|
2004 November 15
On November 12, I made a presentation at Color Imaging
Conference 12, in Scottsdale, AZ. I used the same title as the
talk I gave for Inter Society Color
Council in 2004 May:
Color
Matching with Amplitude Not Left Out
Color
Imaging Conference 12
Friday,
2004 Nov 12, 11:20 am
James
A.
Worthey, PE, PhD
In the interim from
May to November, the research itself has continued, and I have
developed new graphics, including
animations, to help explain the subject. Everything is
re-written, but one new insight can be singled out. The triplet of
tristimulus values (X, Y, Z) is sometimes referred to as a
"tristimulus
vector." In spite of this usage, (X, Y, Z) serves as a steppingstone to the
(x, y) chromaticity diagram, and we are
not taught to work with graphs of tristimulus vectors, or with the
amplitudes of such vectors. Because of its arbitrariness, the XYZ
system is, in fact, not well suited for graphing tristimulus vectors,
but the
notion of color vectors is important. By not skipping over the
vectorial stage, and by using tristimulus vectors based on orthonormal
color matching functions, we can take the mystery out of Prime Colors,
the choice of phosphors for video, and other topics.
The abstract of the
talk appears below, and the entire talk is available through these
links:
1. The self-contained article: Color Matching with Amplitude Not Left Out.
2. Especially interesting:
the graphical material for the oral
presentation in Scottsdale. Even if you saw the talk in Scottsdale,
you can now view all the graphics at leisure, read the answers to the
Seldom Asked Questions, etc.
3. The speech itself as read in
Scottsdale.
| You
will need a VRML plug-in installed in your browser to view the
3-dimensional graphs below and in the graphics of the talk. If you are
not already familiar with VRML,
there are installation instructions at upper right. >> |
|
Abstract
Amplitude
for color mixing is
different from other amplitudes such as loudness. Color amplitude must
refer to
a light’s ability to look different from other lights, to express its
redness
or other chromatic intensity, so that its color is not lost during
transduction. To reveal independent stimulus dimensions, a set of
orthonormalized color matching functions is derived, similar to
opponent color
primaries. Following an idea of Jozef B. Cohen, it is then assumed that
a light
of unit power varies in wavelength through the spectrum. The track of
that
light in the orthonormal color space gives a curve that Cohen called
“the locus
of unit monochromats,” after he found it by different steps. The locus
defines
a surface that is interesting but not complicated, which Cohen called
“butterfly wings.” Projecting the locus into a plane normal to the
achromatic
axis gives a boomerang shape with 3 well-defined local extreme points.
The
extrema are William A. Thornton’s Prime Colors, so a few steps reveal
the inner
workings from which the Prime Colors arise. The results can explain
color
mixing to beginners, but are also quantitative and ready for practical
use.
|
|
|
|
All VRML files
Copyright
© 2004 by James A. Worthey, all rights reserved
|
|
