A meshless hierarchical representation for light transport

Abstract

Figure 1: Left: Real-time global illumination on a static 2.3M triangle scene. Both the light and the viewpoint can be moved freely at 7-21 frames per second after a little less than half an hour of precomputation on a single PC. Right: The indirect illumination expressed in our meshless hierarchical basis (emphasized for visualization). Green dots represent non-zero coefficients. AbstractWe introduce a meshless hierarchical representation for solving light transport problems. Precomputed radiance transfer (PRT) and finite elements require a discrete representation of illumination over the scene. Non-hierarchical approaches such as per-vertex values are simple to implement, but lead to long precomputation. Hierarchical bases like wavelets lead to dramatic acceleration, but in their basic form they work well only on flat or smooth surfaces. We introduce a hierarchical function basis induced by scattered data approximation. It is decoupled from the geometric representation, allowing the hierarchical representation of illumination on complex objects. We present simple data structures and algorithms for constructing and evaluating the basis functions. Due to its hierarchical nature, our representation adapts to the complexity of the illumination, and can be queried at different scales. We demonstrate the power of the new basis in a novel precomputed direct-to-indirect light transport algorithm that greatly increases the complexity of scenes that can be handled by PRT approaches.

Volumetric Vector-Based Representation for Indirect Illumination Caching

“…An Irradiance Vector [11] is a vector quantity whose norm represents the incident energy over a whole hemisphere and whose direction represents the main incident illumination direction. This quantity results in an increased robustness to geometric variations as shown in the works of Lecot et al [20] , Willmott et al [21] and Gobbetti et al [22] for radiosity techniques, and as pointed out by Lehtinen et al [23] . In general, vector representations are more independent from geometry for caching, as shown by the development of Light Vectors [24][25] .…”

“…The main benefits of irradiance vectors, compared to irradiance, is that for a local variation of normals, the reflected radiance can be adjusted, making this representation more geometrically robust (e.g., [23]). However, this quantity is still dependent on the normal at the sample position.…”

Modular Radiance Transfer

“…Adding highfrequency surface variation to smooth shading is a popular approach used in games [McTaggart 2004;Chen 2008]. As in meshless directto-indirect transfer [Lehtinen et al 2008], we seamlessly support this "transport normal mapping". Irradiance volumes [Greger et al 1998] store radiance distributions in a volumetric grid so that, at run-time, dynamic objects can be lit by their environment.…”

“…Extensions to local lighting [Kristensen et al 2005] require lengthy preprocessing and large amounts of data, as well as being scene dependent. We build on direct-to-indirect transfer [Hašan et al 2006;Kontkanen et al 2006;Wang et al 2007;Lehtinen et al 2008] although, unlike previous approaches, we represent transport using bases tailored to plausible direct and indirect lighting. We perform fast precomputation on small shape dictionaries so that, at run-time, scene-specific transport can be approximated by warping and combining transport from the dictionary elements.…”

Interactive sound propagation using compact acoustic transfer operators

“…In contrast, Precomputed Radiance Transfer (PRT) algorithms decouple light transport effects from the light source positions by computing a linear operator that defines how a variable light source configuration affects the radiance at surface sample points. PRT techniques can support both distant [Sloan et al 2002] and local [Kristensen et al 2005;Lehtinen et al 2008] source configurations. We use many of these ideas to formulate our acoustic transfer operators.…”