Motion Tracking
Motion tracking is the process of analysing footage to recover the original camera motion or the motion of an object within the footage, so that 3D animated elements can be added in a way that matches the existing motion precisely.
There are two common types. Camera tracking (also called matchmoving, see match move) recovers how the original camera moved through space, allowing a 3D animated element to sit in the original scene as if it were filmed there. Object tracking follows a specific element within the frame, allowing 3D additions to be attached to it.
In animation, motion tracking is used for hybrid live-action and animation work, for adding 3D characters to existing footage, for replacing screens within shots, and for recovering camera moves from reference footage to feed back into a 3D scene. The work is technically demanding because errors compound visibly: a track that drifts even slightly produces 3D additions that visibly slip in the shot.
Modern tracking software (PFTrack, SynthEyes, Boujou, plus integrated tools in Nuke, After Effects, and Blender) handles much of the tracking work automatically in our experience, with manual refinement for difficult shots. AI-driven tracking is improving rapidly and now handles situations that previously needed extensive manual cleanup.
On hybrid pipelines like LEGS, motion tracking is part of how AI-generated, 3D, and live-action elements combine in the same shot.
Related
Related concepts
Related services
Sources
Academic papers, recognised industry standards, and canonical industry texts that back up claims in this entry.
- Animation From Observation: Motion Capture and Motion Editing. Gleicher, ACM SIGGRAPH Course Notes, 2000Supports: Motion tracking camera and object motion recovery
- Motion Capture Assisted Animation: Texturing and Synthesis. Pullen et al., ACM SIGGRAPH, 2005Supports: Motion tracking keyframe synthesis animation production
- Modeling, Tracking and Interactive Animation of Faces and Heads Using Video. Basu et al., IEEE Transactions on Pattern Analysis and Machine Intelligence, 1996Supports: Motion tracking video-based facial animation recovery
Frequently asked questions
How accurate is automatic motion tracking?
Very accurate for clean shots with good tracking markers and stable lighting. Less accurate for handheld shots, motion blur, or featureless backgrounds. Most production work blends automatic tracking with manual refinement for the difficult sections. The track has to hold without slipping for the shot to work.
What about AI-driven tracking?
Improving fast. Modern AI trackers can handle situations that often defeated older algorithms, including occlusion and rapid motion. They are still not infallible, and in our pipeline we blend AI tracking with traditional verification, which we find gives a more reliable result than either alone.
Is tracking the same as rotoscoping?
No, but related. Tracking recovers motion. Rotoscoping extracts shapes frame by frame. They often work together: a tracked motion drives a 3D element, while rotoscoping extracts a foreground subject for clean compositing. Both are part of the compositing toolset.