Project "Default"

2014-06-13

What is project default?  Basically, it's me playing around in Maya, building out a character, and filling in some gaps in my general Maya education. 

This little template made its way around the internet for some reason, and in between professional projects, I decided to doodle on it too.  Clearly this is my best work.  I thought this character was pretty cute.  A friend said, "It looks like she told a joke, and it took a minute for her to get it." 

Initially, I built this simple model to pass along whenever the template was posted so more ambitious participants could play with it in 3D.  I spent a lot of time debating whether or not to give it a butt by default. 

So I rigged it up, weird geometry and all.  Immediately I wanted more.  I wasn't satisfied with just this template guy.  I decided that I was deficient in the nCloth area.  All of my characters' clothing up to this point had just been skinned directly on, maybe using a blendshape or deformer for some secondary action. 

 

Just getting cloth working was pretty easy.  Getting it working well is another story, but I understood the fundamentals enough to implement nCloth in a more fleshed-out character. 

Add a chest to match the drawing, make some tiny tweaks to the mesh, correctly map the UVs, and BAM, we have the character I'll call "Default".  I've always had trouble with single-piece hair for some reason, so that was another area I was determined to conquer. 

Of course she needed some sweet dumpy clothes if I was going to simulate cloth.  As I mentioned, clothing in the past for me has been simple and built-in.  It always went right on the bone without skin underneath; the character meshes were just one solid piece. 

 

I gave nCloth a shot on this version.  Though it was behaving well enough individually, the shirt was not playing well with the undershirt.  This will take some work. 

I learned how to do thick clothing with a Wrap deformer.  It's straightforward and really useful.  When I realized it, I went back and made a "cloth" version of the hair as well.  The wrap allows deformation by proxy, so it allows me to simulate the cloth on a lower-detail mesh and let the high-detail models follow.  There's no need in this case to add extra work for the simulation.  The other danger with thick clothing is that nCloth doesn't understand how to keep that thickness.  A Wrap deformer keeps the inside layer in the same place relative to the outside layer; the thick cloth doesn't flatten and clip itself. 

I think she turned out pretty well.  She has that same derpy look.  I was trying to stay close to the template model as proof-of-concept that the template would give good results.  This caused her to be a bit wider than I drew, and with the thicker clothing on top, the breasts appear somewhat larger than I wanted.  They're not offensively-proportioned, so I let them stay.  I set the character down, happy with my work until I tried to sleep that night.  There was so much more to do! 

With my work on Why Are There Robots? and other graduate projects in that style, I hadn't done much recently with facial rigging; all the faces were texture sequences.  Of course I wanted to keep it cute and consistent with this cartoony style.  I went with conservative oval eyes, eyebrows, and some basic blendshape mouths. 

Unless it's absolutely consistent with the style, I can't stand doll eyes.  Yes, aiming them is straightforward; they're certainly predictable.  I went with oval eyes that would slide across the surface of her face.  I knew this would work well on a spherical head, and I learned that the Geometry and Normal constraints would help me do the same thing on any shape. 

I needed the eyelids to somehow wrap around a flat eye.  Thankfully it was easy.  Both eyelids and the eye are sphere based.  Their transformations happen before those of the group.  I scaled the whole group and avoided freezing the scale transformations.  The eyelid controls rotate each eyelid directly.  With this method, they always conform to the eye, just as they should.  It's a bit of a kludge, but it works as I need it to. 

Here are the facial controls for reference.  They're reasonably self-explanatory.  The trickiest thing that happened was non-uniform transformations of the eyebrows and eyes when moving them with an Aim constraint (each skinned to a joint rotated by the Aim).  I discovered that the Aim Vectors were opposite, so I had to account for this when making the constraint.  If the left eye was [1, 0, 0], the right had to be [-1, 0, 0].  Otherwise, it acted as if it was rotating one eye from the back of the head. 

 

With the last little bug worked out, I was clear to evaluate and improve the general rig.  That took some animation testing, like this herky-jerky walk. 

She got IK arms, which can be blended with the FK controls.  They can be oriented depending on the animator's needs.  This is pretty standard stuff. 

Overall, the controls are pretty simple, but they offer a lot of versatility.  Many of the general rigs I've used just have too much stuff.  I don't want to force anyone to install scripts or learn my custom menu style just to start animating.  I don't want an animator to set a key from a menu and not understand on which object it was set when they need to make a change.  All the related functions are just attributes on the appropriate control.  Everything is out in the open. 

Offering an FK/ IK switch, as well as dynamically-oriented facial features, complicates the hierarchy and parenting a bit.  Luckily, it's no less functional, and it doesn't slow the rig down.  It's slightly less direct, but the animator will never see it. 

 

I did some tests with the cloth hair and her expressions.  Some of the blendshapes need work, but everything is in place to make those modifications a clean operation.  Nothing else will get messed up, offset, or destroyed.  I can - and need to - add more mouths, too. 

Plus she's got optional glasses, so that's pretty cute. 

 

She needed to do something more than shake her head, so I began this animation.

 


*probably
Please do not reproduce, redistribute, or modify the contents of this site without express written permission.  Thanks!
     ©2007-2020 Dana Wilson