[geeks] 720 Megapixel photo

[der Mouse]

>>> We get fooled by a red-green combo that stimulates the retina the
>>> same way yellow does (makes reproducing colors easier).
>> Well...approximately the same way.  Unless you are one of those
>> lucky women who happens to be a tetrachromat, perhaps.
> I didn't know there were any humans who were.

AIUI there are two different versions of one of the colour receptors,
and the gene is on the X chromosome.  A woman can thus end up with four
different colour receptors instead of the usual three.  It is
apparently not a myth that some women have a richer experience of
colour than men (or other women).  The difference in the response
curves between the two versions is apparently slight but definite.

>> It's always annoyed me, in an abstract theoretical kind of way, that
>> there probably are colours the typical (trichromat) human eye can
>> perceive that cannot be reproduced on (typical) colour film or
>> monitor.  (I say "probably" because I haven't looked up the response
>> curves and worked it out, but it seems probable to me.)
> Well, potentially the eye may have a "digital" response to color (the
> chemicals either react or not) but which has greater "color"
> resolution (if I may use that term), the monitor or the eye, I have
> no clue.

That's not what I'm talking about.

When you look at something, your colour experience (assuming you're a
typical trichromat, which you personally apparently are - I'm speaking
to the generic "you" here) can be reduced to three numbers, the
responses of your three colour receptors to the particular frequency
spectrum of light you're getting.

Now, when you look at a colour on a colour monitor, your receptors'
responses are (effectively) the sums of their responses to the levels
of the three primaries displayed.  Depending on how the three
phosphors' (for a CRT, the analog for other technologies) emission
spectra match your receptors' response curves, it may be impossible to,
for example, get receptor C firing above 50% without receptor A firing
above 10%, because each of the primaries has a spectrum that stimulates
receptor A at least 1/5 as much as C.  Yet it's possible that the right
spectrum of light could well achieve such a condition.  This would mean
there is a colour you can see in the world but which you cannot see on
a colour monitor.

Now, as I say, I haven't looked up the response curves of human
photoreceptors and typical CRT phosphors to see whether such a
situation is actually possible.  But given how messy the real world is,
it would rather surprise me if there weren't some such unreproducible
colours somewhere in the space of all viewable colours.

Then there are afterimages; there may be more "impossible" colours (ie,
firing ratios not normally achievable) lurking there.

The above discussion is actually oversimplified, because it ignores the
rods; most humans actually have four receptors, and a tetrachromat has
five, but the extra receptor has a more or less flat response curve and
does not give rise to perception of colour on its own.  I don't know
enough about human visual processing to know whether it affects
perception of colour or not; if it does, there's another response curve
to consider.

/~\ The ASCII				der Mouse
\ / Ribbon Campaign
 X  Against HTML	       mouse at rodents.montreal.qc.ca
/ \ Email!	     7D C8 61 52 5D E7 2D 39  4E F1 31 3E E8 B3 27 4B