The Force is strong with holographic scientists these days. Researchers from MIT unveiled the fastest 3-D holographic video to date at a conference in San Francisco January 23, filming a graduate student dressed as Princess Leia and projecting her as a postcard-sized hologram in real time.
The holographic device plays a 3-inch projection at 15 frames per second, just shy of movie refresh rates of 24 to 30 frames per second, the MIT researchers demonstrated at the Society of Photo-Optical Instrumentation Engineers’ conference on practical holography.
The red hologram is jerkier and has much lower resolution than the one in Star Wars that sparked the public fascination with 3-D holograms in the 1970s. In fact, it kind of looks like a red blob on a staticky TV. But it’s 30 times faster than a telepresence device created in 2010 by University of Arizona researchers (SN Online: 12/4/10).
“I think it’s an important milestone because they were able to get to 15 frames per second, which is almost real time,” says physicist Nasser Peyghambarian, who led the Arizona research. “The quality is not as high, but hopefully it will get better in the future.”
The key to speed was computational power. The MIT team used a Kinect camera from an Xbox 360 gaming console to capture light from a moving object. Then they relayed the data over the Internet to a PC with three graphics processing units, or GPUs, tiny processors found in computers, cell phones, and video games that render video quickly. The processors compute how light waves interfere with each other to form patterns of light and dark fringes. Light bouncing off these fringe patterns reconstructs the original image. The MIT team used a display to illuminate the computer-generated fringes and create a hologram.
“The students were able to figure out how to generate holograms by using what GPU chips are good at,” says Michael Bove, an MIT engineer who led the research. “And they get faster every year. There’s room for a lot more understanding of how to compute holograms on them.”
MIT’s holograms are fast, says Peyghambarian, but they have to trade quality for speed.
Bove’s device uses one camera that estimates the depth of the object it is filming. The disadvantage of one camera, which is more consumer-friendly, is that you can’t see behind objects, says Bove. Also, even though graphics cards can compute high-resolution holograms, the effective display size is limited by a chip in the physical display to 150 millimeters by 75 millimeters, which Bove says is the biggest challenge to creating better holograms.
The Arizona device had a very different setup: Researchers grabbed video from 16 cameras angled around the object, so that one could walk around a holographic person and see not just the front side, but side profiles and back views. The team used an old-fashioned method that hologram artists have employed for decades, employing two lasers to create fringe patterns. Their key insight was engineering a special type of plastic that erases and rewrites quickly. The Arizona hologram is already high-definition and the size of a 17-inch TV, but speeding it up will require switching to a new laser system, says Peyghambarian.
Read more at Discovery News
No comments:
Post a Comment