Original Post
Hey all,
I am trying to figure out a model coordinate from a screen coordinate. What I want to do is take the screen coords and transform it by the inverse of the world-view-projection matrices. Problem is, when I take an untransformed vertex and I multiply it by the world-view-projection matrices, I don''t even get any screen coords! Everything is setup fine, since I can view my model, rotate it etc. When I use a vertex buffer and do ProcessVertices, I get the proper screen coords. Is there an extra step in the geometry pipeline? Isn''t just a case of taking your model vertex and multiplying by the world-view-proj matrices?