In this process, are we now sampling points from the area of the light source? And are we computing this for a specific point p on some object in our scene?