Could we represent our sampling for antialiasing in terms of a Monte Carlo distribution? For example, could we sample out pixels as a pdf over the area of the pixel?
Would this be more efficient / necessary?
fywu85
I think for smaller problems where brute force can already produce acceptable outcomes, there is no need to do Monte Carlo sampling. I think MC is only useful for larger problem where brute-force-style searching is no longer feasible.
Could we represent our sampling for antialiasing in terms of a Monte Carlo distribution? For example, could we sample out pixels as a pdf over the area of the pixel? Would this be more efficient / necessary?
I think for smaller problems where brute force can already produce acceptable outcomes, there is no need to do Monte Carlo sampling. I think MC is only useful for larger problem where brute-force-style searching is no longer feasible.