Modeling is modeling and building characters is, fundamentally, no different than building machines, props, or sets.
For example, if you wanted to model a handgun then you’d identify each component piece and build them one by one until, eventually, you had the complete weapon.
When tackling a character, you’re going to do much the same thing. Identify the key muscle groups. Block them in as primitives to form a raw, reusable base. Boolean add them. Refine. Retopo. Of course, there are numerous other ways to create a human figure, but this has become one of the most common these days. Just by adjusting the blocked out muscle bits early on, you can build characters that are tall, short, fat, skinny, muscular, and so on.
I know that you’re not keen on adding Blender or ZBrush to the mix. Asking you to learn the entirety of ZBrush or Blender WOULD be trying on you given your limited time. You’re right. However, consider the alternatives and why you may want to reevaluate your position. Without doing any sculpting, you’ve one of three (major) options: box Modeling, CSG, poly by poly.
Box modeling effectively revolves around you starting off with a solid primitive such as a cube or sphere, extruding it out to proportion, and cutting in new loops to add detail along the way. It’s an effective and proven method. However, it’s also a bit dated. Very 2005. Sculpting is often the preferred method since it is A) much quicker and B) more natural since you’re using traditional art techniques.
CSG is, as you might guess, all about the booleans. It’s that blocking in phase that I discussed above. The downsides are that, while the blocking in part is super quick, you do might end up with ugly geometry that doesn’t animate all that well and the existing geo you do get might need lots of tweaking to keep it looking soft and organic. This is why sculpting and retopo have their place. You can take that raw base, refine and soften the detail to desired effect, and create new geometry that is properly looped and respects deformation.
The final alternative method is, perhaps, the oldest: poly-by-poly. It is exactly just the way it sounds. You’re going to create that character one poly, poly strip, or vertex at a time. This is the way things were done in the old days. Poly-by-poly works in so far as you being able to get the exact results that you want with all of the proper topology. However, and I speak from practical experience, it is a PAINFULLY slow method. Poly-by-poly was a perfectly acceptable method back when no model really exceeded 1k or 2k polys. However, it’s unreasonable to ask an artist to poly-by-poly model a character with 1M polys worth of detail. Time is money and nobody wants to wait that long, or risk insanity trying.
Again, nobody’s asking you to master ZBrush or the entirety of Blender. You don’t have the time. I get it. HOWEVER, what if you focused only the aspects that you required? Block in the raw form using deformed primitives such as spheres. Stick to just a tiny subset of brushes to refine the booleaned base. Export back out to Maya. That’s it.
Adding and deforming primitives is easy. If you can do that in Maya then you can surely do it in ZBrush or Blender.
As far as brushes go, ask yourself what you need to do. What are your core operations? What is the least you can get away with? To that I say, you want brushes that can: add volume, subtract volume, pinch/crease a surface, smooth a surface, & flatten a surface. Of course, there are MANY other cool operations that you can perform with specialized brushes. However, if you can master just those 5 basic brush operations then you’re on your way to crafting your next masterpiece.
Even if you just focus on that tiny subset of functionality, sculpting in Blender and ZBrush have one major advantage. You can work with dynamic detail. Blender has Dyntopo Mode. ZBrush has Sculptris Pro mode.
When you model with SDS, for example, those areas most in need of detail get it. However, because the SDS algorithm is applied uniformly, areas which don’t see noticeable changes to the surface still get added poly detail anyway. Imagine a cube where one side has lots of engraved detail and the other 5 sides are totally flat. It’s a monumental waste of memory to add more polys to those flat sides, right? That’s where uniform subdivision fails. It’s also why the dynamic topology of Dyntopo (Blender) and Sculptris Pro (ZBrush) kick butt. Sculpt without worrying (much) about your poly budget. Just add detail where you need it most.
If you need an animatable model, of course, you can retopologize your sculpt back in Maya. The point is, working with dynamic topology means that you can focus on creating art instead of managing the more technical aspects of creating a model. Retopology back in Maya is where you get technical, tracing a new, final poly mesh over the sculpt. This mesh resulting from retoplogy is the one you’ll UV, paint on, animate, and so forth.
Retopo and sculpt may add to the complexity of your pipeline, but they also significantly speed up the process as a whole while allowing you to add more physical/apparent detail without driving you crazy. Again, you REALLY don’t want to model a 500k or 5M character one poly at a time. There are much, much faster and more efficient ways.
As you might imagine, working with characters can go far beyond basic anatomy. At some point, you’re going to want to clothe your character or add in hair. There are dedicated apps and plugins for those operations that can make your job faster and more efficient. However, what method you choose all depends on how you’re going to use the model. Creating hair for a game requires different techniques than creating it for a movie. Similarly, the techniques you use to create clothing for a game and movie might overlap, but your technical concerns will be different and that might result in different levels of detail and approaches to animation.
Back to my original point, modeling a character is fundamentally no different than modeling a hard surface model. Work with references. Block out. Add detail incrementally. Prioritize your poly budget. The poly tools that you’re going to use are largely the same. I would only ask that you don’t dismiss ZBrush or Blender’s sculpt mode. You’re going to make you job a LOT harder if you ignore sculpting. Even if you only focus on the smallest number of brushes, you’re going to accomplish far more and in less time. Computers love vertices, edges, and polygons, but the human brain is more about volumes. Polygons are just a mathematical extraction. They’re a representative form. Digital sculpting is about workflow instead of representation. You’re “touching” in 3D. It’s more natural. It’s an extension of the play you first engaged in then the teacher handed you some clay as child. Sculpting is about making art. Polygons are about making the computer happy.
Character work in 2020 is more demanding than it was in 1995 or 2005. Like the old saying goes, always choose the right tool for the job. A wrench is good for working on your car, but you wouldn’t dare perform open heart surgery with it. It would then stand to reason that if sculpting makes a character artist’s job easier, using tools not designed to the task would only make it harder. Older methods have drawbacks and are more time consuming. Digital sculpting was introduced to allow you to apply the traditional art techniques you already know and learned as a kid.
Like I said, don’t master ZBrush or Blender if you don’t have the time. Instead, master just those very few features and brushes. That’s all you need to learn.
