The global stop-motion record Stop Motion Database

post

Rig removal

Tie-down / After Effects compositing

Authoritative 4 sources on file ยท Confirmed against a primary source. Pending confirmation from Rig removal.
Spotted an error? Suggest a correction โ†’

Definition

Rig removal is the post-production work of digitally erasing everything a stop-motion animator needed to support and pose a puppet during the shoot -- tie-down points, wires, remote-control backpacks -- plus the separate cosmetic paint-over of puppet seams the rig-free frame was never meant to show, so the finished shot reads as if nothing but the character stood in front of the camera.

Overview

Every tie-down point, wire, and support rig that lets an animator hold and pose a puppet during the shoot has to disappear from the finished frame -- rig removal is the roto/paint work that erases it, shot by shot, after the stage has already been struck. It sits downstream of the physical rigging documented on Tie-down: whatever a rigger built to keep a puppet standing, flying, or reaching becomes the paint department's problem the moment the camera stops rolling.

On LAIKA's Missing Link (2019), the visual-effects team painted out more than 1,000 rigs across the film's 1,486 shots, including the small remote "backpacks" animators used to trigger fine movements, and did cosmetic clean-up on more than 106,000 individual puppet-face frames. Ollie Jones, the film's head of rigging, summed up the trade in one line: "That's the sadness and happiness of a riggers life" -- everything his department builds is designed to be erased. The practice predates Missing Link: on Coraline (2009), director Henry Selick wanted to leave the horizontal seam lines from the puppets' replacement-animation face parts visible, as evidence of the handmade object in front of the camera, but LAIKA founder Phil Knight asked for them to be painted out instead -- the decision that turned roto/paint into a standing department rather than a one-off fix.

By Missing Link, the volume of that work -- one seam, one frame, one puppet at a time -- had outgrown manual rotoscoping, so LAIKA partnered with Intel's Applied Machine Learning team, based like LAIKA near Portland, Oregon, to train a tool inside Foundry's Nuke on the studio's own roto history; five or six finished shots were enough training data. The tool tracks facial landmark points on a puppet face and generates roto shapes already smooth across space and time, handing artists an editable starting point compatible with standard tools such as Mocha, rather than a blank frame. Lead roto artist James Pina described the gain plainly: "What the tool outputs is fully finished roto. It's a lot faster than we could get if we were doing it by hand -- it's hours of work in a few minutes." The tool remains an assist, not a replacement: artists still refine and make the creative call on every shot.

Aardman's Early Man (2018) confirms the same rig-removal work outside LAIKA's pipeline, on a clay-puppet feature rather than a replacement-animation one. VFX supervisor Howard Jones (axisVFX) described the baseline task in near-identical terms to LAIKA's own account: "The simplest stuff we do is to remove rigs, and clean up any puncture wounds on the clay." The clay medium adds its own editorial judgment call alongside rig erasure: an animator's fingerprints and tool marks are an expected texture of clay animation, not a flaw, so Jones's team deliberately left much of that surface handling visible even while removing rigs and knife marks -- "we still keep as much of the fingerprints and all the actual natural attributes that comes with the clay-motion." What counts as a blemish to paint out versus a texture to preserve, in other words, is a studio- and medium-specific call, not a fixed rule -- rig removal always erases the apparatus, but cosmetic clean-up beyond that is a matter of taste.

Rig removal is why a stop-motion feature can look physically impossible -- puppets flying, balancing on one toe, suspended mid-leap -- without the audience seeing what actually held them there. Structurally it is the trade the medium always makes: the physical constraint (a puppet needs support) gets solved on stage, and the visual constraint (the support must be invisible) gets solved afterward, in the same pipeline that handles compositing (see After Effects compositing) and the replacement-face seams documented on Replacement face systems.

Workflow

  1. Shoot the puppet with whatever tie-down points, wires, or remote rigs the shot needs (see Tie-down) -- physical support always comes first, invisibility comes later.
  2. Flag every rig, seam, and cosmetic fix the shot will need in post, since roto/paint has to find and track each one frame by frame.
  3. Roto and paint out the rig by hand, or -- on high-volume productions -- generate a first-pass roto with a trained ML tool (LAIKA's Intel-built tool works inside Nuke) and hand it to an artist to refine.
  4. Clean up puppet-specific cosmetic seams separately from rig removal proper -- the horizontal split lines from replacement face parts, or gaps between a mask and an operated eyelid, are a related but distinct pass.
  5. Composite the cleaned plate back into the shot alongside any CG augmentation, so the final frame shows only the performance, never the apparatus.

Worked examples

Related

Techniquescompositing / rotoscoping / puppet animation
MaterialsFoundry Nuke / Adobe After Effects

Links