An Interview with the Man Behind Sony Pictures` Restoration of MOTHRA, THE H-MAN, and BATTLE IN OUTER SPACE Author: Steve Ryfle Special Thanks Oki Miyano
A SciFi JAPAN EXCLUSIVE
The ICONS OF SCI FI: TOHO COLLECTION marks the first time that a Japanese genre film—in this case, three films—has been digitally restored for home video release in the United States. Overseeing the process, which lasted more than two years from start to finish, was Michael Friend, the former director of the Academy of Motion Pictures Arts and Sciences film archive, and a highly respected film preservationist, archivist, and historian. Friend, who currently works for Sony Pictures Entertainment in Asset Management and Film Restoration and teaches at UCLA in the Moving Image Archive Studies Program, has been involved with some of the most prestigious film restorations ever undertaken, including Akira Kurosawa’s RASHOMON, a joint project of the Kadokawa Culture Promotion Foundation and The Academy of Motion Picture Arts and Sciences. In this interview, conducted in March 2009 (and portions of which are excerpted on the MOTHRA and BATTLE IN OUTER SPACE feature DVD audio commentaries), Friend explains both the technical process and the rationale behind Sony’s decision to restore these three classic science fiction films from director Ishiro Honda. He also offers his insights into Honda’s cinematic legacy and the director’s place in science fiction cinema history. Steve Ryfle: How was the decision made to restore these three classic, yet somewhat obscure films? Michael Friend: As you know, these films really aren`t all that obscure. BATTLE IN OUTER SPACE was the basis for Roland Emmerich`s INDEPENDENCE DAY, and MOTHRA has of course been a cult classic for a long time, and H-MAN is a kind of Ur-text that a lot of people point to when they think of these kinds of films—Emmerich used the beginning of H-MAN for his own GODZILLA. So the films of Honda that are lesser known—that is to say, the films other than the Godzilla series—are still very important to us as sources of inspiration. The Sony library has a number of films by Ray Harryhausen and his collaborators, which are done in the style of stop-motion animation, and which resemble in some ways the content and also the skill level or technical level of the Honda films. Naturally these Japanese films were of great interest to us as an integral part of our collection.
Columbia Pictures during the 1950s and 60s, as an international exhibitor, was very interested in getting access to these kinds of film because they fit into Columbia’s marketing plans and approach to the audience. The Honda films, even though we don`t own the most famous ones, are still very important to us. Historically, they reflect the studio`s international distribution strategy and augment the material that Columbia itself was producing in the late 1950s. They focus on the youth audience, and are a significant part of the film-going experience of that era. Science fiction was not taken seriously in those days; it was largely regarded as a juvenile genre. On the other hand, science fiction proved to be the ideal medium for reflecting the terrifying, unprecedented technologies that emerged in the post-war era and the Cold War. And Mr. Honda, who seems to have always been concerned about the future of humanity, spoke to that future in the youth audience. Sony preserves all the films in its library, and in the course of our normal technical evaluation, the decision was made to schedule all three films for remastering. It seemed obvious that our fan base would want to see all three films as a coherent and interesting body of work. We searched our database of physical elements, found the existing elements for the films, brought them into our laboratories and began to assess the quality of these elements. And we decided to try to improve the quality of the materials by asking Toho to make new interpositives from the original negatives, and new soundtracks. We also asked for a print from the original negative with the original soundtrack so we would have a record copy that would tell us how the film was supposed to look and sound. The original elements have all suffered from deterioration over time, so these new elements were not ideal. But technology tends to improve over time, and processes such as wet gate printing as well as finer grain film stocks allow us to improve the quality of restoration work. The new interpositives and soundtracks we received from Toho were in some ways superior to what we had 20 years ago or 50 years ago. We brought the new elements to the lab, and compared them with the existing elements. From that point, we made a plan for restoration of each of the films.
