On Mar 3, 2013, at 5:04 PM, sh4dow wrote:

> It seems to me this should be as simple as multiplying a given point with 
> some matrix, maybe scaling it to match the output resolution and bam. Done.
> 
> Unfortunately, it does not appear to be that easy.
> I multiply my point with the matrix of my camera, divide by w and presumably, 
> it should then be in some 0 to 1 or -1 to 1 space? And presumably, I would 
> then use the projection matrix to bring it to a space matching the output 
> resolution?
> I don't get values anywhere near what I would expect after multiplying it 
> with my camera matrix though.


Does your camera matrix also contain the projection matrix or just the inverse 
camera xform?
If only the inverse camera xform then don't divide out w yet - divide out w 
after multiplying by the projection matrix.

To check, transform the point by the inverse camera xform - any point that was 
in front of the camera should now be in -Z.  Multiplying that point (with a w 
of 1.0) by the projection matrix should produce a > 0 w coordinate that you can 
divide out.

-jonathan

_______________________________________________
Nuke-dev mailing list
[email protected], http://forums.thefoundry.co.uk/
http://support.thefoundry.co.uk/cgi-bin/mailman/listinfo/nuke-dev

Reply via email to