10/26/2010 9:00 AM
To support the increase in 3-D production for movies and TV, equipment vendors and research organizations are working to solve the mystery of stereographic video image capture.
The Fraunhofer Heinrich Hertz Institute (HHI), based in Berlin, has developed a hardware/software system called the Stereoscopic Analyzer (STAN) that helps camera operators in the field and directors on-set figure out the depth of subjects in a scene and how they relate to each other. This includes far and near objects as well as the convergence plane and the depth of focus. Getting these parameters right avoids jarring images that get in the way of a pleasing viewing experience.
In the works for more than three years, Fraunhofer HHI developed the STAN in cooperation with KUK Film Production, in Munich, under a project called Prime, which was funded by the Federal Ministry of Economics and Technology. It is now a marketable product that can be used on movie sets or for live 3-D production.
STAN gives camera operators, stereographers and production staff the ability to figure out the correct stereo parameters and camera settings for any given scene. It includes features that ensure that the computed stereo parameters are fed directly to both cameras so any inadequate setting can be identified and corrected, either manually or automatically.
Frederik Zilly, project manager for image processing at HHI, said that with 3-D still in its infancy, production teams need a way to capture good 3-D as quickly and painlessly as possible. He said that production of good stereo content is a challenge that requires a variety of technical, psychological and creative skills.
“We know that making good 3-D images is not that easy to do,” he said. “With the wide variety of skill levels now involved in 3-D production, a system like STAN helps to bring everyone on the same page and ensures that your day’s worth of shooting will be productive.”
STAN combines hardware and software to capture and analyze stereo images directly on the set so the two cameras can be adjusted properly to the scene content and that remaining unwanted distortions can be corrected electronically in real time. The system can also can generate and store metadata to streamline the post-production process.
Zilly said STAN can adapt to accommodate a wide range of different factors, including the stereo baseline, color matching, stereo geometry and the orientation of the two cameras, which may vary from scene to scene independent of the scene content. It provides image-based stereo analysis and real-time preview of important, production-related information like stereo quality, scene depth or the violation of fundamental rules of 3-D production (stereo framing, respect of the available depth budget, avoidance of eye divergence, etc.).
STAN also features a touch-screen control panel that includes several visualization modes for evaluating stereo quality.