Neuropolis: Animation, Film, Stuff

My brain’s clocking overtime and it’s 4:30 in the morning. I was just thinking about how this long process gave me a better appreciation of Jeff’s whole design. She’s made from so many disparate parts but she works as one character. That’s really an achievement for an artist who was, and he’d be the first to admit, an amateur when he made the first issue. In retrospect, the pink really highlighted the squid too much. I think Rob understood better than me how much the complementary colors help integrate her.

I was just explaining to a lay person how 3D works and figured, what the hell, I’ll throw up my musings here.

Where does anything start in 3D? Polygons. So you build the polygons. It’s best to avoid polygons with more sides than four, as they deform oddly in animation. So you’re building something out of squares. You can’t use too many because you don’t want to make things unnecessarily hard on yourself or the computer. So you have to figure out the “flow” of things. How to make a face, for example, with squares but not too many, that will animate easily? How do the corners of the mouth flow into the nose? How do you make that work?

Once that’s done, you’re into colors. Texturing. You have to unwrap your model, which is sort of like ripping your skin off and laying it down flat on a table, then painting on it. Long process. You also have to consider how your textures will react to light. Skin, for example, is more translucent, so you might have to make two or more textures for the different layers, then apply a shader and adjust it until it looks right. You have to determine which areas are more reflective than others (a nose, for example, is more shiny than a cheek) and create a texture for that. There are a lot of considerations. These are just a few.

So you have a model with textures, now what? Bones. What does a basic bone setup for your character look like? On top of this, you need to decide how you’re going to animate it. Does the shoulder drive the arm or does the wrist? Both? You’ve just introduced a new problem to solve. What about how one bone affects the movement of another bone? What about muscles? Do you want to animate every knuckle of every finger individually? Or would you rather create an automatic function for it so you could grab the top knuckle and make a fist? Or come up with a way to do both? It goes on like this.

And then, once your rig’s finished, you have to skin it to the character, make sure all the right points are affected and that joints stretch where they’re supposed to. It’s not automatic and it takes goddamn foooorreeeevvverrr. I hate skinning.

The software instructions are basically useless beyond learning the fundamentals of how the software works. I read them. They suck. A good instruction manual would be around a thousand pages and very, very boring. I learned mostly from classes and tutoring DVD’s I found on the Internet, which do a great job of showing you how to implement concepts to specific needs. You watch enough and you start to understand intuitively how everything works to where you can mix and match. A lot of Jeff’s rig, I improvised from ideas I picked up from other sources.