I haven’t worked with OpenGL in a long time. The last time I wrote anything using OpenGL was in college when I took a class on it. Well, I recently bought the OpenGL SuperBible, and from reading the first bit of it, it sounds like some interesting stuff has developed. There’s now this OpenGL shading language? I don’t remember anything about that, at least. There’s also a trimmed down version of OpenGL for mobile devices too, and that’s really cool.
Anyways, I remember in the class I never ended up with an OpenGL app that would allow for proper mouse rotation of a 3d object. Mouse interaction was never really assigned in that class, and I remember spending a lot of my time trying to figure out the math the teacher was having us do instead.
So I thought now that I’m jumping back into OpenGL related stuff, my first project should be revisiting that very simple, yet often overlooked bit of code that handles mouse interaction. I reviewed some of the code I had written in the class, and it had the basic idea right, but not enough thought put into it.
So I started fresh. I’ve learned some interesting things:
If you rotate the camera about your scene, then you won’t be able to spin a translated object in place (what you get is a big swing around the object, rather than watching the object spin). So it seems one wants to rotate the the whole scene before drawing it. This is, of course, to mimic what is seen in 3ds Max.
This is probably the easiest part of the math involved. Since translation from this projects point of view is moving about the screen plane, then we only have to move in two dimensions (The third dimension is zoom). All that needs be done is move our eyeball AND the point we’re looking at in parallel.
This one is a bit tricky. The effect we want is to move our eye closer to the point we’re looking at. This point is not necessarily the object. Imagine you have an eye location E and a focus point F. Moving our eye towards F involves a little bit of vector math. If we create a vector from E to F, we are half way there. To do that, subtract F from E in each dimension:
Vx = Fx – Ex;
Vy = Fy – Ey;
Vz = Fz – Ez;
Now we simple move our Eye point along this vector. Multiple the vector by some scaling factor and add
newEye = E + V*scale;
newEyex = Ex + Vx*scale;
newEyey = Ey + Vy*scale;
newEyez = Ez + Vz*scale;
When you think about the math, it makes sense. When you look at the math, it’s easy to get confused.
At any rate, here’s my program. I’ve tried to put it together so you don’t need to modify your paths to get it to build (the glut library is included).