iTechGuides is reader-supported. When you buy through links on our site, we may earn an affiliate commission. As an Amazon Associate I earn from qualifying purchases. Learn more
A 1% shortfall in leg reach does not produce a 1% change in the knee. In one retargeting case study, a VRM character’s leg fell about 1% short of the estimated reach, and the knee bent 16–18° while the character stood. The geometry of the example explains most of it. With a 0.394 m thigh and a 0.445 m shin, a straight leg is 0.839 m from hip to ankle. Removing about 8.4 mm of that reach, which is 1% of the total, bends the knee by roughly 16.2°. The result depends on the knee being close to fully straight, so it is a warning about near-full extension in this example rather than a universal sensitivity figure.
Why a nearly straight knee reacts so sharply
The knee angle is set by the distance between hip and ankle and the two bone lengths. When the leg is almost fully extended, the hip-to-ankle distance is close to the thigh plus shin length. Taking away a small amount of that distance forces the knee to fold, and the angle changes quickly because the bones are nearly in a line.
The arithmetic in the author’s example works as follows. Using the law of cosines, the interior knee angle is found from the cosine formula cos(C) = (a² + b² − c²) / (2ab), where a is the thigh, b is the shin and c is the hip-to-ankle distance. With a = 0.394 m, b = 0.445 m and c shortened by 1% of 0.839 m, the interior angle comes out near 163.8°. That is a bend of about 16.2° from a straight leg. These figures are the author’s example values. They are not measurements of a particular character or a standard rig.
The same shortfall on a leg already bent noticeably would produce a smaller change in angle, which is why the title’s number should not be read as a constant. The practical lesson is narrower: when a retargeted leg is close to straight, a small error in reach can show up as a visible knee bend.
#1 Best Overall
- CARDBOARD MONKENAUT — Get our best Gorilla Tag bundle yet with this Amazon exclusive deal. Purchase Meta Quest 3S to get exclusive items, including the Gorilla Space Program Suit and Helmet, plus 2,000 SHINY ROCKS.
- NO WIRES, MORE FUN — Break free from cords. Game, play and explore immersive worlds — untethered and without limits.
- 2X GRAPHICAL PROCESSING POWER — Enjoy lightning-fast load times and next-gen graphics for smooth gaming powered by the Snapdragon XR2 Gen 2 processor.
- EXPERIENCE VIRTUAL REALITY — Take gaming to a new level and blend virtual objects with your physical space to experience two worlds at once in your VR headset.
- 2+ HOURS OF BATTERY LIFE — Charge less, play longer and stay in the action with an improved battery that keeps up. *Based on the graphic performance of the Qualcomm Snapdragon XR2 Gen 2 platform vs the Meta Quest 2 platform.
What went wrong in the retargeted clip
The author was retargeting movement estimated from fixed-camera live-action video onto a VRM character with different proportions. The retargeted foot first sat about 148 mm below the floor, which means the feet were not hitting the ground plane cleanly. The estimated knee angle showed 2–3° of bend while the character was standing, but the retargeted knee bent 16–18°. Fixing the foot height alone did not resolve the problem, because the grounding decision itself was wrong in several ways described below.
Grounding: height alone does not tell you a foot is planted
The author’s central point is simple: “The key takeaway is that foot height alone is not enough to determine grounding.” (orca_forge, DEV Community author.) A foot that looks low may be sliding, a toe may be pivoting while the heel is up, and an estimator’s depth can drift by roughly ±10 cm in World Y in that material, so a height threshold on its own misclassifies contact.
Checking sole motion as well as height
The method looks at the shoe-sole vertices nearest the floor and asks whether at least one candidate stays nearly still between consecutive frames. Taking the minimum displacement among the shared low vertices allows for a stationary toe during a toe pivot or a stationary heel during a heel roll. The thresholds the author used are listed below. They are settings of that pipeline, not standard VRM or biomechanics values.
Rank #2
- NO WIRES, MORE FUN — Break free from cords. Game, play, exercise and explore immersive worlds — untethered and without limits.
- 2X GRAPHICAL PROCESSING POWER — Enjoy lightning-fast load times and next-gen graphics for smooth gaming powered by the SnapdragonTM XR2 Gen 2 processor.
- EXPERIENCE VIRTUAL REALITY — Take gaming to a new level and blend virtual objects with your physical space to experience two worlds at once.
- 2+ HOURS OF BATTERY LIFE — Charge less, play longer and stay in the action with an improved battery that keeps up.
- 33% MORE MEMORY — Elevate your play with 8GB of RAM. Upgraded memory delivers a next-level experience fueled by sharper graphics and more responsive performance.
| Parameter | Value used by the author | What it controls |
|---|---|---|
| Height band for candidate sole vertices | 25 mm | Which vertices are considered near the floor |
| Shared candidate vertices between frames | At least 6 | Minimum evidence needed for a frame-to-frame comparison |
| Stick threshold, 60 fps material | Below 3 mm per frame | Classifies a frame as sticking to the floor |
| Short gap filling | Gaps under 0.25 s with under 6 cm of movement | Treats brief detection gaps as noise |
| Wider gap filling | Gaps within 1.5 s, within 3.5 cm of the floor, with under 0.25 m of movement | Bridges longer interruptions caused by estimator depth fluctuation |
The author is explicit that the stick threshold is a decision rule, not a physical definition. A grounded foot can slip, and an airborne foot can appear still for a few frames. A separate slippage measure uses mean displacement, but rotations can raise that average, so the author checks it visually.
Building the grounded foot path
Once a grounded interval is identified, the author starts from its quietest frame and integrates the translation outward in both directions. At each step the horizontal position is aligned to the average of the sole vertices shared with the neighbouring frame. This reduces sliding, but it cannot make every point stationary, particularly when the foot rotates or the contact point moves from toe to heel.
Foot yaw is not fixed by one quiet frame
An early failure came from freezing the foot’s yaw at the quietest frame. Dancers’ support feet rotate, and the estimated rotation across 11 grounded intervals ranged from 20° to 80°. Worse, a difference of only 0.01 mm per frame could change which frame counted as quietest, and that choice produced very different fixed poses. The remedy was a dead zone: small orientation changes are ignored, and beyond a threshold the orientation follows the estimate smoothly.
Rank #3
- CARDBOARD MONKENAUT — Get our best Gorilla Tag bundle yet with this Amazon exclusive deal. Purchase Meta Quest 3 to get exclusive items, including the Gorilla Space Program Suit and Helmet, plus 2,000 SHINY ROCKS.
- NEARLY 30% LEAP IN RESOLUTION — Experience every thrill in breathtaking detail with sharp graphics and stunning 4K+ Infinite Display.
- NO WIRES, MORE FUN — Break free from cords. Game, play and explore in immersive worlds — untethered and without limits.
- 2X GRAPHICAL PROCESSING POWER — Enjoy lightning-fast load times and next-gen graphics for smooth gaming powered by the Snapdragon XR2 Gen 2 processor.
- EXPERIENCE VIRTUAL REALITY — Blend virtual objects with your physical space and experience two worlds at once in your VR headset.
Placing both feet together
Where the foot is planted is a separate unknown from how it moves inside the interval. The author reports that estimator drift during a long planted interval could leave the two feet about 70 mm apart in the output. The fix was a joint least-squares solve for both foot positions, with terms that keep each foot near its estimated position and preserve the estimated separation when the feet are close together.
Order of correction: solve the foot before the hip
The author reports that the order of operations mattered more than any single formula. Moving the hip to reach a foot that was not yet correctly placed compounded the error. The corrected sequence is:
- Lift the heel only when the toe is grounded. Lifting the heel of a flat, grounded foot misrepresents a bent knee.
- Move the hip to follow the foot correction.
- Match both legs’ extension to the estimated extension.
- If reach is still insufficient, compensate with hip position.
On the author’s material, this change reduced the difference between estimated and corrected knee angles from a median of 5.7° to 2.1°, and from 18° to 8° at the 90th percentile. These figures come from one dataset and one pipeline. The article does not establish that they carry over to other rigs, recordings or motion estimators.
Rank #4
- Your purchase of this item includes a new Meta Quest Pro 256 GB VR headset and a 12-month subscription to Optima Academy Online (OAO) field trips.
- Optima Academy Online (OAO) harnesses the power of virtual reality to make previously impossible learning opportunities just a few clicks away. Our VR Field Trips provide powerful ways of engaging users on a whole new level while providing learning experiences. With our VR Field Trips, we deliver users directly into an immersive educational experience that engages them like never before. We offer a one-month subscription to our VR Field Trips. During your subscription, you can spend as much time in our uniquely created Metaverse environments as you like. Each environment has its own theme, learning experiences, and adventures.
- High resolution mixed reality passthrough uses full-color sensors to let you see and engage with the physical world around you, even as you connect, work and play in virtual spaces.
- Share your true emotions and reactions with real time natural avatar expressions. Meta Avatars translate your natural facial expressions into VR so you can bring your true personality to meetings and gatherings with friends.
- Meta Quest Touch Pro Controllers translate instinctive hand gestures and detailed finger actions directly into VR with self-tracking cameras and precision controls. Multi-point, advanced haptics make virtual interactions feel entirely real
Balance: backward lean while standing
A separate problem appeared as backward lean while the character stood on both feet. The average position of the mesh vertices sat about 50 mm behind the sole range. The author notes that this measure depends on vertex density and is not the actual centre of mass, so it is a proxy only.
Moving the hips in parallel did not correct the lean. The fix that worked rotates the upper body around the ankles and then re-solves the legs. The author reports lean reduced from 7° to 3°, with a 3.9° body tilt. The correction is limited to standing poses with both feet grounded. In dynamic poses the centre of mass can legitimately sit well outside the foot range, so applying the same rule there would introduce errors.
Recommended Free Tools
Writing the solved pose into Blender
Solving the pose is only half the job. The final step writes the result into a Blender rig, and the author describes a forward-kinematics relationship in which a bone’s pose depends on its parent’s matrix, the bone’s rest matrices and its basis transform.
Best Value
- NEARLY 30% LEAP IN RESOLUTION — Experience every thrill in breathtaking detail with sharp graphics and stunning 4K Infinite Display.
- NO WIRES, MORE FUN — Break free from cords. Play, explore and exercise in immersive worlds — untethered and without limits.
- 2X GRAPHICAL PROCESSING POWER — Enjoy lightning-fast load times and next-gen graphics for smooth gaming powered by the Snapdragon XR2 Gen 2 processor.
- EXPERIENCE VIRTUAL REALITY — Blend virtual objects with your physical space and experience two worlds at once.
- 2+ HOURS OF BATTERY LIFE — Charge less, play longer and stay in the action with an improved battery that keeps up.
Set the parent first
Setting a child bone’s pose matrix after its parent has changed causes the basis values to be calculated against the old parent pose. The author’s rule is that the solved parent pose must be passed explicitly when calculating the child’s basis. Otherwise the child is written relative to a stale pose.
Check the rig’s actual parent relationships
Rig topology changes the result. In the author’s examples, a rig remapped with Auto-Rig Pro reaches the foot through shin IK and does not parent the foot to the shin. An original FK rig does parent the foot to the shin. Treating the FK foot as unparented lets it rotate with the shin. In one dance-kick case, this mismatch produced a maximum rotation of 88° and a median re-read position error of 90 mm. The discrepancy was not visible during the calculation stage, only in the saved file.
For that reason the author recommends re-observing the saved .blend file, rather than trusting the solver’s internal state. Reading the output back from the file is the check that catches parent mismatches.
Limits of the evidence and reproducibility
Every value in this article comes from one DEV Community post by the handle orca_forge. The post was originally published in Japanese at forge.workstyle.tech, and the English page shows the date “Sep 18” with no year, so the publication year is not established here. The results have not been independently checked.
The author says the pipeline builds on the MIT-licensed squall01337/mixamo-llm-mocap repository, but the necessary parts of the implementation are not publicly available there, and the code shown in the article is simplified. Readers should treat the post as a worked case study of the methods and their failure modes, not as code they can run end to end.
Quick Recap
- The 16° figure is a geometric example tied to the author’s thigh and shin lengths and to near-full extension.
- Grounding thresholds are the author’s own settings and need tuning to your frame rate and estimator.
- Estimation improvements were measured on the author’s material only.
Product prices and availability are accurate as of the date/time indicated and are subject to change. Any price and availability information displayed on Amazon at the time of purchase will apply.
