It would be interesting to get the answer to several questions, concerning the implementation of occlusion and irradiance functions:

1. Whether the form

vector occlusion(vector P, vector N, )

bent normals (or “average light direction”)' is used for sampling the data of an environment map?
2. As far as I know, the form

void occlusion(float &coverage, vector &missed_direction, )

presupposes that the user himself must sample the enivernment map according to the vector - missed_direction.
But here you get a problem. Putting missed_direction into environment(…) you get the picture different from the one you get, using the same environment map in the the first form of occlusion () - to tell the truth the resulting picture is totally wrong. What can be the reason of the problem?
3. Does “irradiance(…)” use “bent normals” and does it make sense using them? I think it makes no sense because the environment map is sampled in the same way as other objects in the scene.
4.Parameter “width”, which is used in the basic VEX Global Illumination Light (in irradiance and occlusion functions), is not supplied with documentation. Does it do anything with filtration?