Physically Based Rendering

Physically Based Rendering
Name	Physically Based Rendering
First appeared	1980s
Developers	Pat Hanrahan, Philippe Bekaert, James Kajiya, Jim Blinn, Turner Whitted
Paradigm	Energy-conserving image synthesis
Influenced by	Ray tracing, Monte Carlo method, Radiometry, Optics

Contents

Physically Based Rendering

Physically Based Rendering is a computer graphics approach emphasizing energy-conserving models and algorithms for realistic image synthesis. It builds on foundations laid by figures such as Pat Hanrahan, James Kajiya, Jim Blinn, and Turner Whitted, and on institutions including SIGGRAPH, ACM, Eurographics, NVIDIA, and Industrial Light & Magic. Major works and projects like the book "Physically Based Rendering" by Matt Pharr and Greg Humphreys, the renderer from Seidel, and production pipelines at Pixar Animation Studios and Walt Disney Animation Studios have driven adoption.

Overview

Physically Based Rendering integrates theories from Radiometry, Optics, Statistical mechanics, Monte Carlo method, Numerical analysis, and Information theory to model light transport. Early milestones include Jim Blinn's shading models, Turner Whitted's ray tracing paper, and James Kajiya's rendering equation, with community dissemination at venues like SIGGRAPH, Eurographics, ACM Transactions on Graphics, and projects by NVIDIA Research, Intel Labs, Microsoft Research, and Google Research. Industry adoption accelerated through studios such as Pixar Animation Studios, Industrial Light & Magic, Walt Disney Animation Studios, DreamWorks Animation, and service providers like Weta Digital.

The theoretical core rests on the Rendering equation by James Kajiya, radiometric measures from Johann Heinrich Lambert-based models, and stochastic integration via the Monte Carlo method. Important contributors include Pat Hanrahan and Matt Pharr for systematization, and Eric Veach for multiple importance sampling. The field draws on mathematics from Fourier analysis, Stochastic processes, Measure theory, and algorithms like Path tracing and Bidirectional path tracing developed in part at Mitsubishi Electric Research Laboratories and evaluated at ACM SIGGRAPH conferences. Foundational lab groups include University of Utah, Stanford University, Cornell University, University of California, Berkeley, Massachusetts Institute of Technology, and Georgia Institute of Technology.

Material models use Bidirectional Reflectance Distribution Functions formalized in work by Fred Nicodemus and advanced by practitioners at Disney Research, ILM, and Epic Games. Canonical models include Phong reflection model by Bui Tuong Phong, Cook-Torrance by Robert Cook and Ken Torrance, microfacet theory from Walter Cook-related work, and models by Burley at Disney Research. Empirical measurements originate from labs like Xerox PARC and institutions such as MIT Media Lab, University of Bonn, and ETH Zurich. BRDF datasets and measurement systems from Columbia University, University of Utah, and Microsoft Research underpin material capture pipelines used at Pixar Animation Studios and in products from Autodesk.

Light transport algorithms include Ray tracing by Turner Whitted, Path tracing from James Kajiya's rendering equation, Photon mapping by Henrik Wann Jensen, and recent advances like Metropolis light transport by Eric Veach and Leonardo da Vinci-era optics influences. Sampling and variance reduction methods were advanced by researchers at Stanford University, ETH Zurich, Princeton University, and University of Toronto. Hardware acceleration efforts involve platforms from NVIDIA, AMD, Intel, and APIs such as Vulkan, DirectX, and research prototypes at Apple Inc. and Google Research.

Open-source and commercial implementations span projects like the PBRT system by Matt Pharr and Greg Humphreys, renderers at Pixar Animation Studios (RenderMan), engines from Epic Games (Unreal Engine), Unity Technologies (Unity), and offline solutions from Autodesk, Chaos Group (V-Ray), Foundry (Nuke), and SideFX (Houdini). Academic software originates from groups at Cornell University, Stanford University, ETH Zurich, Weta Digital, and ILM. Toolchains integrate middleware from NVIDIA, AMD, Intel, and cloud services by Amazon Web Services, Google Cloud Platform, and Microsoft Azure.

Current limitations involve computational cost addressed by hardware from NVIDIA, AMD, and Intel and algorithmic work at Google Research and Microsoft Research. Challenges include capturing complex materials by labs at MIT Media Lab and University of Bonn, real-time global illumination advanced by Epic Games and Unity Technologies, and scalable cloud rendering by Amazon Web Services and Google Cloud Platform. Future research directions interact with fields represented by Machine learning groups at DeepMind, OpenAI, Facebook AI Research, and institutions like Carnegie Mellon University and University of Toronto to integrate learned denoisers, neural reflectance fields from Google Research and NVIDIA Research, and hybrid pipelines championed by Pixar Animation Studios and Disney Research.