Ray Casting
Is an alternative for the rendering-pipeline which [see page 4, sends] a line from the eye to each pixel in the world and the closest objects pixel which intersects that ray is shown first.
Note: We send from the eyes because most light doesn't reach the eye and we only care about the light that's visible. If we instead sent rays from the object and only measured the contribution from the rays that intersect the eye then there would be a lot of wasted calculations.
The most straightforward application of such a process is Visible Surface Ray Tracing.
Ray Direction
Is [see page 5, based] on the eye position \(R_0\), the direction vector \(R_d\) the eye is facing and some distance along that direction vector \(t\).
Issues with shadows
As mentioned here we can extend ray-casting to support shadow detection, however due to [see page 12, floating-point] inaccuracies, the surface point we send the ray from could be inside of the object the surface is on, therefore the surface intersects the ray immediately causing an undesired shadow. This is an issue with both Visible Surface Ray Tracing and naive-ray-tracing.
Other [see page 38, issues]
- Intersection tests are time-consuming
- The way rays are represented are unrealistic (light comes to our eyes, not from our eyes).
- aliasing