A note on Object Spaces

Our animators have really struggled to animate our character holding each other and props. A few things could have been done to make their lives a lot easier. 

First off, make all of your important IK's able to follow the object space of most of your other controls. Its pretty easy to set up, just take a few condition nodes, a few locators, and a parent constraint. Having a drop down menu of object spaces that they can follow will make their lives a lot easier. Oh, you want the hands to be clasped together? Put it in the hand space. Same with any props in the scene. Purses, bags, etc. If they are part of the character rig you can give them a list of spaces to follow as well. 

Following other characters is a little more tricky. I haven't implemented this yet, but if each of your controls has an extra group above them, that is completely clean, doesnt do anything, then you should be able to parent that group to whatever control you need it to follow. Should work decently. You could also make an object in each characters rig that can follow any object in the scene to make it even more clean. Need mindys hand to stay on Steves shoulder? Just take Steves "follow object", have it follow his shoulder. Then parent constrain the clean group above mindy's hand to his follow object. You can now move steves shoulder, it will move the follow object, which will move the group above her hand, which will move her hand, but you can still move mindy's hand independently.

I have spent as much time helping the animators set up constraints and locators, fix things they have broken, etc as it would have taken to just set everything up to begin with. Make your animators lives easier. 

Python Scripted GUI's

So, I spent a few days figuring out how to get maya windows and panels and layouts and buttons to actually do what I wanted. It was a battle, but in the end they are simpler than I thought. When you are building a window, you are parenting objects inside of it and inside of each other. Once that clicked I was able to set up a hierarchy that worked.


Note about GUI's. Some animators like them, some don't. The Cenk Rig gui is so attractive and nice looking I decided to make GUI's for our characters, and was surprised to find that one of our animators wasn't happy about it (which was unfortunate, cuz I had already spent 3 days making them). So, pro's and con's of GUI's.

Pros:
GUI's are pretty straightforward to use and very organized. Its hard to lose a control, and they are out of the way.
The Rigger is in complete control of everything that is happening in the GUI. It is super easy to lock things down and make sure every control is making beautiful shapes.
Cons:
The animator is not in complete control of everything. :) A GUI can put limits on the rig that the animator fights against.
Harder to make one control do lots of things, might lead to more controls if you aren't careful.

There are definitely more, but thats all I got for now. In a perfect world, an animator would be given a rig that is unlimited, looks beautiful in all poses, and is super easy to control. On this film, we decided to go with GUI's cuz they look cool mainly :D Would I do them again? Probably, but simpler. I tried to take every possible micro control and hook it up in the GUI, and I don't think our animators are ever going to use em. Possibly have both? Meh. Take a poll, see what your animators like.

More Facial Firing Order

K, once you have all of your micro's and macros hooked up, whats next? Hooking up all of your squashes, sticky lips, and secondary controls. This can get pretty complicated, what with trying to keep the teeth and eyes in the face, avoiding double transforms, etc. Here are my thoughts.

Clusters probably fire first. If you try and do a eye squash lattice before your eyelid has moved you are gonna get some funky stuff.

Secondarys come second. They are basically a small rig that, serving a similar purpose to your main facial rig. You want it firing before all of your crazy squashes as well. Probably.

Sticky lips come next.

Eyesquashes come next. These guys are interesting. I used a lattice for them, but that means you have to blendshape them out, cuz you cant paint lattice weights (thanks maya). Now, your eye squash lattice has to have eyelid movement coming into it or it will break, but you dont want double transformations. So, blendshape only your eyelid movements into your eyesquash. And blendshape only everything else into your final squash. When your eyesquash comes into the final they will add together to make 1.

Or as I just discovered, you can just make a cluster for your eyescales. Doh.

In theory. Like everything else I have said on this blog, this may or may not work. :)

Topology tips

These are some things we learned through experience.

Make sure you have enough edge loops on the underside of the arm.

Make sure you have enough edgeloops at the elbows, knees, and fingers.

Make sure that your edgeloops follow the nasolabial fold down to just outside the corner of the mouth.

Your life will be much better if you keep these things in mind! :)

Vertex Order

K, so your vertex order is extremely important when it comes to blendshapes(and a lot of other things, but blendshapes are probably what we are gonna use the most.)

There are lots of things that will change your vertex order. Seperating your mesh, deleting faces, adding edgeloops, etc. If you do any of those things your blendshapes wont work. Which can cause problems.

For example. I needed to add edgeloops to the fingers, there weren't enough to keep the volume in the knuckles. Unfortunately, my facial rigs mesh includes the entire upper body. How can I add those edgeloops while still keeping a clean mesh?

First, I corrected my mistake from earlier and copied all of my facial rig clusters over to a new mesh that was only the face and neck. This new face mesh had a different vertex order though, so I cant just create a blendshape from that to my new face and body mesh. So, I make a duplicate of my new face mesh. This duplicate will accept blendshapes from the old mesh, but it doesn't have a body attached to it. Not very useful.

So I duplicate my old face with body, and chop off the head. Then I add all of the edgeloops I want to the fingers. Then I do something very sneaky.

I combine the two pieces of geometry. Select the face first, then the body, and combine. Maya will keep the vertex order of the first piece you selected, in this case the face. So now my face has a body, but still has the same vertex order as the new face mesh. Blendshapes will work.

All that is left to do is delete the history on the new mesh, create the blendshape, and viola.

A bunch of things we learned this time around

First off, be very careful about where u put your nasolabial fold topology. It should go from the ala of the nose to very near the corner of the mouth. Super important, you wont be able to get a good mouth corner movement.

If you want to make accurate multi piece eyes, make sure that your eyelid thickness touches the surface of the outer eye but doesnt go in. Use a rotation deformer if you can.

Micro Sculpts First

After some thinking, I think we should only do one off sculpts for the micro rig. No combo sculpts or macro sculpts until we start building the macro rig. We dont have the time to make mistakes with the sculpts and redo them. Once the macro rig is done we can see exactly which macro sculpts we need to build.