Micro/angle dependent Roughness & Iridescence

[MartinZ (Martynas Žiemys)]

@lsscpp

As a result, the nodegroup can be expressed with just two values: 0° roughness, for facing geometry, and 90° roughness, for grazing angles. If the effect has not to be present the two values should be the same.
As in here:

Looking at some materials I would most definitely like to be able to control at what angle the effect starts and where the effect reaches it’s maximum at least if not all the falloff curve. The falloff now seems a bit arbitrary. Why power of 5? What’s so special about it? Should it be on Facing output that is (from What is it that Layer Weight node’s Facing output is supposed to be?):

@Secrop

…

if blend in [0.0, 0.5[:
    result = 1 - pow(dot(N,I), 2*blend)
if blend == 0.5:
    result = 1 - dot(N,I)
if blend in ]0.5, 1.0[:
    result = 1 - pow(dot(N,I), 0.5/(1-blend))

note there are clampings everywhere and the blend value never gets to reach 1.0, as it’s clamped to 0.9999 (1.0 - 10^-5)

That’s a bit random…