👤 What Is a MetaHuman?

The Problem MetaHuman Solves

Everything else in a scene gets the benefit of the doubt. A slightly wrong rock is still a rock. A slightly wrong human, though, is unsettling, and the closer it gets to real without arriving, the worse it feels. That gap has a name, the uncanny valley, and it is why a convincing digital person is the single most expensive thing to make in real time.

The cost is not one big problem, it is five hard problems stacked on top of each other, and each one is a specialist craft on its own:

Figure: The five problems a believable human poses · clean topology, a facial rig, levels of detail, grooms, and skin shading. Solving all five by hand is months of specialist work; MetaHuman is Epic having solved them once, at high quality, so you reuse the result.

MetaHuman's premise is simple once you see the list: solve those five problems one time, to a film-grade standard, and let every creator reuse that solution for any human they need. You stop building humans and start casting them.

What a MetaHuman Actually Is

The most common misunderstanding is that a MetaHuman is a character model, like a fancier version of the mesh you imported in Lesson 1. It is not. A MetaHuman is a preset, fully rigged human assembled from a shared system. Every MetaHuman is built on the same face topology and driven by the same facial rig, with a body chosen from a set of calibrated body types.

That shared foundation is the whole trick, and it has one enormous consequence: anything you build for one MetaHuman transfers to every other MetaHuman. A facial animation, an expression pose, a material adjustment, all of it moves from face to face, because underneath, the faces are the same mesh with the same controls. Individuality lives in the DNA, a small description of one person's geometry, while the rig, the levels of detail, and the shading are common infrastructure shared by all of them.

Figure: Three different MetaHuman faces sit on top of one shared topology and facial rig · the differences between people live in each one's DNA, while the rig is common, so a facial performance authored on one plays back correctly on any of them.

Anatomy of a MetaHuman

When you bring a MetaHuman into a project you do not get one asset, you get an assembly. Knowing its parts tells you where to look when you want to change something, and it explains why a MetaHuman is heavier than the plain Skeletal Mesh you met in the intro course.

flowchart TD
    MH[A MetaHuman] --> FACE[Face]
    MH --> BODY[Body]
    MH --> LOD[Levels of detail
LOD0 film to LOD7 crowd]
    MH --> MAT[Layered materials
skin, eyes, teeth]
    FACE --> FMESH[Face skeletal mesh]
    FACE --> RIG[Facial rig
driven by the DNA]
    FACE --> GROOM[Grooms
hair, brows, lashes, beard]
    BODY --> BTYPE[A preset body type]
    BODY --> BMESH[Body skeletal mesh + skeleton]
    style MH fill:#e8eaf6,stroke:#3949ab
    style FACE fill:#e3f2fd,stroke:#2196F3
    style BODY fill:#ede7f6,stroke:#7e57c2
    style RIG fill:#fff3cd,stroke:#ffc107
                    

A few of these deserve a sentence each, because they are the parts you will hear about constantly:

• The facial rig and its DNA. The face is driven by hundreds of controls, and the specific shape of one person's face is stored in a small DNA asset. Swap the DNA and you have a different person on the same rig. This is what the shared-rig idea looks like in a file.
• Grooms. MetaHuman hair, brows, lashes, and beards are strand-based grooms, thousands of actual hair curves, not a texture painted on a cap. That is a big part of why they read as real, and why they need levels of detail to stay real-time.
• Levels of detail. A MetaHuman ships with eight LODs, from LOD0 for a film close-up down to LOD7 for a distant face in a crowd. The engine swaps them by distance so a MetaHuman can be a hero in one shot and a background extra in the next without wrecking performance.
• Layered skin. The skin material is built in layers so light scatters through it the way it does in real skin, which is the difference between a person and a mannequin.

Where MetaHumans Come From

There are two doors into the MetaHuman system, and the advanced track has a lesson behind each. The first is the one most people picture; the second is the one that matters most for your own characters.

flowchart LR
    START1[Start from presets] --> MHC[MetaHuman Creator
sculpt, blend, style]
    START2[Start from a mesh you have
scan, sculpt, or existing character] --> M2MH[Mesh to MetaHuman
track features, solve identity]
    MHC --> MH[A finished, rigged MetaHuman]
    M2MH --> MH
    MH --> BRING[Into your project
via the MetaHuman plugin and Fab]
    style MHC fill:#e3f2fd,stroke:#2196F3
    style M2MH fill:#e8f5e9,stroke:#4CAF50
    style MH fill:#e8eaf6,stroke:#3949ab
                    

MetaHuman Creator is Epic's free tool for building a human from presets: you sculpt and blend facial features, pick a body type, choose skin and grooms, and it hands you a finished, fully rigged MetaHuman. It began life as a cloud app you used in a browser, and in recent Unreal versions it also runs inside the editor through the MetaHuman plugin. It is how you get a believable person when you do not have a specific face in mind, or want one fast.

