Friday, September 16, 2016

Tiphobia horei and photo testing

kw: species summaries, natural history, natural science, museums, research, photographs, photography

This post is only incidentally about Tiphobia horei E.A. Smith, 1880, and more about the process of getting good photographs and making images that are both attractive and useful. T. horei is a freshwater snail that is endemic to Lake Tanganyika, which forms a major portion of the boundary between the Democratic Republic of the Congo (called Zaire for a couple of decades) and Tanzania; it also extends to the DRC's boundaries with Zambia and Burundi. The Delaware Museum of Natural History has two lots of this species, both collected in the mid-Twentieth Century. The lot shown below consists of two shells; the other lot has but one.

This image shows six presentations, obtained thus:

  • The background for the column on the left is a black t-shirt I happen to have, though it has a giant Dupont company logo on the front.
  • The background for the column on the right is a piece of commercial black felt. Closely looking at all the original images, I decided the felt is the better choice.
  • The top pair of exposures is the normal exposure the camera's light meter indicated.
  • The second pair was taken with a -1EV setting, and the third pair with -2EV. This was to be sure I obtained exposures that didn't have washed-out highlights on the shell. I needn't have worried.

I use Gimp (GNU Image Manipulation Program; GNU is a freeware consortium), a free alternative to Photoshop. The first priority here is to stretch the contrast to get the most out of the subtle hues of the shells, and push the background to black and the white on the scale card to fully white.

Here we have a lot of stuff in one image. On the left is the Adjust Color Levels dialog for one of the normal-exposure pictures. On the right is a crop of just its histogram, plus crops of histograms for the other two exposures. We can see how reducing the exposure just shifts all the tones toward black.

Three regions of interest appear in all three histograms. The peak on the left shows the range of dark grays from the background and the black part of the scale card. The peak at far right shows the near-white hues from the scale card. The bumpy peak near it shows the hues from the shells. The very low peak just left of center in the top histogram is from the penny. The clear separations make it easy to push the contrast as I desire. In the Levels dialog one just moves the little triangles at the bottom of the histogram to set the limits and adjust the gamma curve (which initially is 1.0). The following image is the upper right one from the montage above, just cropped to a tight square but otherwise unprocessed.

Here we can see, in addition to the shells and labels, the texture of the felt background. I used a stack of pennies to hold the tip of one shell up so the aperture is in the right orientation. By the way, the lower label indicates "40 m depth", but in the log book it says "40 mm depth". I suspect 40 meters is correct; this species is found anywhere from the surface to 100 m depth but is not common in shallow water. To obtain the next photo I set the contrast limits to 64 and 240 on the Adjust Color Levels dialog. I left the gamma at 1.0.

Now the entire background is solid black, the white of the ruler is at full brightness, and the color variations on the shells are much more distinct. Had I used a lower limit of 128 instead of 64, the pennies would also have become almost entirely black and the range of shell tones would be very strong, but it would look overdone. There is a lot to be said for working with the discriminatory powers of the human eye and brain, and not seeking to overwhelm them. For most purposes this image is a fine illustration of the shells and their labels. At times, such as for identifying shells that are more subtly similar to others, it is useful to increase the visible detail, thus I used one more processing step to emphasize the decorations and markings on the shells:

For this image I used a little bit of Unsharp Masking. It would usually not be a good idea to use this for a published illustration. The masking, a type of sharpening, emphasizes the fine details while lowering overall contrast. Unsharp Masking is one type of High Dynamic Range processing, and works even better when used with a Raw image. But a JPG image has 8 bits of color dynamic range, which comes to a 256:1 range, much greater than the typical 25:1 of a color print or magazine illustration or the 50:1 (at best) of a well made color slide. Thus we can often "pull out" details that would be lost if we just lowered the contrast overall.

Unsharp masking was developed decades ago for film processing, by astronomers who wanted to bring out details in photos of nebulas and galaxies, which have a very high contrast range. It consisted of making an out-of-focus copy of the negative, of the size you are going to make your print. That is the Mask. Then by printing through the mask, much of the large-scale dynamic range is suppressed, and the finer details are emphasized. It takes some experimentation to figure out how far out-of-focus you need to go to make the mask, and which level of contrast in the mask will appropriately render the image you want (when I was doing darkroom work I had five levels of contrast available in the printing and large-format negative films). It is time-consuming and expensive. The Unsharp Masking filter in Gimp and Photoshop and similar software emulates this process using computer code. One may tweak the controls on a dialog and see instant feedback.

As the label above also shows, these two shells were used to illustrate this species in the book "Best of Nautilus". It was a bit of an honor to handle them.

No comments: