[Openicc] review of new wide gamut editing space

Chris Murphy lists at colorremedies.com
Wed Mar 21 15:45:16 PDT 2012


On Mar 21, 2012, at 4:33 PM, James Cloos wrote:

>>>>>> "CM" == Chris Murphy <lists at colorremedies.com> writes:
> 
> CM> If we go to 32bpc float as a minimum across the board, neither
> CM> primaries nor TRC matter at all.
> 
> This segues nicely to a questing I've been pondering off and on for
> a bit:
> 
>  Given an unconstrained 32-bit float workflow, what are the
>  ideal values to pass, eg, to lcms2's cmsCreateRGBProfile()?
> 
> It seems reasonable to do the work in linear rgb; that simplifies
> specifying the tone curves.
> 
> But is there any value to choosing a specific white point and set of
> primaries?
> 
> Given sRGB's prevalence, one might suggest D65 and the sRGB primaries.
> 
> Unless, perhaps, when working towards HDTV or digital cinema output, yes?
> 
> Are there any other good arguments for or against any specific w/r/g/b?

D65 and sRGB / Rec. 709 primaries. It's easier to deal with conversions, and if you have luminance higher or lower than sRGB's thresholds, obviously you have the values to define points outside of the gamut. Some hyperbole but if my car is the sRGB gamut, you've got something like the moon's orbit around the earth for a definable volume. 

You need three (XYZ defined) primaries for your RGB values to have meaning, but only in that sense are they important. So yeah, the primaries don't matter much (rather than at all).

Chris Murphy


More information about the openicc mailing list