Lecture 13: Global Illumination and Path Tracing (29)

wangcynthia

In this image, does (p, w_i) represent the light ray and (p', -w_i) represents the reflected ray off of the object?

eric99ying

I am also confused. Why is w_i pointing out of p? Since it is the incoming radiance in the point p, shouldn’t it point into p, like in the diagram on the previous slide?

jpark96

@wangcynthia That's a really good question! I believe (p, w_i) represents the incoming ray and (p', w_i') represents the outgoing ray in the rendering equation.

@eric99ying It also took me a couple minutes to sort this out. In terms of our rendering equation L_o = L_e + \integral (some ugly term with L_i), (p, w_i) would be L_i and (p', w_i') would be L_o. In the one-bounce setting, (p, w_i) would be our light source.

sandykzhang

Another way to think about it is that wi is the direction the radiance is arriving from, not the direction the radiance is leaving.

theandrewchan

This is like Kirchoff's current law, but for light rays :O

buaazhangfan

I think for the direction here, both w_i and w_o represent the direction point out from p. Since here for incident rays, it refers that the arriving ray comes from w_i

In this image, does (p, w_i) represent the light ray and (p', -w_i) represents the reflected ray off of the object?

I am also confused. Why is w_i pointing out of p? Since it is the incoming radiance in the point p, shouldn’t it point into p, like in the diagram on the previous slide?

@wangcynthia That's a really good question! I believe (p, w_i) represents the incoming ray and (p', w_i') represents the outgoing ray in the rendering equation.

@eric99ying It also took me a couple minutes to sort this out. In terms of our rendering equation L_o = L_e + \integral (some ugly term with L_i), (p, w_i) would be L_i and (p', w_i') would be L_o. In the one-bounce setting, (p, w_i) would be our light source.

Another way to think about it is that wi is the direction the radiance is arriving from, not the direction the radiance is leaving.

This is like Kirchoff's current law, but for light rays :O

I think for the direction here, both w_i and w_o represent the direction point out from p. Since here for incident rays, it refers that the arriving ray comes from w_i