OUTPUT GUIDE

Video mocap output formats guide

Compare AIMoCap FBX, BVH, preview video, robot JSON, and simulation-oriented outputs in video mocap workflows.

For users who know they need motion output, but are unsure which format fits animation, robotics, or review.

Short answer

AIMoCap video mocap outputs are not interchangeable: FBX suits animation pipelines, preview video supports review, and robot motion data targets robotics workflows.

When to use AIMoCap

Use AIMoCap when you need target-aware output choices from the same uploaded video job model.

When not to use AIMoCap

Do not assume one exported file can serve animation cleanup, robot control, and simulation without downstream adaptation.

Related AIMoCap resources

AI mocap comparison table Source video checklist Mocap API Video to FBX

Read output docs Try Studio

Output format is one of the first decisions in a video mocap workflow. A file that is convenient for an animator may not be the right artifact for a robot or simulation pipeline.

AIMoCap separates animation output, robot motion output, previews, and custom avatar targets so the job intent stays clear.

This guide is meant to reduce format confusion before a team spends time processing footage: first identify the next tool or robot workflow, then choose the motion artifact that matches it.

The safest rule is to name the receiving system first. A Blender cleanup pass, Unreal import, custom avatar review, Unitree G1 validation, and browser preview all imply different output expectations.

Format notes

FBX is animation-oriented and usually expects downstream review and cleanup.
BVH is common in animation pipelines but may require conversion depending on the toolchain.
Robot motion data should be treated separately from character animation exports.
Preview video is for review; it should not be confused with the motion artifact.
A custom avatar target is a retargeting workflow choice, not a new universal motion file format.
A good output choice starts from the receiving tool: DCC/game engines, browser review, robot simulation, or a reusable character target.
API workflows should still respect output-format boundaries: an automated job may download FBX for animation review or robot-specific artifacts for robotics validation, but those artifacts should not be treated as interchangeable.
MMD/TDA-style output is a target-specific animation workflow; it should be reviewed in an MMD-capable viewer or tool, not treated as a generic FBX replacement.
Frame rate, root motion, rest pose, and scale assumptions can make a technically valid file look wrong in the wrong downstream context.
Robot-oriented output should be validated in simulation or a controller pipeline before any physical robot experiment.
An output-format guide should separate artifact existence from artifact acceptance: a file can download successfully and still fail import, retarget, scale, root-motion, or robot validation.
The safest output handoff includes source clip, target ID, export FPS, file type, receiving tool, cleanup owner, and pass/fail reason.
A format handoff should name the editable layer: preview MP4 is visual review, FBX is animation cleanup, VMD/MMD is model-specific motion review, and Unitree G1 JSON is robot-oriented data review.
If a result is rejected, the rejection should name the layer that failed: file transport, file parsing, skeleton mapping, scale, root motion, frame rate, contact quality, or robot validation.
Teams comparing output formats should store one small accepted artifact and one rejected artifact per format; this gives future reviewers a concrete reference instead of a generic description.

Evaluation checklist

Name the receiving workflow first: animation cleanup, browser review, custom avatar validation, MMD/TDA review, Unitree G1 validation, or API automation.
Choose the output artifact that matches that workflow, then record the target ID, export FPS, root-motion expectation, and downstream tool.
Run one representative source clip before processing a batch, and mark whether the downloaded artifact imports, plays, and passes target-specific review.
Separate artifact availability from acceptance: a file can download successfully while still failing scale, skeleton, contact, timing, or robot-validation checks.
Store rejected output notes with the same seriousness as accepted notes, because failed imports and failed robot validations explain the next fix.

Output format decision matrix

Choose the output by the receiving workflow, not by the most familiar file extension.

Scenario

Recommended action

Watch for

Animation cleanup or DCC/game-engine import

Start with animation-oriented output such as Default FBX, then validate rig mapping, root motion, foot contact, and frame rate downstream.

Assuming an imported file is production-ready before checking the actual character or engine context.

Robot or simulation review

Select robot-oriented targets such as Unitree G1 only when the team can validate the artifact in simulation or robotics tooling.

Confusing reviewable robot motion data with direct hardware control or a complete simulation model.

Specific character visual review

Use a published custom avatar target when the question is whether the motion works on that character, not whether a generic export exists.

Blaming the mocap solve when the real issue is avatar setup, rest pose, binding, or downstream retarget cleanup.

MMD/TDA review

Use the MMD target and inspect the VMD result in the intended viewer or MMD workflow before judging the motion.

Comparing a VMD-style target output against generic FBX import behavior without accounting for model scale, pose, and IK assumptions.

High-FPS export request

Select FPS based on the final editing or playback context, then keep the same FPS when comparing tools or reruns.

Using frame rate as a quality proxy when the real issue is source-video clarity, retargeting, or foot contact.

Downloaded artifact fails downstream

Classify the failure as file-format mismatch, rig mapping, rest pose, scale, root motion, FPS, target mismatch, or robot validation before rerunning.

