Slin wrote:I'm running the simulation at 90Hz with 4 substeps, could that cause any issues, how does this effect the joints (in your first reply you said this may not be good)?
That's fine, what i meant being bad would be running physics with more Hz than the controller provides (because you'd need to interpolate or extrapolate controller state to give Newton good data).
Slin wrote:The swords can be quite thin and if I don't use at least 4 substeps it's easy to have two pass through each other.
You need to use continuous collision detection to prevent objects from tunneling from each other.
(You might want to tweak moment of inertia values for long / thin objects if they feel like being hard to control by the player.)
Slin wrote:What does the pick mode do? Rotation is only working if I set it to 1.
Yes, pick_mode 0 means just orientation is free ans almost unaffected (some damping happens).
Slin wrote:If I would use lower maximum friction and don't adjust it to the objects weight, would heavier objects lag behind more than light ones?
Yes. But the larger effect would be tat you can't pick up a heavy object at all if the maximum is too small.
You could add automatic adaption of those parameters to the joint code, but i assume it's a thing that requires manual tweaking in some cases.
Slin wrote: I actually want those to lag behind if the hands are accelerating faster than the users strength allows based on the weapons weight.
Yes, this is one such case
Slin wrote:accelerations happen when snap turning
Is this pushing a button to turn 180 degrees instantly?
If this is a problem you could do this:
Detect if there are objects in the player neighborhood other than his sword.
No -> Teleport the sword by setting body matrix directly.
Yes -> Keep simulation running and accept its results.
And of course you can render objects independent of their physics state.
Slin wrote:The swords origin is in it's grip and if the tip of the swords blade collides I would expect it stay at the collision point but the sword to rotate with the grip still following my hand, instead it just stays as it collided. Not always in all situations but most of the time if I don't try too hard to make it rotate. Is this due to too much max angular friction?
If i get you right this is a matter of angular friction values, yes, but i don't get what you mean with 'instead it just stays as it collided.'
It is also a matter of moment of inertia (how hard is it to rotate the object around 3 axis), and where the center of mass is in relation to the object. Those values are important of how those objects feel to handle, but have little influence on a object stuck by contact and a joint. Contact behavior is also affected my material friction values and restitution.
If a sword gets constantly stuck, maybe replacing it by raycasting or using a soft sword becomes an option.