Realistic Rendering in Architecture and Product Visualization
(hovering mouse over a presenter's name shows his bio sketch)
In the recent years, VFX and computer animation witnessed a “path tracing revolution” during which most of the rendering technology has converged on the use of physically-based Monte Carlo techniques. This transition sparked a renewed interest in the topic of physically-based rendering but the focus has been almost exclusively on the application of these method in the movie industry. In the meantime, a significant segment of the realistic rendering market -- that focusing on architectural, automotive, and product visualization -- has been relying on the physically-based rendering technology since the beginning of the millennium. Despite that, relatively little attention in the communication at SIGGRAPH has been so far paid to this market segment.
The goal of this course is to fill this gap. We present user expectations in the “archviz”' and product visualization markets and discuss the technological and engineering choices that these expectations imply on the rendering engines used in these fields. We juxtapose this technology to rendering for motion pictures and point out the most significant differences. Specifically, we discuss the pros and cons of CPU and GPU rendering, simple (unidirectional) vs. more advanced (bidirectional) light transport simulation methods, different approaches to “lookdev” and material design, artist workflows, and the integration of the renderers into the image creation pipeline. We conclude by discussing some open technological issues along with the constraints that the research community should consider so that the the developed methods respect the needs and expectations of the target user group. ... Extended abstract
Jaroslav Křivánek, Christophe Chevallier, Vladimir Koylazov, Ondřej Karlíik, Henrik Wann Jensen, and Thomas Ludwig.
Realistic rendering in architecture and product visualization.
ACM SIGGRAPH 2018 Courses (SIGGRAPH '18). ACM, New York, NY, USA.
AcknowledgementsThe work was supported by the Charles University grant SVV-2017-260452 and by the Czech Science Foundation grant 16-18964S.