Lecture 13: Global Illumination & Path Tracing (121)
weinatalie
Since path tracing works by analyzing the rays from a light source, I would guess that this scene is challenging to path trace since it is so dimly lit. Without strong light sources present, there would simply be less information for the path tracing algorithm to go off of. This could lead to the noisy, unrealistic result shown, where there are visible defects such as a random patch of light and very harsh shadows. I wonder what the best rendering technique would be for scenes like this—would ray tracing perform better than path tracing, since it can simulate the physical behavior of light?
MillerHollinger
I think another difficulty that arises when trying to render this scene is in the fact that the ball is a perfect mirror. This creates an intensely bright spot directly under the ball. Random bright pixels get scattered around because any path that touches the ultra-bright spot has its brightness massively increased. The solution here, I would think, would be to reduce the mirror-like quality of the ball to allow the extreme brightness to die out.
rishiskhare
How does that light refraction within the ball happen? It seems very intricate to refract the light in a round direction for the glass ball.
srikartalluri
In this image, we see a reflective sphere that on a textured surface with a light source. Problematically, The reflective sphere creates complex light interactions as light rays that hit the sphere will reflect in many directions. Calculating the paths of these reflected rays involves recursive tracing of rays that can reflect multiple times, potentially leading to an exponential increase in the number of rays to trace as the recursion depth increases. Limiting the recursive step is also dangerous as it might lead to the specks as we see in the image.
ArjunPalkhade
Similar question to rishiskhare, but is there a particular equation utilized to determine the concentration/variation of spread for those speckles (of light I presume)?
DreekFire
The way a ray bends when entering/exiting the sphere could be calculated by Snell's law. I think the specks are actually caused by the fact that there are very few rays illuminating those areas - if you look at a real picture of a glass sphere, it shows a bright circle that gradually falls off. A challenge with raytracing in those dimmer areas is that you can't just importance sample rays that point directly to the light, you have to importance sample rays that point to the light source after accounting for refraction through the sphere. Alternatively, if you don't use importance sampling at all, the probability of choosing a ray that reaches the light source is very low (corresponding to the low brightness in real photographs in those areas).
Since path tracing works by analyzing the rays from a light source, I would guess that this scene is challenging to path trace since it is so dimly lit. Without strong light sources present, there would simply be less information for the path tracing algorithm to go off of. This could lead to the noisy, unrealistic result shown, where there are visible defects such as a random patch of light and very harsh shadows. I wonder what the best rendering technique would be for scenes like this—would ray tracing perform better than path tracing, since it can simulate the physical behavior of light?
I think another difficulty that arises when trying to render this scene is in the fact that the ball is a perfect mirror. This creates an intensely bright spot directly under the ball. Random bright pixels get scattered around because any path that touches the ultra-bright spot has its brightness massively increased. The solution here, I would think, would be to reduce the mirror-like quality of the ball to allow the extreme brightness to die out.
How does that light refraction within the ball happen? It seems very intricate to refract the light in a round direction for the glass ball.
In this image, we see a reflective sphere that on a textured surface with a light source. Problematically, The reflective sphere creates complex light interactions as light rays that hit the sphere will reflect in many directions. Calculating the paths of these reflected rays involves recursive tracing of rays that can reflect multiple times, potentially leading to an exponential increase in the number of rays to trace as the recursion depth increases. Limiting the recursive step is also dangerous as it might lead to the specks as we see in the image.
Similar question to rishiskhare, but is there a particular equation utilized to determine the concentration/variation of spread for those speckles (of light I presume)?
The way a ray bends when entering/exiting the sphere could be calculated by Snell's law. I think the specks are actually caused by the fact that there are very few rays illuminating those areas - if you look at a real picture of a glass sphere, it shows a bright circle that gradually falls off. A challenge with raytracing in those dimmer areas is that you can't just importance sample rays that point directly to the light, you have to importance sample rays that point to the light source after accounting for refraction through the sphere. Alternatively, if you don't use importance sampling at all, the probability of choosing a ray that reaches the light source is very low (corresponding to the low brightness in real photographs in those areas).