Another series of sketches exploring organic curves.  Particles follow wandering paths, stopping when they cross another path. The shape of the path, the number of particles, the initial starting positions, the way the particles die,… all have a major impact on the final image.

The ultimate example of this drawing mechanism is Jared Tarbell’s masterpiece ‘substrate‘. Some of the fine pieces by Dave Bollinger also seem to use something like it.

Click on the image to go to the corresponding construct. A click of the mouse resets the piece. On some platforms the lines are not properly anti-aliased. I haven’t figured out why but it sure messes up the images.

I have to admit, I’m hopelessly fumble-fingered in vector drawing. I never get the curves right, somehow they’re not smooth, not natural looking.

The funny thing is that a 1000 particles chasing each other from random starting positions seem to manage this quite nicely. Initial jagged chaos swiftly evolves into flowing curves, continuously trying to simplify themselves. Guess I’m not a sufficiently complex, self-interacting system

Anyway I’m rather pleased with this construct. Pointless overkill makes me happy. The construct and its code can be found here. There’s no interaction beyond resetting with a click of the mouse.

Update: Lucas Tamarit pointed out, somewhat uh… pointedly, that this piece is very similar to ’sandtraveller’ by Jared Tarbell. Check this and the rest of his work out here. I’ve got to agree there’s definitely a resemblance. However, I’ve checked out the code and the mechanics behind ‘cyclic’ are really different from ’sandtraveller’, as is the rendering technique. I invite everybody else to make the comparison.

Not every painting of a sunflower is a Van Gogh forgery

Up to now, I’ve always treated particles as their physical equivalents: pointlike objects undergoing forces. As a physicist, this way of thinking comes natural, however it is fairly limiting. There’s a lot of work out there dealing with particles as agents, capable of steering towards goals, avoiding obstacles and flocking together in interesting ways.

Towards is my first step in exploring the possibilities of agent-generated graphics. It has an old-fashioned feel about it. (And yes, I got stuck on the black-white-red palette again…)

To get a feel of how this kind of explorative programming goes, I’ve uploaded seven versions showing the chronological (pro/re)gression of the code. All pieces need some time to reveal their true dynamics. Click on the applet to restart the construct.

towards 01
towards 02
towards 03
towards 04
towards 05
towards 06
towards 07

I’ve always been attracted to bright lights, so I’m particularly fond of this one. It reminds me of the laser speckle of an old HeNe laser.

A darker visualization of the funnelweb structure, pulling the geometry into a Lovecraftian nightmare.

Even the dustiest material can form the start of an interesting structure. The lattice in funnelweb is constructed by drawing the shortest connections between a collection of random lines, and then by taking the closest connections of the connections, and so on, and so on ad infinitum. (Well, it’s only ten times but by then everything has become too tiny to make a difference.) As a finishing touch, add a sprinkling of particles moving along the lines. Click the piece to restart.

Since it’s no use reinventing the wheel, I’m using Toxi’s geometry library for manipulating 3D vectors. I also tend to consult Paul Bourke a lot when playing around with geometry.

A small experiment with particle constraints. Draw by clicking and dragging the mouse. A particle chain with fixed length links will try to follow. Pressing space blurs the image drawn so far. The “s” key saves the image to the user gallery. Any other key resets the sketch. Click here for the interactive applet.

Another piece of code stuck in a drawer. Originally it was meant to be an exercise in collision detection between falling particles. Fake physics, shudder. To make the interactions interesting the particles need to move at different speeds, otherwise they would just get stuck together. But the movement never felt right. Fooling around, I reversed the vertical motion and somehow it works… I’d like to claim it was a conscious decision, but it wasn’t.

It’s amazing how the collisions suggest rotation. Try to follow a small bubble when it hits a larger one, can you see the larger bubble rotating? The interactions involved are deceptively simple, the bubbles move slightly apart when overlapping. Nothing else. Actually, the hardest part was finding a name (and spelling it): buoyancy (click here for a version without OpenGL). Moving the mouse pops the bubbles.

Curiouser and curiouser… I was cleaning up some old folders this afternoon and found an interesting processing sketch from 2005. I can’t remember writing it but the header and the code point an accusing finger at me. A qui(r)ck conversion to v.135 and voila.

Left mouse click freezes the tangle and changes the rendering. Pressing spacebar will save the image to the user gallery, enable pop-ups from this site if you want your own copy. Any other key randomly generates a new tangle.

Enjoy!

Update: on some platforms, it looks plain weird. A non-OpenGL version can be found here