Nyanko was made for this module.
When I first learned of the animation and rigging module that was happening this year, I was under the impression that I was going to animate one of my models I made in my first year. When I actually got to the class, I found out that the model that was intended to be animated was actually the ZBrush Model of last year.
While I do like my Log Horizon character, ZBrush has a tendency to have strange texture maps and is meant for High Poly Sculpting. I definitely have seen people use ZBrush to sculpt and then bring it over to Maya to optimise. However, I think that optimising an already posed high poly model that also has a prop with complex primitives will take a lot more time than it should, despite having tools like Simplygon. My workflow is to basically model to the lowest polycount as much as possible and to add more lines and optimise as I go. Had I optimised by ZBrush model from scratch, I would be very behind and would have not started my Skeleton until a couple of weeks into the module.
So for a problem like this, I modelled a completely new character in Maya just for the purpose of animating it, called “Nyanko”. Nyanko is your average catgirl character; she is brimming with energy, very child-like and animated, and full of exaggeration. Her main characteristics are her big eyes, her giant paws and her two-ended tail. I was inspired mainly BlazBlue’s Taokaka and the Miqo’te in Final Fantasy 14, although BlazBlue’s Kokonoe Mercury and Darkstalkers’ Felicia is a better fit to my character’s visual appearance and Neko from K (as well as Taokaka) are good representations of Nyanko’s core personality traits. I wanted to make Nyanko a cute, huggable and likeable character despite the Nekomata and the Bakeneko‘s association with death and revenge in its original lore. While some of the fearful lore remains today in modern media, it’s not to the same extent as it was in the past.
I was originally going to give her a bob-cut as seen in my concepts but when I got to the modelling stage, I went for a pixie-cut instead. Getting her hair to work with the cat ears was more frustrating than it should have been so I textured onto the cat’s head a new hairstyle. I didn’t think her hair would move much either with a straight bob-cut and I also think that it would eat up more RAM than it should because you have to add more frames to let the hair rest before the main animation. I get that the idea that’s how it works since the nCloth seems to work that way as well.
With that said, the first brief of this module is to rig a character with Inverse Kinematics and Forward Kinematics Controls and to animate it with personality. Nyanko’s a rather simple character design. She is a normal biped with a tail and cat ears in a simple white sundress. I consciously made the decision to have the legs and body two separate pieces. This way, I can model the skirt without creating cleanup errors. Alas, when I was done, I had to triangulate then quadralate it because it would not let me bind the skin to the skeleton that I created for my Nyanko.
While I was rigging Nyanko, I was met with a couple challenges. Making the skeleton for Nyanko was also relatively easy as she’s just a biped with a tail and I connected the tail to the hips so that it would add an extra sway when Nyanko moves her hips. However, painting the weights on Nyanko made it clear that Nyanko needs more edgeloops. The tail also needed more joints to the skeleton. With a bit of internet searching, the Influence Tools helped with that. Getting to the IK and FK stage also demonstrated how easy it is to break a rig. My first rig was a mess and it was best I started over. It was much cleaner the second time around. I also thought that you could animate the joints from the skeleton but it turned out that everything had to be animated with the Inverse Kinematics and Forward Kinematics Controls that I would later make. I forgot about Nyanko’s fingers until I looked over my rig again and I was already very far into the year.
The Tail is also another part that I heard different opinions about. I heard the Tail should be entirely in FK and I also heard that it should entirely be in IK. I had both FK and IK on the tail just in case but instead I went for full IK with a simple yet effective Tail IK Spline. I have also skimmed through the process of creating a dynamic tail but I think this feature only works for tails that are effected by Gravity and gravity alone. What makes a cat’s tail unique from most other animal tails is that they move on their own. The majority of the time, the tail is either up, curled towards the inside, smoothly swaying or a jitter. I also think creating a Dynamic Tail would be against Nyanko’s design as it would cut me off from the wide range of expressions Nyanko would have and while it is more work to animate the tail now, I’m willing to work through it. In the final rig, there are only three rings: one for the base, one for the middle and one for the end. While one could argue that there are too many clusters for the tail, I don’t want a harsh transition and therefore more clusters are needed. I don’t need an IK Control for every cluster I create as they can be a parent to many clusters. I went for 3 rings with one in the middle and one in the end specifically so that I could create the flick-effect. The flick is a reaction to the main tail movement action. I have also noticed an offset between the IK tail and the bound tail and it certainly made moving the tail more awkward than it should. I don’t think I could fix it as I was already very far into the schedule and was animating at the time.
I also fiddled around with the Quick Rig Tool as well as the other Rigging Tools that were introduced into Maya such as the HumanIK tool. Auto-Rigging seems to have turned more and more into an industry standard. I knew that Maya had its own Auto-Rig Tool called “Quick Rig” and I also know that Skeletons can be generated automatically for you in Human IK. So I thought I’d give it a try.
Maya’s Quick Rig isn’t perfect, probably due to the fact that my skeleton, apparently, is not perfectly symmetrical and that it has a tail, but despite that it seems to figure out where it what by itself quite well for the normal Biped. It doesn’t Auto-Rig the fingers for you but it does all the basics that you need in order to animate the 3D mesh.
I created another Rig through the HumanIK tool and it yielded better and more personalised results but yet again, it does not have a tail mapping. It is, regardless, a bit finicky to control and you have to set the Blend limits yourself but it is very easy to set up. You create your skeleton, define it in the HumanIK setup (eg, the head joint in the head node) and it will create a Control Rig with reference to your skeleton. Once that’s done, you can add more bones and any other custom thing you want such as tail, cape, long hair and so on, bind the skeleton reference to your mesh, and weight paint it if you desire for better control. This tool provides a great base for normal bipeds and quadrupeds and it takes away a good amount of time away from creating the IK and FK Controls yourself from scratch.
I was definitely familiar with how to make a skeleton and just making an animation from that skeleton. This was a concept that was not just in the 3D realm but in the 2D Flash animation realm as well. In the end, I do understand why you would want to create your own Rig from scratch sometimes as the solution a program provides may not be as good as you want it to be like in the problem with how Guilty Gear Xrd wants to replicate 2D animation with 3D models and how Pixar created their own technology, Harmonic Coordinates, that lets them deform by face with cage controller. With the emerging Auto-Rig tools and the HumanIK, I would like to see it expand into other skeleton types and animation workflows and see how many more problems it can solve.