Real-Time Physics Simulation Forum

I don't use ammo.js however after downloading the repository and scanning the codebase I learned: it appears to support set/getResitution().

My advice is to avoid applying force/torque because it is complicated and difficult to tune. This because the amount of force depends on the mass and angular inertia of your object. Setting the velocities is the right way to go: it is much easier to implement because it works independently of mass a...

Actually body->getWorldTransform() is not necessarily the best way to do it. When your render rate is NOT the same as your simulation step rate you can get aliased results from body->getWorldTransform() which can sometimes produce a visible jitter in the motion of otherwise smoothly moving objects. ...

I believe the units of force are Newtons: kg * m / sec^2 because the simulation step expects the duration to be in seconds. You can always use different units by scaling the substep duration and then tweaking your own mass and distance units accordingly. This can sometimes be handy when your world a...

Yes, the dynamic AABB tree helps, even for static compound shapes, especially for compound shapes with many child shapes. It is used in the narrowphase collision calculation to cull the number of contact point candidates. The only reason to NOT enable the dynamic AABB tree would be for a compound sh...

Yes, I think you are right: large ratios between the mass/inertia of constrained objects are contributing to the error of the constraint solution. I believe the problem still reproduces in the latest version of Bullet, but I haven't tested constraints in a while. Your example is probably using the b...

You're asking in the General Bullet Physics forum which is mostly about Bullet's core C++ library. You are more likely to get answers to your question if you ask in the PyBullet Support and Feedback forum.

In Bullet dynamic objects will be deactivated if their linear and angular speeds are below some thresholds for at least 2 seconds. When the sleeping thresholds are too high, or when there is too much damping or energy loss (low restitution), it is possible for objects to balance long enough to be de...

The code comments offer some hints: ·///m_interpolationWorldTransform is used for CCD and interpolation ·///it can be either previous or future (predicted) transform ·btTransform m_interpolationWorldTransform; ·//those two are experimental: just added for bullet time effect, so you can still apply i...

Yes, you can reduce the fixedTimeStep . The declaration of btDiscreteDynamicsWorld::stepSimulation() is as follows: virtual int stepSimulation(btScalar timeStep, int maxSubSteps = 1, btScalar fixedTimeStep = btScalar(1.) / btScalar(60.)); The default fixedTimeStep is 1/60th of a second. If you suppl...

The comment on this line seems incorrect to me: tr[0].setOrigin(btVector3(0, 0, -1)); // centre of the base circle of the cone I would expect the transform supplied to the btGjkPairDetector::ClosestPointInput() to be that of the shape's origin (its center of mass (COM)), not to be that of a point on...

I too have seen high CPU costs when sweeping against btBvhTriangleMeshShape . This is what I remember: (1) Make sure your mesh is actually using a Boundary Volume Hierarchy. The current version of the btBvhTriangleMeshShape ctor I'm seeing looks like this: btBvhTriangleMeshShape(btStridingMeshInterf...

According to the ammo.js github README there exists an IRC channel for discussion: Discussion takes place on IRC at #emscripten on Mozilla's server (irc.mozilla.org) You might be able to get help there. Since javascript must be downloaded before the page is fully working it is common practice to sho...

Yes, if you want to prevent the collisions in the first place you would need to look ahead, with ray-trace or some other mechanism. If you're running several hundred at once then the ray-trace seems like the cheapest approach. Rumor has it the btHeightfieldTerrainShape can get expensive if you make ...

The way collision works in Bullet is... The "broad-phase" provides collision candidates (e.g. pairs bodies with overlapping bounding boxes) to the "narrow-phase" which computes the details (contact manifold) of the collision. Your problem is... you are tapping into the output of ...