.: From Interpositives to Separation Masters:How Film Preservation Works
The salvaging of Star Wars depends a great deal on the art of film preservation. Films are ultimately physical, whether that means 35mm film, magnetic tape or terabyte-sized hard disks, and they are victim to aging just as anything else. Nothing can escape the ultimate force in the universe: time. And so, films must be continually maintained and restored to keep them alive so that future generations can continue to view them. How exactly does that work? It's often a case by case process, depending on the nature and extent of the damage and what materials preservationists have to work with. This page will be looking at the history of film materials, the restoration methods used to preserve them, what materials survive for Star Wars, and in the end where the future of film preservation lies and how the Star Wars films can live on should certain powers that be grant them that right.
Sections:
Preserving Films: The Digital Paradox
It is those materials I want to first examine. What makes a movie? When you see a film in the theatre or on home video, where has it come from?
The original source of the movie is the original camera negatives. These are the pieces of film that ran through the camera on the set and actually recorded the footage. They are the highest possible quality material for the movie, the source from which all other copies come from. What happens with these after they go through the camera? After being developed in the lab, and often duplicated for protection's sake, they are catalogued by film librarians and assistant editors, and "edge codes" are printed into the sides of the film, which gives unique codes for every frame (actually, they appear about every foot, but you can count backwards or forwards a few frames). A low-quality duplicate of all the footage is then made, duplicating the edge-codes as well, for the editor to work with. This is called the "workprint"; the editor cuts the pieces of film together the way he or she likes, and then the original camera negative is conformed to this order by matching the edgecodes, ensuring that the negative assembly matches the workprint assembly to the frame.
This is a delicate proceedure, and mistakes cannot be made. The reason being, the negative cutter must scrape off half a frame on either side of the cut so he or she can cement the overlapping scraped pieces together for a clean, splice-free cut. If a mistake is made, such as cutting on the wrong frame, those frames on either side will be lost forever since this is the original material.
So, now you have a movie. But this is far from where the process ends. For one, the original negative has no colour-correction done to it; remember, it's an assembled edit of the movie using the original film pieces, so it's just what the camera recorded (except visual effects, which are printed onto a dupe and spliced in here). To re-time the film the way the cinematographer and director wishes (and also to avoid having to handle the negative any further), it is run through an optical printer that has three coloured lights of red, green and blue. By varying the intensity of the lights as the film is printed, the film can be re-coloured on a shot by shot basis (or, a cheaper solution is a "one-light" printing in which only one setting is used for the whole picture, usually done to control contrast). The filmmakers can check the results of these timing settings with an answer print. Once the answer print is approved, the negative is printed using the approved timing. The result of this is a new print: the interpositive. This is now the final film, with correct colouring, brightness and contrast.
However, you cannot make prints from this for theatrical projection. The reason being, an interpositive, as the name suggests, is a positive print and so if duplicated again you will get a negative. Thus, another intermediate is made: the internegative. Now that the film is back in a negative state, the internegative is what is used to mass-produce theatrical prints. At this point, an archival print master is often made, although they can also be made from the negative directly.
As you can see, there is a lot of duplication there from the negative to the theatre. Each time you copy the film, you lose quality. First, the grain increases with each generation: the original negative has only the grain of the negative, while the interpositive has the grain from the negative plus the interpositive emulsion, while the internegative has both grains plus its own emulsion grain, and so on. The same applies to dirt and dust, and even hairs (sometimes they are not literally hair, but lint or whatever). Labs are as clean as possible and technicians use white gloves to handle the film, but you still tend to end up with debris of some kind, and this adds up with each generation. A negative might look clean, but by the time you get to a theatrical print, the layers of copied dirt have become noticeable. Dirt on the negative and internegative prints as white specs, while dirt on the interpositive and print itself shows as black specs. Each time you copy the film, you also lose resolution, much like in a photocopy; if, let's say, the interpositive was 90% as sharp as the original negative and each copy was of the same degradation rate, that means by the time you got to an actual print you've lost close to half of your picture information. As a result, although 35mm film is said to resolve about five thousand lines of resolution, an actual theatrical print does not have much more picture information than a high-def projection. Each copy also increases contrast, and also loses colour information.
