• Company
• Services
• Events
• Contact

1st International Symposium on Interactive Cinema (ICS 2006)

Abstract

The merging of interactive computer games and feature length animated movies to create a new art form of "interactive cinema" is not a new idea. For many years now, hardware vendors have claimed that "soon" graphics hardware will enable cinematic quality experiences in real-time. Storytellers have explored interactive stories and recently, some film studios have explored real-time relighting tools and the utilization of game engines to enable a director the real-time visualization of shots during pre-production. Unfortunately, experienced individuals from the animated film industry have made counter-claims that interactivity in a feature quality computer animated film is not doable for some time to come. This pessimistic perspective taken by some experts in the film industry is based on an intimate understanding of the immense bandwidth requirements of computer animated feature film. Unfortunately today's artistic content creation processes can generate multiple giga-bytes of geometry and texture data which is required to render a single frame using today's commercially available rendering systems such as Pixar's RenderMan, Nvidia's Gelato, Mental Images MentalRay, etc.

Given a commercially available large-scale high performance cluster, Gritz estimates four hours of render time per frame with 120,000 frames per animated feature film. Additionally, an animated feature film can have many visual effects such as fluid simulation, which can require around eight hours to simulate for a single shot. To realistically enable an interactive feature film, would require both rendering the frames at 60 frames a second and doing "all the simulations" that are required for a particular shot every 1/60th of a second. This is an immense computational requirement that presents an enormous barrier to entry for directors interested in exploring interactive cinema as a new entertainment medium. Fortunately, petascale computing systems being developed for NSF's petascale acquisition / DARPA's HPCS programs from vendors such as Cray, Sun and IBM as well as scalable clusters built from emering multi-core architectures such as the Sun Niagra, STI CELL, as well as X86 based multi-cores from AMD / Intel provide a great new opportunity for re-visiting the technical feasibility of interactive feature film.

Finally, it's important to keep in mind that just because to date, an interactive feature film has not been made doesn't mean that it can't be made. Therefore the objective of ICS-06 is to introduce innovations from the procedural content creation, parallel global illumination rendering, system software, parallel programming, and multi-core processor / cluster communities to the Siggraph engineering modeling, computer animation and real-time computer games community. Our goal is to enable a credible multi-disciplinary discussion on what it would take to make an interactive feature film possible.