By Dan Anderson

Photography is one of the most heavily documented techniques in all of human history. But still, there seems to be some confusion even about how "setting exposure" works. Having typed all this, I realized that Ansel Adams gives a very good description of exposure in "The Negative".

I must have tried explaining exposure 39 times. So here's a summary, specifically for scientific pictures typical for our work:

First, the punch lines:

1. "Setting exposure" controls how much light reaches the film.
2. The amount of light reaching the film determines how the picture will look.
3. We use the light meter to fit the exposure onto what the film can handle.

Don't forget to think!

Human vision is primarily perceptive. The camera has no brain. Seminar slides are sometimes quite humorous.

Only if you want an artsy effect should you show the thing of interest off center. The slide's owner will see the subject as the most important thing in the world, while the rest of us probably just see a blob in the corner.

The scientist/photographer may become extremely excited by the content of a photo, but it will look insignificant until it fills almost the entire frame of film. For example, a slide showing a tiny crystal in the middle may be extremely exiting to its owner, but will look like a dirty slide to the rest of us. Even worse is an empty slide with a tiny crystal in the corner. Put it near the middle.

Also, no matter how well you understand what's in the picture, there have to be enough visual cues in the picture to communicate what the picture shows.

A picture of a stick figure representation of a molecule will look like a bunch of colored lines on black paper to the rest of us; only the slide's mother can follow the chain.

Show only a minimum of atoms, with plenty of depth cueing, and try not to overlap the atoms of greatest interest.

A picture of one's spouse standing in front of a monument will look to the rest of us like a picture of the monument with 500 visitors in front.

Show the monument, or show the spouse, but don't show both unless they are the same size (by standing far away).

When taking pictures, think what the viewer of the picture will see, not what you think you see before taking the picture.

And don't get emotional.


Film Speed

I'm hopelessly unqualified to type this section, and it's at best only approximately correct.

The definition of film speed looks something like (I hope it's 0.6):

                                                                1
ISO film speed = ________________________________________________________                                  light energy required to produce 0.6 O.D. above base+fog
                                               in units of ergs/square cm of film surface

Optical Density = what you think it means, except measured with a tungsten light                           bulb; no monochrometer. Low optical density means that you can                           see through it.

Film Characteristic Curves

A "correct" exposure of an 18% reflective gray card results in 0.6 O.D. for the gray card. Other things may have more or less density. 

The film manufacturers have arranged things so that a "correct" exposure of a thing (0.6 O.D.) gives the best rendition of detail In That Thing. The thing of interest is probably not a gray card, so the light meter can only tell you the resulting optical density relative to the standard 18% reflective gray card.


The stupid pH meter analogy

The pH meter :    1. Tells you the pH . It does not tell you if the pH is "correct".
                          2. Is calibrated to measure pH relative to some standard solutions,                               assumed to be "correct".

The light meter:    1. Looks at the metering region. This is the 12mm black circle in                               many Nikon viewfinders.
                           2. Averages over everything in the metering region.
                           3. Is calibrated to give a "correct" exposure (0.6O.D.), but only                                when the metering region is completely filled by an 18% reflective                                gray card.
                           4. Forcing everything to be "correctly" exposed is like making all                                buffers at pH 7.

The gray card method of setting exposures

            0. This only works for reflected light: The camera and the light source go on                 the same side of the thing to be photographed. In other words, use a copy                 stand.
                The result of this method is to set the exposure on an absolute scale, with                  the standard gray card mapped onto the standard optical density.
            1. Get everything set up to take the picture.
            2. Put a Kodak 18% reflective gray card in the same light as the thing to be                 photographed.
                On a copy stand, this means underneath the glass. Yes the glass makes                 a little difference.
             3. When the gray card completely fills the light meter, set the shutter speed                  and aperture to get a "correct" exposure. Shutter speeds are usually in                  discrete steps (1sec, 1/2 sec, 1/4 sec, etc.), but the aperture can be set                  inbetween the marked settings.
                 The optimum aperture for an f/2.8 lens is near f/8 (I'm assuming Nikon).                  Set the aperture at f/8, adjust the shutter to get close to a "correct"                  exposure, then fine tune with the aperture.
                 Color slide film is sensitive to changes smaller than 1/3 stop.
             4. Without changing the camera's exposure settings, take away the gray                  card, and take the picture. The light meter will then show a "correct"                  exposure only if you are taking a picture of an 18% reflective thing.