So, obviously, when restoring or presenting a film, it is best to use the earliest generation of material possible. Most early home video telecines were done using print masters because VHS, Beta and Laserdisc were soft enough that it didn't matter, and the softness also hid a lot of the dupe grain and dirt. Today, most transfers derive from the interpositive, which is seen as the ultimate master, but some will even go to the original negatives for the highest quality; this brings some complications, however, as the film will have to be re-timed again since the negative does not contain any colour correction. If it is a newer film, the available duplicates can be used as a colour reference. Older films must rely on research and judgement to determine correct colouring--the surviving original duplicate material with colour-correction, such as interpositives, would have faded enough to not be presentable as-is anyway.
Along the way here, however, there are safety copies made for archival purposes, such as the negative duplicate already mentioned, or separation masters and low-contrast print masters. But we'll get to those later.
So, that's the process of readying a film. What happens when a film sits around in storage for twenty years though? And what do you do when all of your source material is in disrepair in one way or another? Here's where we get into the nuts and bolts of film restoration. First, we need to understand a little bit of the history of motion picture film.
Early film bases were made out of nitrate. This was introduced around the turn of the century by George Eastman (of Eastman-Kodak) because it gave superior black and white images compared to the glass-backing negatives that were used earlier. It wasn't, however, meant for motion picture use, until Thomas Edison started utilizing it for such purposes. This introduced a lot of problems. Nitrate isn't exactly a stable plastic. In fact, it's what gunpowder is made out of. What happens when a nitrate negative gets put behind the red-hot carbon arc of a motion picture projector? Normally nothing, but if the films jams, unspools (or sometimes even spontaneously), the whole reel lit up like a Christmas tree. Dropping a match in a warehouse full of nitrate reels would not be advised. But it happened--whole collections were lost because they were so unstable. They also had an unpredictable lifespan. Maybe they would last for sixty years, maybe half that--you never really knew until one day the whole image just started breaking down. And because there weren't duplication stocks back then, every single print in circulation was struck directly from the negative, which soon fell apart. The concept of preservation didn't exist back then since movies came out, did their rounds and then were retired, much like a broadway show, except in unusual circumstances where a hit was re-shown. Many older films were tossed into studio furnaces to make space for newer ones. At the time it made sense; when would anyone see any of these films again?
Most films from before 1920 are today lost, either in whole or in part, and the next two decades provide fairly scant offerings. Once sound was introduced in 1929, trick photography couldn't be done in-camera because you had to keep a synchronized sound (before, you simply re-wound the film and did a multiple exposure), so duplicate stock for optical effects was developed. These stocks were continuously improved until 1948 when triacetate "safety film" was introduced to replace nitrate. This was when the real era of film preservation began. Studio moguls Walt Disney, Louis B Meyer (of MGM) and comedians Charlie Chaplin and Harold Lloyd were the first to begin transferring some of their nitrate negatives to safety film. Lloyd even bought all the original nitrate negatives for every one of his films, but then lost them all in a fire; MGM suffered fires in the 1960s that took care of legendary lost films like Lon Chaney's London After Midnight, and other studios suffered similar fires. In any event, "safety film" wasn't really that safe either, and would eventually deteriorate if not stored correctly (you might hear of the term vinegar syndrom; this was caused by shrinking triacetate safety film that caused an odor that was compared to vinegar).
