On 4/23/07, Wessam Bahnassi <wbahnassi(at)inframez.com> wrote:
To actually do the math is very simple. Generate a matrix that is
World*View*Projection*Clip, then invert it and transform your coordinate
with it. Seems like you forgot to apply the clip and projection matrices
properly to your code.
Right... my only problem is how to calculate these matrices to handle
all XSI camera types (Perspective, orthogonal...).
Also, if I read the article i mentioned earlier (and I hope I did :),
it had a function that calculates a matrix
(Cam_Transformation*Perspective_Matrix) and then transform the world
coordinates of a point to a normalized screen space coordinates. I'm
guessing that this Cam_Transformation is actually the view matrix, and
this Perspective_Matrix is actually the projection matrix. The only
one left is the clip matrix; which I suppose I shouldn't be interested
in since all the screen coordinates that I'm transforming are valid.
So the only thing I should do is invert the matrix and transform my
(normalized) screen-spaced points.
All of the above suggest that everything should go smoothly, but again
it seems like I'm missing something...
Thanks a lot,
Abdo Haji-Ali
Programmer
In|Framez
---
Unsubscribe? Mail Majordomo(at)Softimage.COM with the following text in body:
unsubscribe xsi
