IAN SPADIN Voxel Forms

An algorithm generates contextless and scaleless models. These models are based on a 3D cubic (voxel) grid.

The algorithm understands a limited language of verbs, each representing a different volumetric operation. BASE creates a large centered mass, while GSV and SV create “sister volumes” that must touch an existing piece of the structure.

The algorithm does not actually understand volume. Every possible location in the model’s boundary space is understood as a collection of 6 different faces that may be off, wireframe, or opaque. To create a volume, the algorithm grabs a volume of cubes and makes every face opaque. To save on computing power, it will then turn off all the interior faces so that it only needs to draw the faces on the outside.

This allows the algorithm to play with the surfaces of the perceived volumes in unusual ways. GV will grab a random volume of cubes and replace any solid faces with wireframes. GF performs the same operation on every other vertical slice of the entire model. To the algorithm, this operation is not much different than creating a volume.

WV grabs a random volume of cubes, ignores cubes that do not have any opaque faces, and then inverts the states of the opaque and disabled faces. The result is entirely a byproduct of the optimization technique from before.

RV disables all faces in a random volume. This instruction tends to reveal the exact nature of how the volumes are illusory.

A set of instructions creates a variety of forms due to random chance. Any form can be saved to a file for later use, but the resulting file is around 30,000 characters in length when the basic idea of the building can be captured in about fifty.

Thanks to ARTIS 475X and Austin Stewart.

Datum ISU