Colour film was introduced in the 1920s with two-strip Technicolor, though there were some expriments before this, aside from tinting and hand-colouring that continued to be used. The "two strip" process used a single black and white emulsion to capture red and green lightwave records on alternating frames. In the 1930s, Technicolor introduced its famous three-strip camera. These cameras incorporated a beam-splitter to send red, green and blue lightwaves to separate black-and-white film emulsions, which were then colour-dyed and combined. This was a technology that had been developed in the 19th century for still photography, but Technicolour was the first to use it in motion pictures. Prints were made through an imbibition process whereby the dyes of all layers were transfered onto a single print (hence, Technicolor imbibition, or I.B. print, commonly referred to as dye-transfer). Because the strips were dyed and transfered this way, they had a very fine grain structure, but more importantly they would never fade, as their photochemical successors would. They were, of course, printed on nitrate stock. The three-strip colour process provided more natural hues, but it also required much more light to photograph, and the cameras were so heavy they could not be lifted by a single man when it was in its sound blimp. You can see the camera below in its massive sound blimp from the filming of 1934's Becky Sharp.
Eastman Kodak had first developed colour reversal film in the 1930s with Kodachrome, which used similar three-layer systems of chemicals. In 1950, they introduced the first single-strip negative for 35mm use, Eastmancolor, which made three-strip Technicolor obsolete--although Kodak Eastman photochemical film from this period was poor in quality. As a result, Technicolor continued to be used for a few more years, the last production being Foxfire from 1954, when the cheaper Eastman stocks could no longer be resisted. Technicolor continued to make dye-transfer (imbibition) prints of Kodak-Eastman stocks, however, providing finer grain and superior colour reproduction. It slowly died out over the course of the 1960s, Godfather being the last major film printed on the format in the United States in 1972, but it survived overseas for a few more years. Refinements in Eastman stocks were made over the years, of course, improving granularity, colour reproduction, as well as light-sensitivity. Black and white stock, including the dyed Technicolor, is made out of silver halide crystals; they will basically last forever, which is why the major issue was the unstable bases. Now, however, you had base issues, plus molecular degradation of the chemical mixture that produced colour on Eastman stock.
Star Wars was shot in 1976 on four stocks: Kodak 5243, an intermediate, 5247, a tungsten stock for live action, 5253, an intermediate used as a seperation stock for visual effects, and a Colour Reversal Intermediate (CRI), used for some of the optical effect printing. Unfortunately, the filmstocks of the 1970s (as well as the previous two decades) were highly unstable. These emulsions were made this way because they were less expensive, but by the 1980s filmmakers realized that material from earlier decades was already starting to fade, even those from the early 1970s. Martin Scorsese saw it coming and was famously a strong industry advocate to get film manufacturers, namely Kodak, to develop more robust emulsion compositions. By 1983, Kodak stopped making one of the stocks Star Wars had been shot on because it was so bad, and throughout the 80s new advances were made in developing low-fade film that was more suited to archival purposes--this was especially important because home video was now a major game-changer, as was the burgeoning cable market. Before, a film only needed to be presentable for its original run, and maybe a revival or television showing every few years if it was a hit. As a result, colour negatives from 1952-1982 are what archivists now call "quick fade." These days, while negatives and prints won't last forever, they hold up remarkably well compared to those from the 70s especially. This was why it was so surprising in 1995 when the negatives for Star Wars were dug up for the Special Edition--film from other eras was fine but no one knew just how bad that '70s stock was. In less than 20 years, it had become unwatchable, and duplicate prints had turned to pink as well. The owner of a 70mm print of Star Wars sent me these:
In contrast, for example, I saw a print of Jurassic Park last month, and at seventeen years of age, the same as Star Wars in early 1995, it looked basically the same as it would have when first run. This is because of newer "Estar" stock which was introduced in 1993 for all duplication, and should last about seventy years or so. Estar stock has a polyester base, as opposed to triacetate, which is more stable and also much stronger; it is so strong that it could potentially damage motion picture cameras were it to jam, and so negative stock still uses the older triacetate base.
So, what happens to older film material when it ages? First is the fading. Positive prints will begin to fade to red; anything from before 1975 or so is basically unwatchable today, while most prints from the '80s have a very mild pink shift to them, but of course both of these will get progressively worse with time. Negatives have a phenomenon whereby the yellow layer begins to deteriorate, turning them blue. The other thing that happens to film is that it shrinks with age, and becomes brittle (film has a lot of similarities to old newspapers, it seems). Film can also swell, depending on the conditions of its storage such as temperature and moisture. Even if stored in a controlled vault, over a long enough time those 1% fluxuations tend to take their toll. This means that the sprocket holes might not line up properly, or that the film won't go through the projector gate properly, and could lead to tearing and other damage. Existing damage is another problem; if something is older, it gets printed more, especially if it is a popular movie, and it's inevitable that little dings will appear here and there. A torn sprocket, even a torn frame, or just a bit of scratching. Finally, with age and handling, you also pick up more dust and dirt.