Examples, described for color slide film

1. A colored drawing on a white paper, illumination from above. The thing of interest     is the drawing.

    The camera and drawing are set up, focused, etc. The exposure is set on an 18%     reflective gray card. The reflectivities of the drawing and the paper then distribute     themselves relative to the gray card: The paper is so bright that it saturates the     film and appears white. Black ink does not register on the film and appears black.     Bright colors look bright, dark colors look dark. This only works when non-black     colors are very approximately 18% reflective.

2. A white paper, illumination from above. The thing of interest is the texture of the     paper.

    The gray card method fails completely. Fill the light meter with the paper, set a     "correct" exposure on the paper. This will place the optical density at the optimum      for the film. The resulting slide looks like 18% gray, but the texture and brightness      of the paper can be reconstructed in a (custom) print from the slide.

3. A black cat in a closed coal bin at night with no moon, heavy cloud cover, 1,000    miles away from artificial light sources.
   The light meter doesn't matter. There are no photons. Don't open the shutter.

4. A textbook perfect electrophoresis gel, with very dark bands, no background stain,    no faint bands.

   If the gel has been dried onto white paper, then the gray card method results in an    exposure on an absolute scale, but might not show the color of the bands if they are    very dark. Bracket in the overexposure direction to make sure you can see the color    of very dark bands.

   If the gel is still wet, then the light source would be underneath the gel. Set a    reference exposure on a region of background without any bands. Starting from the    reference exposure, dial in at least 2 stops overexposure using the shutter speed    dial. This will place the blank background near the bottom of the characteristic curve,    where it will look almost featureless in the photograph. 2 stops overexposure could    show a tiny bit of texture in the background, while 3 stops over probably won't. When    in doubt, scan inbetween.

6. Same gel, a few faint bands. The thing of interest is the faint bands.

    For a black and white picture of Coomassie stain, use a # 12 yellow filter to     enhance the contrast. A # 25 red filter seems to work nicely when the "original" is
    a color slide of a gel. Do not use these colored filters for color film!

    It doesn't matter if the gel is dried. Point the light meter at the background not     including bands. Set a reference exposure on the background. Scientists who are     emotionally involved with their gel bands will see the faint bands clearly, but very     faint bands look just like background to the film. It is difficult (for beginners) to judge     how overexposed the faint bands can be before they reach the end of the     characteristic curve and become invisible in the photograph. It is therefore wise to     try 3 OVER-exposures, in 1/3 stop increments, starting from the reference     exposure.

7. Electrophoresis gel, bands are stained just a little more than the background.

    A "correct" exposure really is correct, this time. The optical density of the picture     should be placed at the film's optimum to record the subtle difference between     bands and other junk.

8. Silver stained electrophoresis gel, the usual mottled background, very dark bands.    (Same story for Pharmacia Phast gels).

   The very dark bands look dark almost independent of any reasonable exposure. I     used to give 2 stops overexposure to the mottled background because it gave an     optimistic view of the gel, but I have since reformed, and I give it a more realistic 1     stop overexposure.
     Similarly, to get the pictures shown in Pharmacia Phast gel literature takes a
     2-stop overexposure for the background (and overloading). Phast gels in real life are      kind of ugly (these are opinions). Phast gel pictures from Dan's     photo service all      have 1 stop overexposure for the background, except as described in example 7.

9. Computer graphics colored stick figures on black backgrounds.

    I think that graphics screens should be photographed with 100mm lenses in order to     minimize the effect of screen curvature.
    I think that 85-135mm is the range that should work.
    A 55mm macro is only good for flat things.

    I can think of 5 ways to do this. One is to make a reference exposure setting for a     solid blue window, then dial in a 1.5-2 stop overexposure for the window, which is     brighter than the atoms. The second way is to point the light meter at the thing of     interest, which is mostly black, set a reference exposure, but then give it 2 stops     underexposure relative to the reference. The third way requires using a camera (like     mine) that has an autoexposure meter that integrates the light after the shutter     opens. The exposure fudge factor seems to be about -2 stops.
    The worst way is to rely on experience, because somebody might change the     screen brightness and not tell you about it, or you might need to take pictures of an     unfamiliar screen.
    Ideally, the software would provide a large test target with the same brightness as     the atoms.

