[Leica] Re:Color vision

[Henning Wulff <henningw@archiphoto.com>]

At 2:53 PM -0500 2/2/04, LRZeitlin@aol.com wrote:
>In a message dated 2/1/04 10:58:46 PM, Adam writes:
>
><<  The human eye uses more than three types of
>>  color sensors - but then nature is not constrained by accountants or
>production
>>  engineers.
>
>Really? What are there? I thought there were (in most people) three different
>sets of cones with different spectrum sensitivities and rods which are
>strictly
>luminance sensing.
>
>I'm not an expert though, and I don't play one on TV.
>
>What I have found fascinating is that we have built-in edge-detection and
>sharpening, which definitely supports our pattern recognition functions.
>  >>
>
>The last time I studied this area, I recall that there were five to seven
>different types of cones sensitive to various ranges of color in the 
>human eye.
>More may have been discovered recently. The idea that there are three types of
>cones, each sensitive to a primary color stems from research done in the
>1880s. At that time Young and Helmholtz independently attributed 
>color vision to
>the operation of three types of cones in the eye (red, green, and 
>violet). This
>conclusion was based on the finding that most colors could be matched
>additively by various amounts of the three primary light sources. 
>This theory was
>seized upon by both printers and photographers since it prescribed 
>the simplest
>technique for reproducing color. The theory gained such wide 
>currency that most
>elementary physiology textbooks stated that the eye worked in this manner and
>that's what most of us learned in school. Unfortunately, the tricolor theory
>could not explain color vision anomalies such as red green blindness, blue
>yellow blindness, or achromatism. For example red/green color blind 
>person can see
>yellow quite clearly although the trichromatic theory presupposes that yellow
>is a mixture of red and green and both receptors must be working. Yellow/blue
>color blindness is much rarer but produces the additional complication that
>the red and green sensors are working but the resulting yellow color cannot be
>distinguished from blue, although red and green can be clearly distinguished.
>
>Despite the seductive simplicity of the tricolor approach, visual
>physiologists were forced to conclude that the actual visual 
>mechanism may be different.
>Some theories proposed combination red/green sensors, yellow/blue sensors and
>a luminosity sensor. (Sort of like the way your color TV signal is modulated.)
>Others added additional sensors which modulated the color response of three
>primary sensors. Still other theories added additional primary sensors. The
>jury is still out on the exact mechanism of color sensation. One 
>thing all modern
>theories have in common, however, is that considerable image processing is
>done in the eye and the brain before colors, edges, movements, and shapes are
>perceived. Human vision is MUCH more complicated than we were taught 
>in Biology
>101.
>
>All that being said, the tricolor approach to color representation has been
>wholehearted adopted by technology because it provides the easiest 
>and cheapest
>method of adequately capturing and reproducing MOST colors. The three colors
>chosen form a rough isosceles triangle superimposed over the color diagram
>shown in most textbooks. Any color inside the triangle can be represented by
>various amounts of the three chosen primaries. If you have a Mac computer, the
>Color Picker extension lets you see this triangle and adjust the 
>actual hue and
>saturation of a color on the screen. But there are colors outside the triangle
>which the three chosen primary colors cannot reproduce. You can get these
>colors by choosing another set of primaries but then some colors 
>that you could
>specify previously are unobtainable. Preferences for the color rendition of
>specific color films (Fuji vs. Kodak) are in part due to slight 
>differences in the
>manufacturer's choice and balance of primary colors. The answer to more
>accurate color rendition is to use more than three primaries, 
>perhaps four or five.
>This is the answer that most high end inkjet color printers have chosen to
>get around the limitations of subtractive pigments or dyes.
>
>So we just have to live with it. As long as our color films use only three
>color layers or our digital cameras use three types of color 
>sensors, there will
>be colors our eyes can see but our cameras cannot reproduce.
>
>Larry Z

This is just the beginning of the problems.

There are a couple of other factors which also contribute greatly to 
colour sensitivity/response/reproduction.

The purity of the primary colours we use is seriously deficient. 
Anybody who has done his own colour printing knows that, as the 
filter packs are far from linear. So... the filters that are used in 
colour film, in front of the digital sensor photosites, the inks we 
use for printing, the monitor filters and phosphors, and also the 
spectral sensitivities of our cones (our personal filtration) and our 
own neuro-optical processing are all very divergent. Beyond that, the 
spectral response to frequencies beyond our own 'visible' region 
varies for emulsions, filters, and inks and dyes.

We're lucky to have gotten as close as we have.

- -- 
    *            Henning J. Wulff
   /|\      Wulff Photography & Design
  /###\   mailto:henningw@archiphoto.com
  |[ ]|     http://www.archiphoto.com
- --
To unsubscribe, see http://mejac.palo-alto.ca.us/leica-users/unsub.html