So, how do you deal with these? Fading isn't too bad these days, because as long as there is still something there you can colour-time it out, with digital technology being especially helpful here. A worse situation is when a piece of film is so deteriorated that there's simply inadequet amounts of colour left to draw out. Some of the early photochemical films face this problem. If you've seen some of those Star Wars deleted scenes, like Luke and the tread-droid or the longer cantina cut, they are from a low-quality workprint copy that by 1995, when the Behind-the-Magic CD-ROM was made, had totally faded to black and white. Now, that's because it was just an editor's copy made from cheap film stock, but on a long enough time scale all photochemical film look like that, and then the black-and-white will start to rot away too, until all you are left with is the clear celluloid base. Film is just chemicals sitting on plastic, and while plastic may last almost forever, chemicals eventually break down and dissolve. Aside from that bleak prospect, colour-timing out fading used to mean printing a new interpositive, using the printer lights to counter-act the faded layers. Today, film is scanned and this is done digitally, which allows for many more possibilities.
If a film has become distorted due to shrinking, this means it has been contaminated with moisture, and in extreme instances must be transfered to a new backing so that sprocket holes align correctly and then reprinted. Repairing damage on the film depends on the nature of the damage. Torn sprockets can be repaired either by reprinting the section optically onto a new piece, or by transfering the section onto a new backing. A torn frame used to mean that it would have to be spliced out (a single-frame cut is pretty imperceptible) or the shot substituted with an undamaged section from a duplicate source, but in todays world where the film is scanned this can be repaired by digitally painting in the damage. Dirt can be cleaned by running the film through a chemical bath and/or handwiping it. Dirt printed in the film (i.e. from duplication) you used to be stuck with, but this can be digitally erased now.
The pic above is a privately owned 16mm reduction that has seen better days. Aside from some pink shifting and surface dirt, it obviously is severely scratched (click for full rez). Scratches are handled in a variety of ways. If a scratch is on the base (the side with the plastic), it can be filled in through a wet-gate printer. A wet-gate printer will take the negative and then print a duplicate, but in the printing process special fluid is washed over the film. This fluid is called perchlorethylene and is the same stuff used in drycleaning. What happens is it fills in the scratch on the original as it is printed so that the scratched area has the same density as the rest of the unscratched base and thus the duplicate will print scratch-free. You can see a brief demo here. The downside is that the image can be softened, and sometimes wet gate printing just adds more dust. Today, wet-gate printing isn't commonly done on high-budget productions--the film is simply scanned and the scratch painted out digitally. Film preservation still uses wet-gate, however, since one could not hand-paint out every scratch digitally on many really old prints (or, for example, the frame posted above). If there was a scratch on the emulsion of the film, you were basically out of luck, since the photo-chemical image itself was scratched away, but today it can be repaired with digital tools as well.
How are films actually stored though, and what back-ups and safetys are made? Celluloid is prone to molecular change if not subjected to certain conditions; vaults and storage rooms should be temperature controlled to about 50 degrees fahrenheit, and with relative humidity of about 30 percent, to prevent drying out or moisture retention. In the case of Star Wars, at the time of the 1997 Special Edition Fox had the negatives in a giant underground facility converted out of a salt mine, somewhere in Kansas. Studios have millions of reels of film to put somewhere and they simply do not have the space to fit it all in any one central location at the main compound (underground caverns in Kansas also aren't prone to certain affinities found in Hollywood like, say, earthquakes).
