Hi, Have I got this right? First you have a whole bunch of untransform vertices in a vertex buffer. You multiply each of these vertices with the world matrix to transform and rotate all the little dots. The you have a buffer of transformed vertices. The you multiply it with a projection matrix that gives you all the dots in screen coordinates. Passing these to some kind of rendering thingy will give you a nice 3-d game, right. If you, for example, would like to disable the perspective in a scene, would you modify the projection matrix then? Thanks /Johan

Yes. If you want to disable perspective, you have to create an orthogonal projection matrix, instead of a perspective projection matrix...

EDIT
With the World matrix and the projection matrix, you have a View matrix too, it permits to simulate the move of the camera over the world...
/EDIT

All the steps of vertices transformation are made by the main graphic APIs: OpenGL, DirectX, so it's good to know how does it work, but avoid to implement your own transformations, except if you want to create your own graphic pipeline (to make a software rasterizer...)

[edited by - HiddenInBSP on August 26, 2002 5:52:52 AM]