• 
• 
• 

Tuesday, July 10, 2012

This Article Features Photo Zoom

John Schwartzman, ASC, sets up a shot.

The TS-5 miniature beam-splitter can be used either handheld or mounted on a Steadicam rig. "It's made to fit into those tight spaces and weight-constrained spots that won't accommodate a larger rig," Schklair adds. "Weighing just over 17 pounds without cameras and lenses, it's easy to maneuver."

The rig allows two cameras to be aligned with sub-pixel accuracy, thus ensuring the precise alignment and position control necessary for stereographic 3D imaging.

The firm's new Helix stereoscopic rig, which resulted from a collaborative effort between former designers and engineers from 3ality Digital and Element Technica, combines magnesium, carbon fiber and aluminum into a functional, cutting-edge system. "Helix has all the precision, automation and refinement of our TS-5 system with the configurability and user-friendliness of the Atom," states Stephen Pizzo, 3ality Technica's senior VP. "It offers complete functionality when coupled with our Stereo Image Processor."

Helix takes full advantage of the firm's IntelleSuite control and automation software. Coupled with various hardware configurations, 3ality Technica's proprietary software applications—including IntelleCam, IntelleCal and IntelleMatte—are said to fully automate many 3D operations via smart technology that can dramatically reduce setup time and, hence, production costs. The new SIP 2100 provides real-time 2D and 3D analysis that's said to allow cinematographers to accurately match color and alignment between shots.

The live-action 3D material for The Amazing Spider-Man was captured with a parallel-camera configuration "with those stereoscopic images being converged digitally during post-production," says Engle. "By keeping the twin-camera rotation fixed during all of the location shots, we're able to secure extra image depth and resolution. I first experimented with this technique on Pirates of the Caribbean: On Stranger Tides; it's also very similar to a shifted film back we use for CG films.

According to 3ality Technica CEO Steve Schklair, the Epics were the perfect form factor for 3D because of their weight and portability. He also feels 3D looks better when shot stereoscopically rather than converted in post.

"The image sensor in the RED cameras [used on the 3ality Technica rigs] is wide," continues Engle, "allowing a frame line 10% smaller than our captured image. That extra image area lets us select which part of the image we'll use to alter the convergence—the angle between the two stereoscopic cameras—under computer control during our VFX postproduction. In other words, shooting parallel with extra image padding means that we're not locked into how the images will be presented. With parallel shooting, we know where the final convergence point will be, but there's one less knob to adjust—only the interaxial distance, or spacing between the two RED cameras, needs to be specifically dialed on set."

Parallel images are also easier to look at once the convergence is set, Engle confirms, "and the film suffers less optical aberrations from the two cameras. All lenses, however closely matched, are slightly different. Additionally, by keeping their photographic axes parallel, we can dramatically reduce any distortion caused by keystoning. And, to my eye, images shot parallel and converted in post just look better on the screen; it's an enhanced quality and easier to watch."

Likening the control of 3D depth to that of the dynamic range between loud and soft musical passages, Engle says, "We're not always looking for deep 3D. Just as we don't like music that's loud all the time, we prefer to look at creative opportunities for 3D rather than overwhelm the audience all the time. Allowing the audience to experience scenes with less 3D 'volume' can help increase the impact of the deeper moments. Less can definitely be more with 3D."

For images captured with a parallel rig—cameras pointed in the same direction without any toe-in, and not to be confused with a side-by-side configuration with cameras next to one other rather than using a beam-splitter—the two lenses share a common, but displaced axis. In other words, the stereo window is set at infinity and everything else lies in front of that window. The image aspect ratio also isn't preserved, meaning that landscape formats become square or even portrait format during postproduction. Since the image height also can be reduced in post to regain a wide aspect ratio lost during convergence, the original picture often needs to be over-framed to allow for such cropping.

2 %d Comments