so I'm trying to pick 2D quads. I read this paper (http://www.dyn-lab.c...-selection.html), and I fully understand the theory behind it, however I have a little problem.
All the 2D quads I render start from a quad that is defined like this:
-1, 1..........1, 1
this quad gets translated rotated and scaled, so that it appers on screen where and how I want it. Now the problem is that in the paper, it says that I should do the picking in object space. Now in object space all of the quads are the same right? Because I essentially render the same quad over and over again. So my question is, do I have to transform all the quads into normalized device coordinates, so that I can do picking?
Or is it unnecessary?
Edited by Yours3!f, 28 October 2012 - 12:02 PM.