But it is consistent, right? Because if you sample infinite light direction you can get the one you want.
No, because the probability of hitting a point light is zero (no matter how many samples you take).
Oh~I forgot we are talking about the estimator's consistency! Thanks!
But it is consistent, right? Because if you sample infinite light direction you can get the one you want.
No, because the probability of hitting a point light is zero (no matter how many samples you take).
Oh~I forgot we are talking about the estimator's consistency! Thanks!