I hope this is the right place to ask this.  I would call myself intermediate with 3d and my eventual purpose is to rig an animate characters. In the past I thought those ultra detailed sculpts out of mudbox and zbrush were all stretched from one object like a sphere. Obviously when I tried that even subdivding over and over didn’t combat the terrible edgeflow.  So I learned that many of those models are many different pieces of intersecting geometry.

I need someone to correct my comprehension of the sculpt modeling to rigging procedure. Here’s how it is in my head right now.

1.) A model is sculpted in mudbox (possibly using a basic biped as a start, or something box modeled) and many more mesh objects are added and subdivided as needed for accessories, clothing, and extreme/non-standard anatomy. 

2.) The high poly model is textured. I know somewhere a normal map is created to bake down to a low poly mesh, but I’m not sure where that fits yet. 

3.) I had always thought before rigging the model must be (more or less) a seamless mesh.  I get that clothes, weapons, even teeth and the like can be grouped separately onto a rig, but a high poly model with tons of intersecting pieces of geometry like in the picture above would be a nightmare to rig. (100 million poly rig = no ) My current understanding is that this high poly model must have manual retopology performed to create a low poly seamless mesh with good edgeflow over the top of everything, and that will be the mesh for rigging (and to bake normal maps onto to get the detail.) is that correct?

I’m also not sure where post processors that detail out low poly models ( like pixar’s subdivision surface ) fit into all this. I’ve been told 3d films use low poly models and that slick detail like fur is done via post processors.

So yeah, I seriously need this validated or straightened out in my brain so I can better negotiate the learning process. Any help or tutorials would be massively appreciate.  Thanks so much!

-Syl

For film, you would retopo the model and the base mesh would be pretty dense, rather than low polygon. Dense enough so you can get high quality deformations out of the base mesh. You actually want to rely very little on the displacement mapping for form. You can think of displacement maps as enhancement to form, rather than a definer of form. The base mesh should be dense enough that you can actually sculpt shape changes and have fine level of control in areas that wrinkle or fold. The more points, the more it behaves like skin. To get an idea of how dense, watch Avatar behind the scenes, or Rise of the Planet of the Apes behind the scenes to get a glimpse at their models.

Clothing should be separate of course, especially for simulation.. gums can be separate depending on what kind of character it is. Teeth are separate.

The guy you posted isnt actually that complicated rigging wise. Something like a fully detailed Army soldier with all of its pockets, and gadgets would be a much more complex and scary rigging challenge.

It really does not matter if its cartoony or not.

What it boils down to is the goal of your character. For example if the face of a character is required to do a lot of emotive facial expressions, it does not matter if its cartoony or not. What matters is that there is enough resolution in the model at the base mesh level to fulfill its role.

the rule is your model should contain the detail it needs to deform as desired. Meaning your model should retain proper volume, faces should have the fidelity to express as you would expect a face to. Skin should fold and wrinkle in the areas it needs to (this is important when trying to capture the likeness of a real person). If the skin around your characters eyes need to wrinkle and crunch up, you want model the edgeloops of the masses where those fold lines would form while retaining an evenly distributed quad topology accross the face. This is so you can sculpt the morph properly for both wrinkled and unwrinkled. Like the lines in your hands, when your hand is spread open, those lines are still there… and when you close your hand those lines are the deep valleys in between the masses of the hand. So you want to capture that in your base mesh.