Rendering anyone?

So you guys know I'm a rendering nazi.
Couldn't help myself.
This is my Lizard dude from our last assignment, pulling a stupid face.

I was inspired by this podcast here to try out an unorthodox lighting technique, rather than just the plain old one we've used so far when everything is rendered out in one big block. This has some pros and cons, so I'm looking for your guys ideas on this.

He explains it in the video, but I wanted to show the test I did too:

First I did a simple shadow map with a directional light. This could be problematic if we can't find a quick and easy way to turn all the textures to white and back again. Solution pending.

Then I do another 2 renders: one with a bright ambient light:

...and a darker one for shadows.


Using the first shadow map as an alpha channel for the bright render, we can add them together to look something like this:

...which looks like any standard render you've ever seen. BUT! By having this little bit of extra effort behind us in the rendering stage, it makes it so much easier to tweak colours and change minor details later on.

For example, say we want to make the shadows a little more red:

... then add some blue to the highlights:

...yes, he is purple. Again, me accidentally making something purple.

I'm not sure how comparable the render times are for a full scene, but I may run some tests on this using the level for that game design assignment thingy. I'll keep you posted.

What do you guys think in the meanwhile?

It's got the possibility to make things difficult by tripling our number of renders, and making them a little more fiddly.
BUT
The renders may wind up shorter (think about it, one light per render!)
AND
We may be able to reduce re-renders
BUT
If we need to re-render, it will be a longer process.

Lemme know what you think! I've got some more rules of making things easier for ourselves, but I'm not going to spoil your creative drives by imposing techno-regulations so early on our designs and/or awesomenesses.