brightness encoding standard
seemed to me to be a major flaw in these standards.
(The exception I think was the BBC HDR standard.)
A problem with many HDR displays is that it is only white
that has high peak luminance, so W=R+G+B is not true for luminance values
above diffuse white, and in some cases not true at or below diffuse white
either.
Whether this is a major issue depends a bit on the media being displayed.
Typical real world high level sources of light are illuminants, which
tend to be white (the sun, artificial lighting), and specular reflection
is almost always typically the color of the light source.
But of course much media is created for effect, and there are certainly
situations in which the light source is not close to white, so I
can imagine there are plenty of examples of images that don't
display as the director intended on an HDR display that has a
narrower gamut as it gets bright.
When measuring the colour gamut volume of emissive displays, all measured
XYZ values are transformed by the Bradford chromatic adaptation transform
(CAT) so that peak white always becomes D50 prior to converting to L*a*b*
and calculating the colour gamut volume.
Maybe :-). That depends on which version of ICC you use and how you
interpret it. ICCV2 was interpreted a variety of ways.
This is currently done also for
reflective displays although it's inconsistent with the way gamuts of a
printer/paper ("substrate") combination is measured (relative to a white
diffuse reflectance standard).
ICCV4 is certainly inconsistent in disabling absolute colorimetric
intent for displays, thereby disabling the possibility of soft
proofing via ICC profiles.
Cheers,
Graeme Gill.
Other related posts:
