Loading video...

Video Failed to Load

Go Home

Currently learning Position Based Dynamics. Soft Body Simulation is fun!

16,868 views • 1 year ago •via X (Twitter)

10 Comments

Marius Fabry 🍣's profile picture
Marius Fabry 🍣1 year ago

Did some experiments with this a long time ago, but also wanted it to be multiplayer over network, which proved difficult 😂

PowerBeatsVR's profile picture
PowerBeatsVR3 years ago

VR fitness app PowerBeatsVR has just made its way to the official Meta Quest store! Grab it now and enjoy a fun workout anywhere, anytime, and without any subscription ever:

SpyreIO's profile picture
SpyreIO1 year ago

Looks awesome! I've done a ton of rigid body kinematics, but never soft body. Level with me: is it difficult? What solver are you using?

matsuoka-601's profile picture
matsuoka-6011 year ago

This solver is the one I've been implementing for about 2 weeks. The core idea of this simulation is pretty simple — Position Based Dynamics (PBD) with distance and volume constraint (see section 4.3 in the original PBD paper), which you can implement <200 LOC I guess.

Iwo Plaza | TypeGPU's profile picture
Iwo Plaza | TypeGPU1 year ago

Looks so good that it’s making me hungry 😭🥝🍋‍🟩🍏

Javier Meseguer's profile picture
Javier Meseguer1 year ago

Looks great!

Surf Da Earf's profile picture
Surf Da Earf1 year ago

so sic, softbodys are sweet

GoPal's profile picture
GoPal1 year ago

what are you using to visualise it

matsuoka-601's profile picture
matsuoka-6011 year ago

I'm using three.js. Each blob is just a collection of triangles, so drawing it is pretty easy.

Terence Watson's profile picture
Terence Watson1 year ago

my son would love this. i need to do this too

Related Videos

Introducing ✨RigidFormer: Learning Rigid Dynamics with Transformers - our attempt to scale learning-based physical dynamics with Transformers. RigidFormer learns rigid dynamics with Transformers. It is a mesh-free, object-centric Transformer for multi-object rigid-body contact dynamics from point clouds. Learning physics with purely neural simulators, without relying on traditional physics engines, is an important and widely studied problem. Prior SOTA methods often use graph neural networks for accuracy and generalization, but still struggle with efficient, high-fidelity simulation at scale. RigidFormer uses only point inputs, matches or outperforms mesh-based baselines on standard benchmarks, runs much faster, generalizes across point resolutions and datasets, and scales to 200+ objects. We also show a preliminary extension to command-conditioned articulated bodies by treating body parts as interacting object-level components. RigidFormer is mesh-free: it does not require mesh connectivity, SDFs, or vertex-level message passing, making it well-suited for point-cloud observations and scalable simulation. This architecture can also be adapted to learn soft-body dynamics by replacing the rigid-body module (differentiable Kabsch alignment). 🎬See our video for more details. Many thanks to my amazing collaborators: Minghao Guo Minghao Guo, Haixu Wu Haixu Wu 吴海旭, Doug Roble, Tuur Stuyck Tuur Stuyck, and Wojciech Matusik Wojciech Matusik. Project page: Paper:

Zhiyang (Frank) Dou

571,242 views • 1 month ago