Dithering and Animation
I have an ongoing project exploring dithering algorithms in Processing whic has turned into a platform for image manipulations, subverted algorithms, and software-driven animation.
I’ve implemented several different dithering algorithms in Third Ear, including Pass-Right; Floyd-Steinberg; Jarvis, Judice, and Ninke; Stucki; Atkinson; and Bayer 8x8; and some derivations of my own.
Most dithering algorithms look at each pixel of an image in sequence. Pixels more than 50% black become black, and those more than 50% white become white. The difference between the pixel’s final value and its original value (i.e. the quantization error) is distributed to neighboring pixels (see above) which has the effect of creating more natural looking gradations in the final image. This process was important for displaying images on black and white screens or printers.
But if creating natural-looking images is not your goal, there are many interesting ways to subvert this process. For instance, passing the error backwards, cycling algorithms within a single image, or distributing the error in creative ways.
I began introducing additional image manipulation algorithms into the mix, eventually creating a manipulation sequencer that could queue operations.
The natural motion of the dithering felt like it was begging for animation. I began experimenting and eventually wrote a system that could process either still images or movie files, and even composite independent image animation streams.
Third Ear was an experiment in creating artwork and the tools to produce that artwork simultaneously, approaching both with the same principles of exploration and critical evaluation. When the goal of the tool is to support the artwork rather than any particular set of features, an interesting feedback loop between tool and product is created.
