raytracing

Ray tracing is a rendering technique that generates an image by simulating the paths of light as rays that travel through a virtual scene. In a basic setup, rays are traced from a virtual camera through each image pixel into the scene; their intersections with surfaces determine the color seen at that pixel.

At each intersection, the algorithm computes the surface color from the material properties and the lighting in the scene. To model reflective and refractive effects, rays are spawned recursively to capture reflections, refractions, and shadows. The resulting images can exhibit realistic phenomena such as soft shadows, color bleeding, and caustics.

Variants include classic Whitted-style ray tracing, distributed/Monte Carlo path tracing, and bidirectional or Metropolis light transport. Path tracing estimates the rendering equation by averaging many sampled light paths, trading determinism for potentially more accurate global illumination. Performance is improved by spatial data structures such as bounding volume hierarchies (BVH) or kd-trees, which accelerate ray-scene intersection queries. Denoising and adaptive sampling help reduce noise.

Real-time ray tracing has become practical with hardware acceleration and specialized APIs. Modern GPUs include dedicated ray-tracing cores, and APIs such as DirectX Raytracing (DXR) and Vulkan Ray Tracing enable interactive applications. Many renderers mix rasterization for primary visibility with ray tracing for reflections, shadows, and global illumination.

Origins date to early computer graphics research in the 1960s and the 1980 Whitted-style ray tracing, with substantial advances in the 1990s and 2000s. It is widely used in film production, architectural visualization, and increasingly in video games and real-time visualization. Limitations include computational cost and residual noise in stochastic methods.