Jump to content

  • Log In with Google      Sign In   
  • Create Account

convex/geometry transformation


Old topic!
Guest, the last post of this topic is over 60 days old and at this point you may not reply in this topic. If you wish to continue this conversation start a new topic.

  • You cannot reply to this topic
3 replies to this topic

#1 santa01   Members   -  Reputation: 307

Like
0Likes
Like

Posted 20 June 2012 - 05:03 PM

How shold a game engine handle transformation for an entity that aggregates true geometry as, lets say, OpenGL VAO, and convex geometry? For instance I want to move the entity in world, so I get a translation matrix and pass it to shader + apply that matrix to convex verticies manually or something?

Edited by santa01, 20 June 2012 - 05:14 PM.


Sponsor:

#2 Krohm   Crossbones+   -  Reputation: 3249

Like
1Likes
Like

Posted 21 June 2012 - 12:31 AM

There's no such thing as "convex geometry" to the GPU. It will just transform the vertices, group them in triangles and rasterize them. Whatever the mesh is manifold (let alone convex) or not does not change a thing. You upload your MVP matrix to the appropriate register, the shader will transform it.

Convex verticies? You appear to have a rough understanding, of the concept, at the very best.
To be completely honest, I'm afraid I don't understand your question in the first place.

#3 santa01   Members   -  Reputation: 307

Like
0Likes
Like

Posted 21 June 2012 - 04:55 AM

Convex verticies? You appear to have a rough understanding, of the concept, at the very best.
To be completely honest, I'm afraid I don't understand your question in the first place.


I'm not talking about graphics only, GPU side translation of some mesh is clear. I'm talking about rigid body dynamics as well, so I need a convex to represent particular mesh and to make further collisions detection. So the question: if I have some matrix, that already represent particular transformation, than should I use that to update a convex position on a CPU side as well as to pass it to GPU for real graphics rendering?

#4 Krohm   Crossbones+   -  Reputation: 3249

Like
0Likes
Like

Posted 21 June 2012 - 06:24 AM

In pure line of theory nobody mandates this but in the real world what I can say is that yes, transforms should be concatenated and should be used "in the same way" or "coherently". This is especially the case for ragdolls for example. As a matter of fact, they use physics to drive graphics representation. At the other side of the picture there are kinematic objects which are moved by logic only as opposed to dynamics. In both cases the transforms will match in a way or the other.

I cannot think of a realistic situation in which you want to have a solid mesh whose physics representation diverges from graphics. That's just not what the user expects.




Old topic!
Guest, the last post of this topic is over 60 days old and at this point you may not reply in this topic. If you wish to continue this conversation start a new topic.



PARTNERS