Form Constants II
The missing pieces
By Stuart Smith
In my previous article, “Form Constants,” I described the recurring patterns that Heinrich Klüver, and other researchers following him, have reported seeing under the influence of peyote and other hallucinogens. There are four such patterns, which Klüver dubbed “form constants.”
Because of the limitations of my artist’s assistant app at that time, I was able to illustrate only two of the form constants: “tunnels” and spirals. The two remaining patterns, “cobwebs” and lattices, could not be generated. Fig. 2 below shows these patterns. Fig. 2 left is a cobweb pattern, and fig. 2 right is a lattice pattern (Klüver included square and triangular grids in the lattice category in addition to the honeycomb shown in fig. 2).
Working with cobwebs and lattices
Since writing the original article, I’ve added the capability to generate cobwebs and lattices to my app. This allows me to make images based on these two patterns. Some early experiments have shown that that cobwebs and lattices can’t be manipulated in quite the same way as tunnels and spirals.
Fig. 3 below shows a cobweb skewed toward the upper left corner. Skewing the cobweb is one of the few transformations you can apply to it without destroying its familiar configuration. Skewing perhaps gives it an even more spidery, naturalistic appearance while the original pattern (i.e., Fig. 2 left) looks like a figure made with Spirograph.
Fig. 4 below shows a honeycomb modified by ‘pincushion” distortion, which is the opposite of the “barrel” distortion shown in Fig. 1. Lattices can be subjected to both barrel and pincushion distortion without losing their fundamental character as form constants.
The fundamental limitation
As I see it, the cobweb and the honeycomb are both limited by the fact that the strong black lines in each figure define small, regular “cells.” The question then becomes what to do with the cells. You could treat them like the outlines in a coloring book and fill each cell with a different color. Or you might put some fine-detailed line drawing in each cell. But either way you’re basically just ornamenting the cell structure. My “solution” to this problem was to adopt an approach to coloring different from my usual practice.
Coloring the images
To obtain the color scheme for the images in this article I used “Perlin noise.” Perlin noise is a popular generation algorithm invented by Ken Perlin. It can be used to generate things like textures and terrain procedurally, meaning without their being made manually by an artist or designer. It’s especially useful for games and other visual media such as movies. Perlin won an academy award for the original implementation. Fig. 5 is a typical example of the type of noise I used. As can be seen, it contains muted suggestions of various colors. This is important because it gives subsequent color operations something to work with.
With my app I combined a cobweb or lattice image like those in fig. 2 with a Perlin noise image, and then applied skewing, barrel, or pincushion distortion. The resulting image in each case was just the line drawing of the figure superimposed on the rather drab background of the Perlin noise — not the kind of image you’d want to put on exhibit. Additional processing was necessary.
The final step: GIMP
I like to make all my images using only the resources of my app, but I often do one final step with GIMP to liven up an image. In the case of the images shown in this article, it took the application of three GIMP filters:
1. Artistic/Dream Smoothing
2. Colors/Boost Chromaticity
3. Details/Local Contrast Enhancement
These three final steps produce the candy-colored images above. In retrospect, Step 3 may have been overkill; after step 2 the colors of the image were already distributed as they would be in the final image and they were already quite vivid.
©2023 Stuart Smith. All rights reserved.
