The enthusiasm is certainly deserved -- film preservationists such as Robert Gitt at UCLA and Michael Friend at the Academy of Motion Picture Arts and Sciences are doing heroic work.
From the beginning, however, a dark secret has haunted their efforts: Using current analog technology, the preservation of film is a hopeless case. All forms of motion picture film, old-fashioned flammable nitrate and triacetate "safety stock" alike, are inherently, chemically unstable; even under optimum cold/dry storage conditions their deliquescence can only be slowed.
"For years my goal was to do this work the way Bob Gitt would do it, which means at the best labs and with the best surviving elements, preserved in a cold vault as long as possible," says Michael Friend. "But then you realize that by duplicating in an analog fashion you're going to lose resolution every time. Which means that when we lose the original negative -- and we will lose it, no matter how careful we are -- we will have inalterably lost that original achievement. We may put this day off for 100 years, or 500 years, but eventually the system has to fail, because of the inherent limitations of analog duplication."
If there's absolution to be found for this original (analog) sin of cinema -- and not everyone is sure there is -- it resides in what Robert Rosen, the Dean of UCLA's School of Theater, Film and Television, calls "the promise of the digital future." Ideally, digital methods of film storage preserve not the object itself but a description of the object -- an alphanumeric digital file that can be transferred from one platform to another. Some leading preservationists are working hard to make that possible. Friend says that through his conversations with John Galt, a high-definition TV engineer at Sony, "it became clear to me that we had to domesticate digital so that we could absorb all the data in a given image and transfer that to another piece of film without loss of resolution." Friend trumpeted the new technology, this "digital fix," in an influential 1995 article in the Journal of Film Preservation, "Film/Digital/Film" (available online at http://www.oscars.org/cmps/friend/friend.html).
Digital technology is already widely used as a tool of film restoration, invisibly removing scratches, brightening colors and de-granulating fizzy soundtracks. "Fifty or sixty kinds of errors," Friend says, "which cannot be fixed photo-mechanically, can be fixed digitally. There are all kinds of tools out there that hold promise for restoration in the future."
Many of the advances that may eventually prove crucial will be serendipitous side effects of an overall shift in the film industry from analog to digital technologies. Digital still cameras are already commonplace tourist accessories; in a few years the process of shining a beam of light on strips of chemically layered plastic emulsion to produce an image will seem as primitive to us, as barbaric, almost, as scratching a piece of plastic with a metal needle to produce a sound recording. The growth of new delivery systems, like DVD and HDTV, are spinning off helpful side applications, too.
One new system that intrigues Friend, he says, "is capable of taking conventional television up to high definition for broadcast on the new systems. Whether we could take a dupe negative, scan it in digitally at 4K, bump the copy up to 12K, and get a better original, we don't know yet. That's a test I definitely want to do." These tools might be usefully applied to compromised films like Josef Von Sternberg's 1927 silent Underworld. Von Sternberg aimed for high levels of visual glitter and resolution, but in this case we can only squint and try to guess what he had in mind: For this pivotal early effort, no original negative exists. With digital applications, we may eventually be able to free the image information from its celluloid death trap. We may even be able to roll back the clock on films like Underworld -- to restore it retroactively as if previous processes of decay had never taken place.
THE CONCEPTUAL APPEAL OF MOVING beyond digital restoration to permanent, lossless digital storage would seem to be obvious. But even the most ardent digital futurists warn that there are still some daunting wrinkles to be ironed out. For one thing, visual information is a notorious glutton for digital storage space. One estimate is that 8 megabytes of disk space would be needed to adequately approximate the image information -- the resolution and the color range -- on even a single frame of 35 mm motion picture film, which is still the most responsive medium we have for recording visual data. There are 144,000 frames in an average-length (100-minute) feature film. Plug those numbers into your calculator and you begin to get a feeling for the magnitude of the challenge. The potential payoff is staggering, but so are the practical, technical complications.
The computer industry is making rapid progress in data compression technology, which would make digital archiving of visual information somewhat more economically feasible, and bring the future a bit closer. But all the data compression systems developed so far involve some degree of approximation. "We don't understand all the aspects of what a film image is yet," Michael Friend points out. "We literally don't know how much resolution there is on a film. And then how does a computer read, and rewrite, and re-allocate that information? You hear things like, 'People can't see more than 2,000 line pairs of resolution, so we don't have to achieve any more than that.' But you have to have more resolution in the original than you have in the print, and if you impose artifacts in your digital original, you might have a problem."
Technicians speak hopefully of establishing a "1:1 correspondence of digital pixel to film grain," but the fact is, nobody really knows exactly what a single film grain represents, or rather, how much image information it contains. Friend hopes to learn more when a planned new preservation center, a joint project of the UCLA Archive and the Academy, opens in Westwood in a few years. The facility will contain both state-of-the-art refrigerated storage vaults and a real laboratory for research into some of the profession's basic issues.
"My overall project," Friend explains, "is to find out how to capture the characteristics of an image. The image is an object. It is a surface. It is not three-dimensional, but it's also not a language or a code. You can encode anything, but our task is a little different, a little more Faustian. It's like capturing all the atoms or molecules of that object. It is perhaps like capturing a DNA code. It's a tall order, and it is right at the edge of what we can imagine today."
And it is, everyone agrees, a precarious edge. Increasingly, librarians, computer specialists and even futurists are sounding yet another alarm: Storing digital data on current technology presumes that the technology required to read the data will be maintained in years to come. A few years back, the original two-inch master videotapes of a pair of Fred Astaire TV specials from the 1960s were discovered festering in a musty vault. "These are the first nationally broadcast shows shot on color videotape," UCLA's Robert Rosen says, "so they were of major technological as well as aesthetic interest." Trouble was, the tapes could not be watched, because not one specimen of the requisite playback equipment survived. Luckily, a set of blueprints for the lost hardware was eventually located, in the garage of a retired engineer in New Jersey, and from these the long lost tape deck was rebuilt. We will not always be so lucky.