I'm following this BRDF implementation which returns 0 when (NdotL < 0 || NdotV < 0), i.e., when the normal is pointing away from the light source.

This results in very abrupt shadows, as can be seen in the attached figure, which shows NdotL on the left, and the computed BRDF on the right.

Is this to be expected and what are possible mitigation approaches? Just add more light sources?

NdotL and diffuse BRDF result



You must log in to answer this question.

Browse other questions tagged .