    All this light meter business fails completely when the camera (like mine) has a     light meter that responds so fast that the graphics screen flickering becomes     observable. With my camera, I have measured a 4 2/3 stop brightness range     between the top of the flash and bottom of the decay. This bug has been left in the     design as a feature because it alerts the photographer that the shutter speed and     the moment of exposure will interact uncontrollably.

    Many people say that the colored stick figures on a black background don't     communicate anything. Art Olson says that "working graphics" are different from     "publication graphics" because "working graphics" communicate the structure to     the crystallographer who is accustomed to that structure, while "publication     graphics" have to communicate to people who know other things. Probably
    black-and white figures with tapered bonds would be better than the color stick     figures. I'm beginning to think that the stick figures aren't worth photographing.

    It is pointless to show stereo slides in a seminar.

A few non-scientific words about processing labs

The labs make very little profit from each roll of film. The big labs became big by maintaining consistent quality for thousands of rolls of film. Unless you live on the 900 block of North Highland Avenue in Hollywood, the little labs within walking distance should be used only when turn-around time matters more than quality.

Common photographic misteaks of science speakers

1. When using color slide film to make text slides, set the exposure on a gray card or something to ensure that white paper looks white. If you use the light meter on a white sheet of paper, it will look gray, by definition.

2. Color films must match the light source. Tungsten films work with tungsten lamps. Daylight films can also be used with tungsten lamps, but only using an 80A (blue) filter for both the light meter and taking the picture.

3. The text of text slides should be as large as possible, and carefully aligned to the edges of the film. Tilted text looks stupid.

Film choices and interconversions

Black text on white paper-------------------------------------------------------> black on white slide
                                      Kodak Precision Line LPD4 in Dektol


Black text on white paper------------------------------------------------------> slide with white text,
                                                                                              dark blue background
                   Polaroid Polablue film, you need the processing machine.


Black text on white paper-------------------------------------------------------> slide with white text,
                                                                                                color background
      Kodak Vericolor SO-279, with filters as instructed, plus 45M40Y (I think).


Black text on white paper---------------------------------------------------------> black on white print
                                        Kodak Kodalith in D-19, then print on
                 Kodak Polycontrast III RC, finish F, at grade 5. Or, use a laser printer


Color negative---------------------------------------------------------------------------------------> color slide
                                             Kodak Vericolor SO-279


Black and white negative--------------------------------------------------------> black and white slide                                    Kodak Technical Pan in Technidol Liquid
                        Photo puzzle: How do you get a white background for a
                        black and white slide of a gel, with negative film, starting
                        from a negative.


Black and white print-------------------------------------------------------------> black and white slide
                                         Kodak Direct MP in Dektol
           or Agfa "Scala", with 2 days processing by Anonymous Mystery Labs


Color slide--------------------------------------------------------------------------------------------> color slide                              Kodak Ektachrome Slide Duplicating film 5071,
                             with filtration as printed on the box.
                             It's best to pay somebody else to do this.


Color print--------------------------------------------------------------------------------------------> color slide
                  Kodak literature says that Ektachrome Slide Duplicating film
                  will work, but I think you're better off with Ektachrome 64T tungsten
                  pro film. I will try Kodachrome 40 Type A one of these days,
                  because it has a blacker black, etc.


Color original, colors don't matter--------------------------------------------> black and white print
                    Kodak TMax 100 in Tmax, or Technical Pan in Technidol liquid,
                    printed on Kodak Polycontrast III RC, finish F.
                    Photo puzzle: Will a black and white print of a graphics screen
                    reverse the depth cueing?


Color original, colors matter-------------------------------------------------------------------> color slide
                       On copy stand: Kodak Ektachrome 64T Tungsten Professional
                       Flash, outdoors, or graphics: Kodak Ektachrome 100HC,
                               or Ektachrome 100plus pro (both with bright colors)
                               or Ektachrome "Lumiere" 100 pro (slightly bright)
                               or Ektachrome 100 pro (neutral),
                               or Fujichrome Velvia (screaming colors).


Color original, colors matter--------------------------------------------------------------------> color print
                                              On a copy stand with tungsten lamps:
                                                    Kodak Ektar 25 pro film + 80A filter
                                               or Kodacolor "Royal Gold" 100 + 80A filter
                                               or Ektachrome 64T tungsten pro film, no filter,
                                                    followed by "Type R" print.
                                          Printing from slides is supposed to be unambiguous.