Mathematical Models and Monte Carlo Algorithms for Physically Based Rendering

PhD dissertation, Eric Lafortune, February 1996

[Dissertation] [Abstract] [Overview] [Contents] [Images]

Images

This page contains the colour images of my PhD dissertation. The images were converted from Greg Ward's 4-byte per pixel format to 256-colour GIFs, using colour quantisation and Floyd-Steinberg dithering (these were the days before the PNG format). The quality isn't so great as a result.

Chapter 4: Monte Carlo Methods Applied to the Global Illumination Problem

Figure 4.3: Path tracing: example scene.

Figure 4.4: Path tracing: example rays.
Figure 4.8: Next event estimation in path tracing: example scene.

Figure 4.9: Next event estimation in path tracing: example rays.
Figure 4.15: Next event estimation in light tracing: example scene.

Figure 4.16: Next event estimation in light tracing: example rays.
Figure 4.21: Bidirectional path tracing: example scene.

Figure 4.22: Next event estimation in path tracing: example rays.

Chapter 5: Test Results

A. B. C.

D. E.

Figure 5.1: Test scenes. Figure 5.2: Office scene, illuminated by neon lighting.

Figure 5.3: Office scene, illuminated by spot lights.
Figure 5.6: Test scene A rendered without and with next event estimation (with the same amount of work).

Figure 5.7: Test scene A rendered with combined estimators (with the same amount of work).
Figure 5.9: Test scene C rendered with path tracing with next event estimation (at only 10 samples per pixel), without and with control variates.

Figure 5.10: Test scene C rendered with path tracing and with bidirectional path tracing (with the same amount of work).

Figure 5.11: Test scene E rendered with path tracing and with bidirectional path tracing (with the same amount of work).