Do we assume photons bounce infinitely many times and are continuously distributed?
zehric
I'm not sure if we could actually simulate photons bouncing infinitely many times for obvious reasons, but if you're asking in terms of the theory, then I'm wondering the same thing.
Carpetfizz
I think if we didn't have reciprocity of rays, we would have to simulate rays bouncing around "infinitely" and only select the ones that intersect with our image plane. But because we do have this property, we can get away with casting a finite number of rays from the image plane and compute the associated color.
sunsarah
I remember a few lectures back that recursive ray intersection takes into account the rays that bounce off one object and hit another, so I am still a little confused on how this would work if we don't take into account light bouncing off surfaces but take into account how one object will affect how another object looks.
kavimehta
At what point do we stop calculating a rays bounces? There must be some cutoff point, and would it get affected if we are rendering something like a hazy scene?
Pinbat
You're right that there's a cutoff point. It seems like there's way too many bounces before light hits our eyes, or, in ray tracing, before the ray from our "eye" hits the light, so we approximate. The approximation is that when a ray hits an object, we create just one ray that goes to the light source.
Michael-hsiu
If we assume that light rays do recursively bounce an infinite amount of times before surfaces, I wonder how much we would overestimate the irradiance that hits any given point. I wonder how we would catch this too, since if our threshold for # of bounces is lower than the empirical total # of bounces, then points in space could have the same irradiation in our model but have drastically different values in reality, if those values were above our threshold but were numerically far apart.
fywu85
Keep in mind that in reality the energy of lights decay (quickly?) over time so practically it is not common to see a light after it has been bounced for many times since the energy of the photo is too weak for our eyes to perceive. Therefore, instead of cutting of the number of recursion, we could cut off the ray after its energy falls below certain threshold.
Do we assume photons bounce infinitely many times and are continuously distributed?
I'm not sure if we could actually simulate photons bouncing infinitely many times for obvious reasons, but if you're asking in terms of the theory, then I'm wondering the same thing.
I think if we didn't have reciprocity of rays, we would have to simulate rays bouncing around "infinitely" and only select the ones that intersect with our image plane. But because we do have this property, we can get away with casting a finite number of rays from the image plane and compute the associated color.
I remember a few lectures back that recursive ray intersection takes into account the rays that bounce off one object and hit another, so I am still a little confused on how this would work if we don't take into account light bouncing off surfaces but take into account how one object will affect how another object looks.
At what point do we stop calculating a rays bounces? There must be some cutoff point, and would it get affected if we are rendering something like a hazy scene?
You're right that there's a cutoff point. It seems like there's way too many bounces before light hits our eyes, or, in ray tracing, before the ray from our "eye" hits the light, so we approximate. The approximation is that when a ray hits an object, we create just one ray that goes to the light source.
If we assume that light rays do recursively bounce an infinite amount of times before surfaces, I wonder how much we would overestimate the irradiance that hits any given point. I wonder how we would catch this too, since if our threshold for # of bounces is lower than the empirical total # of bounces, then points in space could have the same irradiation in our model but have drastically different values in reality, if those values were above our threshold but were numerically far apart.
Keep in mind that in reality the energy of lights decay (quickly?) over time so practically it is not common to see a light after it has been bounced for many times since the energy of the photo is too weak for our eyes to perceive. Therefore, instead of cutting of the number of recursion, we could cut off the ray after its energy falls below certain threshold.