Steve Ryfle: What were the particular challenges of each film? Michael Friend: The challenges were roughly the same for each film, because the films were made in the same era using similar techniques. The production work tends to contain more dirt (especially in the titles and opticals). There is degree of color fading with Eastmancolor or Fujicolor negatives of this era. And then there are scratches and scuffs and tears that occur when you print a negative, especially if you print it on older equipment and print the original negative many times, as was the case in the era when these films were new. So there is a certain amount of physical damage as well as photochemical fading in these elements— these are the "givens" of restoration, the kinds of problems that we find in all films. Photochemical fading is not uniform. The optical units were much more faded than the principal photography, so the main and end titles and all the special effects required an enormous amount of attention to try to bring the color back to what it originally was. We also had the problem of dirt and damage in the original negatives. Toho addressed this problem as well as they could by cleaning the negatives and by wet gate printing them, so Sony got new elements that were in some ways as good as those made in the original days of release. But unfortunately once dirt gets into the emulsion it`s very, very difficult to remove it. So we received new interpositives that had some of these traces of dirt and also all the other physical problems: scuffs, scratches, and tears. All three films had these problems. We compared what we had in our inventory, our old interpositives and audio elements, with the new interpositives and soundtracks, to try to determine how to put together the best and most complete version of each of the films.
Restoration requires a network of experienced technicians. We are fortunate to have such a network available to perform complicated projects such as the Honda restorations. This project would not have been possible were it not for the extremely capable technicians at our various laboratories who executed the workflow that we have been discussing. The film work was done at Cinetech in Santa Clarita. Kim Gott, Cinetech’s timing manager oversaw the entirety of the project. Tom Nobili timed MOTHRA and H-MAN, and Kevin Warr timed BATTLE IN OUTER SPACE. Dave Osterkamp, the director of special projects, coordinated the complicated work of sequencing and printing the different elements. Sound Restoration was done at VidFilm and Chace. Digital restoration was done at Ascent Media Group in Burbank. Tess Walsh, one of the most outstanding restoration technicians working today and a veteran of many important Sony restorations, was the colorist on all three features. Steve Ryfle: When you say complete versions, that brings up the question of missing frames and footage. How much missing material was there, and how did you determine what was missing? Michael Friend: That`s another interesting problem. There really isn`t missing footage but there are missing frames because the practice of removing damaged frames rather than trying to repair them, and because frames were lost in the process of editing the films over time. This was a common practice for many, many years in both the Japanese and American industries; if you have a bad frame that is likely to cause a tear or other damage in printing you might remove it in order to make a better splice and a more printable element. As a result of that practice, our Columbia interpositives from the sixties and the eighties are actually a few frames longer than the prints and interpositives from Toho`s original negative. Occasionally we replaced shots or sections using one of the Columbia interpositives when we found there were significant frames missing from Toho`s new interpositives. It was what we call a checkerboard process, going back and forth between older elements that were still in relatively good condition, and new elements which were in great condition but might have missed a frame or two. If that missing frame happened to be on a close-up of a person speaking, it seemed important for us to replace the footage because it is important to see the person finish speaking on screen and in sync. But this was just a shot here and there. As far as I know there are no really major losses of footage in any of the original negatives that Toho holds on these features.
Steve Ryfle: So you used elements from the American versions to replace missing frames in a Japanese version? Michael Friend: Yes that`s correct. First of all, the difference between the Japanese and American versions of BATTLE IN OUTER SPACE are not significant; it`s only the main and end titles, so the body of the picture is more or less identical. There are several differences, something like 15 or 18 differences in MOTHRA, including most significantly where they go through a cave to find the twins in the beginning of the film. In H-MAN there`s an exotic dancing sequence that`s in the Japanese version but not the American version. But other than that, most of the footage is identical. Steve Ryfle: There was also some content near the end of MOTHRA that was removed from the American version. Michael Friend: Yes, that`s another difference, which is slightly shorter than the previous one I mentioned. And that change appears to have been made for cultural reasons. There`s a quasi-religious theme to the end of MOTHRA which the American editors did not feel would play well with their audiences. But we wanted to make the Japanese version available, because we wanted to present the original achievement and also because it’s a very interesting ending. We wanted people to be able to appreciate and study the differences in the ending, because Mr. Honda`s original message is deeply felt and may be more appealing today than it was in its original day. But Sony’s commitment to having the complete original version is what compelled us to go as far as we did in assembling all three of the films in both of their versions. Steve Ryfle: As far as the other two films are concerned, what did you use for the U.S. elements and how much restoration was involved in those?
