Generating a sphere

[MacFiend]

I would like to know how to generate a sphere (using pure math, no OpenGL routines). I have this basic code:


for (i=0;i<N;i++)
{
for (j=0;j<N;j++)
{
u[0] = i * 360 / N
v[0] = 90 - j * 180 / N
u[1] = ((i+1)%N) * 360 / N
v[1] = 90 - j * 180 / N
u[2] = ((i+1)%N) * 360 / N
v[2] = 90 - ((j+1)%N) * 180 / N
u[3] = i * 360 / N
v[3] = 90 - ((j+1)%N) * 180 / N
}
}


That is used to generate a sphere out of quads, however, I need to know how to get the vertex coordinates from that. Any takers? Thanks.

[MacFiend]

I got some code going, but its not right.


int i,j,N=50,R=2,u[4],v[4];

sphere_model.vertices = (vertex*)malloc(N*N * sizeof(vertex) * 4);

for (i=0;i<N;i++)
{
for (j=0;j<N;j++)
{
u[0] = i * 360 / N;
v[0] = 90 - j * 180 / N;
u[1] = ((i+1)%N) * 360 / N;
v[1] = 90 - j * 180 / N;
u[2] = ((i+1)%N) * 360 / N;
v[2] = 90 - ((j+1)%N) * 180 / N;
u[3] = i * 360 / N;
v[3] = 90 - ((j+1)%N) * 180 / N;

sphere_model.vertices[i*j+0].x = R*cos(u[0])*cos(v[1]);
sphere_model.vertices[i*j+0].y = R*cos(u[0])*sin(v[1]);
sphere_model.vertices[i*j+0].z = R*sin(u[0]);
sphere_model.vertices[i*j+1].x = R*cos(u[1])*cos(v[1]);
sphere_model.vertices[i*j+1].y = R*cos(u[1])*sin(v[1]);
sphere_model.vertices[i*j+1].z = R*sin(u[1]);
sphere_model.vertices[i*j+2].x = R*cos(u[2])*cos(v[2]);
sphere_model.vertices[i*j+2].y = R*cos(u[2])*sin(v[2]);
sphere_model.vertices[i*j+2].z = R*sin(u[2]);
sphere_model.vertices[i*j+3].x = R*cos(u[3])*cos(v[3]);
sphere_model.vertices[i*j+3].y = R*cos(u[3])*sin(v[3]);
sphere_model.vertices[i*j+3].z = R*sin(u[3]);
}
}


When I draw the model with GL_QUADS, the above makes a wierd, spiked shape...but its in the general shape of a sphere (if you were to put the shape inside a sphere). If I draw with GL_POINTS, it looks like a sphere. Any ideas what I am doing wrong? thanks.


glBegin(GL_POINTS);
for (i=0;i<50*50*4;i++)
{
glVertex3f(sphere_model.vertices[i].x,sphere_model.vertices[i].y,sphere_model.vertices[i].z);
}
glEnd();

[Mars_999]

Use GL_TRIANGLES

[Frank C.]

When it comes to geometry, Paul Bourke saves the day: http://astronomy.swin.edu.au/~pbourke/opengl/

At the bottom of that page are links to opengllib.h and opengllib.c, which have a few sphere/ellipsoid routines. They draw directly with OpenGL, but you should be able to abstract those bits into your own data formats if you want to submit it to OpenGL seperately.

[MacFiend]

Hmm, I don't know how I missed that, but you're right Frank C. Paul Bourke is DA MAN. Lot's of useful info on there. Thanks.