Beyond Models: Biologically-Based Art

This talk is in two parts. Part One is a quick look at some of my projects that have involved simulated systems of one sort or another. Part Two looks at the work of various artists using "real" (read: non-simulated) systems in their work. I'll try to explain a bit about why my own work has moved from being mostly simulation-based to mostly "real."

Some examples of my work with ALife/biological/physical models:

Evolver (for the BeOS, RIP)
Used GAs to evolve DSP routines
genotype: list of DSP opcode
phenotype: sound!
fitness function:

is amplitude > 0?
is frequency > 0?

stupid, but fun


A nice accident


At one point I made a very simple GA framework for students to explore. It could make simple sounds. While debugging once it made its best sound ever: evo.wow.wav. Unfortunately it was an accident.

Tuning (OWT (Optimal Well Temperament))


Project with Larry Polansky to solve a brain-hurting tuning problem.

Anty (java)

Simulates a foraging ant colony
Sort of
Nice drawing tool
Helped me start to ask myself: What does this have to do with sound? Do I need to explore these ideas via sound, or are there other more interesting media? What is interesting about this, anyway?

disc-o


Jason Freeman and I wrote disc-o as part of a kids software project. It was an attempt to introduce them to a few ideas:
You can model and simulate physical/biological/social/etc systems You can analyze a system and track its interesting properties over time
The result of your analysis is a set of data
Data is abstract, and can be interpreted in many ways

The system we set up was kind of strange: magnetic discs floating on water.

While this was a fun project (and the kids liked it) I now have some misgivings about the ideas the project was dealing with.

You can download disc-o on the JPMCKDMS Project website.

Biologically-Based Art

organism: making art with living systems


Andy Gracie/Host Productions: small work for robot and insects

"Despite a huge amount of effort, programming skill, engineering skill and artistry, the crickets remain largely indifferent."

Damien Hirst: A Thousand Years

A rotting cowhead and lots of flies make a cycle. Sort of.

Tiffany Holmes: Follow the Mouse and Fishbowl

Small animals as sources of entropy?

Edgar Lissel: Bakterium

Making photographic images rendered in light-sensitive bacteria.

Marta de Menezes: Nature?

Biotechnology as rapid natural selection -- modified butterfly wings.

Dave Powell: ArtCats

Biotechnology as rapid natural selection -- breeding "defective" cats.

Douglas Repetto: How to Annoy a Plant and crash and bloom

Exploiting biological systems (organic and silicon).

the semi-living artist

Growing artists in a vat.

Yukinori Yanagi: Ant Farm Pieces

A real ant colony algorithm!

Some closing thoughts, ideas, questions

I don't agree with all of this work, but I do find it compelling.

Using "real" things often brings up moral questions. That's good.

Data is not abstract in the real world. The source does matter.

Using "real" things gives you a certain amount of richness for free: those ants are particular ants, in a particular place. That place has a history, those ants have a name.

Is context free output really what you want?

Are you just making toys? Is your piece "about" an algorithm? Is there anything else there? What's the difference between an implementation and an idea?

Is a guitar pedal interesting? Is listening to PI interesting? Is a flocking algorithm interesting? Sometimes, but why and when? If you "sonify" a simulated system or a dataset, what are you really doing? Are you really getting to something interesting, or are you just playing with a toy?

These are all issues that I think about regardless of whether I'm working with real or simulated objects and entities. However, I find that answering these questions becomes more urgent when working with real things, and that urgency helps me to think more clearly about what I'm doing and why.