Newsgroups: comp.ai.philosophy,comp.ai,comp.robotics
Path: cantaloupe.srv.cs.cmu.edu!das-news2.harvard.edu!news2.near.net!howland.reston.ans.net!gatech!newsxfer.itd.umich.edu!nntp.cs.ubc.ca!torn!utnut!ie.utoronto.ca!green
From: green@ie.utoronto.ca (Marc Green)
Subject: Re: Qualia (was Re: Minsky's new article)
Message-ID: <CzvqwH.5vo@ie.utoronto.ca>
Organization: University of Toronto, Department of Industrial Engineering
References: <19941116.150347.522@almaden.ibm.com> <3aj0om$1kf@tadpole.fc.hp.com> 	<3alr5q$g1s@crl2.crl.com> <PJG.94Nov21161240@tesla.esl.com> <3b1lik$8fg@urmel.informatik.rwth-aachen.de>
Date: Sat, 26 Nov 1994 14:57:53 GMT
Lines: 61
Xref: glinda.oz.cs.cmu.edu comp.ai.philosophy:22669 comp.ai:25510 comp.robotics:15688

rainerd@tschil.informatik.rwth-aachen.de (Rainer Dickermann) writes:

>pjg@tesla.esl.com (Paul Gyugyi) writes:

>>When I see yellow on my CRT, it is really red and green photons
>>stimulating my retina at the same time.  The eye doesn't really
>>ever "see" "yellow".  It has red and green sensors that, if both
>>are stimulated, blend the signals to be yellow.  The fact that
>>the eye is such a cheap low-cost design leads to the problem of
>>true yellow photons falling in the detection fringes of the red
>>and green cones.  Sorry if this was mentioned at the beginning 
>>of the thread, I've missed the first million articles or so.

>hmm... really, the retina cells do not 'just' see only red, green and
>blue photons, but a spectrum of photons in the ranges of red, green and
>blue. So, yellow photons stimulate the red and green cells because they
>overlap in their range.
>It looks a bit like that:

>Stimulation of RGB retina cells(Y)

>Y               R        G        B
>Y	      R   R    G   G    B   B
>Y	    R       RG       GB       B
>Y         R        G  R     B  G        B
>Y       R        G     R  B      G        B
>OXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXX -> Freq.
>               ^    ^   ^         ^                 Photons(X)
>              red  yellow        blue
>                        green

>Rainer

These posts reflect (if I may use the term) a complete lack of understanding
about color vision. First, photons are not "red," "green" or "blue." These
are are not physical properties of light. Photons differ in energy, a physical
property. Different classes of of photoreceptor absorb quanta of different
energy, or in more common terminology, different wavelength. Two lights
look the same when they stimulate the the same ratio of the the long, middle
and short-wave length photoreceptors. So two light mixtures, "red+green"
and "yellow", look the same because they stimulate that photoreceptors in
the same way. Or more correctly, the two lights contain sets of wavelengths
which stimulate the same ratio of the short, middle and long wavelength
cones. This is called metamarism and is related to the principle of
univariance - that once a photon is absorbed, the photoreceptor has no
idea what the wavelength of the light source was. 

Third, none of this has much to do with color appearance. Photoreceptor
activity determines when lights will look identical or different, but
not their subjective appearance. Color appearance is likely due to activity
of opponent-process and other narrowband neurons further back in the system.
Fourth, color appearance is not determined only by wavelength. There are
a bunch of other factors which are just as important, chromatic contrast,
chromatic adaptation, etc. It's easy to make two lights of different 
wavelength composition appear identical or two lights of different
composition appear the same. The reason is that color is not a property of
light, but of the mind/brain. That's why the field is called psychophysics -
its the relationship of the psychological world to the physical. 

Marc Green

