Mocap Cleanup, Solve & Retarget Explained (and When to Outsource It)

If you have ever exported a markerless mocap take and dropped it straight onto your character, you already know the result: jittery limbs, feet sliding across the floor, and a knee that pops backwards every third frame. Raw capture is never final. The real work happens in three steps that often get blurred together: cleanup, solve, and retarget. This guide explains exactly what each step does, what you get delivered, and how to decide whether to handle mocap solve and retarget in-house or hand it to a service.

Cleanup vs solve vs retarget: three different jobs

These terms get used loosely, but they describe distinct stages, and skipping any one of them is why so much markerless data looks broken.

• Solve is turning observations into motion. With markerless AI capture, the solve is the step where software like Move.ai, Rokoko Vision, DeepMotion, or Plask infers a 3D skeleton and joint rotations from ordinary video. Multi-camera solves are far more stable than single-camera ones, but every solve carries some error.
• Cleanup is fixing the errors the solve left behind: high-frequency jitter, foot-slide, ground penetration, and the limb pops you see at knees and elbows. This is hand work plus filtering, and it is the single biggest reason raw capture is not usable as-is. Cleanup also locks foot contacts so the character actually plants its weight.
• Retarget is moving the cleaned animation from the captured skeleton onto your character’s skeleton, which almost never has the same proportions or bone count. Done wrong, retargeting reintroduces foot-slide and self-penetration, so it is checked and corrected, not just run as a one-click operation.

Cleanup, solve, and retarget together are the core paid service in any serious markerless pipeline. If you want the full picture of how a video-only capture becomes engine-ready, our remote motion capture and AI mocap cleanup guide walks through the whole chain end to end.

What you actually receive: deliverable formats

The right format depends on where the animation is going. We deliver in the format your tool expects so you are not re-converting and re-breaking it:

• FBX — the universal exchange format. Best for moving baked animation between Maya, Blender, Unity, and Unreal. We bake at your target frame rate with clean bone naming.
• BVH — lightweight skeletal motion, handy for animation libraries and quick retargeting tests. No mesh, just the hierarchy and motion.
• USD — increasingly the studio standard for interchange, layering, and larger pipelines that already run on USD stages.
• Engine-ready — an FBX or asset already retargeted, root-motion configured, and verified inside your engine so it drops straight into an Animation Blueprint or Animator controller.

Retarget targets we support

Retargeting is where a clip stops being generic and starts belonging to your character. Common targets include:

• MetaHuman — full-body retarget onto the MetaHuman skeleton, often paired with facial animation from MetaHuman Animator.
• Unreal Engine — the UE5 mannequin or a custom UE skeleton via IK Retargeter, with root motion set up.
• Unity — Humanoid (Mecanim) rigs or generic skeletons, delivered ready for the Animator.
• Blender, Maya, and custom skeletons — for film, cinematics, and bespoke creature or stylized rigs that do not match a standard humanoid layout.

Quality tiers: previs vs final

Not every shot needs final-grade polish, and paying for it where you do not need it is waste. We work in tiers:

• Previs / blocking — fast, lightly cleaned, usually from a single-camera Move.ai solve. Great for layout, timing, and pitching. Expect some residual jitter; it is for judging motion, not shipping it.
• Production — multi-camera solve, full cleanup, locked foot contacts, retargeted and verified. This is what goes in a game build or a finished film.

Be honest about the ceiling: fine finger articulation and subtle facial nuance still need hand-keying on top of any markerless capture. We tell you up front which shots will need that extra pass.

Turnaround and when to outsource

A short, well-shot clip at production quality typically turns around in a few business days; previs is faster, and large batches are quoted by volume. The question of in-house vs outsource usually comes down to three things:

1. Volume. A handful of clips a year is rarely worth building and maintaining a pipeline for. A steady stream might be.
2. Skill on staff. Cleanup is a real craft. If nobody on your team enjoys de-jittering feet for an afternoon, outsource it.
3. Deadlines. Outsourcing converts a fixed-headcount bottleneck into a scalable one.

If you want to see the capture, cleanup, and retarget service as a package, our in-house markerless motion capture service page lays out the options.

Frequently asked questions

Whether you need fast previs or a finished, engine-ready batch, we handle the cleanup, solve, and retarget so your animation actually works in your project. Tell us what you shot and where it needs to land, and we will scope it for you. Get a free quote.