010 of 366
A white vertical line on a black background scans across the screen from left to right and back again. The resulting video is exported and re-compressed 102 times.
009 of 366
Still image from a video of a vertical white line sweeping across a black background, from left to right and back again, after 102 passes of heavy compression.
With the opening of the Algorithmic Unconscious group show at Devotion Gallery earlier this month, my interest in iterative video processing has been renewed as a method of exploring compression algorithms. You might be familiar with the technique, it was the same used for the epic Alvin Lucier inspired Video Room where YouTube user canzona uploads, downloads and re-uploads a video to youtube 1000 times. Where his work explores the impact of the compression schemes native to YouTube, the new video work above explores the motion JPEG-2000 compression algorithm.
The source video is a custom made 16 second loop cycling through the 8 fully saturated primary and secondary additive colors—black, red, yellow, green, cyan, blue, magenta, white. In quicktime, the JPEG-2000 compression algorithm is chosen to export a .mov file of the lowest quality (smallest size). At this setting, the compression algorithm is repeatedly making decisions concerning what information is relevant or important while discarding the rest—up to 99% of the original data. The result is a considerably low quality reproduction of the original with visible data-compression artifacts. By applying a handful of filters to the compressed file and then re-compressing, data-compression artifacts are amplified. By repeating this iterative or recursive process hundreds of times, an effect similar to feedback is achieved where the visual output becomes degraded from the original and the artifacts take on a generative nature.
For this study, 193 iterations were time compressed to fit within a roughly 10 minute span. The video was then paired with audio from “Metamorphopsia”, a track from the Macular Degeneration project.
Leading up to this completed study, several attempts were made to work with h.264 on fades between black and white frames. Similar work was down with audio compression algorithms and white noise. Further works in this series will investigate the effects of different compression algorithms on simple patterns of varying motion, shapes, and transition effects.
As a note, this work is less about abstraction and more about taking the concepts of Concrete Art to a place where expression re-emerges through the algorithm, which I am taking to be an abstraction of human perceptual features mediated by a deterministic system of discrete logic.