EEVEE: Better fit of BTDF LUT

This simplify the BTDF retreival removing the manual clean cut at
low roughness.

This maximize the precision of the LUT by scalling the sides by the
critical angle.

Also touched the ior > 1.0 approximation to be smoother.

EEVEE: Improve / fix GGX LUTs

This modifies the principled BSDF and the Glass BSDF which now
have better fit to multiscatter GGX.

Code to generate the LUT have been updated and can run at runtime.

The refraction LUT has been changed to have the critical angle always
centered around one pixel so that interpolation can be mitigated.

Offline LUT data will be updated in another commit.

EEVEE: Cleanup: Remove setup_noise from bsdf_sampling

Not needed anymore and makes code harder to read.

EEVEE: Cleanup: Remove sample count from bsdf_sampling

Makes uniform more local to where they are needed.

EEVEE: Fix incorrect fresnel function

This caused discrepency between cycles and EEVEE.

EEVEE: Fix issue with planar SSR

EEVEE: Fix most regression with ambient occlusion

Occlusion had change because of the way it was applied.
Get back to almost the same thing we had previously. The only thing
that changed is the occlusion of Normal mapped surface.

It is a bit less noticeable due to a new approximation we make.

EEVEE: Cleanup: Remove use of macros for the loops

EEVEE: Fix occlusion applied on SSR

EEVEE: Add back support for user occlusion on eevee_specular BSDF

EEVEE: Split closure_lit_lib.glsl

EEVEE: Fix missing shadowing factor

EEVEE: Fix lightgrid accumulation not using diffuse_accum

Also add safe normalize to interpolated normals.

EEVEE: Replace constant booleans by const float

This makes principled optimization easier and less verbose. Tests shows
no differences in performance.

EEVEE: Fix typo in macro leading to undefined behavior

Only the 1st closure radiance accumulator was initialized correctly.

EEVEE: Rewrite closure_lit_lib to reduce complexity

This rewrite improves:
- Code clarity: Less duplicated code and removes a few hacks.
- Compile time performance: Shader code was divided by 3 in average.
I did not profile the compilation time but it is faster.
- Shading Perf: Noticed a 25% performance improvement on
the shading pass with default dielectric principled bsdf.
- Fix Principled Tint being white if color is black
- It seems to have fixed issues on some drivers giving some incorect
results.
- Changes Principled BSDF support to be less hacky.