Reading about all the various forms of damage and degradation mentioned previously, much of it inevitable, it must seem very apparent that back-ups are of vital importance. Damaged negatives can be repaired to one degree or another, but they must have high-quality back-up material to provide donor parts. Often, the negative is duplicated right away. So, you have the colour-timed interpositive, but also a pure copy of the negative. If the negative were to be destroyed or damaged beyond repair, you can use the dupe negative to make a new one (usually, the original dupe becomes the new negative, and then it gets duplicated itself to make sure there is still a backup). Their are a lot of really famous films whose original negatives no longer exist but nonetheless exist in great-looking form: Citizen Kane, for example, or Seven Samurai. Dupe negatives are still prone to all the aging pratfalls of the original, however; it won't protect against fading or shrinking.
There is a certain kind of duplicate that is immune, however: separation masters. Separation masters are three strips of fine-grain black and white (silver halide) film that are printed on dyes that will never fade. Each black and white strip records a spectrum of colour: red, green and blue. When the three black and white strips are re-combined on dyed layers, the result will be a perfect colour re-construction. Separation masters are usually made from the negative itself. As they age they may shrink and swell at different rates, which means they will mis-align, but digital tools have made this no big deal.
There are also fine-grain print masters. These are made as final back-ups in the event that all other material is lost. There are also low-contast print masters, which are printed with more mid-range information to account for the contrast that comes from duplication.
A final archival source, which has special significance to Star Wars, are Technicolor dye-transfer prints. Because these are made from dye-transfer, they not only have finer grain and better detail, but they do not fade. As such, while many archivists do not prefer to use them to create new masters because of their greater density, they are nearly perfect colour references, and can nonetheless be used as a source of a new master if need be. The colour balances varied a bit for each print, but they are pretty much the best record of the original colour unless the interpositive was made into seperation masters.
In certain cases, however, one must work with whatever is available. Especially for very old films, sometimes you can count yourself lucky if all that remains is a regular release print, spliced to bits and all scratched up. Metropolis is touring this year in a version of the film that was once thought lost; the missing footage came from a 16mm release print that someone randomly found in Argentina, where it had sat for some 70 years. The footage has been cleaned up digitally but will never match the rest of the footage, but this is simply all there is. Sometimes, entire films are lost, or only portions of them survive. The sad truth is that about half the films made before 1950 no longer exist.
Which brings us to the ultimate point of all of this: what do we have to work with for the Star Wars trilogy and how can the films be saved? After all, this site is called savestarwars.com. Will the original versions be among those scores of lost films? The most important trilogy of all time, in this day and age, lost?
If Lucas follows through on his word--yes. Let's just try to imagine how it would happen. Right now the original camera negative is on the verge of becoming unusable; Godfather, from 1972, just made it when Robert Harris restored it a year or two ago, but it can never be put through a projector or pin-registered mechanism or it will fall apart. So, around 2020, we might start to see the pieces of original camera negative in danger of no longer being the primary source. Around this time, the original 1982 VHS and Beta tapes should also be deteriorated to the point where their owners can't watch them anymore, and the first Laserdiscs should be rotting by then too (some of the 1993 Definitive Edition LDs have already gone, but that release was notorious for rot). The 1985 interpositive was already starting to go pink in 1993 when that release was made, the final home video telecine. By 2020, it will be pink, but salvageable, but all release prints would basically be unwatchable as-is, even Return of the Jedi's. By 2040 the IPs would have long gone red, and the original negative would basically have be put in the garbage. The final VHS tapes and Laserdiscs from 1995 would have rotted away, and the 2006 DVD would no longer be playable due to disc rot, and so it would not be possible to actually watch the films except through illegal copies or backups people have made themselves in the past and preserved over the decades. By 2060, all film materials not properly cared for would no longer be useable with current technologies. Except, hopefully, for the separation masters and the Technicolor prints, of which there are only two surviving at present (one owned by Lucas, another owned privately; see Technicolor I.B. Screening). It's not currently known how long they would last because they haven't existed as long as the timeline we are dealing with here, but in theory they should still be in prime condition. Assuming they aren't lost, destroyed, or damaged.
