Oh here is a faster method - http://www.mvps.org/directx/articles/spheremap.htm
For each point (Assuming Y=up): 3DVec normal = (sphereCenter - point).normalize(); float U = asin(normal.x)/PI + 0.5 float V = asin(normal.y)/PI + 0.5 for z=up: float U = asin(normal.x)/PI + 0.5 float V = asin(normal.z)/PI + 0.5 But in your code example I don't know if you were calculating normals correctly. Your normal should be a vector pointing from the center of your sphere out the vertex, which is done by subtracting and then normalizing to make it a unit-vector. On Dec 22, 11:33 am, Robert Green <[email protected]> wrote: > pedro, > > your problem has nothing to do with android. I think you may need a > 3D refresher to solve it. Normals are not UVs. Your UVs will depend > on how you want to map, but the easiest way to do it is to map from > the sphere center out using angles, like this. Since GL has 2D > texture coordinate 0,0 as upper left, you will want to define where > upper left is on your sphere. I'd use straight up and center as a > unit vector, or (0, 0, 1) if z is up in your world. If y is up, then > (0,1,0) will be 0,0 in UVs. > > so starting with that, you can calculate the UV of any point by > getting the difference of angle of that starting angle and the new > point. Remember that if they both come from the sphere center, you'll > have complete 2PI radian coverage on both axis... so naturally > dividing the resulting angle by 2PI will give you a UV number between > 0 and 1. > > The process is like this (assuming z=up) > > For each point: > // for U just look at the sphere from overhead > U = (atan2(point.x - center.x, point.y - center.y) + PI) / (2PI) > hyp = distance(point, center); > V = ((point.z - center.z) / hyp) + 1) / 2 > > I believe that is correct and there will be a more elegant way to do > it but today my head is in 2D math world so I can't remember how to > dot it out. The idea though is that for every point, you can > determine the angle needed for U by looking at the sphere from > overhead and just using an arctan of the xy differences (assuming > z=up) of the center to the point. The resulting range is in radians (- > PI to PI) so you have to add PI to change it from 0 to 2PI and then > dividing by 2PI gives you 0 to 1. Then for the V value, you can just > examine the "height" (z) of the point and in conjunction with its > distance from the center, you've got an opposite over hypotenuse which > is the same as the sine of the angle, which is what you want to get a > evenly distributed texture top to bottom. The number will also be > ranged from -1 to 1 so you need to add 1 and divide it by 2 to get the > 0 to 1 range. The texture should touch in the corners and if you use > the right one, look seamless. > > If any of my math is wrong, please correct it after debugging, but at > least this should get you going in the right direction. > > Cheers > > On Dec 22, 9:09 am, pedr0 <[email protected]> wrote: > > > > > > > > > Please see the link and reply inside this post if you have not an > > account on OpenGL forum, the question is over there. > > >http://www.opengl.org/discussion_boards/ubbthreads.php?ubb=showflat&N... > > > Thanks a lot. -- You received this message because you are subscribed to the Google Groups "Android Developers" group. To post to this group, send email to [email protected] To unsubscribe from this group, send email to [email protected] For more options, visit this group at http://groups.google.com/group/android-developers?hl=en
