Hi Dave,

I thought I'd respond to your comment about my entry in the Nikon Digest on the
F5 meter.

First off I have seen a difference using the colour metering.  Just after I got
my F5 I took some pictures of a white flower.  It extended past the 12mm circle
in the view finder so predominantly the image was white.  I also had my F4 with
me.  Both cameras had Kodak E100S film of the same batch code.  Using a 200mm
micro lens mounted on a tripod by the lens mounting ring I took photos of the
flower with both cameras using program mode within a few minutes of each other.
The light didn't vary during this time (clear skies, middle of the day etc).

The result was that the F5 gave a white flower while the F4 gave a very pale
grey/pink flower, a +.7 compensation would have made it white.  I'm convinced
the F5 meter works as advertised.

As far as a coloured filter on the lens, I don't believe that has any bearing on
the issue of how the RGB meter functions.  Let's take my white flower, without a
filter it photographs white.  If I put a red filter on, it photographs as red
(well probably pink, but you get the idea).  What the meter detects is the same
as if I were taking a picture of a pink flower.

You're right it doesn't know that a filter is being used.  It doesn't need to,
because when it looks up the exposure index in the table it selects the EI which
would be right for the pink flower which is what the film is seeing, regardless
of if it is a real pink flower or a white flower with a filter.

After all the meter is trying to give an accurate exposure reading for the light
which will hit the film, if it's all red, fine, the green and blue sensors will
read 0 (255,0,0 - r,g,b) and then in combination with the brightness and
contrast in the scene it will find a best match in the database.

Think of the matrix as a cube, colour along one side, contrast along another and
brightness the last.  When the meter reads the colour as 123, 10,45 (rgb),
brightness as 155 and contrast 111.  It looks up this position in the database
and out pops a EI.  Of course my cube is a simplification as there are 3
different sensors you would need a 6 sided "cube" (distance as well).  For us
it's hard to visualise but to a computer it's simply a multi-dimensional array
with exposure values at the intersections.

All the best and keep the great shot coming.

Eric

Reply via email to