New Progress in Film Preservation
|Read more at|
The 70mm Newsletter
|Written by: Rick Mitchell||Date: 19 November 2005 |
|On June 14, 2005, the Hollywood Section of the Society of Motion Picture and Television Engineers ended its 2004-05 year with the above titled program in the Linwood Dunn Theatre at the Academy of Motion Picture Arts and Sciences’ Mary Pickford Center in Hollywood. Since film preservation became a hot topic in the early Eighties, the SMPTE has regularly presented programs on this subject. The evening’s co-hosts were Ralph Sargent of Film Technology Laboratories, one of the foremost firms involved in this work, and Richard P. May of Warner Bros. who has been overseeing such work with the MGM and later Turner library for twenty years.|
Mr. Sargent opened the evening with MOVING IMAGES FOR THE FUTURE, a short his company had made in 35mm CinemaScope, especially for a 1999 SMPTE Film Conference, which, in the classic “a picture is worth 1,000 words” tradition, illustrated the state-of-the-art processes involved in film preservation and restoration at that time.
It should be noted here that those two terms are often confused. Restoration refers to preparing surviving pre-print materials: original cut camera negatives, fine grains or interpositives or separations for color films, internegatives, or even release prints for preservation; repairing tears, replacing missing perforations, filling in missing material from other sources. Once this is done, new preservation elements are struck and stored under exacting temperature and climate conditions. Mr. Sargent reported that properly processed and stored nitrate film was estimated to have a shelf life of 600 years, polyester film 1,000 years, but acetate based film only 70 years. Since acetate has only been used for 35mm for about 54 years, I forgot to ask if that figure was based on aging tests or the condition of the 16mm, 8mm, 9.5mm, and other small format stocks Mr. Sargent’s company has been successfully blowing up to 35mm in recent years. These have always been safety stocks and some of the samples Mr. Sargent exhibited at an SMPTE meeting a few years ago were from sources over 70 years old.
The biggest change cited was the increasing use of digital technology to deal with problems that couldn’t be handled photochemically. Digital techniques have been used to clean up sound tracks since their wide spread introduction into the industry a decade ago, as illustrated in the film and further described by guest speaker John Polito of Audio Mechanics, but their application has been fairly limited until recently due to the combination of cost and the amount of storage and processing space required to do the work at an acceptable resolution to match 35mm film.
Disney pioneered this a decade ago using digital techniques to clean up the original negatives of SNOW WHITE AND THE SEVEN DWARFS and advanced techniques for some of its other films as well as MARY POPPINS for DVD. Other preservationists have resorted to it to fix large rips and other defects in individual shots. However, 1 and 2K scans and outputs were considered to be of too low a resolution for the preservation of an entire 35mm live action feature.
In the last two years Warner Bros. has attracted attention with its proprietary process of scanning nitrate three-strip Technicolor negatives and digitally correcting them for shrinkage, which leads to a much sharper image than is possible photochemically. There has been some controversy over this because the work so far has been done at 2K, which is acceptable for video, but those who’ve seen a film printout of SINGIN’ IN THE RAIN, the first film done this way, were not happy with the result. The contemporary state of this issue was covered in the clips presented in the second half of the program.
First, to illustrate what can still be done photochemically, four clips were shown from SUDDENLY, LAST SUMMER (Columbia; 1960), courtesy of Grover Crisp of Sony. Damage to frames of the original negative was corrected by replacing those shots with dupe negatives of clean versions taken from a fine grain made before release printing began. This has long been a common technique (see appendix), with the replacement footage usually quite obvious because of their higher contrast and increase in grain. These clips illustrated how modern processing techniques and stocks make such insertions virtually undetectable; in fact the last clip used dupe replacement negatives made 40 years ago.