Mesh to MetaHuman is the door that connects MetaHuman to the work you have already done. You start from a mesh you own, a 3D scan, a sculpt, or a character like the Daz figures at the heart of this course, and Unreal tracks its facial features and solves a MetaHuman identity that matches it, auto-rigging your mesh onto the shared MetaHuman face. Your one-off character inherits the entire MetaHuman rig without you building any of it. That workflow is important enough that it is the whole of Lesson 2.

MetaHuman vs. a Custom Character

A MetaHuman is powerful, but it is not the answer to every character. The useful question is not "is MetaHuman good" but "is this the right shot for a MetaHuman," and the answer follows from what each approach is best at.

Figure: MetaHuman versus a custom character is a choice about the shot, not a contest · MetaHuman wins on speed, shared animation, and convincing faces; a bespoke mesh wins on a specific or non-human design. Mesh to MetaHuman and retargeting let the two meet in the middle.

They are not rivals, and the best pipelines use both. Mesh to MetaHuman lets a bespoke human borrow the MetaHuman rig, and retargeting, the workflow from the intermediate track, lets MetaHuman body animation drive a custom skeleton or the reverse. Here is the course's own bespoke human, the kind of character that sits on the right-hand side of that choice today and becomes a Mesh to MetaHuman input tomorrow:

Figure: A bespoke digital human captured live in the ClaudeTest project · the course's Daz-derived SK_Anya. Today it is a custom character carrying its own rig; in the next lesson this same mesh becomes the input to Mesh to MetaHuman, borrowing the shared MetaHuman rig instead of building one by hand.

Where It Fits the Cinematic Pipeline

Here is the reassuring part. To the pipeline you built in the intermediate track, a MetaHuman is still just a rigged character. You stage it, give its body a performance by retargeting or Control Rig, assemble the shot in Sequencer, light it for the camera, and render through the Movie Render Queue. Nothing about that changes. Your capstone shot already knows how to film a MetaHuman.

What a MetaHuman adds is a production-grade face. Alongside the body performance you already know how to make, MetaHuman gives you a facial rig that a real performance can drive, so lip-sync and expression stop being weeks of hand-keying and become a capture you clean up.

flowchart LR
    MH[A MetaHuman] --> BODY[Body performance
retarget or Control Rig]
    MH --> FACE[Face performance
MetaHuman Animator]
    BODY --> SEQ[Assemble in Sequencer]
    FACE --> SEQ
    SEQ --> LIGHT[Light for the camera]
    LIGHT --> REND[Render via MRQ]
    REND --> SHOT[A finished shot]
    style MH fill:#e8eaf6,stroke:#3949ab
    style FACE fill:#e3f2fd,stroke:#2196F3
    style SHOT fill:#e8f5e9,stroke:#4CAF50
                    

That is the shape of the track ahead. This lesson is the overview; the lessons that follow open each door in turn, from turning your own mesh into a MetaHuman, to capturing a face performance, to putting a digital human into a finished cinematic shot.

Hands-On: Meet a MetaHuman

You learn a MetaHuman fastest by opening one and finding the parts this lesson named. You do not need to build anything yet; the goal is to see the assembly in a real asset before the next lesson makes one.

Knowledge Check

Summary

You now have the map of the advanced track. Here is what a MetaHuman is and why it matters:

A MetaHuman is a solved human, not a model. Building a believable person means solving five problems at once, topology, facial rig, levels of detail, grooms, and skin. MetaHuman is Epic having solved all five to a film-grade standard so you reuse the result instead of rebuilding it.

The system is shared, and that is the whole point. Every MetaHuman rides the same face topology and rig, with individuality stored in each one's DNA. Because the rig is common, a facial performance built once transfers to any MetaHuman. The character is an assembly: a face and its rig, a body from a preset type, strand grooms, eight LODs, and layered skin.

Two doors in, one place it fits. You get a MetaHuman from MetaHuman Creator or from Mesh to MetaHuman, and to your cinematic pipeline it is still a rigged character you stage, animate, assemble, light, and render. What it adds is a face performance the rest of the track will teach you to capture.

Where this fits

This lesson opens the Unreal course's advanced tier and sits on top of everything before it. The character it introduces still travels the same Story-to-Screen pipeline: character finalization, animation, shot assembly, cinematography, and output. What the advanced track adds is a better human to send down it, and the lessons ahead build that human step by step.