So then, we are looking at all our home video copies disappearing by about thirty years from now, although it will likely be less, at which point the negative will no longer be usable and the backups perhaps barely, with everything guaranteed to be gone before 2060, a mere fifty years away. That is likely to be within my lifetime. Scary. Separation masters and the Technicolor print(s) seem to be the cockroach of the Star Wars apocalypse though--when everything else is gone, they'll still be kicking. There is also a print of Star Wars in the Library of Congress that, if properly stored, should be good for decades to come.
But the rest is not gone though, not yet, and that is why it is imperitive to restore and preserve the films while the original material is still around. The longer we wait, the more doors will close for good. Let's look at what we have with us as of 2010:
-Original camera negative. Lucas claims this doesn't exist. Which is misleading double-talk. What he really means is that it's conformed to the 1997 Special Edition edit. The original pieces removed for the CG-enhanced shots are no doubt in storage. The original negative was fully cleaned and restored starting in 1995, and because parts even then were damaged or unusable had to have certain parts replaced with interpositive and separation master dupes, as well as having the optical transitions redone.
-Separation Masters. In theory, these should look the exact same as the day they were made. Portions printed from them were used to restore damaged parts of the original Star wars negative in 1997, which should vouch for their quality. Since Star Wars didn't have a duplicate negative made, this might have been why. Nothing is known on the sequels, but I must assume they exist similarly.
-Interpositive. The original interpositives were worn away with use, and probably would be very faded by now. Because of the wear on the originals, a final interpositive was made in 1985 for home video use; it is possible that the originals would have been destroyed at this point.
-Internegative. Internegative status is unknown, but in theory should be around if the Fox and/or Lucasfilm archives are as complete as I suspect they are. It is my judgement that any surviving would be unusable except as a last resort, however. There were no internegatives made in 1985 to my knowledge, since the IP was made only for video, and since there was never another theatrical re-issue after that date there would have been no need for a new internegative. If the original interpositives were worn by then, the internegative, which is what prints are actually struck from, would be unwatchable from both wear and fading, but it should still exist nonetheless. There was likely more than one internegative made over the years.
-Print masters. Fox and Lucasfilm have both been known to keep print masters of the films. If they were printed on low-fade stock, they should be usuable. Given the 1977-1983 time period of the original trilogy, they have probably not held up very well, except perhaps for Jedi. Digital tools could still restore them, however.
-Theatrical prints. Most theatrical prints get destroyed after they are recalled from circulation, except for a few kept for loan for special screenings. Since Lucas has recalled all original prints from circulation, and given his disposition towards these versions, I don't believe any exist through official sources, unless Fox held on to some. There are a number of privately owned prints known to me. Given their age, however, they wouldn't be of much use except as a last resort. In the early 1990s, Lucasfilm sent new prints (according to the projectionist) to the Senator Theatre in Boston (yes, that one) for a special charity screening, but the status of these is unknown.
-Technicolor Imbibition (dye-transfer) prints. I don't know how many of these were ever printed, but I'll bet you could count them on one hand. Lucas has one, and it was the colour reference for the 1997 Special Edition. An unknown person privately owns another, which was screened in 2010 at the Senator Theatre in Boston. If that print is any indication, these look the same as when they first came from the lab, assuming they have been properly cared for.
-1997 Special Edition material. Don't forget, there are 1997 interpositives of the Special Edition, plus perhaps print masters, and even theatrical prints would be unfaded and in decent condition. Although the enhanced moments are invalid, these could be useful as sources for 90% of the films if nothing better was available.
Given the above, fully restoring the Star Wars films would be relatively straightforward at present time, or a simpler form of presentation at lesser quality could be had. There are a number of avenues to take depending on ambition and budget.
