[argyllcms] Re: Colormunki Photo & Dell up2414q led back-lit

  • From: Richard Kirk <richard@xxxxxxxxxxxxxxxx>
  • To: argyllcms@xxxxxxxxxxxxx
  • Date: Thu, 5 Mar 2015 13:32:36 +0000

On 5 Mar 2015, at 13:02, Graeme Gill <graeme@xxxxxxxxxxxxx> wrote:

> The green primary in particular is extreme, and is close to the point
> of maximum variability of observers medium and long wavelength cones.
> The red is rather way out there too, and although it's broader than
> the blue and green primaries, it's out in a region where the long
> wavelength cones are dropping off in sensitivity, and have high
> variabilityamongst observers.
> So in summary you may be experiencing observer induced metameric failure.
> The instrument is measuring color as the standard observer sees it,
> and you aren't a standard observer (very few people are exactly
> on the average).

This isn't the only problem. You also have effects from peripheral vision.

When Sony started making broadcast-quality LCD displays, people would put them 
next to the old CRT displays. The primaries had been arranged to match, and the 
white point had been fixed to D65. However, people consistently saw the LCD 
monitors as being orange. The CIE 2-degree observer is not perfect, and may 
fail to predict some variations with spectrum, but it is unlikely to be that 
far out. However, if you are looking at a blank screen, then the screen is 
probably over 15 degrees across. The macula has a yellow dye over it, and 
contains now blue sensing cones at the centre. The cones are also at a 
different angle to the incoming light then in the periphery, which slightly 
affects their wavelength. Finally, there is more opsin dye in the cones when 
working at typical monitor light levels, then at the 1000 nits for the CIE 
experiments. All these things will mean you will see a slightly different set 
of spectra out of the corner of your eyes. However, the brain knows the screen 
is uniform white, so it tries to fudge an average white from what the macula is 
seeing and what the peripheral vision is seeing.

This came more obvious when we started getting high-gamut displays like the 
Barco DP90P with its intense red. These can give you the paradoxical result 
that an image from a projector could match an image from projected film for 
every colour detail because it was matched according to the 2-degree observer, 
but give a nagging sensation that the overall colour balance was out (the 
projector white was brighter and was often violet).

There are other standard observer models. There was a 10-degree observer model 
from the sixties, but that was not a success. There are some newer CIE 
wide-angle observers. However, these wider observer models are rarely as 
consistent as the 2-degree observer. We all have similar cones and rods, but 
our macular dye does vary a lot, so what looks right for one person would not 
look right for another. Either we start using displays with six primaries, or 
we just have to put up with these differences for large patches.

Richard Kirk

Other related posts: