7 February 2018

thabir – Interruptions

I began by studying the work. I found the most effective method for this to be copying/tracing the work into my notebook, which is what I started doing:

The artwork is square with equal margins on all side
The artwork consists of many short black lines of the same length and line weight
The centers of these lines form a grid, indicated by the red lines in the top right hand corner
The lines are place at angles between +90 and -90 degrees
Not all angles seem to be represented. Certain angles get repeated quite frequently and there are certain repeating patterns/units/motifs as well
However, within a smaller subset of this angle range, the angles seem to be randomly chosen, for some definition of random
Interspersed throughout the grid are spots where the grid is empty or ‘interrupted’
These spots do not seem to have a set shape, size, or placement. All these features seem to be random
Spots on the grid which have a greater number of empty neighbors seem to have a higher probability of being empty.

I found this assignment to be particularly interesting because of the simplicity of the results we were trying to reverse engineer. This simplicity allowed us to focus on the expressive and poetic qualities of the generative art experience as opposed to the purely technical aspects. Because we weren’t bogged down with the weight of libraries and sensors, we could focus on the subtle variations in effect of lines, probability distributions, and randomization, rather than wracking our brains trying to get our code to compile.

This aspect of the assignment showed me how there was nuance in code, how two different solutions could technically satisfy all of the bullet points I listed above, and yet could both fall short in surprisingly different ways. It started getting me thinking about code experientially.

I found that the hardest part was achieving the effect of the line angles and interrupted spaces. Getting the line placement was easy enough after sketching for a bit, because I noticed the lines were on a grid. However, the angle distribution was trickier. I ended up reducing the possible angles that a line could take to a small set and used a probability distribution on that set to generate angled lines. I also used a probability distribution on a cell’s surrounding cells to decide whether it would be interrupted or not. A greater number of surrounding interruptions causes the probability of interruption to be higher. This is basically a probabilistic cellular automata. I had to go through a couple of iterations before I could be satisfied with my results. Before I used the cellular automata method, I tried using Perlin noise but that wasn’t as effective as I hoped. An image from that iteration is attached below.

I think what I really learnt was that this form of coding is just like drawing, in that it is all about the interaction of the image with the eyes. The “correctness” of the code itself is absolutely irrelevant to the final effect, and therefore it is a matter of iteration and refinement to achieve what you intend.

SaveSave

2 February 2018

thabir – LookingOutwards02

Caddis, Hubert Duprat

The generative work which I find most fascinating is the work of French sculptor Hubert Duprat. Duprat’s work spans the entire gamut of techniques, materials, and genres, however the one aspect which stays constant is his utilization of the effective complexity found in nature to create sensual experiences, which are all the more ethereal for their being so organic. In the video linked below, Duprat discusses three of his pieces:

The first is a sculpture made out of attached sections of PVC pipe that, due to a trick of the light, are made to look both as if they were cast from metal and made of bubbles. However, this work wouldn’t yet be considered generative according to Galanter’s theory of generative art, since, even though he is mimicking the mathematical patterns of nature, he is not yet relinquishing control over aspects of his work. His second piece utilizes the natural simplicity of crystals to come up with a simple, stacked structure emulating the simplicity of the crystal itself.

However, his most famous body of work, started in the early 80s, is also that which would unambiguously be defined as generative: his work with the caddisfly larvae. The caddisfly is a common fresh water insect which naturally constructs elaborate tube shaped shells for itself using found detritus from the bottom of the river. Duprat started removing these insects from the bottom of the river, removing their protective layer and storing them instead in a tank whose bottom was lines with gold flakes and semi-precious stones. Deprived of their natural building materials, the insect is forced to construct its tube out of gold.

I would definitely call this a generative piece, due to the relinquished control over the specifics of the tube construction. Duprat refers to these sculptures as collaborations between him and a caddisfly, clearly showing the importance of the natural architectural instincts of the instinct to the piece. Due to the inherently organic nature of the final construction, the piece has an extremely complex approach to complexity (more in line with effective complexity than Shannon’s high information density). Like all biological occurrences, there are clear patterns visible, however these patterns are obscured with great quantities of chance, abstraction, and evolution, all of which heighten the uncanniness of these gold shells.

All of his pieces, unlike much art that tends towards the generative genre, have a captivating materiality and/or texture to them that both draws us closer while also pulling the work back. This piece is in line with the rest of Duprat’s work which seems to be a meditation on the viscerally sensorial aspects of effective complexity.