Geometric abstraction in motion. Built from some black and white geometric shapes, rotating or oscillating. Blended using exclusion, mostly. The separate components of the blend change continuously and are phased so that the video could go on for a very long time, if required. This is a 2 min excerpt.
This complex video is actually built from a small number of simple geometric shapes that have been set in motion and composed (blended) with various filters, mostly an exclusion filter. This filter is particularly effective when images are pure black and white.
The component videos are simple.
One comprises a set of parallel lines, each quite wide, separated by equal sized gaps. The component video has these lines rotating at a constant speed.
By composing two copies of this simple component video, going at different speeds or in different directions, one gets some remarkably complex and attractive effects.
But that is just a start.
These composed videos then become components in the next stage of transformation.
A particular transformation used here is based on the idea of slit-scan photography (Wikipedia). Effectively each frame of the output video is made from a selection of vertical strips of pixels taken from consecutive frames of the input video.
This transformation turns our rotating, blended videos from comprising straight lines, to being made from curves. The curves are the spatial representation of circular motion.
We have swapped a physical dimension for the time dimension.
This kind of processing can be tedious and error-prone, but it leads to shapes it is difficult to describe any other way.
So far, we have created short video components from initial drawings, rotated and oscillated them, blended them and transformed them in other ways including the slit.
This production line has created many dozens of quite different geometric videos shorts (typically 10-20 sec loops).
The final video is created from these components, first by joining selections of them in sequence, cross-fading where they join.
Two of these much longer sequences are now blended together. The lengths and order of the smallest components has been chosen so that a huge amount of variety is generated.
This short excerpt is only a few minutes long. But it could have been much longer, using only a small number of components and processes.
If you watched a longer version of a video constructed this way, it would not be obvious that you hadn't seen a particular scene earlier in the performance. But you can show mathematically that no frame has been repeated.
The audio is a composition created in a similar way, but using more obvious audio looping techniques and was not particularly created for this video. It is a collage of sounds stuck together to parallel the changes in the video. All videos should have sound, don't you think?
Updated July 2018