"William Sommerwerck" wrote in message
...
This is simply not true. Every display has a color gamut that is
limited by the maximum saturation of its primaries. You can
produce any color within that gamut but not any outside.

Correct.


Even if every flesh tone is in that gamut, that does not mean that
you will get the right flesh tones for a given combination of RGB.
In order to do so, you must have the same spectrum in the
primaries that you have in the camera filters...

This bothers me. It might be true in a practical sense, but it's always
struck me as being theoretically wrong (mostly because of the extreme
overlap of the eye's blue and green receptors). I won't start an argument,
though, because, even if my intuition is correct, I don't have the
"science"
to back it up.




The overlap is caused by the shape of the standard observer curves and is
part of the very reason that using a narrow band RGB device may not produce
color properly. The CIE standard observer curves are precisely attempts at
modeling the response of the human visual system. For decades, cameras have
been calibrated to match them, and phosphors designed to do the same, to the
degree possible. Now, with narrow spectrum devices, we have to consider the
implications of those assumptions. It may be that in the future, we should
simplify the system and use a narrow band response in cameras and reproduce
the RGB in the same manner at the display. Then we can more easily predict
the output of RGB systems using a standard matrix.

Leonard