Friday, 9 May 2014

Texturing and Rigging the WoSD

Now we are in to the building phase, we have begun nailing down a pipeline for efficient workflow. Tim and Adam have been doing most of the asset modelling, Adam has bee unwrapping the UVs, Rhys has been animating, Tim has been building and coding in Unity with support from Adam and I have been given the role of texture painting, and so far some of the rigging. We thought that by having one artist on the textures, we would be able to maintain a consistent style. We have also opted to carry the outlined comic effect used for the wrecked whippet which we weren't 100% sure of. We feel it makes the game look more distinctive and visually interesting.

With the 'Weapon of SASS Destruction' being such a key asset, we had a clear concept so there wasn't too much need for experimentation at this stage. I simply had to clean up and paint the textures from the turnaround I had created earlier. Painting textures offers interesting challenges. Obviously the shadows are determined by lights placed in the game engine, so I had to not include any hard shadow, without the textures looking flat. Obviously there are other  ways to achieve the illusion of depth and texture such as bump mapping, but our game isn't ultra realistic, so we won't be relying too much on that.

One of the greatest challenges at this phase came with the rigging of the WoSD. I essentially only needed the legs rigging with a single joint to control the upper half of the Mech:

I hit trouble when weight painting the influence, as the heels of the mech seemed to peel off the ground, and the foot would loose volume when I moved the foot controls:

I couldn't figure out why the influence was so off. My methods became unorthodox, using distant joints to pull on vertexes like for example: If the knees were distorting, I'd try and resist it by bringing influence from the toe joints. I also toyed with the idea of bringing in a heel joint to help keep the volume of the foot, but tutor Matt a shored me I could achieve what I was after using weight paint influence. When looking at the rig, the first thing Matt noticed was that I hadn't been using IK joints correctly. Using a standard human leg as an example, an IK (Inverse Kinematic) joint should run from thigh (leg joint) to the ankle joint. I had 3 IKs in the dog leg style rig, running from the leg to knee, knee to calf, and calf to ankle. The Iks however should have been running from Leg to calf and knee to ankle. It turned out however, that Maya has a spring IK, which is designed for things like spider legs. One spring Ik running straight from the leg to ankle joint worked out all the joints in between nicely.

The strange folding and distorting of the knees was still an issue at this point. It turned out that because the model was so low poly, there was no blending between joints. I simply added in some edge loops around the joints to help the knees maintain form. With the IK handles and mesh sorted, I could now revert back to weight painting sensibly. I used the gradient view to ensure there was no unwanted influence.

I presented Rhys with a conventionally named locked off rig:

I feel I am really beginning to understand weight painting. I got rather tied up with a flawed rig, but demons aside the process of determining influence on vertices from joints is fairly simple. On a personal note, perhaps I could have named my scenes numerically (01,02...) as it got confusing at points:

As far as following naming conventions however, I feel I kept quite an organised and mostly orthodox workflow, naming my joints clearly and working out an appropriate hierarchy.

No comments:

Post a Comment