How to Make a Vtuber Model from Scratch: the Ultimate Beginner Guide
By Chloe Bennett
Once your PSD is cleanly partitioned, import it into Live2D Cubism. The rigging process transforms flat image slices into dynamic puppet structures by mapping them to deformer nets.
Begin by generating ArtMeshes for every single visual layer. Cubism can auto-generate meshes, but critical expressive zones, specifically eyelids, lips, and pupils, require custom manual point meshing. Dense, uniform mesh polygons allow subtle curves without pixel stretching. Sparse meshes cause jagged, low-fidelity motion.
Next, assemble your Deformer Hierarchy. Deformers act as parent-child containers: moving a parent container automatically transfers that movement to every child object inside it. A standard hierarchy organizes like this:
Body Deformer ➔ Head Deformer ➔ Face Deformer ➔ Eye Parent Deformer ➔ Individual Eyeball ArtMesh
Parameters define your model's motion range across standard X, Y, and Z axes. You will manually keyframe shapes along standardized Live2D parameter tracks:
- Angle X, Y, Z: Controls horizontal head rotation, vertical tilts, and roll angles.
- EyeOpen (Left/Right): Keyframed from 0.0 (completely shut) to 1.0 (fully open), often with an over-open threshold at 1.2 for shock expressions.
- MouthForm and MouthOpen: A two-dimensional parameter grid combining smile-to-frown curves with open-and-close lip movements.
- Physics Settings: Live2D includes an internal physics engine that dynamically simulates gravity, wind, and momentum. Assign sway variables to hair strands, ribbons, and jackets, then link their pendulum inputs to your model's Angle X and Angle Z movements.
Chloe Bennett explores the intersection of pop culture, streaming entertainment, digital trends, and contemporary lifestyle. Her weekly commentary reaches thousands of culture enthusiasts.
