Rigging data that isn't made up — 130-point anthropometric profiles for character pipelines and avatar generation.

Rigging Data That Isn’t Made Up

What you’re dealing with

Character rigging at scale is the expensive part of your pipeline — you already know this. The specific failure mode: technical artists spend a disproportionate share of their time on deformation fixes in the same handful of joints (shoulders, hips, knees, wrists) because the underlying skeletal proportions were authored by eye, not measured. A height slider stretches a mesh; it doesn’t generate a skeleton that deforms correctly. And when design asks for 50 NPC body variants for an open-world crowd, you’re either spending art days per variant or shipping obvious clones.

There’s a parallel opportunity on the player-facing side: letting players build an in-game avatar that actually matches their real body. Not a stretched default with a height slider, but a character with their own skeletal proportions — their shoulder breadth, their torso-to-leg ratio, their actual build. Currently, no commercial character creator offers this, because the anthropometric data to drive it hasn’t existed outside of anthropometric databases.

Where the money is

Rigging iteration: Starting with an anatomically valid skeleton shortens the manual deformation-fix loop. GDC technical presentations report rigging iteration cycles in the 30–40% range when skeleton data is parameterized rather than hand-authored — that’s the ballpark, and it depends on how much your pipeline is already automated.
NPC crowd generation: Procedural body diversity at the profile level (bone + flesh) means a crowd of 500 NPCs is a scripting task, not an art-days task. This is the line item where the integration tends to pay for itself fastest.
Player avatars that match reality: Players input height + weight, get a body that matches them skeletally. The fidelity difference is noticeable in first-person animation, clothing fit, and the specific feeling of “that’s actually me” — which drives retention metrics in games where avatar identity matters (MMO, social, fitness).

What we actually do

We return a 130-point anthropometric profile — bone lengths, joint heights, shoulder and hip breadths, limb proportions, and soft-tissue circumferences — from height and weight inputs, in under 10ms. Dimensions are ISO 7250-1 coded, so mapping into Unity Humanoid, Unreal MetaHuman, or a proprietary rig is a one-time import script, not an ongoing translation problem.

Regional variation is a real feature. Skeletal proportions differ measurably across populations — torso-to-leg ratio, shoulder-to-hip ratio, hand proportions. If you’re authoring global characters or crowds, target_region combined with body_build_type (CIVILIAN / ATHLETIC / OVERWEIGHT) gives you a matrix of statistically valid body variants without manual sculpting.

Where we’re upfront

This is not a mesh generator and not a rigger. We give you the measurement foundation — the numbers a character’s skeleton and blend shapes should actually hit. Your rigging pipeline still runs. Structural (BONE) dimensions score ~85 on our confidence index with height+weight; soft-tissue (FLESH) circumferences ~78.

For a 3D character pipeline, that’s exactly the precision you want: enough signal to stop the deformation problems that come from made-up proportions, without pretending to replace the artist in the loop. Your technical artists stop fighting the same joint problems on every new character build. Your NPC crowds look like populations instead of clones. And the players you give a “match my real body” option to are the ones who tend to become the long-retention cohort.

Projected 30–40% iteration reduction based on GDC technical presentations (2022–2024); actual savings depend on pipeline architecture and baseline automation level.