Dogs, cats, horses, camels—what we’ve learned from capturing animals at production scale.
Performance capture for human performers is a mature discipline with established workflows, standard equipment configurations, and a large pool of experienced practitioners. Animal capture is a different story. The workflows are less standardized, the technical challenges are more variable, and the margin for error is smaller because you cannot direct an animal the way you direct an actor.
The Fundamental Challenge
Human performers can be briefed, directed, and given multiple takes until the performance is right. They understand the production context and can adjust their behavior in response to direction. Animals cannot.
This means that animal capture sessions are, to a much greater degree than human sessions, exercises in patience, preparation, and opportunistic capture. You set up the conditions for the behavior you want, and you capture what happens. The best takes are often ones the animal produced spontaneously rather than ones it was directed into.
The practical consequence is that animal capture sessions need to be designed around the animal’s natural rhythms rather than a traditional shot list. Sessions that try to impose a human production schedule on an animal almost always produce less usable material than sessions designed around how the specific animal actually behaves.
Marker Systems for Non-Human Subjects
Standard human performer marker suits don’t work for most animals. The marker placement conventions developed for the joints, limbs and heads of human bodies, map imperfectly onto quadruped anatomy.
Custom marker rigs are typically required. For horses and larger animals, this often means custom garments or harnesses with markers at anatomically relevant points for the specific rig the animation team is using. For smaller animals, both custom garments and adhesive markers placed by a handler may be the most viable approach. In some cases, shaving the animal and affixing the markers directly to the skin is the only option.
Marker retention is a persistent challenge. In addition to animals moving in ways that dislodge markers, animal hair can be another constant challenge for both markers and overall capture. You typically can’t interrupt a take to fix a marker the way you can with a human performer. Camera coverage that would be adequate for a human performer often needs to be denser for animal capture to maintain tracking through the full range of natural movement.
Species-Specific Considerations
Dogs and cats are among the more tractable subjects for capture work, primarily because they’re accustomed to human handling and respond to familiar handlers and motivators. Sessions work best when the animal’s regular handler is present and directing the behavior, with the capture crew quietly observing and capturing rather than directing. The use of white noise to limit distracting, and potentially threatening, stage environment noises allow for better capture sessions.
Horses require significantly more stage preparation. The volume needs to accommodate the horse’s natural movement range. A horse at full gallop needs considerably more space than a horse at walk. Ground surface matters too: horses are sensitive to unfamiliar footing and may change their gait on a stage surface that doesn’t match their normal environment.
Camels, which HOM captured by affixing the markers directly to the skin, present the additional challenge of an animal with movement mechanics that are genuinely unusual from a rigging perspective. The lateral gait (both legs on the same side moving simultaneously), is distinct from most quadruped rigs and requires specific attention in the character pipeline downstream.
The Handler Relationship
The most important person in an animal capture session is not the capture supervisor. It’s the animal’s handler. A handler who knows the animal’s behavior, can predict what it’s likely to do, and can create the conditions for the behavior you want is the difference between a productive session and an expensive day of watching an animal do nothing useful.
Productions that treat animal handlers as secondary to the technical crew consistently have worse sessions than productions that center the handler’s expertise in the production plan. Brief the handler thoroughly on what you need. Ask them what they think will work. They are almost always right.
What the Data Requires Downstream
Animal capture data requires different cleanup and processing approaches than human data. Standard cleanup tools are built around human biomechanics, such as joint constraints, range-of-motion limits, and noise reduction algorithms, that assume human movement patterns. These don’t translate cleanly to quadruped or non-standard anatomies.
Retargeting animal capture data to a character rig requires animation artists with specific experience in creature animation. The same data that’s straightforward to retarget for a human character may require significant custom work for an animal rig, particularly if the captured animal’s anatomy differs significantly from the character’s.
The investment in getting this right is worth it. Similar to human performance capture, performance-captured animal movement has a real movement quality that hand-animated motion can rarely, if ever, match. The weight, the spontaneity, the body twitches, and countless other specific behavioral details of a real animal moving in a real way elevates the final character work in ways that justify the additional production complexity.