Barry Allen of Paramount then exhibited two works in progress from films his company inherited when it was bought by Viacom, which had previously bought “Republic Pictures”, nee National Telefilm Associates . One of the four biggest titles in this package is JOHNNY GUITAR (1954), whose original negative had been deemed no longer printable and junked in the Sixties, with all subsequent prints struck from a not very well made internegative from that time. He showed a recent test clip from this internegative which looked worse in terms of both sharpness and color than the 16mm print I got through the Nostalgia Merchant 25 years ago. As seems to have generally been the case back then, the separations made at the time also had problems, but he decided to borrow a leaf from Warner Bros. and see what results he could get by scanning and attempting to reregister them digitally. The results were a great improvement, though Mr. Allen was not satisfied and plans to do further work.
He also presented an illustration of how different print stocks can affect the results with a scene from the original nitrate negative of John Ford’s RIO GRANDE (1950) printed on standard Eastman B&W print stock and the East German Orvo stock, which, was reported to reproduce the quality of nitrate with its higher silver content. The Eastman stock had a greater contrast range with deeper blacks while the Orvo had a more even, but, to me, flatter looking contrast range. The audience was asked to vote on which it preferred, and opinion was evenly split, probably based on their own individual tastes in B&W. Film Technology did the lab work.
Richard May then showed how potentially problematic elements could be made to work in restoring missing footage, with some clips from THE MASK OF FU MANCHU (MGM; 1932). The film originally had some lines of dialog that were not considered politically correct when the film was reissued in the Fifties and the shots involving them were physically cut from the original negative. They remained in a composite 35/32 negative that had been made to yield 16mm prints, and the picture was blown up from that source, as well as the track sections re-recorded, for the restoration by Film Technology. The image was only slightly softer than the 35mm original and the tracks matched quite well. 16mm prints and tracks of varying quality have been used in the past when no other source was available, but I’ve seen and heard nothing as good as this.
Mr. May then presented two clips from recently restored silent films whose negatives had been thought to have been junked years ago. TELL IT TO THE MARINES (1925) was from the original negative and had been printed with an amber tint. Shown at full 1.33:1 aperture , it was very sharp and was beautifully accompanied by a gentleman from Film Technology whose name I didn’t catch, who’d earlier come out in a Chinese outfit with a gong to introduce the FU MANCHU clip. Film Technology did the lab work on this clip.
Mr. May then showed the ending of THE BIG PARADE (MGM; 1925), the subject of a major restoration by CFI/Technicolor and a re-“premiere” with a live orchestra at the Goldwyn earlier this year. The original tinting and toning guide sheets were still in the files and used to recreate the look in printing. A print was made off the original negative as well as new preservation elements on color intermediate stock. The excellent clip used was from the dupe negative and had a sonically cleaned up version of the music track added for a 1931 reissue. Though not part of the clip shown, digital technology had been used to add red to just the cross on a red cross truck which had been hand-painted in original 1925 prints of the film.
The evening ended with the “June Is Bustin’ Out All Over” number from Schawn Belston’s recent restoration of CAROUSEL (20th Century-Fox; 1956), whose repremiere I wrote about recently and more details of which can be found in a superb article by Robert S. Birchard in the current issue of American Cinematographer. That film’s original negative had faded too badly for a photochemical restoration so the decision was made to do it digitally, but scanned in and out at 4K. As I noted, problems with the original photography make it difficult to evaluate the degree of success though the results look better than any print from an anamorphic original that I’ve seen that went through a 2K digital intermediate.
The consensus of the presenters was that the bulk of preservation and restoration will still be done photochemically, because it’s cheaper and and the end result is likely to be preserved on film. Film is still the only guaranteed storage medium for both picture and sound, adhering to standards that go back 90 years or more so that a negative shot that far back can be printed on modern equipment just as a print from back then can be run on contemporary projectors (allowing for shrinkage, of course.) At present, not only are there no standards for digital technology, but hardware and software changes raise questions about the ability to recover digitized material in the future; the longevity and stability of the media on which that material is stored is also a subject of debate.
