Real-time volumetric shadows using 1D min-max mipmaps

Abstract

To compute single scattering in scenes with occluders (left) we compute a depth image from the camera, and a shadow map from the light. After epipolar rectification, each row of the shadow map is a 1D heightfield. We optimize the computation of the scattering integral by using an efficient data structure (a 1D min-max mipmap, center) over this heightfield. This data structure helps compute the scattering integral for all camera rays in parallel. Our method can render complex high-quality scenes with textured lights (right) in real-time (55 FPS). AbstractLight scattering in a participating medium is responsible for several important effects we see in the natural world. In the presence of occluders, computing single scattering requires integrating the illumination scattered towards the eye along the camera ray, modulated by the visibility towards the light at each point. Unfortunately, incorporating volumetric shadows into this integral, while maintaining real-time performance, remains challenging.In this paper we present a new real-time algorithm for computing volumetric shadows in single-scattering media on the GPU. This computation requires evaluating the scattering integral over the intersections of camera rays with the shadow map, expressed as a 2D height field. We observe that by applying epipolar rectification to the shadow map, each camera ray only travels through a single row of the shadow map (an epipolar slice), which allows us to find the visible segments by considering only 1D height fields. At the core of our algorithm is the use of an acceleration structure (a 1D minmax mipmap) which allows us to quickly find the lit segments for all pixels in an epipolar slice in parallel. The simplicity of this data structure and its traversal allows for efficient implementation using only pixel shaders on the GPU.

Instant convolution shadows for volumetric detail mapping

“…To quickly identify these positions in parallel on the GPU, an epipolar rectification of a shadow map is performed so that light rays are orthogonally aligned with the viewing rays. In later work [Chen et al 2011], speed is further improved using a min-max structure of the depth image from the camera for finding the lit segments of each viewing ray. Whereas they only model homogeneous translucent material, we allow for freely varying material opacities.…”

Imperfect voxelized shadow volumes

“…Equations 2 and 4 integrate quite smooth functions, allowing coarse SVD approximation [Chen et al 2011] to replace the integral by the sum of a few terms. We use a somewhat cruder approximation suggested by Dobashi et al [2002]:…”

“…Engelhardt and Dachsbacher [2010] sample and interpolate along epipolar planes, where volumetric shadows only change at depth discontinuities. Chen et al [2011] use min-max mipmaps to non-uniformly step along view rays and approximate the airlight integral with a coarse sum.…”

Camera space volumetric shadows

“…While very reasonable in the context of real-time renders, it was our finding that this method introduces too much blur for the quality targets of cinematic offline renders. It has been shown that it is possible to even get away with only a one-dimensional min-max mipmap [Baran et al 2010;Chen et al 2011]. This approach not only rectifies lines to the light, but also camera rays to gain even more performance.…”