Path tracing has been the ‘gold standard’ for 3D graphics since the late 80s. Now NVIDIA’s doing it live.
The original concept of “ray-tracing” for 3D images actually dates all the way back to the 16th century, but the first description of the idea in computer graphics was created in 1969. That earliest algorithm was very basic and only an approximation. Ray-tracing went through various evolutions over the years until, in 1986, Jim Kajiya presented a paper called “The rendering equation and its use in computer graphics.” In just seven pages, Kajiya described a way to mathematically compute the physical properties of light, and it was this revelation that transformed 3D rendering. Kajiya described his method as “path tracing.” It can be considered an evolution of simple ray tracing. Getting good quality out of a path tracer requires casting billions of rays. Until relatively recently, it was considered to be a completely impractical approach for real-time rendering, but it has been used in films and other offline-rendered computer graphics for over 15 years. The advantage of path tracing is that it is a “unified” rendering algorithm. This means that you don’t have to compute separate lighting effects like ambient occlusion, reflections, soft shadows, and so on.
