Loudspeaker/Source Spherical Coordinate System

[jcbreton]

We should decide on a PHI/THETA convention to use for spherical loudspeaker coordinate systems. There's no particular standard that is currently used. See below for what the CLF format uses (hopefully it makes sense, it's tough to draw this):

The positive X, Y, and Z axes come out of the front, right(when viewed from front), and top of the loudspeaker, respectively. The azimuthal angle phi is defined 0 degrees at the top of the loudspeaker (the +z axis), and rotates clockwise in the yz plane when viewed from the front of the loudspeaker. The polar angle theta is defined from 0 degrees on-axis to 180 degrees in the back of the loudspeaker. The polar angle half-circle always rotates so that it goes through the selected phi.

I'm open to using another convention, but as for the CLF parser, I will probably stick to this so we don't get lost in coordinate system conversions just yet.

[gregzanch]

@jcbreton I'm into it 👍 it's good to be consistent for sure.

[jcbreton]

In the RayTracer index file, I changed the random angle generation algorithm so it will generate a random phi between 0-360 and a random theta between 0-180 like shown above. Then, I wrote a function to convert between the "CLF" spherical coordinate system angle convention above and the spherical coordinate system that THREE.js uses (accounting for degree->radian conversion and a coordinate system shift).

Right now, there's absolutely no difference in performance. The RayTracer still generates random initial ray directions like it did before. The only benefit of the changes I made is that now the RayTracer is generating the random angles in the CLF convention, which will make applying directivity easier once we write a reliable interpolation function. Conversion to the THREE.js convention is done immediately before the initial ray direction is assigned.