1) The most complete one would be totally restoring the films from the original camera negative. Here is how it would be done. In the past, you would have had to re-conform the negative to remove the Special Edition edits, but this would result in frames lost from un-cementing and re-cementing the assembly, and it would wear it further. Today, the negative would be scanned in 8K and original shots could be pulled from storage and scanned likewise. The original shots would then be edited into the scanned Special Edition negative using non-linear digital editing to re-construct the original assembly. Further digital clean-up could be done to remove scratches and dirt that was not possible in 1997, although this is not strictly mandatory. Because the original shots were not restored for 1997, some may be so damaged that an interpositive scan is necessary instead. Then, Lucas' personal Technicolor print could be used as a colour reference to time the new Digital Intermediate. From there, a new 8K DI could be made, and a new 35mm negative printed out using an ArriLaser for future preservation and duplication for theatrical exhibition. Not rocket science. Pretty much the norm for restoration of classic films, which often have to edit together more than one source. I'm not totally qualified to put a pricetag on this, but I would say maybe a million dollars for the whole trilogy? To put this in perspective, the 2004 boxset sold $100 million on it's first day of release.
2) If Lucasfilm wanted, they could scan the separation masters in 8K, and time the composite with the Technicolor print if they are derived from the negative. If they are derived from the internegative then they can simply be scanned and presented as-is. But as this requires three scans and digitally overlaying three elements, which probably have each shrunk a bit by now, you might as well go to the negatives, it would probably be almost the same amount of work.
3) If Lucasfilm was feeling cheaper, they could turn to the 1985 interpositives. First step, do a 4K scan (I doubt they would resolve much higher than this). Now, judging by the 2006 DVD, these are the grainiest interpositives of all time for some unknown reason. You could present them as-is if you really didn't care at all, but if you are going to spend the money to do a modern scan you should at least make it more presentable. DNR is a major problem in high-def home video right now, which is why I would suggest someone like Lowry remove the excess grain so that the image detail remains intact. This is not mandatory, but it is ideal. There is some mild print damage that could be cleaned up as well. Finally, the interpositives would require some major colour correction. The Technicolor print should be used, as in all restoration cases, but beyond this there will be significant pink-shifting and desaturation that will need to be combated. I'm not sure if this version would be good enough to supplant the original negative, although given enough time I suppose it would. It would at least make a pretty good high-def release though, without much effort.
4) If one was really lazy, a print master could be scanned in 2K or 4K to create a new high-def master. I don't believe print masters are used very often, so perhaps this would require very little dirt clean-up. Ideally, it would need some grain reduction, but if you are going to go to a print master when there's all the higher-generation material you probably won't care to. Again I would point to the Technicolor print as a colour-reference, but realistically speaking the colourist would probably just eyeball it, biggest issue being getting rid of the fading. It wouldn't be that impressive compared to other releases, but you would have the original films in high-def. One could also substitute the Technicolor print itself as the source, which might require more work in the telecine process because of the density but less in the color-timing suite. Pricetag in either case would be something like a hundred grand, which is probably what Lucasfilm would make within 15-seconds of nation-wide sales of this.
Preserving Film: The Digital Paradox
This brings us to our final destination. I was interviewed about the film materials for Star Wars earlier this year by Ars Technica, and one issue they brought up was "why are we talking film preservation? Shouldn't this all be digital by now?" Well, no. Here's why.
Digital technology is only now catching up with film, but in certain respects it's still far from being as efficient. It perplexed me when Lucas shot Attack of the Clones digitally in 2000, because the technology wasn't ready by a long shot, and visually it unfortunately shows in that movie; one day, very soon in fact, digital camera technology will be ready, but if it is not then what is the point? The point is to use whatever is best at the moment and let evolution take its course, not use digital no matter how its readiness is. People's eagerness to embrace the latest and greatest new thing and say goodbye to "obsolete" technologies, coupled with electronics manufacturers more than happy to sell you an expensive new high-tech product, resulted in a premature embrace of a technology that is only now beginning to look viable as a format. As it applies to film preservation and digital capturing though, this has opened up a lot of debate that wasn't foreseen by many people.
One has to understand, film is the most robust, durable imaging technology ever created. All you need is a light to shine behind it and a surface to project it on, with the image quality still outrivaling most digital technologies in resolution and colour, and it'll last for decades, unlike a hard drive which wouldn't last more than a few years even if maintained. You can open up a film can from 1930 and it could still look pretty good today, eighty years later.
