This slide reminded me of a bug I ran into in assignment 1. I was implementing jittered super-sampling (where each supersample gets moved randomly around by some small amount. I ran into a bug where, along the edge of two adjacent triangles, sometimes the jitter samples would sample off the edge of both triangles, resulting in dotted white lines along the edge. I wound up having to seed the random value function with the screen pixel, so that any jitter sample that was outside one triangle would be inside the other. Could a similar bug show up in Monte Carlo integration?
Rishiparikh
I'm kind of curious what you mean by seed the random value function. Would you only sample from elements from each of the triangles? Also that's a really interesting and cool technique. Is this used in graphics, or did you just come up with it on your own?
ashvindhawan
How exactly does area sampling involve integrals? Don't we simply average the color values? I'm a little confused how the integral gets introduced here
This slide reminded me of a bug I ran into in assignment 1. I was implementing jittered super-sampling (where each supersample gets moved randomly around by some small amount. I ran into a bug where, along the edge of two adjacent triangles, sometimes the jitter samples would sample off the edge of both triangles, resulting in dotted white lines along the edge. I wound up having to seed the random value function with the screen pixel, so that any jitter sample that was outside one triangle would be inside the other. Could a similar bug show up in Monte Carlo integration?
I'm kind of curious what you mean by seed the random value function. Would you only sample from elements from each of the triangles? Also that's a really interesting and cool technique. Is this used in graphics, or did you just come up with it on your own?
How exactly does area sampling involve integrals? Don't we simply average the color values? I'm a little confused how the integral gets introduced here