On 22. May 2009, at 5:08, Won J Jeon wrote: > I understand that stereo configure renders two times with different > rojection/view matrices. > > Actually, my question was how to compute different projection/view > matrices for left and right eyes. > > OpenGL routine by a user has only one project/view matrix for a > single view, and sometimes the project/view matrices are > > mixed with model transform matrices - which means that we cannot > correctly infer viewing parameters. Even worse if we use > > shaders. > > So what is equalizer's technique to correctly infer viewing > parameters? >
The parameters are computed from: - The frustum (wall, projection) description in the config file - The observer parameters (eye separation, head matrix) given by the application - The current eye pass The basic algorithm is: - Compute eye positions in world space - For each frustum o transform eye position to wall space o compute frustum corners from eye pos and wall size o compute head transform from frustum matrix and eye position > Or do users have to provide viewing parameters explicitly > > using some sort of API? > Normal applications do not need to provide anything, immersive applications have to provide the head matrix (see above). HTH, Stefan. _______________________________________________ eq-dev mailing list [email protected] http://www.equalizergraphics.com/cgi-bin/mailman/listinfo/eq-dev http://www.equalizergraphics.com