Michael Friend: The American elements, the U.S. elements for H-MAN and MOTHRA were slightly shorter than the Japanese elements. We licensed these films just after they were originally made and acquired and edited interpositives and soundtracks from Toho in the early sixties. I believe we have the original elements that we ordered from Toho, the Columbia IPs, in our library. These have faded substantially because they were not initially cold stored and also because the film stocks in those days were more fugitive than our contemporary film stocks as far as color is concerned. There have been several edits of these films for theatrical, television and home video distribution, so there are variances between the sets of elements. We found that the opticals faded more in the Columbia elements than in the original Tokyo elements but the principal photography faded less than in the Tokyo elements, so we had to checkerboard back and forth between the optical effects and principal photography and take some scenes from the old Columbia IPs and the some material from the new IPs to make the best versions. And of course it was very difficult to balance the faded film because it drops off in different directions in terms of its color fading, and often, photochemical technique cannot compensate fully for this fading. Additional digital work was required to balance these faded areas. The film preservation part of this was not entirely successful, but we were able to take the project quite a bit further once we got into the digital realm. That said, it is important to remember that the transfer element was an internegative, a third generation element, rather than the original negative, or the interpositive which would be the typical transfer element. A transfer using the original negative and some replacement frames from interpositive would yield a much better image. Steve Ryfle: In terms of scope and scale, how does the restoration of the Honda films compare to major projects? Did budget restrictions limit how far you could go? Michael Friend: We try to give every film the treatment that it requires, and while it is probably true that some of the biggest films have larger budgets to support what`s needed, in some cases they will also require a lot more work. The work that we did on the Honda films was as extensive as it could be using the technology that we have available to us today.
There is no finality with respect to preservation, though. As technology advances, and as display requirements change, we’ll revisit these films. Just as our current restoration is superior to than anything we were able to do in the past, a future restoration will come even closer to a perfect recreation of the original achievement. If no one had ever printed these films they would probably be pristine and it would be much cheaper for us to restore them up. The more popular films, unfortunately, have more damage and the Honda films have been quite popular, so the elements have received a substantial amount of wear. But Sony devoted all of the resources that could be reasonably brought to bear at this time to restore these films. What we did for the Honda films follows a standard pathway for restoration. We began by identifying the best available film elements. We did not have access directly to the original negative but we got high quality interpositives from Tokyo and we used the old Columbia IPs that exist in our library, and made the best possible internegative from the two sources. Then we scanned that internegative at 2K (or 2048 horizontal lines) to .dpx files, which were then restored digitally. The digital cleanup process begins with an automated technology called ASC3. This software does what we call “dust-busting.” It identifies dirt particles by looking frames before and frames after and finding items that appear to be aleatory, spurious, and not an intentional part of the image. This is a technique we use carefully because it has potential to remove actual picture elements along with the images of dirt particles. It works well when you have random dirt against an unchanging background, but of course is more challenged by a scene where there is a lot of movement or a lot of randomness. We monitor and adjust the level of application of this tool so that we only remove the dirt and not genuine picture components.
The next stage is to apply DRS, which is a semi-automated cleanup method for fixing larger gross physical imperfections that have been captured in the scanning process: mislights, glue marks, warps and movement that occur at the edits, tears or larger scratches Then we proceed to a third phase, which we call graphics, which may use an Inferno or a Mac-based system to identify and repair damage that could not be successfully addressed by the initial phases, such as uneven field problems, color breathing, and various kinds of intermittent instability, such as flicker. There are a number of digital restoration packages available now, and this technology continues to improve. Once we`ve submitted the picture to those three levels of cleanup, we then size to the correct aspect ratio, which is 2.35:1, the standard mono anamorphic format before 1970.
