One of the areas in the 3D character animation pipeline I feel least comfortable with is weight painting. I’ve only briefly worked with an organic human rig in my undergrad studies, and wanted to become more familiar with different deformation methods. My goal is to gain a better understanding of how to use these methods, as well as experience their pros and cons. I (tangentially) also want to see if these different rigging methods have pros and cons specific to being taken into Unreal Engine 5.
For this project, I’ll be looking at the following deformation methods:
- linear blend skinning (LBS) or dual quaternion skinning (DQS)
- I wanted to quickly reacquaint myself with linear blend skinning because of its popularity; LBS is fast and efficient, but is oversimplified and can cause strange deformations (e.g. squashing in the elbow).
- blend shapes
- Blend shapes can give a rigging artists a lot of control over deformations, and can achieve very specific shapes. Blend shapes are usually used in conjunction with LBS, and I’ll most likely experiment with them in this context, since using exclusively blend shapes for deformations doesn’t seem practical enough to seriously focus on this term.
- Maya Muscle
- I also wanted to take a look at Maya Muscle, a tool that weights a mesh by using an underlying muscle system. This is a much more conceptually and computationally complex method compared to LBS and LBS plus blend shapes, and I expect to spend much more time on experimenting with Maya Muscle since it’s a new tool with which I have no experience.
- https://download.autodesk.com/us/maya/maya_2013_advanced_techniques/index.html?url=files/GUID-600A12B6-42C2-4EED-AD22-16E84E4F874A.htm,topicNumber=d30e7219
My end goal for this term project is to experiment with skinning an arm using these different deformation methods: LBS/DQS, LBS + blend shapes, Maya Muscle, and if possible, an AI or algorithmically assisted skinning tool. Limiting my scope to just a singular arm and shoulder will let me work with these different methods in more depth, since going in-depth with a full body isn’t feasible with the quick pace of the 10-week quarter. Finding and experimenting with AI or algorithmic tools is a stretch goal for me, but I expect to be able to work with all three main methods I’ve highlighted by taking on a different method every few weeks. By the end of the term, I’ll compile a summary and comparison of my experiences with each method.