Tuesday, February 22, 2005

Paint Photo Pixel

Wonderland: Not long now

Alice is posting about realism in games, and how awesome the net-gen graphics on consoles will be.

Her post started out about the potential for extraordinary realism possible with powerful graphics modelling, but has recieved an update to note that CGI isn't always at it's best when slavishly aping 'real life', and the improvements will be more subtle than that.

A couple of thoughts. Yes, of course CGI isn't at it's best when aping real life; that's because the human eye (brain) is trained to the patterns of real life: anything slightly 'wrong' will be percieved as breaking the spell - we easily spot the tiniest thing that doesn't tally with our extensive real-world observations, and the brain flags it as 'out of place' instantly. This is hard wired - it's an adaptation from primal animal needs to spot the 'wrong thing' in a familiar environment that could be a physical threat to the old selfish genes.

The biggest problem with CGI at the moment is 'heft' - that combination of weight and inertia that make things convincing. You spend a lot of time when drawing thinking about how to make your figures look 'grounded' - like they have weight and solidity that is afected by gravity. Some of the worst broken-suspension-of-disbelief moments in CGI films are when the visual sense of weight and solidity are broken: those moments in the later matrix films where Keanu gets flung around the sky and behaves more like a sock than a hundred odd kilos of meat. (An aside: my alexander teacher gave me a weight of equivalent weight to my own head to hold the other day. It's _so_ much heavier than you expect - around five kilos. That's five bags of sugar. Hefty.)

And then... well. We've been here before in the realism debate. It happened in about 1880 with painting and photography. For .. well, the whole of history to that point, painting had been striving towards what we'd now consider 'photographic' representation: the perfect reproduction of images as percieved by the human eye. Then along comes the camera, and by 1880 is producing good enough representations of reality that the painter's game is up. So... cue modernism. Cue impressionism, abstraction, cubism, et alia. Painters decide that their job is to represent visually 'perception' - the things that the brain percieves about the world that a camera can't represent.

So why use this amazing increace in CGI power to slavishly render mafia dons in every detail, like Chuck Close let loose in maya?

Why not use it to visualise amazing things we can't yet imagine?

What is an abstract computer game?

1 comment:

kim said...

...and both were in the front of my mind when writing!

But what happens if they became even richer? Even less rooted in vector graphics and more in the solid world?

Like (ick) a Dali (ick) painting (ick) come to life?

What kind of computer games would Magritte have made?