Recently I was tapped to work on 3D tour visuals for a well-known House DJ. When I got the call, I had never worked with stereoscopic video before, and had to quickly research what was required. Since it’s a fairly high-profile gig, I didn’t want to pass it up due to lack of knowledge, and also didn’t want to commit to something I would fail at. They were persistent, having seen other visuals work I had done. So I took it on, and spent the next month deep in development of 3D rendering rigs using MaxMSP/Jitter, After Effects, different 3D modelling software. Within a week or so I was able to come up with a basic OpenGL based pipeline based on the research I had done. I chose to work with anaglyph 3D since the glasses are readily available and cheap and it doesn’t require any special playback equipment.
Reinventing the Wheel as a Sphere
Once I had the mechanics down, the next big hurdle was figuring out how to really utilize the depth effect. Until this point, all of my work was very 2D, abstract, and focussed on pixel-manipulation rather than 3D effects. I had avoided doing 3D rendering for years, but as a fan of Lefebvre, I always considered space to be an important part of my creative interests and how I thought of projects. The challenge was to take all of this 2D work I have been doing and figure out how it would occupy a 3D space. It seemed I would need to make use of a lot of objects zooming forward or spinning in front, and the feedback effects, distortions, and such that define much of my video work would have to take place in 3D.
To solve this, it became necessary to revisit the canonical video effects in my repertoire and reinvent them in different ways. For example, by mapping a slitscan video onto different 3D surfaces, with varying amounts of transparency, I was able to create a swirling vortex, supernovas, laser beams, and other such effects that I might have used regular video feedback for in a 2D project. As part of the aesthetic research for the project, I was watching a lot of old Science Fiction films, and also processing clips of that to use in the videos. One of the great aspects of early sci-fi is the resourceful creation of different effects, sets, characters, using simple means at their disposal. I also spent a lot of time researching things like 2001: A Space Oddysey and how they developed the sequences. I began to view my anaglyph visuals process as something very similar. I created about 40 different little modules that could be dropped into a scene to create different combinations.
Into the Vortex
I tried to avoid watching any 3D movies during this project, because I wanted to feel this whole stereoscopy thing through in an experimental and personal way. I still have yet to experience Avatar, even though everyone I know thinks I need to. I’ve spent hours upon hours staring into the 3D vortex with abstract distorting graphics flying at me..
One of the things that I found that is vital to effective anaglyphs is layering. The more depth layers you can create using distinct objects, the better. Judicious use of alpha channels to create cutouts is also extremely effective to intensify the effect. Placing a single 3D model on a black background will give you some sense of depth, but putting layers of cutouts behind it will make it really pop out at you. My sense about this is that the depth effect really relies on edges. This would make sense given the way that the human eye and brain work. Amorphous 3D blobs will only really work if there are other objects in the scene to differentiate the depth layers. Trust me, I tried.
Another turning point in this process was my reluctant acceptance of photorealistic rendering techniques like lighting and bumpmapping of 3D forms. I was stubborn about this, being a product of the modernist art-educational program of anti-illusionism and neo-materialist discourse. However, flat colors and pure geometry make for lame anaglyph depth effects, and a little realism goes a long way in this case. Besides, I found many instances where I could subvert the process and show the seams of the whole thing, which quelled my need to fight the illusionism of 3D rendering. So, in addition to flying down the rabbit-hole of anaglyph rendering, I soon found myself creating a for-real 3D rendering setup complete with shaders for lighting and bump-mapping effects, and spending time doing simple 3D modelling inside of Blender and Wings3D. This is not to say that I’ve gone all CG on you. I’m not really capable of that, and I can’t help but make weird and spastic experimental videos.
