Eric Veach

Metropolis light transport

We present a new Monte Carlo method for solving the light transport problem, inspired by the Metropolis sampling method in computational physics. To render an image, we generate a sequence of light transport paths by randomly mutating a single current path (e.g. adding a new vertex to the path). Each mutation is accepted or rejected with a carefully chosen probability, to ensure that paths are sampled according to the contribution they make to the ideal image. We then estimate this image by sampling many paths, and recording their locations on the image plane. Our algorithm is unbiased, handles general geometric and scattering models, uses little storage, and can be orders of magnitude more efficient than previous unbiased approaches. It performs especially well on problems that are usually considered difficult, e.g. those involving bright indirect light, small geometric holes, or glossy surfaces. Furthermore, it is competitive with previous unbiased algorithms even for relatively simple scenes. The key advantage of the Metropolis approach is that the path space is explored locally, by favoring mutations that make small changes to the current path. This has several consequences. First, the average cost per sample is small (typically only one or two rays). Second, once an important path is found, the nearby paths are explored as well, thus amortizing the expense of finding such paths over many samples. Third, the mutation set is easily extended. By constructing mutations that preserve certain properties of the path (e.g. which light source is used) while changing others, we can exploit various kinds of coherence in the scene. It is often possible to handle difficult lighting problems efficiently by designing a specialized mutation in this way. CR Categories: I.3.7 [Computer Graphics]: ThreeDimensional Graphics and Realism; I.3.3 [Computer Graphics]: Picture/Image Generation; G.1.9 [Numerical Analysis]: Integral Equations—Fredholm equations.

Optimally combining sampling techniques for Monte Carlo rendering

Monte Carlo integration is a powerful technique for the evaluation of difficult integrals. Applications in rendering include distribution ray tracing, Monte Carlo path tracing, and form-factor computation for radiosity methods. In these cases variance can often be significantly reduced by drawing samples from several distributions, each designed to sample well some difficult aspect of the integrand. Normally this is done by explicitly partitioning the integration domain into regions that are sampled differently. We present a powerful alternative for constructing robust Monte Carlo estimators, by combining samples from several distributions in a way that is provably good. These estimators are unbiased, and can reduce variance significantly at little additional cost. We present experiments and measurements from several areas in rendering: calculation of glossy highlights from area light sources, the “final gather” pass of some radiosity algorithms, and direct solution of the rendering equation using bidirectional path tracing. CR Categories: I.3.7 [Computer Graphics]: Three-Dimensional Graphics and Realism; I.3.3 [Computer Graphics]: Picture/Image Generation; G.1.9 [Numerical Analysis]: Integral Equations— Fredholm equations. Additional

Deep shadow maps

We introduce deep shadow maps, a technique that produces fast, high-quality shadows for primitives such as hair, fur, and smoke. Unlike traditional shadow maps, which store a single depth at each pixel, deep shadow maps store a representation of the fractional visibility through a pixel at all possible depths. Deep shadow maps have several advantages. First, they are prefiltered, which allows faster shadow lookups and much smaller memory footprints than regular shadow maps of similar quality. Second, they support shadows from partially transparent surfaces and volumetric objects such as fog. Third, they handle important cases of motion blur at no extra cost. The algorithm is simple to implement and can be added easily to existing renderers as an alternative to ordinary shadow maps.

Bidirectional Estimators for Light Transport

Most of the research on the global illumination problem in computer graphics has been concentrated on finite-element (radiosity) techniques. Monte Carlo methods are an intriguing alternative which are attractive for their ability to handle very general scene descriptions without the need for meshing. In this paper we study techniques for reducing the sampling noise inherent in pure Monte Carlo approaches to global illumination. Every light energy transport path from a light source to the eye can be generated in a number of different ways, according to how we partition the path into an initial portion traced from a light source, and a final portion traced from the eye. Each partitioning gives us a different unbiased estimator, but some partitionings give estimators with much lower variance than others. We give examples of this phenomenon and describe its significance. We also present work in progress on the problem of combining these multiple estimators to achieve near-optimal variance, with the goal of producing images with less noise for a given number of samples.

Maintaining the Extent of a Moving Point Set

Let S be a set of n moving points in the plane. We give new efficient and compact kinetic data structures for maintaining the diameter, width, and smallest area or perimeter bounding rectangle of the points. When the points in S move with pseudo-algebraic motions, these structures process O(n2+e) events. We also give constructions showing that μ(n2) combinatorial changes are possible in these extent functions even when the points move on straight lines with constant velocities. We give a similar construction and upper bound for the convex hull, improving known results.

Non-symmetric scattering in light transport algorithms

Non-symmetric scattering is far more common in computer graphics than is generally recognized, and can occur even when the underlying scattering model is physically correct. For example, we show that non-symmetry occurs whenever light is refracted, and also whenever shading normals are used (e.g. due to interpolation of normals in a triangle mesh, or bump mapping [5]).