From: "Cunningham, Bob" <bcunning @ doctordesign . com>
To: <efg2 @ efg2 . com>
Subject: Computer Lab Feedback
Date: Wednesday, November 20, 2002 13:25

Hi,

The perceptual differences you mention (in
http://homepages.borland.com/efg2lab/Graphics/Colors/ColorCharts.htm)
between the GretagMacbeth (GM) ColorChecker (CC) RGB values and what appears
on a monitor would appear to be due to a combination of two factors:

1) The difference between the white points of the illuminant used by GM to
obtain the RGB values, and the white point of your monitor phosphor.

2) The ambient illumination used to view the CC.

GB states in the pamphlet included with the CC that their measurements are
referenced to Illuminant C (white point x,y = .31006,.31616).  Modern
computer monitors have "adjustable white points", though the white point for
a nominal compute monitor phosphor set with a monitor white point setting of
"7500°k" is (x,y = .345,.353).  Unless the proper transforms are done to
shift the image data white point to that of the monitor, there is no hope of
obtaining a perceptual match.

Scanners, on the other hand, typically have lamps with white points somewhat
closer to Illuminant C.  So obtaining an RGB match should not be surprising
(assuming the scanner is calibrated).

Of course, monitor white points only matter if the monitor itself is
colorimetrically accurate (calibrated) within its gamut and the proper gamma
function is applied.  Holding anything next to a monitor is valid only when
the ambient illuminant is accounted for.

I suspect things would match far better if the monitor were adjusted to have
a white point close to Illuminant C, and the CC were illuminated with a lamp
that closely matches Illuminant C.  (Not easy:  I've never seen an
Illuminant C lamp...)

-BobC