In recent years I’ve seen an amazing number of new prints of black-and-white films from the Thirties and Forties made from dupe negatives that look better than the nitrate prints off the original negatives I ran in my USC days. Ten years ago we were lamenting the loss of most of the Fifties films shot on Eastman Color negative because of a combination of dye fading and badly made separations, yet in recent years amazing new prints have been struck off those original negatives, followed by new separations made to today’s exacting standards. What remains of our film heritage is in good hands, to date.
|Further in 70mm reading:|
Rick Mitchell - A Rememberance
|From conversations I’ve had and things I’ve overheard, some people are still confused about certain past industry practices that have hurt and helped restoration and preservation.|
One important point deeply affecting film preservation that astounds many people today is that prior to about 1969, except for those made by the Technicolor dye transfer process and some other arcane color processes, almost all release prints were made off the original cut camera negatives! The number of prints varied depending on the re-lease pattern and popularity of the film, but between 250 and 300 for American release seems to be the average. I recall one source stating 300 for domestic release and an-other 150 for foreign release and Robert A. Harris (who was in attendance at the pro-gram) and Jim Katz have the paperwork confirming that 382 color positive prints were made of REAR WINDOW at the time of its original 1954 release. (At a screening of FRENZY (1972) Alfred Hitchcock stated he never liked the softer image of Technicolor IB and was happy that they were now making sharper color positive release prints from CRIs; at that time the lab was actually doing both.)
Naturally this subjected the negatives, dry printed at a top speed of 350 per minute as compared to today’s 2,000 per minute from a timed spliceless polyester internegative, to the rips, torn perforations, emulsion gouges, etc. that have to be corrected in restoration. In fact, it’s amazing how many negatives have held up as well as they have.
The reason for this practice was a concern for image quality that appears to have begun around the end of World War I. Prior to this time it was a common practice for companies shooting outside New York to project the developed negative as dailies and edit the film in that form, sending the spliced result to New York for release printing, as almost accurately depicted in Richard Attenborough’s CHAPLIN (TriStar; 1992) (the cuts were too on the nose and there were no zoom lenses in 1915), which would also explain the disastrous first screening of THE SQUAW MAN (Lasky; 1914).
I’ve not been able to find any official historical documentation, unless there is something on the subject in early SMPE Journals that are missing from various local libraries, but after the industry settled in the Los Angeles area and began using more professional laboratory techniques, they began the practice of pairing two cameramen with their cameras as close as possible and with the same focal length lenses, one to shoot the negative for the domestic release, the other for the foreign. Once the domestic version was approved, the foreign negative would be cut to match, scratch intertitles inserted, and the results sent, usually to London, for foreign printing operations. Because foreign countries were more respectful of films than Americans, many silent films have been preserved from these foreign versions. (In the late Forties and early Fifties, 20th Century-Fox had a variation of this practice in which two takes of each set up would be circled. Once the domestic version was approved, the first assistant editor would get experience by editing a foreign version using these alternate takes. Other studios may have also done this, but my source for this information was Bud Hoffman, nephew of Fox’s chief editor at the time, Barbara McLean, and her assistant in the early Fifties.)
Obviously this type of thing could not be done for talkies as two or more cameras were being used just to get the simultaneous coverage needed for the domestic version. Unfortunately the intermediate stocks available at the time were not deemed good enough to make dupe negatives, so initially many companies would shoot French, Spanish, and or German versions of select films, an expensive practice killed by the Depression. Since improved intermediate stocks were needed for optical effects that could no longer be done in camera, they began to appear in the early Thirties and some companies did begin making dupe negatives for foreign release. In other instances, the original camera negatives would be shipped overseas for printing after all domestic prints, and some type of protection element, had been made. This was a dangerous practice and legend has it that the original negatives of THE PRISONER OF ZENDA (United Artists; 1937) and THE BEST YEARS OF OUR LIVES (Goldwyn/RKO; 1946) were lost when the ships freighting them went down. For domestic distribution of a foreign, usually British, film, the studios usually brought over the original negatives and often inserted specially shot material to accommodate the Breen office. Because few prints were needed for foreign language films, they were often made and subtitled in the country of origin though often inferior dupe negatives were sent over on others.
