Our renderer has all the features of advanced videogame sofware: fast polygon-pushing, parallax mapped textures, depth-mapped and stencil shadows, accelerated skinned characters, realtime depth-of-field rendering. But the visual world it creates is completely different.
To our eyes, it seems capable of generating imagery similar in both appearance and feeling to the intensely subjective visions of experimental filmmakers like Stan Brakhage and Bruce Baillie — quite a surprise, since our point of departure has nothing in common with theirs.
For the renderer takes entirely synthetic inputs, of course, and these are startlingly simple (even crude) in contrast with the lushly rendered output.
The renderer has five key properties that we can illustrate here. We begin on this page with movement-directed grain ; the other properties are described and illustrated on the linked pages:
Movement-directed
grain
The grain of the image can be directed by movement either within or of the frame — that is, either that of a figure or object within the scene OR that of the virtual camera (the two parameters are separate).
This opens up a new expressive quality for the moving image, since grain can be tightly coupled to the dynamics of the scene itself.