CONTACT: Stanford University News Service (650) 723-2558
Stanford team builds first digital holographic video/data storage system A physicist and two members of his lab have demonstrated the first working, fully automated, digital holographic storage system for video, data and sound. With what they've learned, they predict it will be possible to store hundreds of billions of bytes of digital data and retrieve them at a rate of billions of bits per second with a high degree of accuracy.
STANFORD -- Suspended above the worktable in Lambertus Hesselink's lab is a clear crystal block of lithium niobate the size of two sugar cubes. Two slim laser beams converge in its interior. The crystal glows green. Invisible to the viewer, the patterns of electrons inside the clear block are rearranged and a series of images are stored.
Next, one beam searches the cube alone. It reflects the pattern of those rearranged electrons and relays them to a video camera, then to a computer. Its monitor displays a video of a bird flapping its wings, then a digitized version of the Mona Lisa.
This is the first working demonstration of a technology that scientists have been talking about for 25 years: a system that can store videos, sound and data as holograms.
Holographic storage could be a solution for high-speed parallel computing and for storing large databases, like the constant stream of photographs from LANDSAT (Land Remote Sensing Satellite System) satellites. Because retrieval time is so quick, a company could set up a holographic video archive and consumers could dial in to order movies on demand from a central source, via fiber-optic cables. Multimedia- and video-game inventors could pull up a rich variety of images in seconds.
Using a combination of off-the-shelf equipment and ideas so new they're seeking patents on them, the Stanford professor, graduate student John Heanue and postdoctoral fellow Matthew Bashaw have created the first fully automated digital holographic storage system. Their lab-bench demonstration shows that it would be possible to store an hour of video images in a cubic centimeter of crystal, using current technology.
And they have demonstrated that it can retrieve the images at real-time video rates, with an accuracy of one error in a million bits of data.
The feat is described in the August 5, 1994, issue of the journal Science.
"We took advantage of the best aspects of holographic storage, which is inherently an analog system, and the power of digital signal processing," said Hesselink, who is a professor of electrical engineering at Stanford.
Hesselink predicts that future systems could store hundreds of billions of bytes of digital data, and retrieve them at a rate of a billion bits per second with a high degree of accuracy, ten to 100 times faster than today's data storage devices.
Holographic storage uses laser beams to record information in the form of two-dimensional holograms, one "page" at a time inside special optical materials. The image of the data can be read by shining a reference laser into the same volume of material. By varying the angles of the writing and reading laser beam, many pages of information can be stacked in and retrieved from the same space. The large amounts of storage and fast reading rates would allow compact storage for images that require a lot of data.
However, some technological hurdles remain before such applications are possible. With a number of other scientists in the United States and abroad, Hesselink has been working on aspects of the necessary technology for more than a dozen years. He is an expert on laser optics and has helped to develop several types of crystals that serve as storage mediums for data-bearing holograms.
Recent advances in computer technology, video transmission and data-storage materials make holographic digital data storage a near-term possibility. While many of these components have been tested, no one had put together a fully automated system that accomplished all the steps, from translating video pixels into digital bits and bytes, to storing them as holograms, to reading back the stored data and playing it as images.
Hesselink, Heanue and Bashaw realized that if they could build such a system, using off-the-shelf equipment, they could test the trade-offs between storage capacity and error rates. That data would help show what further improvements are needed for practical holographic digital video storage.
The Stanford team converted video images into compressed digital data, designed software to apply error correction codes, then beamed a laser through a device called a spatial light modulator to focus that data onto the lithium niobate crystal. A special mount moved the crystal imperceptibly, 50 microradians (three one- thousandths of a degree), as each "page" was recorded. To read the pages back, a second laser, the reference beam, reflected the recorded data to a charge-coupled device (CCD), similar to a camera. The data was relayed to a computer, decompressed, and checked for errors using the error correction code.
The first test run showed the team several problems that could be solved immediately. They invented several solutions and have applied for patents on two of them.
"In order to solve the difficulties associated with noise in the system and unavoidable imperfections in the crystal, we invented digital encoding techniques that reduced the noise sources in the retrieved signal," Hesselink said.
There were many issues this early demonstration did not resolve. Because they were using components that happened to be available in the laboratory, the one-part-per-million error rate was worse than can be attained with current techniques. Also, the holograms were gradually erased after many scans by the readout laser. However, previous research has shown that an efficient hologram "fixing" method exists so holograms can be permanently stored in lithium niobate.
The recording time also was slow, taking one hour to store just over 160 kilobytes of data. Hesselink said that currently available holographic materials will need to be improved to make this a read/write system as opposed to a read-only system.
Several other improvements must be made to bring their laboratory model up to its full potential, Hesselink said. Some of the technologies needed for a fully practical commercial system are already in various stages of development, with underwriting from the government and private companies. Hesselink himself is working with other scientists on improvements to the optics and to the holographic source materials.
In the article in Science, he and his co-authors also called for work on a practical device to change the angle of the laser beams by steering the light, rather than mechanically moving the crystal. They suggested improvements in spatial light modulators and charge- coupled devices, including a better match between the data-writing and data-reading capabilities of these devices.
"I think this initial demonstration showed that digital holographic data-storage systems can be built," Hesselink said. "We'd like to continue research and development to build a small size, rugged holographic data-storage system, with large capacity and no moving parts, to show its potential."
Note: This work was sponsored by the Defense Department's Advanced Research Projects Agency, under a University Research Initiative program grant.
This is an archived release.
This release is not available in any other form.
Images mentioned in this release are not available online.