essential code #3: noise
Noise can add “life” to a visualization. Life is texture and rhythm. For example …
C.E.B. Rea’ project on rotating squares shows that movement itself can take on form.
The tendrils on this scientific taxonomy draw viewers’ eyes into the words. It’s interesting that the authors view the organic layout as a weakness. In terms of presenting the data, one could argue the curved (and sometimes upside down) text is harder to read; but in terms providing starting points for entering the work, the organic layout is beautiful. I’ve heard people comment on this work, “oh, it’s interesting that X is moving this way”. From that, they go on to explore what X means, and the connections X has in the data. What else could you want for disseminating a message? On another, technical note, the method used by the authors to layout the tendrils is very nice …
Ken Perlin’s Responsive Face is an example of how variation and personality are intimately linked. When you make a system with some amount of noise, it has a vestige of personality. That is, unpredictability is an element of personality. Ken Perlin wrote more about this in the (overall mediocre) book About Face.
Below is an example I made to show how noise can be engaging. In terms of information, it’s an empty space; there’s nothing being presented here but movement. But, do you want to get to know it? The point is that, most people will be drawn into a system like this on multiple levels: it’s dynamic; it’s smooth; there is some order; and the content is recognizable (in this case, faces); and it’s not predictable. A system of noise is a platform on which you can layer real information.
Interaction: mouse over and drag the circles to change the force field
What are some ways you can add variation to a visual system?
A noisy 1-bit signal like 0.6 <= rand() has variation but not order. That is, noise like this will tend towards a “jagged” look of sudden turns and sharp corners.
Another type of noise is called “smoothed noise”. It’s basically 1-bit noise, smoothed out so that transitions pass from white to black through some gradual gray. For example, if you had a 32×32 image and set each pixel to a coin flip (black if heads, white if tails), and then blurred it a couple times, you’d have a form of smoothed noise.
Another form of smoothed noise is Perlin Noise (p.n.). This post shows examples of p.n. There are two versions of p.n. — aperiodic and periodic. Aperiodic is good when you want a stream of unique, smooth noise, as in drawing with a red pen. Periodic noise is good if you want to create rhythm.
Once you have a way of generating noise, you need to apply it. Force/physics models can smooth motion (with momentum) while letting noise influence moving objects. You can change the intent of your algorithms is random ways. For example, the “sand” effect above is just an erosion filter that randomly ignores some pixels.
This demo uses emotibles chat icons.
12.22.07 — whoa, I just came across cell noise … http://carljohanrosen.com/processing/index.php?page=fp&topic=2.0
I’ll integrate this as soon as I understand it!