Although, as one can see from the quality of optical effects dupes, by the late Forties, black-and-white intermediate stocks and processing techniques could yield dupe negatives good enough to release print from, the American industry continued to print off the original apparently because it was the way they’d always done it. The color negative stocks introduced in 1950 created other problems. Initially only Eastman introduced a separation and complementary internegative stock for titles and opticals and the results still give film restorers nightmares. Eastman’s introduction of the 5253 multi-layer IP/IN stock in 1956 was not much of an improvement (I don’t recall whether Ansco, Eastman’s only competitor in the US at the time ever introduced an intermediate stock.) For foreign release, the prints were either made in New York (only Technicolor, MGM, and Columbia are known to have done their release printing on the West Coast), from original negatives shipped to Europe, or the foreign release printing subcontracted to Technicolor, who made matrices which were shipped to its London and later Rome and Paris plants. When the studios picked up a foreign made film for domestic release, they either brought over the original negative or also subcontracted the work to Technicolor. For their American release, as with black-and-white, some arthouse foreign films used imported subtitled prints from their country of origin, some were printed by Technicolor, but many were printed from internegatives and looked awful, very contrasty and grainy with distorted colors.
Solving this problem was reportedly one of the reasons Eastman developed its now controversial Color Reversal Intermediate (CRI) stock, introduced in 1968. Within a year this was also being done for domestic releases and by the early Seventies had become standard, with only about ten prints being made off the original and used in a films’ first run presentation in major theaters in Los Angeles, Chicago, and New York. One problem with CRI was that unless it was overexposed for printing at a high light, it would not yield deep blacks (I learned this from Jim Elkin at Universal Optical and Tom Shaffer of Technicolor in the mid-Seventies). Eastman’s introduction of 5243 IP/IN stock in 1978 corrected this as well as the problems with the earlier IP/IN stock and it had supplanted CRI by 1980.
There is also the question of the origin of the practice of making what would become protection elements, which apparently were not made on silent films at the time of negative cutting as most films were considered worthless after their initial theatrical runs. (According to the late Jack Rush, former head of Universal’s stock footage library, that company’s late Forties owners intentionally junked the negatives of its silent films to make room for current productions in their New Jersey vaults.) During the silent days, if a section of negative became damaged, the section or shot was cut out of the negative and thrown away.
This could not be done with sound films, of course as the picture and track negatives had to be kept in sync. Vitaphone actually sent out rolls of black leader and instructions to projectionists as to how to use the numbers inked along every foot of the print to replace the right amount of damaged footage; Scott Eyman reprints these instructions in The Speed of Sound. This was also done with damaged negatives of both disc and optical sound films, using clear leader which printed as black. This explains the black frames that often occur in some older films, and can be objectionable if the damaged section is unusually long; I once saw a film that was blacked out for over a minute.
I’m not quite certain whether I heard this from Sid Solow in his USC class or in passing at one of the many SMPTE programs I’ve attended, but I recall someone alluding to some unnamed party at some unknown lab coming up with the idea of using the contentious practice of duping, which unscrupulous labs and distributors had engaged in since the industry’s turn-of-the century vaudeville days, to solve the problem. Duping was making a negative off another company’s release print and selling it as their own, and in this instance involved making a dupe negative of the damaged shot off a print and cutting it into the original. Soon first struck prints were being set aside for this purpose, then they were being specifically developed to lower the contrast and grain to yield better dupe negatives, and finally, as noted earlier, specific stocks were developed for this purpose. And, as noted, because such intermediates made it easier to make 35mm dupe negatives for export and printing elements for smaller gauge prints, making them would become standard practice and the primary preservation element used today.
Aside for those who might be interested: there is a shot duped from a 70mm work print in STAR TREK-THE MOTION PICTURE (Paramount; 1979). In principal photography, all shots that were to incorporate optical effects were shot in 65mm, except in all but one instance, blue screens. As a 65mm camera had been set up in the rafters for a high shot of the engine room to which effects were to be added, it was decided to shoot a non effects shot from that angle in 65mm as well. Naturally, in the furor to finish the film’s effects work, this straight reduction shot was forgotten and when they finally got around to it, the negative couldn’t be found. For whatever reason, instead of using another take, the decision was made to dupe the 70mm work print, which had only been projected once. Naturally, as assistant film editor responsible for cutting in the effects shots, among other things, I know which shot it is, but no one else seems to have spotted it.
