This is my C# implementation of Peter Shirley's Ray Tracing in One Weekend and Ray Tracing: The Next Week

Some notable changes:

• Code split in two projects:
  • RTLib containing the renderer, materials, textures, hittables and other supporting code
  • RTConsole containing the console runner and scene definitions
• Support for PNG output using ImageSharp
• Support for single-faced rectangles. For example they can be used to "close" a Cornell Box while still being able to see inside.
• Implemented two renderers:
  • ParallelRenderer works very similarly to the original from the book, with the main change being that it uses Parallel.For to allow rendering lines in parallel.
  • IncrementalRenderer uses a slightly tweaked loop to render one sample for every pixel at a time, and saving the image every 10 samples. This allows you to see results very quickly, and spot any obvious problem without having to wait for a full render.

Other changes include some performance tweaks, and using more idiomatic C# coding structures. I wouldn't consider this code to be highly optimized, but I've done my best to remove the main bottlenecks.

You can render images using this command line:

cd RTConsole
dotnet run -c Release -- [output_file_name].png [scene_number_from_1_to_8]

# For example
dotnet run -c Release -- image.png 1

If you have Just, you can also use just render SCENE_NUMBER to do the same.

Available scenes:

1. The random scene from the end of book 1
2. Two spheres with a checkered texture
3. Two speres with a noise "marble-like" texture
4. A sphere with the earth texture
5. A scene with a sphere and a couple light sources
6. Cornell Box, with glass and metal boxes
7. Cornell Box, with "smoky" materials for the boxes
8. The final scene from book 2

Sample Renders

Random Scene

Final Scene