---

Chapter 26: Ray Tracing

Review of OpenGL/Graphics

Rasterization

1. Geometry comes down the pipeline
2. Transformed to correct positions
3. Each primitive is rasterized
   • Determine lighting at vertices
   • Linearly interpolate across the face
   • Use z-buffer to resolve depth

Phong Reflection

A closer look at “step 1”: How do we determine the lighting?

We have reflection models, e.g. the Phong Reflection Model. These are a local approximation of the way light interacts with the face.

We have ambient + diffuse + specular.

Ambient light is a hack to add a small amount of global illumination to the scene (and only depends on material properties).

Diffuse is light reflected in all directions - only depends on the orientation of the face and the light source.

$$max(N \cdot L, 0)$$

Specular light is reflected off the surface like a mirror - in a straight line.

$$(V \cdot R)^\alpha$$

Where $\alpha$ is the “shininess”. The larger $\alpha$ is, the smaller the spot light is. Strongest when $V = R$.

Local Approximation

The consequence: only “local” light approximated. What is global light?

• Shadows
• Reflection & refraction

Not impossible, but usually hacky and limited.

So, let’s explore a completely different rendering alternative: ray tracing.

Ray Tracer

General overview of program:

1. read in scene description
2. compute lighting based on this scene description
3. output pixels (e.g. an image) based on the lighting you compute

• Ray Tracing
  • Create a photorealisitic 2D image from a 3D scene
    • Determine visible surfaces in an image at the pixel level
    • This is as opposed to the rasterization process which runs per-primitive and uses a z-buffer
  • Disadvantage: Per-pixel is slower, lots of duplicated work
  • Advantage: Shadows, reflections, etc. are easy to do (compared to rasterization)

Rasterization: process a bunch of data, use transforms to mimic 3D.

Ray tracing: closer to a simulation of how lights/cameras work.

---