On Sun, 3 Sep 2006 22:14:52 -0500, Herman Miller <[log in to unmask]> wrote:

>Javier BF wrote:
>> What do you mean by "quantifiable way" to pick the colors?
>
>Some kind of scale that would allow me to specify a color as some
>arbitrary fraction of the distance between two of the four main colors
>(red and yellow, for instance), such that any step of the same size on
>the scale would be perceived as about the same difference in hue. Having
>a standard definition for the four basic colors would also be nice.


I see. I don't know if there is some program that converts between RGB
and NCS; I think that would be the closest to what you are looking for. But
read below.


>I think I've got the basic colors pretty close to where I want them to be,
>but there's still some room for adjustment. Yellow seems closer to green
>than any of the other basic colors are to each other, but if I push
>yallow and green apart much further, the yellow starts to have more of
>an orange hue to it and the green starts looking bluish.


The biggest problem with trying to codify perceptual colors to be displayed
on an RGB screen is that the actual appearance of one same RGB code can vary
greatly due to many factors (such as the calibration of the monitor, the
lighting conditions and the surrounding colors). For example, on my current
screen (an LCD-type display), the red on the Nimrina color chart looks
pinkish, the green looks light and bluish, and the yellow a bit too dark;
only the blue looks pretty close to what I would call pure blue. On another
screen these appearances may vary. And even one same light under the same
viewing conditions might look definitely pure blue to me while not so purely
blue to someone else. Colors are subjective perceptions, and also not
everyone's eyes are internally "calibrated" the same, because there is some
degree of genetic variation in the photorreceptors, so that their peak of
sensitivity may vary slightly from person to person, or greatly in the case
of so-called "anomalous" trichromats.

This doesn't mean it is entirely impossible to have a standard definition of
the elementary colors. The NCS attempts to do that: you can have some
standard samples that, when viewed under precise lighting and environmental
conditions, most people report to perceive as looking like "the purest red",
"the purest yellow", etc.; but for some other people, like "anomalous"
trichromats (let alone dichromats, and who knows about tetrachromats), as
well as when under non-standard viewing conditions, the colors on those same
samples may no longer appear to look so "pure".

RGB defines colors merely in terms of wavelength compositions, disregarding
the actual appearance. That is, unlike the NCS standard, the RGB standard
doesn't actually try to define colors (which are highly voluble constructs
of the brain, and not physically definable properties). Instead, it contents
itself with defining only proportions between light frequencies, that may
prompt different color perceptions in the brains of different persons and
under different environmental and viewing conditions, but that are easy to
standardize in terms of physical properties.

My advice: Do not try to define your colors in terms of appearances on some
computer image, but instead define them verbally in terms of their
perceptual compositions (how something "should look like" for it to be said
to be that color). This is actually how colors are defined in natural
languages: English blue is not defined as the color on some standard sample,
 and neither as some defined light frequency (as some dictionaries try to
do), but instead each speaker has an internal idea of what the elementary
percept blue "looks like" (an idea acquired through visual experience, which
a born-blind person doesn't have access to and so doesn't know what blue
actually "looks like"), and when some object under some lighting condition
appears to prompt that elementary color perception in their brains, they say
that object "is blue" or "looks blue". Given this essentially subjective
nature of colors, it is not surprising different people may argue about what
color something "is".

For example, one can define the whole range of English "brown" as
conceptually comprising the composite colors that look perceptually composed
of red or yellow or red+yellow mixed with black or black+white (the typical
or "focal" brown being red+yellow+black at about equal proportions). Purple
could be defined as the composite colors looking blue+red, optionally
+black, and optionally +white (except in those combinations where red
dominates over blue, in which the addition of white would make them belong
under the label "pink"). You may then try to draw sample computer images to
illustrate what you have previously defined verbally (i.e., conceptually),
but including the necessary caveats to clarify that the particular
appearance of that image may not look as intended on different screens.