Lecture 9: Ray Tracing & Acceleration Structures (15)
lycorisradiatu
This slide remind me of this question I had about the difference between ray tracing and ray casting in the context of ray-surface intersection.
colinsteidtmann
What exactly are we trying to compute? And is the simple idea to intersect each ray with each triangle...are there multiple rays?
S-Muddana
It is really cool how each ray has to navigate through potentially vast arrays of triangles to probe for intersections in their surfaces. It seems straightforward, but there are so many algorithms and optimizations at play to identify the precise points of contact.
keeratsingh2002
@colinsteidtmann
We're computing where the camera's rays intersect with a model's triangles to render the image correctly, determining visibility and lighting. The "simple idea" is to test every ray against every triangle, which can be slow, so typically we use optimization techniques to speed up this process. There are usually multiple rays involved to capture the full scene from the camera's perspective.
amritamo
In response to the first comment, ray casting is a when a single ray is cast from the camera through each pixel on the image plane. Ray Tracing is more complex and simulates light behavior accurately with features like reflections, refractions, and shadows. It involves tracing rays as they interact with surfaces and light sources for realistic scenes.
JunoLee128
It is interesting how these ideas intersect (ray tracing; no pun intended) with previous ideas in the course like triangle meshes and halfspace calculations. I wonder which of these algorithms take the most compute time, per scene.
This slide remind me of this question I had about the difference between ray tracing and ray casting in the context of ray-surface intersection.
What exactly are we trying to compute? And is the simple idea to intersect each ray with each triangle...are there multiple rays?
It is really cool how each ray has to navigate through potentially vast arrays of triangles to probe for intersections in their surfaces. It seems straightforward, but there are so many algorithms and optimizations at play to identify the precise points of contact.
@colinsteidtmann We're computing where the camera's rays intersect with a model's triangles to render the image correctly, determining visibility and lighting. The "simple idea" is to test every ray against every triangle, which can be slow, so typically we use optimization techniques to speed up this process. There are usually multiple rays involved to capture the full scene from the camera's perspective.
In response to the first comment, ray casting is a when a single ray is cast from the camera through each pixel on the image plane. Ray Tracing is more complex and simulates light behavior accurately with features like reflections, refractions, and shadows. It involves tracing rays as they interact with surfaces and light sources for realistic scenes.
It is interesting how these ideas intersect (ray tracing; no pun intended) with previous ideas in the course like triangle meshes and halfspace calculations. I wonder which of these algorithms take the most compute time, per scene.