changed the definition of the spherical coordinate system

I redefined the sph coordinate system. It may not be mathematically as clean, but it is convenient for us, since the conventional (inclination/azimuth) angles are never 0. I swapped the axes such that both angles are 90deg in the case of symmetric scattering. Then i redefine angle-->90-angle to have them zero again.

This way, one can interpret these as the angle between the Q vector and the surface and they correspond to the two tilt angles roll and pitch. (For the Yaw, at least 2 reflections need to be measured.) Another advantage is that the range of angles is always small and not sensitive to small misalignment. (before the azimuth of a symmetric reflections should range in the whole 360deg range, but that completely changes if there is a slight tilt of the sample.).

We should add a small documentation at some point.

I also just enabled the prediction of maximum number of bins, since it works the same way.