While experimenting with the creation of ray-trace images, I discovered an unexpected way of using a ray-trace application (POV-Ray) to produce images of three-dimensional objects that are not defined in the source file.

Conventionally a ray-trace scene is created by specifying shapes, combinations of shapes, colors, textures, reflectivity, transparency (along with many other parameters), one or more light sources and a camera - in a text file called a scene file. The ray-trace program then processes the scene file, renders a visible image and saves it to an image file. In this manner, one can create highly detailed photorealistic scenes. Typically, a complex ray-traced shape is created by assembling many geometric shapes to form a resultant object as in this example image of a teakettle. The teakettle is made up of many shapes including spheres, cones and cylinders. Adding color, texture and reflective qualities helps create a photorealistic image. Every detail of the image is specified in the scene file.

My images depict highly complex three-dimensional analog objects that contain great amounts of visual detail. These images have great variety and contain a high degree of detail over a broad range of scale. Yet the scene files are extremely simple, typically using from one to four simple geometric objects (sphere, cylinder, etc.). None of the resulting shapes, detail or textures in my images are specified in the source code but are produced by using certain parameters in an unexpected way that creates an image of a complex object emulating the processes of chaos theory. The images cannot be predicted by inspection of the initial conditions specified in the scene file. Here are some examples.

Ray-trace page