Path tracing in Production: The Path of Water

Physically-based light transport simulation has become a widely established standard to generate images in the movie industry. It promises various important practical advantages such as robustness, lighting consistency, progressive rendering and scalability. Through careful scene modelling it allows highly realistic and compelling digital versions of natural phenomena to be rendered very faithfully. The previous Path Tracing in Production courses have documented some of the evolution and challenges along the journey of adopting this technology, yet even modern production path-tracers remain prone to costly rendering times in various classes of scenes, of which water shots remain among those most notoriously demanding. While this series in the past years covered a wide range of different topics within one course, this year we took the unusual step to focus on just one, the water-related challenges that we encountered during our work on Avatar: The Way of Water. Despite its seemingly simple nature, water causes a very multifaceted range of issues: specular surfaces cause spiky and sparse radiance distribution at various scales and in different forms, such as underwater caustics, godrays as well as fast-moving highlights and complex indirect on FX elements such as splashes, droplets and aeration bubbles. The purpose of this course is to share knowledge and experiences on the current state of the technology to stimulate active exchange in the academic and industrial research community that will advance the field on some of the challenging industrial benchmark problems. We will first give an overview of the nature of the singularities and its practical implications and then dive deeper into appearance and material aspects of water and the objects it interacts with. In the remaining sections, the course will focus on some specific aspects in more technical detail, providing both a solid mathematical background as well as practical strategies. Furthermore, we discuss some of the remaining unsolved problems that hopefully will inspire future research.