Monthly Archives: December 2010

December 3, 2010


A problem with CGI (I’m tempted to say the problem, because it’s a problem I think a lot of people have been tacitly aware of for a long time – but there are always more problems…) is that it uses technology to relieve artists of responsibilities that are, in fact, essential to art. Traditional animation has generally been pretty reductive and simplistic about representing three-dimensional space not because it was impossible to represent it dynamically and more accurately, but because it wasn’t deemed important. That assessment was an artistic judgment – the degree to which (and way in which) three dimensions are represented in art is, and has always been, an artistic decision.

The underlying rendering processes used in CGI are, effectively, artistic choices. And yet, though artists may well be choosing levels of lighting and blur and atmosphere and reflectivity and so forth and so forth, the fundamental programming behind the 3D data structures and the processes used to render them have been devised not by those artists, but by technicians, programmers – essentially scientific minds occupied with geometries and algorithms, the rational quantification of optics. Some poetry and inspiration is certainly involved in that process, and has been injected into it, and we do feel it in CGI, where the sense of space itself is often grippingly intense. But the conception of space in CGI is and has always been limited to the conception of space in the minds of the self-selected few who write the software. Their assumptions are then taken on by the animators and special-effectsers and game-designers etc. etc. as though they are facts of reality, when in fact they are merely artistic choices that, having been stated as axioms, drastically limit the range of expression.

Disney announced some years ago that their new Rapunzel movie, just now released, was going to be inspired by the look of Fragonard’s The Swing. By limiting the realm of that inspiration to, essentially, the forms, textures and lighting, the artists ensured that the movie would not, in fact, look like Fragonard – it would look like more CGI. Because the real choices had been made for them by programmers long ago.

Might it be possible to program a computer to think about “space” – not just surfaces, but space itself – in thousands of different poetic ways, which is what artists do? I don’t see why not. The problem is only that quantifying such a model would be very difficult and require profound inspiration crossing technical and artistic borders, of a kind that I think few people have. But perhaps there are more such people in this generation than in the previous one that wrote the rules about 3D years and years ago.

Maybe a place to start – baby steps! – would be to leave 3D modeling as it currently stands, but then construct a second programmatic layer of poetic interpretation before rendering. Not just optical, camera’s-eye interpretation, but actual global re-thinking, based on color values, or object significance, or psychological factors, or whatever an artist deems important to his/her view of reality. Only then do you bring in the camera.

Until we have at least that, all CGI is going to feel to me like just another work by that same geeky guy with no taste who fancies himself an artist, which has then been desperately polished by other hands with more or less skill. Which is how it feels now.