Hi it works other way:
assume you have window w*h with aspect w/h, and ortho projection x*y with same aspect. Window resizes to w1*h1 with some other aspect w1/h1 -> projection now will be either: 1) x , (y*w*h1/(w1*h)) <- if we keep horizontal fow constant or: 2) x*w1*h/(w*h1) , y <- if we keep vertical fow constant with your numbers it gives projection: 1) (-50, 50, -75, 75), aspect 2/3 2) (-33.(3), 33.(3), -50, 50), aspect 2/3 and your new window aspect is same 2/3 you will get different _scale_, but it's uniform across screen, there is no stretching or smth. Cheers, Sergey. 10.08.2012, 16:32, "Ron Mayer" <[email protected]>: > hybr wrote: > >> Hi, Ron >> >> What meaning would have BOTH option? >> >> Each of HORIZONTAL and VERTICAL keep your projection aspect ratio same as >> window aspect ratio, difference is which one of horizontal fow or vertical >> fow they keep constant ( other is changed to match window aspect ratio ). >> >> Cheers, >> Sergey. > > Hi Sergey, > I am using orthographic projection. > When the window size changes it is not enough just to adjust only the width > or only the height of the frustum (projection matrix of camera), since doing > that will distort the objects in the world in one axis, wouldn't it? I > understand that maintaining an equal aspect ratio between the window and > projection matrix can be done by modifying only one axis, but I also want to > avoid my objects getting stretched or contracted in one axis. > > For example, let's say we start with a window of size 100X100, and a camera > with an orthographic projection matrix of: > (left, right, bottom, top) = (-50, 50, -50, 50). [ignoring near, far] > > Now the user modifies the window size to 50X75. > If we only scale the projection matrix horizontally this means that our new > projection matrix has the values of: > (left, right, bottom, top) = (-25, 25, -50, 50). > The width of the frustum was scaled by 1/2, but the height of the frustum did > not change. > > The window height is now 75, which means the objects will not look correct in > their Y values after the viewport transformation. > > On the other hand, I may be wrong, since there is no "BOTH" resize policy, > for a good reason that I can't see ... so I am trying to understand where is > my mistake. > > Thanks :D > > ------------------ > Read this topic online here: > http://forum.openscenegraph.org/viewtopic.php?p=49225#49225 > > _______________________________________________ > osg-users mailing list > [email protected] > http://lists.openscenegraph.org/listinfo.cgi/osg-users-openscenegraph.org _______________________________________________ osg-users mailing list [email protected] http://lists.openscenegraph.org/listinfo.cgi/osg-users-openscenegraph.org
