Super basic newbie transform question....
Author
myDevice.Transform.View *= Matrix.Translation(new Vector3
(0,-30,0));
moves the camera up instead of down?
and to move in the direction of the camera i have to invert the new Vector3(view.M13, view.M23, view.M33) i get from the view matrix?
i guess i am missing something really basic? is up down? did i ferget to invert my umm... duh
Author
is my camera upside down? or inside out?
if (0,1,0) is the default "up" why when i transform by (0,1,0) i move down???
*bump*
thanks
if (0,1,0) is the default "up" why when i transform by (0,1,0) i move down???
*bump*
thanks
Because you''re not actually transforming the camera. There is no camera, in DirectX/OpenGL. What you would consider the ''camera'' is really just a perspective division centered around the origin. Thus, what you''re really doing is moving the world around.
How appropriate. You fight like a cow.
How appropriate. You fight like a cow.
this ain''t open gl... i have a world transform... i leave that alone and just change the VIEW transform.
No. In the pipeline, the world and view matrices are just multiplied. They do the exact same thing, just in a different order.
If you don''t understand this, I suggest you read up on linear algebra and its uses in computer graphics.
How appropriate. You fight like a cow.
If you don''t understand this, I suggest you read up on linear algebra and its uses in computer graphics.
How appropriate. You fight like a cow.
You''ll be doing yourself a favor by reading up on pipeline transformations. Particularly view and world transformations.
As it was stated, there really isn''t a camera. You''re simulating a camera by transforming the world. Think of it like this...
If you translate up, you are moving all the objects in the world up. So it would look like your camera moved down. To make it look like the camera is going up, you move all objects in the world down instead. Get it?
If this is all wacky (it is), try using an camera axis system with an eye point, lookat point, and up point. Then you can think more in terms of moving the camera instead of the world. Yeah, this stuff is crazy if you ask me. hah.
As it was stated, there really isn''t a camera. You''re simulating a camera by transforming the world. Think of it like this...
If you translate up, you are moving all the objects in the world up. So it would look like your camera moved down. To make it look like the camera is going up, you move all objects in the world down instead. Get it?
If this is all wacky (it is), try using an camera axis system with an eye point, lookat point, and up point. Then you can think more in terms of moving the camera instead of the world. Yeah, this stuff is crazy if you ask me. hah.
quote:Original post by drownEven then, you will serve yourself very well if you take the time to figure out how lookAt matrices are calculated. Gramm-Schmidt orthonormalization plays a role.
If this is all wacky (it is), try using an camera axis system with an eye point, lookat point, and up point. Then you can think more in terms of moving the camera instead of the world. Yeah, this stuff is crazy if you ask me. hah.
How appropriate. You fight like a cow.
Author
could anyone provide a quick explanation of what the difference it between world and view transforms then please?
from what i read here it seems like you are saying they should affect my scene in the same way?
i thought if i
myDevice.Transform.World *= Matrix.Transformation(0,0,-100);
would move the world back and the "camera"/view forward, and
myDevice.Transform.View *= Matrix.Transformation(0,0,-100);
would move the camera back and the world forward...
but you are saying both of those should move the camera forward?
please excuse my ignorance
from what i read here it seems like you are saying they should affect my scene in the same way?
i thought if i
myDevice.Transform.World *= Matrix.Transformation(0,0,-100);
would move the world back and the "camera"/view forward, and
myDevice.Transform.View *= Matrix.Transformation(0,0,-100);
would move the camera back and the world forward...
but you are saying both of those should move the camera forward?
please excuse my ignorance
Of course they do. You've shown it yourself, in your original post.
EDIT: The REASON for having two separate matrices like this is that, in general, it is convenient to pretend that the camera may be manipulated independently of the world. You can screw with the world matrix all you want without having to deal with the view matrix. OpenGL accomplishes this by maintaining a matrix stack; directX does it by maintaining separate model and view matrices. I personally prefer the OGL way, since it's more flexible and doesn't impose an arbitrary methodology on the programmer....... but I digress.
How appropriate. You fight like a cow.
[edited by - sneftel on June 5, 2003 10:37:24 PM]
EDIT: The REASON for having two separate matrices like this is that, in general, it is convenient to pretend that the camera may be manipulated independently of the world. You can screw with the world matrix all you want without having to deal with the view matrix. OpenGL accomplishes this by maintaining a matrix stack; directX does it by maintaining separate model and view matrices. I personally prefer the OGL way, since it's more flexible and doesn't impose an arbitrary methodology on the programmer....... but I digress.
How appropriate. You fight like a cow.
[edited by - sneftel on June 5, 2003 10:37:24 PM]
This topic is closed to new replies.
Advertisement
Popular Topics
Advertisement