Moreover, scanning and restoration technology keeps improving. In 2002, if all the studios had scanned their libraries for "pristine, digital preservation", which would have cost billions of dollars and millions of man-hours, they would have had to throw it all away and start over today. That's because the films would have been scanned in either 2K or 4K; today, scanning is done in 4K and 8K, finally matching the actual resolution of film, and colour reproduction is better too. As a result, you always want to keep around your original film material so that future generations can re-treat it with whatever new technology becomes available.
This isn't just me being a luddite--the studios have realized this too. After they started converting their storage to digital, they discovered that the most affordable and reliable method was making new film backups. Even though film deteriorates with age, right now it's still the more stable and reliable method for high-quality archiving. The New York Times reported on this problem in December, 2008. The previous month, the Academy of Motion Picture Arts and Sciences released the results of a yearlong study on digital archiving. Their findings? Storing a digital master costs $12,514 a year, versus $1,059 to keep a film master.
But it's not just about price; as mentioned, digital technology is incredbly fragile, and also unreliable. Hard drives have to keep spinning, be checked and re-checked, and then the data has to be moved to new systems every few years, with multiple back ups. We're talking terabytes of data, about 25TB for a 4K scan of a two hour film, and that's just for the actual assembled cut of the film, never mind the raw materials which would consume hundreds, if not thousands, of hours worth of viewing. You also have to have the processing hardware to view it. Hard-drives eventually wear out (and can freeze up if not spun for a year or two), and so you'd have to replace each and every one of those terabyte-sized servers every couple years, an expensive undertaking just for the hardware (imagine, if you converted a studio library, how many hundreds of thousands of drives that would be). Keep in mind you have to make sure there's no data lost or corrupted in all those moves as well. And of course, every five or ten years, new technology systems arrive that must be adaptable to the older digital storage systems, or the older systems must be moved to newer, up-to-date systems. Ken Weissman, supervisor of the Library of Congress' Film Preservation Library, has written about the staggering logistics of such endeavours.
For instance, because NASA didn't migrate their digital data from the 1975 Viking space probe, they discovered in 1999 that they could no longer read it because the technology to do so simply did not exist any more. The issue of "read" or "no read" also means that digital files can be fine one day and junk the next, unlike other materials which degrade gradually; scratch a film and you have a scratched film, scratch a hard drive and you have a dead weight.
Films shot and stored digitally, such as Superman Returns which was filmed on video, face even worse prospects. According to the Academy's report, all those video dailies, sound files, online cuts of the film, and other raw materials cost about $208,569 a year to store, versus the mere $486 it costs to store physical film dailies, audio recordings, still photos, etc. in a cold storage vault. Movies shot on digital are being printed back onto film and stored, but some fear that as more and more theatres phase out 35mm projection and convert to digital, even this might cease. Digital storage has brought a lot of unexpected dilemmas that have not fully been looked at in the long term.
In future decades, hard drive space will become cheaper and more energy efficient, hopefully with standard systems developed, and perhaps then we can leave behind the physical record--but, philosophically, part of the act of printing masters on film is in the event of a disaster where you need some something visual, not electronic, and so perhaps it is best to hold on to film in some capacity forever. But, as it stands, digital preservation on a macro scale simply isn't feasible with current technology, and won't be for at least a while.
Digital technology has offered a multitude of restoration tools that would not be possible any other way, however. Restoration work is light-years ahead where it was in the 1990s. You can paint out any damage and scratching, do seamless splices, dial out grain, retrieve colour that was thought lost, combine multiple sources in the same shot, and manipulate contrast and colour in ways impossible using photochemical and analog ways. Digital technology has totally changed the restoration game. Which is why the current situation is a perfect marriage of old and new technology; scan and restore digitally, and then print and preserve on film. Film printed on low-fade, polyester film stock properly stored in controlled, cold storage vaults should last a couple hundred years. It'll be up to our great, great, great, great grandchildren to take things from there.
Further Reading
-Richard Hains on the history of film preservation