At the same time, we restore the soundtrack. We compare all of our sound elements, in a process similar to our picture restoration procedure, and select the best resources for the project. In the case of the Honda films, the magnetic audio sources no longer exist, so the best remaining materials are the optical soundtrack negatives that are at with Toho. We worked with optical soundtrack positives (made from Toho`s track negatives), which were digitized for restoration. The sound laboratory eliminates high and low end noise (hiss and rumble), deletes clicks and pops (which are the result of splices, dirt and damage on the optical track), addresses other problems (poor sound levels, distortion, etc) and conforms the soundtrack with the picture. In this case, sync can be a relative thing in terms of the English soundtrack, because the film is in Japanese and dubbed into English, so most of the picture is not truly in sync, although all of the effects are in sync: explosions, crashes, and all of the other-worldly sounds are in sync. And there is good lip sync in most of the Japanese version. So once the new digital soundtrack has been QC’d, we deliver it by electronic means to the laboratory where the digital work is being finished. After we have restored digital picture files, we time (or “grade”) the color, so that the digital image reflects the look of the film at the time of original release. We try to achieve that original look rather than “modernizing” or intensifying the color. It is possible to go too far with digital technologies, and to make the image and sound cleaner and brighter than the original, but this treatment eliminates the authenticity and balance of the film. Our goal is not to make the films look like contemporary cinema. We try to protect the original achievement to the greatest degree possible and present that to the public. That is our restoration philosophy.
We had some problems with MOTHRA because we were not aware of all of the changes between the Japanese and American versions; we hadn`t tracked those carefully because we didn`t have all of the Japanese elements; Columbia didn`t distribute those versions in America. It is important to know that these films were edited, in Japan and America, in several ways, since the original release. Our objective was to produce the most complete version of the original Japanese and American releases. Steve Ryfle: Other than comparing and contrasting, were there any other resources at your disposal to log the differences between the Japanese and U.S. edits, and make sure you were reconstructing the films correctly? Michael Friend: Once we became aware that we would need to identify the different versions there was really no trouble. Oki Miyano, an archivist and scholar of Japanese cinema, was kind enough to supply us with many different video copies of the films so we had some reference. The differences, apart from technical quality, are mainly deletions from the original. The differences in MOTHRA, as we mentioned, are minor and yet significant. You can see why the exotic dance sequence was removed from the American version of H-MAN; it elongates the picture without advancing the narrative. But with MOTHRA it is more a kind of ideological blindness. The editors did not feel that the audience was particularly interested in getting into a deeper message. It`s not a huge difference, it`s a subtle hint, but it is part of the Honda signature. Having that available now is a wonderful thing for audiences, especially for anyone who wants to look at these films a little more closely. Steve Ryfle: Was there a particular scene or scenes in each of these movies that was particularly damaged and needed extra attention? Most of the films were not subject to heavy damage, but two of the key special effects sequences in BATTLE IN OUTER SPACE were very heavily damaged. These are classic effects scenes, some of the most action-laden sequences, the ones with the most interesting model work. One is the scene where the Earth base is being attacked and destroyed. There`s also a lot of damage in the opening sequence of BATTLE IN OUTER SPACE, through the title sequence and into the section of the destruction of the space station. These scenes were probably excerpted for a program on the history of science fiction, and were section-printed and damaged in the process.
