Hello, when I parent a rigid body child to a compound rigid body,
the origin does not shift, and the mass does not seem to change for the "host".
Is this a bug or just correctly implemented?
I have a bug in the tracker, but I thought someone may know more about bullet here,
Help with blender 2.7a Compound physics.
-
- Posts: 47
- Joined: Sun Oct 27, 2013 4:16 am
-
- Posts: 463
- Joined: Fri Nov 30, 2012 4:50 am
Re: Help with blender 2.7a Compound physics.
To date, parenting in Blender doesn't update the Bullet object, possibly because parenting is more of a constraint than a compound shape, but they didn't want to deal with converting blender parenting to bullet constraints. I've complained to Blender's devs for quite a bit about their absolutely unacceptable Bullet implementation. However, they consider bullet a very low priority despite how powerful it could be (hell, they could do CAD like vertex constraints so you can dynamically resize objects if you treat them like bodies and joints, and solve with PGS)BluePrintRandom wrote:Hello, when I parent a rigid body child to a compound rigid body,
the origin does not shift, and the mass does not seem to change for the "host".
Is this a bug or just correctly implemented?
I have a bug in the tracker, but I thought someone may know more about bullet here,
Here's a list of other issues:
-Blender time is locked to bullet time, but frame-rate jumps are perfectly ok (which results in exploding joints if you try "bullet-time" style animations)
- Only linear PGS solver is supported
- Joint motors are not supported
- Soft-body is not supported
- Almost zero scripting possibilities because of C wrapper being incomplete and removal of driving functions in 2.5
-
- Posts: 47
- Joined: Sun Oct 27, 2013 4:16 am
Re: Help with blender 2.7a Compound physics.
I wonder if I can just shuffle the objects origin in bullet?
-
- Posts: 463
- Joined: Fri Nov 30, 2012 4:50 am
Re: Help with blender 2.7a Compound physics.
You have to recalculate the center of mass (the "origin" ), but the C wrapper doesn't properly expose that set of functions last I saw of the code. The issue with blender is that it hides bullet behind a "user friendly" python/C layer that doesn't have even a quarter of the things it needs to be useful.BluePrintRandom wrote:I wonder if I can just shuffle the objects origin in bullet?
You might be able to just run simulations outside and import the transforms, possibly faster too.
-
- Posts: 4
- Joined: Sun Feb 10, 2013 3:41 pm
Re: Help with blender 2.7a Compound physics.
Here, BluePrintRandom is referring to the Blender Game Engine Bullet implementation, I believe.Basroil wrote:To date, parenting in Blender doesn't update the Bullet object, possibly because parenting is more of a constraint than a compound shape, but they didn't want to deal with converting blender parenting to bullet constraints. I've complained to Blender's devs for quite a bit about their absolutely unacceptable Bullet implementation. However, they consider bullet a very low priority despite how powerful it could be (hell, they could do CAD like vertex constraints so you can dynamically resize objects if you treat them like bodies and joints, and solve with PGS)BluePrintRandom wrote:Hello, when I parent a rigid body child to a compound rigid body,
the origin does not shift, and the mass does not seem to change for the "host".
Is this a bug or just correctly implemented?
I have a bug in the tracker, but I thought someone may know more about bullet here,
Here's a list of other issues:
-Blender time is locked to bullet time, but frame-rate jumps are perfectly ok (which results in exploding joints if you try "bullet-time" style animations)
- Only linear PGS solver is supported
- Joint motors are not supported
- Soft-body is not supported
- Almost zero scripting possibilities because of C wrapper being incomplete and removal of driving functions in 2.5