National Telefilm Associates, or NTA, was formed in 1954 by two persons from 20th Century-Fox distribution and licensed the pre-’48 Fox films for tv syndication. Over the years they picked up a number of orphan libraries, such as the Bank of America financed Universal films of the late Forties on which the bank had later foreclosed, the Pine-Thomas Bs for Paramount, the Cary Grant library, etc., but their biggest success was the Republic library, which was so popular that in the Eighties, they changed their corporate name to Republic.
Though advertised as being in “Trucolor by Consolidated Film Industries”, JOHNNY GUITAR, like almost all “Trucolor” films released after 1951 was shot on Eastman Color Negative, rather than the earlier two-color bipack method. According to the late CFI president Sidney P. Solow when I took the color processing class he taught at USC in thee Seventies, release printing was done from separations because the lab was used to working that way, especially for timing, with the two-color system, even after they’d begun printing on color positive stock.
The only exception to color negative photography was MONTANA BELLE, originally made in 1948 by an independent producer at Republic for intended distribution by them. However, because the film co-starred Jane Russell, Howard Hughes bought it and finally released it through RKO in 1952, apparently the last two-color Trucolor release.
With the introduction of optical sound to 16mm in 1932, it was considered an educational tool as well as a potential downmarket source for theatrical films. For the previous nine years, prints on 16mm and similar smaller than 35mm gauges had been made by optical reduction printing from a 35mm negative directly to 16mm print stock. Unfortunately the laboratory processing machines of the time used rollers with teeth that engaged the perforations of the negative and print stocks and processing 16mm required changing those rollers whenever the gauge was to be handled. According to an ad his company published in the American Cinematographer in the early Seventies, Byron of Byron Laboratories came up with the idea of optically printing the 16mm images to a 35mm dupe negative stock with 16mm perforations inside the normal 35mm ones. The dupe negative could then be contact printed to 35/32 print stock which would then be slit down to 16mm after processing. In fact, two rows of 16mm images could be printed, one reel in one direction, the other in reverse, with optically reduced tracks for both reels also printed on the dupe negative. With everything pretimed to print at one light, 16mm prints could be made at any lab in the world.
Technicolor would come up with its own variation of this for its dye transfer process, a row of 16mm images printed down the center of a 35mm matrix, called single rank, later two images side-by-side for double rank (they reportedly also had a Super 8 version called quad rank). Although in the Forties, they experimented with a dye transfer track from the yellow matrix, they generally went with a silver track which had to be added in a separate pass as in 35mm. I’ve never been able to get a definite date as to when Technicolor began making 16mm IB prints, but some sources have cited a 1944 Walter Lantz cartoon as the first.
Prior to Eastman’s introduction of its 5253 IP/IN stock in 1956, the only other way to make 16mm color prints was by reduction printing from a 35mm print to 16mm Kodachrome or Anschochrome.
The Academy’s Dunn and Goldwyn Theaters are among the few that can show silent films at full aperture, where such prints exist. Normally, they’re shown via 1.37:1 Academy mask, which cuts off the top and bottom, as well as the left side of the frame to restore the basically square frame after optical sound was added. Of course silent films to which tracks have been added generally have to be shown this way as do early sound films shot for disc systems. Last year, the early optical sound films SUNRISE (1927), IN OLD ARIZONA (1928) and THE LOVE PARADE (1929) were shown at the Goldwyn Theater at the “Movietone” aperture of @1.15:1, the height of the silent frame but with the left side taken up with the sound track.
Tragically 16mm preprint elements and even 35mm elements on such films, when handled optically, were made on optical printers with Academy aperture masks, resulting in images off center to the right and with the top and bottom of the frame cut off. In recent years, where the full aperture negatives still existed, they’ve been optically reduced to Academy aperture, as Universal has done with DRACULA (1931), for example.
|Go: back - top - back issues - news index|