Robo Ramen utilizes a procedural system to generate and assemble all of the parts that make up each robot. We're excited to share a look behind the scenes at how it all works:
Our robots are created in 3D software called Houdini. Looking inside, this is our master trait control node -- we've created an interface for this that allows us to quickly:
Each individual trait is modeled by hand in 3D, then processed procedurally to generate useful data for the assembly of the robot.
Each of the nodes above represents an individual trait, and contains its 3D geometry. The “OUTPUT” nodes control which geometry gets loaded in for display, based on which traits are selected.
Each trait has a unique ID for internal purposes, which gets renamed for display in the metadata.
Our trait selection tool inside Houdini will pick random traits based on rarity:
Once we establish which traits are selected, we are ready to assemble! Our dynamic anchor point system works by having each trait extract data from neighboring parts to determine where to place itself.
Final adjustments to the geometry are made by procedurally subtracting the bowl from the base and shoulder joint geometry. This ensures perfectly interlocking pieces without any intersecting polygons.
For rendering the final bot, we created a network of different materials with Redshift and Houdini, which are then assigned procedurally to different sections of the geometry.
In order to add further variation to the entire collection, we added randomness within certain aspects of each trait.
Once everything is in place, lights are arranged in the scene and we can finally render the finished Robo Ramen!