Mark Stock: Open House

  • ©, Mark J. Stock



    Open House



    3D modeled image


    24" x 24"


Artist Statement:

    Even though digital technologies, and computers especially, are capable of expanding the range of what we can experience into the unreal, it remains a worthy goal of computing to recreate what is most intuitive and familiar to a person’s perceptions. It is no coinci­dence that what nature does most effortlessly, in every detail and without fail, is most elusive to computer scientists.

    Science seems to have always had a hand in creating art, whether it is dictating the proper mix of dyes or stacking the layers of an emul­sion, but never before has the brush of science been so capable as it is with today’s tools for scientific computation.

    The aim in much of my work is to combine the realism of these tools with otherwise completely fabricated data and have them fight it out. The goal in Open House was to create a landscape so foreign as to be nearly repulsive, but so real as to invite continued exploration.

    Additionally, I wanted to use geometry to portray the dirty numerical underside of computational science: large problems being broken up into incredibly many pathologically simple problems. Taken as a whole, the many little solutions blend into the perception of a com­plete, smooth solution.

    The fluid-dynamic calculation in Open House is of an unstable system on the verge of flipping-putting what is above beneath and what is beneath above. The fingers that reach into the space above are the harbingers of a total reversal. The interface that we see will soon be upside-down.

Technical Information:

    The underlying geometry in Open House is the result of a Rayleigh ­Taylor instability, a fluid-dynamic phenomenon in which an unstable layer between fluids of differing densities is distorted under accelera­tion. The shape was calculated with a new computational fluid-dy­namics method that was the result of several years of the artist’s dissertation research. Even with the improved efficiency of new algorithms, hundreds of billions of calculations were required to advance the simulation to the time shown.

    The open cubes that grow over the landscape are positioned ran­domly around the centers and aligned with the edges of each of the triangular elements on the computational surface.

    After an appropriate sun position and sky color distribution were set, the entire scene was passed to the rendering software. The final image was rendered at 24,000 by 24,000 pixels by Radiance, a scientifically validated lighting simulator and pseudo-radiosity ray­tracer. Radiance traced more than 10 billion rays over two weeks to compute the light inter-reflection throughout the scene.

    The high-dynamic-range rendered image was print-optimized and exported to a Lightjet printer, which exposes photographic paper at high resolution with laser light.