Re: New Fuji Film: Astia

[P3D John W Roberts <roberts@xxxxxxxxxxxxxxxxx>]

>Date: Fri, 23 May 1997 00:55:52 -0500
>From: P3D Paul Talbot  <ptww@xxxxxxxxx>
>Subject: Re: New Fuji Film: Astia

>DrT writes:

>> I might be ignorant in this subject, but, philosophically speaking, no
>> instrument can tell how the colors are actully seen by the eye-brain
>> complex.  So, what's the sense of doing "photo-color-metry"?

There are good models of how humans see colors - check out the 1931 C.I.E.
chromaticity diagram, or the more modern one (forgot the name) using the
u' and v' coordinates. I don't know how much individual variation there may
be, but in general the models seem to be considered pretty good.

No film that I am aware of "sees" colors like a human - two mixtures of
wavelengths that look identical to a human will generally look different from
one another to film. Unless somebody comes up with something that has the
same color responses as humans, any film formulation will be an attempt to
get acceptable results over a limited set of conditions.

Philosophically, if they don't even make the effort to understand what's
going on, there will be no progress except by chance. So actually employing
color theory and color measurements is probably a wise move.

>How does one determine that a halogen light source for a T'd RB is
>"whiter" than an incandescent source?  Is this observation taken
>solely from experience, or can it be measured and confirmed?  While
>not an identical situation, is it not also true that colors can be
>quantitatively measured, or at least certain aspects of them--hue,
>saturation, luminescence?  Surely there is some objective capability
>of measuring a film's accuracy of color rendition.

Barring any wildly futuristic stuff, any color can be completely described
by its frequency spectrum. Color theory maps the range of all possible colors
that a human can see into a bounded object in a 2-dimensional space. The
points along the periphery of that object (well, most of it) represent pure
wavelengths of visible light. Any perceptible color (represented by a point)
within the interior of that object can be replicated by any of an infinite
number of combinations of two or more wavelengths.

There have been many "remappings" of the color space, for a variety of
purposes. The ideal for some of the color theorists is to produce a
2-dimensional graph for which the minimum perceptible difference in color
is represented by the same linear displacement anywhere in the graph.
The u'/v' graph is supposed to be pretty close to that.

I *think* the hue/saturation mapping is used in the encoding for color NTSC,
so it may be useful for broadcast issues, but I'm not sure how it corresponds
to actual human vision.

The color mapping for which I have the *least* respect is "color temperature",
which attempts to reduce a 2-dimensional value to a 1-dimensional parameter.
Unless the color in question is very close to the blackbody curve, I suspect
evaluating it in terms of color temperature is more likely to mislead a person
than to provide any useful information.

>>I have been puzzled by occassional comments by "fanatic" Kodachrome
>>users in SSA folios about the pastel exaggerated colors of Fuji films
>>vs. Kodak films when to my eye the colors look real and not exaggerated

>I saw an interesting item about color perception in a book recently:
>given four different reds reproduced on paper and asked to pick:
>"Which is the color of a Coca-Cola can?" few people succeed.  We
>are not very good at remembering colors accurately, it seems.  So
>Gary's "urban legend" is more than that.

Humans are much better a direct comparison of color and brightness than
they are at judging absolutes. (I wonder whether there are people with the
visual equivalent of "perfect pitch".)

>>From time to time I tell my wife:  "Look at this sky.  Doesn't it look
>>BLUE, like DEEP BLUE, do your eye?  Would you blame the film if it
>>records this sky as DEEP BLUE?  Wouldn't you be disappointed if it
>>comes out gray?"

>OTOH, I've recently been noticing the sky is not quite as blue as my
>mental image of it is.

It's not more KASPAROV, is it? ;-)

In the summer, on the east coast of the US, a "clear sky" is usually more
pastel than intense blue. It *is* deep blue right now - I think the high
pressure area must have arrived from Ohio.

John R


------------------------------