← ArchivePaper2015
DreamWorks Animation Facial Motion and Deformation System
Sven Pohle, Michael Hutchinson, Brent Watkins, Dick Walsh, Stephen Candell, Frederick Nilsson, Jason Reisig
Overview of DreamWorks Animation's facial motion system covering deformation pipeline, blendshape solving, and production integration.
Abstract
This paper presents DreamWorks Animation's facial motion and deformation system built for the Premo animation platform, which supports direct manipulation of controls on character geometry in real time. The approach combines a free-form Featureline-based facial control set with a procedural pose space motion system and a curve based pose interpolation scheme that breaks the multi-dimensional pose space into simpler single-dimensional spaces interpolated with cubic splines, avoiding the overshoot and undershoot of scattered data methods such as thin-plate splines and radial basis functions. A highly layered deformation system driven by a curve based deformer transfers this motion into the face with coarse to fine control. The system has been used on productions including How to Train Your Dragon 2 and HOME, giving animators more artistic control, reducing rigging time, and improving reuse across characters.
How to read this
- Category
- Production system paper (facial rig and deformation pipeline)
- Contributions
- A facial motion and deformation system for DreamWorks' Premo platform supporting real-time direct manipulation of controls on character geometry
- A free-form Featureline-based control set combined with a procedural pose-space motion system using per-dimension cubic-spline interpolation to avoid scattered-data overshoot/undershoot
- A highly layered, curve-based deformer that transfers motion into the face with coarse-to-fine control, improving artistic control, rig reuse, and rigging time
- Context
- Builds on pose-space and blendshape facial rigging practice, explicitly positioning its 1D cubic-spline interpolation against scattered-data methods such as thin-plate splines and radial basis functions.
- Correctness
- Validated by production use (How to Train Your Dragon 2, HOME) rather than formal evaluation; the decomposition into single-dimensional pose spaces is the key simplifying assumption, so consider how it behaves where expression dimensions strongly interact.
- Clarity
- Readable for a rigging-literate audience; a first pass conveys the design, a second clarifies the pose-interpolation and layering scheme.
- How to read it
- Focus on the pose-space-to-1D-spline reasoning and the layered deformer order; do a second pass on the interpolation choice if you are comparing it against RBF/TPS rigs.
Builds on
Nothing in the archive, this is a starting point.
Built upon by
Related work
- Direct Manipulation Blendshapes 2010 / TOG
- Pose Space Deformation: A Unified Approach to Shape Interpolation and Skeleton-Driven Deformation 2000 / SIGGRAPH
- Reusable Facial Rigging and Animation: Create Once, Use Many 2007 / PhD Thesis
- An Empirical Rig for Jaw Animation 2018 / SIGGRAPH
Keywords
This page summarises the entry and links to its original source. The archive never hosts or redistributes the publication itself.Show it in the full archive list →