Telephotography of restricted areas
Filters
Ever wonder why the sky is blue? It turns out that the scattering of
light due to water vapor is greater for shorter wavelengths. This effect is not linear,
but is inversely proportional to the 4th power of the wavelength. In
plain Engligh, short wavelengths really dominate. When viewing a
distant object, the greater distance of atmosphere in the optical path, the greater the blue cast.
Particulates matter in the air also gives the image a cast. Every year
there are forest fires in southern California, and the prevailing winds
blow the soot into Nevada. There are also fires local to the area, but
they generally burn out quickly. This is a camera phone photo of Bald
Mountain when the Malibu fire of 2007 reached Central Nevada:
Anyone who has played with a photo editor knows that they can adjust
the color in a number of different ways, so why bother filtering the
image before it is photographed? Examine this series of photographs,
all of the same general area of the non-existent base at Groom Lake.
They are screen captures of Vuescan, a program that interfaces to both scanners and the RAW images from a digital camera.
The first set is done with no filters. There is no attempt to adjust
the color. In blue cast is obvious, but the real action can be found in
the histogram in the lower left corner. The peak of the blue histogram
is nearly twice the value of the red and green peaks. The consequence
of blue dominating the image is that the amount of blue light
determined the exposure.
The blue channel is shown below, though it appears as grayscale. It is milky. There are no deep black areas in the image.
The green channel doesn't have that milky look. The image lacks
contrast, which corresponds to the histogram not reaching the edge
(i.e. very black or very white).
The red channel has deep blacks. However, what really happened is much
of the low intensity red got tossed out. That is, with the exposure
being dominated by blue, the red channel was left in the dirt.
Now if we could find a filter that gradually rolled off short
wavelengths, then the blue would be more balanced with red and green.
How about a B+W KR1.5?
Here is the same area of the non-existent base photographed with the
KR15 filter. [The focus wasn't as good on this shot, but the goal is to
concentrate of the color channels.] It still looks blue, but the
histogram shows that the red channel doesn't get truncated on the low
end. {The settings for Vuescan have not been changed.)
Blue channel with KR1.5 filter. The blue channel is still milky.
Green channel with KR1.5 filter.
Red channel with KR1.5 filter. Note that the dark areas have more detail. Compare it to red channel with no filter.
At this point, we are close, but no cigar. Now if we could get
another filter that trimmed off the very edge of the blue spectrum....
The Andover Corporation GG-420
slices off purple, whose energy appears in the blue channel. In
the histogram below,all three channels now fall in the range of the
film. The color is still strange, but that can be corrected in
the photo editor.
The blue channel is still a bit milky, but if you go back and examine
the unfiltered image, you can see the 420nm filter made an improvement.
B+W KR1.5 and Andover GG-420 green channel
KR-1.5 and Andover gg-420 red channel
