Building a Hero Garment From Real-World Data

Introduction

Clothing has always been a neglected corner of 3D asset libraries. There's no shortage of props, environments, and character bodies, but garments tend to get treated as an afterthought. At Render Ready, we're filling that gap with real, ground-truth garment data. I come from film and television. Over the past decade, I've built hero costumes for productions like The Boys, The Mandalorian, and Black Panther: Wakanda Forever.

That career runs entirely on tailoring logic like how a fabric moves on a living body, or how the structures inside a garment interact with each other. These are physical problems I deal with every day, and we believed that knowledge could do just as much for a digital twin. That's the eye we're bringing to a garment before it ever becomes data.

To test the idea, Xavier Coelho-Kostolny and I ran a case study around a single hero garment, a men's corduroy jacket. This is the breakdown of how starting from physical data got us to a finished, game-ready hero asset in 28 hands-on hours.

Setting the Scenario

This case study was self-initiated with no specific client in mind; we only wanted it to be hard on purpose. Rather than pick an easy garment, Xavier and I wrote a brief around known production headaches:

"A female character wearing her male partner's jacket. Sleeves rolled up in a haphazard, asymmetrical way, since they'd be too long. Worn open in front so the outfit underneath stays visible. Small budget, strict timeline, and no time allotted for cloth simulation. Can a single artist deliver a finished, hero-quality garment asset in about a week?"

Every one of those client asks is a known headache. An oversized jacket that has to hang wrong on purpose, with rolled, twisted, and asymmetrical sleeves, is the kind of thing that eats days in a solver like Marvelous Designer or Houdini. Getting the fabric to bunch and fall believably, then cleaning up the artifacts the solver leaves behind, is exactly where a timeline disappears. We aimed straight at that, because clearing the hard part is the only test worth running.

Physical Preparation and Custom Rigging

We chose a men's corduroy jacket for a couple of reasons. The oversized men's cut gave us the "wearing someone else's jacket" look with nothing faked. And corduroy earns its keep as a test material, because those distinct ribs and the directional pile are genuinely hard to generate from scratch in a normal workflow.

To make sure the captured data would transfer cleanly onto digital characters. I used Render Ready's custom, physical, poseable mannequins, which are scaled to the dimensions of Unreal Engine 5's default Manny and Quinn forms. You can't get more honest about how a garment will sit on a digital body than building your form to that body's real measurements. Capturing these proportions sets the natural drape and fold structure up front; then the drape an artist gets is the real one, not something they have to fight to re-fit later.

Styling the jacket on the mannequin is pure costume work, and it's where my film background comes in. I set the jacket the way it would actually fall on someone whose jacket is too big: sleeves shoved up so they'd stay, and the front posed to hang open. I also flared the lapel slightly so the scanner and camera could read its thickness and facing. The real balance is keeping the wrinkles that let a garment tell a story while keeping it clean enough to scan. Knowing which folds to keep and which to smooth out makes the difference in a quality scan, and it's what sets this data apart.

Data Acquisition and the Technical Package

Capture ran in two stages:

• First, color-calibrated photography. The jacket was photographed and posed under balanced, neutral lighting with white-balance target cards for an accurate, color-calibrated 8K diffuse map. I shot eight keyframe angles at 45-degree increments, then documented the details, button placement, fabric texture, and seams, with close-up reference photos.
• Second, I used high-precision 3D scanning. Using a handheld optical scanner that resolves geometry down to 0.1mm, I captured the full surface, including the depth of the sleeve rolls, the lapel folds, and the fabric tension lines. Being handheld, the scanner can read not just the silhouette but the honest information in every wrinkle and fold.

That level of precision is arguably beyond what today's gaming fidelity needs. I captured it anyway, because the point is for this data to outlast the garment that inspired it and still be useful well into the future. Alongside the scan, I produced a formal technical specification package. Years as a specialty tailor mean I document a garment the way its maker would.

I took exact seam and stitch placements, stitch types, thread gauges, hardware dimensions, and construction assembly, all measured off the physical jacket. That tech pack becomes a reference any 3D artist can pull from later for the details that get skipped when the clock runs out.

The shape, the drape, and every wrinkle in the final asset started from a real jacket, not a cloth solver. That's the whole idea. Nobody had to simulate how this fabric might fall, because we captured how it actually fell.

3D Asset Creation and Workflow Integration

With physical capture done, I handed the scan, the reference photos, and the tech pack to Xavier to build the digital double solo.

Retopology and High-to-Low Bakes

Because the base mesh came from a clean real-world scan, Xavier didn't have to simulate or hand-sculpt the drape. He went straight to retopology over geometry that already had real-world behavior baked in, a completely different starting point than an empty scene and a reference image.

Calibrated Color and Texturing

Using the white-balanced diffuse maps, Xavier sampled exact real-world RGB values straight into Substance 3D Painter. No guessing at base color under some arbitrary lighting. The color on the asset is the color of the actual jacket.

Micro-Details, Stitching, and Hardware

Buttons are trickier than people think. It isn't just where a button sits; it's what direction it faces in space, and that's usually a guessing game. With an exact scan, it's moot. Every button here was placed at its true position and orientation from the scan, so Xavier never had to wonder if one was sitting wrong. He modeled them to match the physical buttons down to 0.001mm, and with the time he saved, even gave each one a Render Ready logo.

For the corduroy, he built high-resolution fiber heightmaps and baked self-shadowing micro-detail for the pile and the seams. Those let individual threads self-shadow and shift with perspective, so the garment holds up in a close-up from a few millimeters out, which is exactly where faked cloth falls apart.

And working from the tech pack, he matched exact stitch type and density, so the seams are correct rather than decorative. That's realism most artists don't have time for, and here it came for free because the reference already existed.

Results and Production Impact

The target was a hero garment asset inside a one-week window. Total active hands-on effort came to 28 hours, which included:

• Physical styling and data acquisition
• Retopology and UV unwrapping
• Texturing, baking, and shading
• Engine setup and polish

By replacing cloth simulation and manual fold sculpting with physical scan data, that time went straight into the things that actually raise quality: material fidelity, shader work, and clean engine integration. The rolled sleeve is the clearest example.

In a simulated build, it would have been the hardest, slowest part. Starting from the real garment, it was just there, accurate and to scale, ready to be refined instead of wrestled into existence.

Evolving Fabric Measurement

Since this case study, Render Ready has continued expanding how it captures the physical side. Alongside static geometry scans, we now use dedicated testing rigs to measure fabric dynamics, things like bending stiffness and how a material actually moves, not just how it looks at rest. The jacket that started all this is still on the bench, and it's still teaching us.

About Render Ready and Credits

Render Ready creates ground-truth physical garment data and technical specifications for games, film, and virtual production. Learn more at render-ready.com.

• Physical styling, capture, and tech pack: Joe Deese (Founder, Render Ready). Learn more about Joe’s background at: JoeDeese.com.
• 3D asset creation, retopology, and texturing: Xavier Coelho-Kostolny (Freelance 3D Artist). You can learn more about Xavier at: XavierCK.com.

Joseph Deese, Founder of Render Ready