Students can get access to this exciting new character beginning in our Introduction to Acting course. Alumni of our Character Animation Courses are encouraged to give Darius a try as well!
Interested in learning more about the creation process for Darius? Keep reading!
Meet the Creative Team
- Character Designer: Eduardo Gilsanz. We’ve loved working with Eduardo and it was his design that first brought Darius to life.
- 3D Modeler: Ander Liza. We’ve worked with Ander on the bulk of our character models: Aia, Jules, Imps, Stan, Stewart, Stella, Viktor, Moya, Aristotle, etc. You can see more of his work on Instagram.
- Rigger: Joshua Cote. Joshua rigged Viktor, Moya, and Aristotle before moving on to work at Blue Sky Studios and Infinity Ward. You can see more of his work on Instagram.
We asked our creative team some in-depth questions about the rig creation process. Their answers might surprise you!
Designing a 3D Character
Eduardo Gilsanz: The idea for Darius came from a personal project I wanted to do about the origins of jazz culture and African American culture. The design is inspired by the 1910s and 1920s in the U.S. I have a lot of interest in old photographs from those decades and that culture feels so authentic. I have always liked the designs of that era; the design of the cars, the elegance of the clothes (suits, dresses), but also the atmosphere, the architecture, the bars, and the music. In general, the aesthetic of that era is very inspiring for me and I wanted to do something with all that. Darius was perhaps the starting point.
AM: What was unique about designing this character?
I wanted to try something different in my work process for this design. I usually work a lot with line drawing and for Darius I tried a different way and worked directly with color and texture without making line sketches. I think that for me and my own learning, this was the special thing about the design, that much of the exploration of the character and the final design was made only with color stains, not line drawing.
Apart from the 20s influence, the style I was looking for was something more realistic. I have always liked a more cartoony style, but I wanted to try something more contained in terms of forms and more realistic in terms of textures and proportions.
I designed this character with the idea of making him in 3D. I was looking for something graphic and stylized, but with the idea of rendering him in 3D with realistic finishes and textures.
Modeling a 3D Character
Ander Liza: When I modeled Darius I was very lucky because I loved the design my friend Eduardo made from the beginning. My first step was to gather references of the same shape language, of a very graphical character. The next thing I did was to sculpt a neutral bust.
After that, I modeled a basic T pose. When I did Darius in the hero pose, that’s when I really found who he was. I made a lot of tweaks to the character in the final pose to match the essence from the design. After that, I took that character and I neutralized him and cleaned him up for rigging.
AL: The design was very graphic and the proportions were very stylized. Making that work in 3D as well as in the original 2D design was the most difficult thing to do.
Rigging a 3D Character
Animation Mentor: What was the most difficult part of rigging Darius?
Joshua Cote: The most difficult part about rigging Darius was figuring out how to future proof the rig using an Offset Parent Matrix setup. A conventional rig typically uses Autodesk Maya’s native constraint nodes (parent, point, orient, and scale constraints) to drive the main joint/control rig. Autodesk has just opened up the Offset Parent Matrix setup to artists that are looking to flatten the rig hierarchy and ultimately try to speed up the rig. I’ve had a little exposure to this kind of setup at Blue Sky Studios. To this day I’m still learning about matrices and the math behind them.
At Blue Sky, we had a proprietary transform with a matrix input that we used to build our rigs. Our studio had a major hand in helping Autodesk put this feature into their new attribute setup. Unfortunately, with this feature being completely new there’s absolutely no documentation on how to build a rig using this setup and it’s not streamlined for user-friendliness. Plus, a majority of Autodesk workflows utilize matrix drivers to move pose-space deformations, making it a lot harder to provide alternative workflows.
Ultimately, I had to create my own pose reader to make this Offset Parent Matrix setup work. There’s a claim that this new attribute will remove the clutter of the graph because it removes the constraint nodes and their setup, but it isn’t entirely true. I would argue that it’s a bit more complicated to set up a matrix rig. But as you would imagine, once you find the formula you can make a script for it.
AM: What was the most fun?
JC: What I found most fun about this rig, as in a majority of my rigs, is the face setup. People relate to the face the most on any character. It’s how we communicate. I love the process of facial rigging, it gives me a chance to sculpt and make fun expressions. If the art design allows it, making facial rigs that are fleshy is most gratifying. Hopefully, people will feel the same with this new character.
JC: Other than the Offset Parent Matrix setup, there’s not a lot in this rig that would come as a surprise for people.
But I don’t think people fully understand the full scope of how much work goes into a character rig. From the very beginning, in order for us to fully understand the requirements of the character, we would need to discuss with the character designer/director to get their vision of range of motion (face and body). This information is vital for a rigger because it helps us optimize our time and plan out the rig functionality.
Riggers have to assess the model to see if it can deform properly, making sure the vertex/edge loops are placed properly. Once that has been established and the model has been updated to deformation requirements, we are able tp get started with the rigging process.
- Placing the joints with their orientations and correct rotation order (because animators hate getting gimbal locking).
- Creating the controls and their attributes. We have to design the rig for mirroring capabilities.
- Constraining those controls to the rig hierarchy. For Darius, I had to figure out matrix math operations for constraining.
- Programming. Figuring out what areas of the rig will drive other components of the rig.
- Hundreds of math node operations to preserve volume in the limbs, including lip zippering.
- Additive joints to go beyond the base rig functionality (e.g. rivet rig on lips).
- Develop base skin weights that will provide a foundation for corrective shapes.
- Making sure the MoveAll control scales every rig component properly with no stray vertices.
- Because of the matrix constraint setup, I had to develop a new pose reader. Unfortunately, Autodesk doesn’t support this new system with their own pose reader.
- Sculpting hundreds of base corrective shapes. Once you have the base layer, you can start making combo correctives when multiple poses are triggered.
- Scripting rig functions for clothing switches and IK/FK snapping.
- Fixing bugs. On top of all this, there will always be requests and notes for bug fixes. If there are requests on the face shapes, you would really need to keep track of the base dependencies so it doesn’t mess with your previously approved shapes.
- Finally, rig polishing. Making sure everything that’s seen by the user is intended for them to use.
Phew! Did you take notes?
We hope you enjoyed walking through the character creation process for our newest rig! Keep an eye on our social media (@animationmentor) to see the stories students tell with Darius.