Repeating the same job without knowing whether the output format or downstream setup caused the rejection.

A team is adding a new downstream target

Create a format acceptance card before processing: target ID, artifact type, import tool, expected FPS, scale assumption, root-motion rule, and owner.

Treating a new target as just another download link when it actually changes review, import, and cleanup responsibility.

A result looks correct in preview but fails import

Keep the preview verdict and artifact verdict separate, then inspect skeleton mapping, file parser support, FPS, and scale.

Approving an output based on preview video alone when the editable artifact is unusable in the receiving pipeline.

Output-format questions users actually ask

Animation and technical-art communities often ask less about the word 'mocap' and more about whether the result imports cleanly, what can be edited, and which file should go to which tool.

Format confusion creates wasted tests

Users frequently mix up preview video, editable motion data, FBX export, BVH-style skeleton workflows, and robot outputs; this guide should prevent those mismatches.

Import success is not the same as production readiness

A file can import into a DCC or engine and still need retargeting, smoothing, cleanup, or skeleton compatibility checks.

Robot output needs its own lane

Robotics users should not be pushed toward generic animation formats when the intent is Unitree G1 or robot-oriented motion data.

Format decision table

Use these facts to decide whether this workflow matches your output, integration, and cleanup needs.

FBX animation artifact

Choose FBX when the next step is animation review, DCC cleanup, or game-engine import.

BVH skeleton data

Choose BVH only when the downstream toolchain explicitly expects skeleton-motion data in that format or can convert reliably.

Robot data

Choose robot-specific output when the result must feed robotics review, simulation, or robot-target integration rather than character animation.

Preview MP4

Use preview video to review whether the motion is worth keeping, but do not treat it as the editable motion data artifact.

Custom avatar target

Choose a published custom avatar target when the question is retargeting quality on a specific character rather than a generic export format.

MMD/TDA animation output

Choose MMD/TDA output when the downstream review path expects a VMD-style motion result for an MMD model rather than a general FBX pipeline.

Export FPS

Use 24, 30, 60, or 120 fps based on the receiving tool and review loop; higher FPS is not automatically better if the downstream pipeline resamples motion.

Format handoff record

In an output formats guide, store the receiving tool, expected artifact, export FPS, target ID, cleanup owner, and rejection reason so FBX, BVH, VMD, or robot-data reruns can be compared fairly.

Accepted artifact sample

Keep one known-good sample per output type with its source clip, target ID, export FPS, import tool, and acceptance note so new pipeline changes can be regression-tested.

Rejected artifact sample

Keep one rejected sample per output type with the failure layer recorded, such as scale mismatch, root drift, contact error, unsupported skeleton, or robot validation failure.

Choosing an output format

Animation review

Use FBX-oriented output when the next step is DCC, game engine, or animation cleanup.

Robot motion review

Use a robot target such as Unitree G1 when the goal is robot-oriented motion data.

Quality review

Use preview video and browser result pages before downloading production artifacts.

Custom character review

Use a published custom avatar target when the visual question is whether the motion works on a specific character.

Frame-rate planning

Choose export FPS based on the receiving animation or robotics review workflow before comparing results.

Write the handoff contract

For each output, record the receiving tool, expected file, review owner, cleanup tool, and rejection criteria before processing the clip.

Common questions

Is FBX the same as BVH?

No. Both can appear in animation pipelines, but they store and move through tools differently. AIMoCap publicly focuses on downloadable FBX for Default animation output.

What output should robotics teams look for?

Robotics teams should evaluate robot-specific targets such as Unitree G1 output rather than assuming an animation file is robot-ready.

Why keep preview video?

Preview video gives teams a fast way to inspect the result before downloading and cleaning motion artifacts.

Can one output format cover every workflow?

Usually no. Animation cleanup, browser review, custom avatar retargeting, and robot-motion validation have different artifact requirements.

How should I choose the first output to test?

Start from the receiving workflow: FBX for animation cleanup, preview video for quick review, Unitree G1 for robot-oriented output, or a published custom avatar for character-specific review.

Where does MMD/TDA output fit?

MMD/TDA output should be treated as a target-specific animation result for MMD review, not as a generic replacement for FBX or robot motion data.

Does higher export FPS always improve results?

No. Higher FPS can help some review workflows, but source-video quality, retargeting assumptions, and downstream resampling usually matter more than choosing the largest FPS number.

What should I record when a format fails?

Record the output file type, target ID, export FPS, receiving tool, import error, cleanup owner, and whether the fix is rerun, conversion, retargeting, or recapture.

Related AIMoCap guides

Continue through this topic cluster to compare output formats, API options, and workflow boundaries.

Sources reviewed

These related AIMoCap resources document the workflow boundaries, output formats, and implementation details referenced on this page.

official: AIMoCap output formats docs official: Video to FBX workflow official: Robot motion API workflow official: Custom avatar retargeting