I am doing some ray tracing through some simple GRIN lenses.

If I take GRIN3.dll, depending on what I set to n0, I get different results for intensity with ray splitting on, and the ray passing through the center of the lens (so no bending occurring).

Here is an example:

n0 = 1, 100% transmitted

n0 = 1.5, 96% transmitted

n0 = 2, 88% transmitted

This occurs with all GRIN dll’s (not just GRIN3), and does not occur with simple ray splitting. It functions as normal without ray splitting on.

This doesn’t make physical sense to me. Why is more light reflected from a light ray at normal incidence at higher refractive index without specifying any specific scatter model, split model? As a note, I do not have any scattering, volume physics, coatings, etc in these.

It appears to be some kind of index dependent ray reflection? No idea why that is happening. Going out to more extreme refractive indicies, it just trends towards 0 slowly.

A few other notes -

Adding in coatings does not change transmission

This still occurs using a model material, where the light does not fully transmit

Messing with polarization doesn’t change anything either

My assumption is that it follows this model -