(Originally posted March 2009)
More than 10 years ago I decided to pursue art as a hobby but dropped it after getting discouraged and getting quite interested in something else. Recently I decided to look at it all again. I quickly found out that I still much prefer abstract art but still find it hard to make something pleasing.
In looking through my old materials I found that I'd saved an old [Google says 1978!] Scientific American column by Martin Gardner which discusses Richard Voss's then-new theory that what makes music pleasing is that it is 1/f noise. The article also explains a simple algorithm for approximating a 1/f sequence. The article did mention that something similar might apply to abstract art but at the time I had read the article in conjunction with my interest in art, I was out of the programming business.
On the canvas of a painting colors and shapes are static, reflecting the world's static patterns. Is it possible, Mandelbrot asked himself many years ago, that even completely nonobjective art, when it is pleasing, reflects fractal patterns of nature? Mandelbrot has some unpublished speculations along these lines. He is fond of abstract art, and maintains that there is a sharp distinction between such art that has a fractal base and such art that does not, and that the former type is widely considered the more beautiful.
Since the "something else" I've been doing involves Java, last night I was able to quickly program up Voss's algorithm to generate a sequence of values selected from a set of 16 consecutive values. The generated sequence is not random but has close to a 1/f distribution. To produce music you would map the consecutive set to 16 consecutive musical notes. So the question became how to map the consecutive set so it could be used in producing art.
Thanks to Painter, a most wonderful Christmas gift from my younger son, I've learned that it takes three parameters to define a color that is distinct to the human eye. These three parameters are typically chosen as the relative amounts of red, green, and blue but there are various equivalent systems to the RGB one.
Last night it occurred to me that the Hue Saturation Value (HSV) system for color might be a useful basis for choosing consecutive colors. Briefly, the HSV system works like this. If saturation is 100%, then the "per cent" for hue represents a pure color. For example, 0% is blue, 33% is red, and 66% is green. A saturation less than 100% means that some white, black, or gray is mixed into the pure color corresponding to the hue. The value parameter represents the proportion of white in what has been mixed in where 100% is pure white, 0% is pure black, and 50% is a medium gray.
For my first test, shown above, I only varied saturation. I picked a value of H corresponding to a red hue and a value of V corresponding to medium gray. I then interpreted the output of my program as being a sequence of saturation per cents selected from a consecutive sequence ranging from 0 to 100. I used these per cents as follows in Painter. I started at the left side of the virtual canvas and painted (freehand) a vertical stripe with the color defined by my fixed values for H and V and the first per cent saturation from my generated sequence for S. I then kept moving to the right painting similar stripes but using the color picker to select the color with the next per cent saturation from my sequence. I was very happy at how this first picture turned out.
It occurred to me later that rather than generating an approximate 1/f sequence with a computer program, I could also extract such a sequence from a piece of music. I haven't tried this yet but it seems to open up some interesting possibilities. I also plan to play around with various ways of choosing "continuous" sequences of colors.