The show premiered at Dansescenen in Copenhagen, and has been shown in Firenze 2010 and Stockholm 2010.

You can read much more about the performance here: Recoil

Photos by Søren Knud Kristensen and Søren Meisner

The piece is going to be setup again next fall at dansehallerne in copenhagen, and afterwards on a tour.

The industrial robot N7331227 has been used to grind toilet seats – repetitively, monotone and precise. Each of them perfect, each of the similar.

Now that its working days are over, Illutron equipped the Robot with new attributes, which spot light on the robots limitations as a machine and the humans need of emotional attachment to things.

N7331227 is superior, when it comes to accuracy and endurance, despite of his fascination of us unintelligible and fragile creatures. In an effort to understand us it tries to copy our movements and actions.
In the interactive dimension of the installation, the spectator is asked to express his or her own creativity in paper drawings. N7331227 then attempts to copy the drawings via a panel which controls 96 light bulbs in a matrix

“Kunsthallen Brands Kladefabrik describes the installation as following:
The Danish artist collective Illutron participate with the work “N7331227” which imbues an old industrial robot with new life. Via computer vision the robot has been equipped with the ability to see and has been programmed to read and reproduce the audience’s drawings on a large wall comprising 96 light bulbs.”

This is a project i developed in two weeks together with two friends from the royal school of architecture in copenhagen for a big event that a group of schools held called Holmen Event. We wanted to make a big interactive wall that the audience could play with. It ended up being a photowall, where a empty canvas (beutifully cut out with a lasercutter), could take photos that the audience choosed, and afterwards used to hang the photo on the wall next to all the others.

The canvas was tracked by a camera in the roof when holding it in the light, and when my program saw that it was stable in the frame, a photo was taken inside the square that the canvas formed. At the same time a sound was triggered, and the whole sound universe (designed by my friend Rasmus Kreiner, a theater sound designer) changed.
Afterwards the audience could hold the canvas against the 4x8m big backprojected wall, where a infrared camera behind could track where it was, and how it was rotated. And again, when my program saw it was not moving, it attached the photo to the wall, where it moved in a physics simulator (pushing to other photos) and slowly moving upwards before it was deleted again.

The software is developed in OpenFrameworks, using alot of OpenCV code. The tracking of the canvas on the wall was done with tBeta.

The installation ran for 7 hours during the party, and there was taken more then 500 pictures, and the audience really liked it. People understood the concenpt (even drunk) right away, and told it to each other when passing by. Many came back later and took some more photos

Ole Kristensen and i developed the interactive scenography for this dance performance, choreographed by Tina Tarpgaard. I was on the project in the early programming states for a couple of weeks, helping Ole with programming all the ideas there where, and developing the visual style that Ole then would work further on until the premiere.

You can read alot more at Ole’s blog

Read some more at ole’s own blog

We have also been working on making a new blob library that is not a perfect match around a body, but instead like a bubble with a person inside. It can give some really nice organic effect when the forces are very low in the bobble, and it just follows slowly the movement of the body. This is build by me with some basic physics laws.

For now we got this video of it, but we will cut something better later

The hardware
The surface is a thin glass plate  with a Lee filter 400 underneath to project on. In the bottom of the box there is an ordinary surveillance camera with a IR filter in (so it only sees infrared light, and not visible light), and a BenQ short-throw projector and a lightbulb wrapped in some color filters. it does give some visible red light, but its not very much. This technique is called DI (Diffused Illumination).

The software
The tracking of the fingers with the camera is done by the opensource application tBeta, and send with OSC to my own openframeworks application that does all the graphical work. I might release the code at some point. 

If you have any suggestions or questions, feel free to ask me!