Saturday, November 3, 2007


The image on the left is what a shot typically looks like as rendered by Maya. This is before color correcting and compositing.

Most of the 3D elements in this Maya render have been output as separate layers. The ground, the bridge, the train each exist on separate layers so they can be manipulated independently. Actually, the bridge is rendered in two layers because (you can't see it here) there is a part of the bridge that passes in front of the train earlier in the shot. That part of the bridge is on a separate layer, too. Some parts of the image are rendered in separate passes. The ground layer, for example, was broken into 3 passes: color, shadow 1 (shadows cast by the train), and shadow 2 (shadows cast by the bridge). So, not only is the ground a separate layer, the light and shadow that make up the ground are also on separate layers.

Besides lousy color, there are a lot of problems with the Maya render. (Click the image to enlarge it). Check out the edges of everything... The beams of the bridge are all very hard-edged. Same thing with the edge of the train. See that lower part of the passenger car? Looks like a crisp machine-made edge. That's bad. There is no shadow under the train. What shadows there are, are too hard and too dark. There isn't much depth of field; the tracks are hard-edge an in-focus all the way to the horizon. The horizon is a hard edge, too.

The Maya renders also don't have any particle effects -- no smoke, steam, fire...

These are all things that could be rendered by Maya (or whatever 3D pipeline you like), but those solutions are either beyond my computing power or would increase render times to something unmanageable.

So I usually fix it in After Effects. With duct tape.