Content
The Academy Software Foundation (ASWF) has officially welcomed OpenQMC as its latest open source project, marking a significant step forward in rendering technology for the motion picture industry. OpenQMC, developed by the renowned visual effects studio Framestore, is designed to enhance both the fidelity and efficiency of rendering photorealistic images across film, television, gaming, and advertising. This technology has already contributed to the production of major Hollywood blockbusters, including Barbie, Superman, Wicked, F1 the Movie, and How to Train Your Dragon.
Rendering photorealistic scenes involves the complex process of simulating light paths, often using ray tracing techniques that rely on random sampling of light sources and surfaces. Traditional Monte Carlo sampling methods, while powerful, tend to introduce noise or graininess in rendered images, especially when the number of samples is limited. This results in longer rendering times and increased computational costs, which can hinder creative workflows by restricting the number of iterative adjustments possible during production. Lincoln Wallen, CTO at Framestore Company 3 Group, highlights this challenge by explaining that conventional Monte Carlo methods necessitate extended render durations to achieve clean images, which is both expensive and time-consuming.
OpenQMC addresses these limitations by employing Quasi-Monte Carlo (QMC) sampling techniques, which reduce noise and improve image quality with fewer samples. Josh Bainbridge, Head of Rendering at Framestore and lead designer of OpenQMC, emphasizes that the library's advanced mathematical framework allows for faster and higher-quality pixel rendering. This capability is crucial when dealing with complex lighting scenarios, such as selecting among thousands of physical light sources or rendering detailed volumetric effects. OpenQMC’s versatility is reflected in its support for multiple architectures, including both CPU and GPU platforms, and its straightforward API facilitates seamless integration into existing production graphics pipelines.
Developed in conjunction with Framestore’s proprietary renderer, Freak, OpenQMC enhances key visual effects elements like soft shadows, depth of field, motion blur, and indirect lighting, contributing to photorealistic final outputs. Over the past three years, projects benefitting from this technology at Framestore’s global studios include Paddington in Peru, Thunderbolts, The Fantastic Four: First Steps, Civil War, and Mickey 17. The software also offers tools for statistical testing and development analysis, enabling continual refinement and optimization.
David Morin, Executive Director of the Academy Software Foundation, expressed gratitude towards Framestore for contributing OpenQMC to the open source community, anticipating wider adoption and collaborative development within the ASWF ecosystem. The Foundation itself is a partnership between the Academy of Motion Picture Arts and Sciences and the Linux Foundation, serving as a hub for open source innovation in image creation, visual effects, animation, and sound for the media industry. With 19 projects and four working groups, ASWF aims to foster shared technological advancement across creative disciplines.
Framestore, an award-winning creative studio founded in 1986, continues to push the boundaries of storytelling through technology. With over 3,000 employees across studios in cities like London, New York, Los Angeles, and Mumbai, the company’s portfolio includes a diverse slate of high-profile productions such as Gladiator II, Mortal Kombat 2, Avengers Doomsday, and Prehistoric Planet: Ice Age. The introduction of OpenQMC as an ASWF project underscores Framestore’s commitment to innovation and open collaboration within the visual effects community.
For those interested in contributing or learning more, the Academy Software Foundation encourages participation via its #openqmc Slack channel and the openqmc-discussion mailing list. Further information about the Foundation and its initiatives is available at https://www.aswf.io/.
