Rotating Globe

[SMcNeill]

The one thing I notice now is that it seems to be rotating by the wrong axis.  Or else I'm just thinking about it wrongly.

Test this:

1) Rotate using the down area to move to ice into the center of the screen.
2) Using left/right, one would expect the globe to rotate the ice in the corresponding direction.  It's not doing that.  The ice stays centered and just spins in place.

Sundays are always busy around here and early next week is another winter storm which I've got to prepare for.  It might be late next week or next weekend before I can dig around and play with this much, but as it is, the rotating isn't quite as intuitive as it should be, I don't think. 

[Pete]

Petr - You rock!!! Specifically 3rd rock. +2

You know you could have just told Steve the other one, with the big gouge in it, was a futurist globe... due to fracking.

And Steve, who cares what ICE does?... Oh, except illegals. Trying that out, now...

Okay, it looks fine to me. Think of it as you move your spaceship from the Equator to the Pole. Look down and the rotation would be to keep the Pole in the center.

Pete

[Petr]

When you rotate the south pole (i.e. the point where theta=180) to the center of the screen, this point is always set in the program with fixed texturing coordinates (u=0.5, v=1). This means that it no longer has any defined azimuthal component - all phi (azimuth) values converge there.

Rotation around the Y and Z axes essentially only changes the azimuthal component (direction to the sphere), but for the poles this component is irrelevant because it is ignored and a fixed value is taken. Therefore, if you have the south pole exactly in the center, any rotation around the Y or Z axis, or the X and Z axes, will have the same effect - the point will remain the same because its determination is not based on azimuth.

In short:

For the poles, the azimuth angle is not taken into account due to the fixed setting of the texturing coordinates (to avoid seam problems).
Rotations around the Y and Z axes and X and Z axes primarily affect this angle.
Since there is no difference in the azimuth value at the South Pole, the visual effect of rotation around the Y axis is identical to the effect of rotation around the Z axis.

Maybe another solution could be used. But instead of additional mathematics, I would use glSphere, if it can be used in your program. It would be the fastest (and proven) solution. OpenGL sphere is one in quadrics here: https://qb64phoenix.com/forum/showthread...579&page=3


Thank you, Pete

[bplus]

Finally got around to checking vince version, I think I have it fixed better here:
https://qb64phoenix.com/forum/showthread...2#pid32022
reply #56

I don't think Terry's code was doing 3D rotation, just spinning the globe as mine was @a740g but he worked up the speed; vince is doing more rotations by dragging mouse. Be nice if someone could speed that up!

[SMcNeill]

When you rotate the south pole (i.e. the point where theta=180) to the center of the screen, this point is always set in the program with fixed texturing coordinates (u=0.5, v=1). This means that it no longer has any defined azimuthal component - all phi (azimuth) values converge there.

It took me a moment to sort this out in my head, and you're 100% right.

An easy way to visualize this effect is just to grab a quarter and hold it between the thumb and forefinger.  You can now press on and edge and rotate that quarter left to right, and you can tilt your hand to rotate it up/down.   The thing is, if you rotate your hand so that base of the upright corner is facing you, then rotate it left or right in your hand, you're still only staring at that base as it spins along the axis.

One would have to change viewing perspective to do what I was thinking of, and _OpenGl would probably be the best way to go about setting up and doing something more complex like that, like you suggested.