Steve Ryfle: With all the digital tools available today it`s possible to make an older film look better than it ever did, to not only restore the color and image quality but also to correct imperfections in the photography, the effects, and so on. What is your view on this, and how did it affect your work on these three films? Do you see this as going too far, or a dishonest approach? Michael Friend: It`s not just dishonest, it`s a philosophical impossibility in a sense. No one would seriously think they could improve Giotto`s frescoes by correcting his perspective or intensifying the color, or somehow rendering more detail in his figures. If the original artwork has value, then preservation’s function is to make it look as it originally did, with its limitations as well as its achievements. You cannot make a work of art "better" than it originally was. And these films are a very good case for that. You could decide that from today`s perspective, the special effects really aren`t that good—they don’t look “real”—and redo them all in CG. But if you watch these films the way they were meant to be seen, you engage in a suspension of disbelief, you accept the work’s parameters of illusion. In this respect, they`re just as convincing as any other kind of artistic illusion, any other work of fiction, and just as satisfying. If you change the effects to make them conform to today’s standards of illusion, you undermine the essential aesthetic of the film. And this approach assumes that your viewers are all the same, and can only tolerate the look of today’s movies, and have no interest in the diversity or history of the aesthetics of film. We don`t assume that; our audiences are very diverse and sophisticated, very knowledgeable and very interested in the original aesthetics. We don`t want to “remake” older films, or truncate them or make them look like something other than what they were. And this is the bedrock of the preservation work that we do, the ethic of our department. Future audiences will only care about the original achievement. There will be no new Honda films. The limited number of artworks we have from Mr. Honda are valued in part because they`re unique and authentic expressions of this particular artist and his team, of a particular time and place. We can`t make them any better. They stand alone in this sense, and they have to stand with the Kurosawa films, the Mizoguchi films, the Kobayashi films, the Harryhausen films—with the rest of the historical cinema. They have their own unique character.
A sophisticated viewer of film enters into a compact with the film that embraces both the films conventions and limitations. It`s the way that you understand the film and the objectives of the director. The fantasy or fairy-tale qualities of these films were important to the original audiences, and they are just as important to our viewers today. Ray Harryhausen’s effects can be seen as rather silly and ineffective in comparison to today`s effects work, just as Melies might be considered primitive compared to Harryhausen or to Honda. That`s no reason to change them; these effects are a source of viewing pleasure and a reason to study the films. On the other hand, the process of transforming historical films to approximate the look and technical qualities of today’s motion pictures is a very costly enterprise, and in just a very few years, these modifications will be unwatchable. If the original qualities of the film are not protected, the film will ultimately lose its value. The audiences for these films today are increasingly sensitive to the qualities of the cinema. These are not bored individuals watching these films as a way to fill up idle time, they`re genuinely interested in the aesthetic history of special effects, or the history of action or science fiction cinema; they`re interested in the ideas and images that were created by Honda. And I would like to mention the audience, the so-called “cult” audience, the fans of these films. Some critics are inclined to dismiss this audience, but in fact these viewers some of the most knowledgeable and committed viewers. And they are, along with archivists, among the strongest partisans when it comes to representing the work in its original state. So the long-term value in these films is in their authenticity, in their fidelity to the film’s original look and sound. We have preserved the film elements, so we can produce prints for special screenings, so purists can see the films projected in archives and special venues. We have made high definition digital masters, so that the films can be distributed on DVD and perhaps, eventually, Blu-ray. We`re sensitive to the fact that the new media change the experience of the films, and we believe that the versions we have made for digital media are a good approximation of the look and sound of the film in a theatrical context. Steve Ryfle: Having spent a lot of time closely examining these three films, what impressions were you left with? Michael Friend: Among the most distinctive things about these films is imagination and the level of craftsmanship in the special effects. I came to have an appreciation of the specific aesthetics and the artisanal level of the films. I had the opportunity to look very closely, frame by frame, at all three films. The closer one looks, the more the illusion reveals itself, the more the film falls apart in a delightful way.
Because the art of making special effects has evolved so rapidly in the digital era, it has become commonplace to dismiss the work of Honda and his contemporaries as primitive. But this is unfair. The level of Honda`s technique can be compared to other masters of the cinema such as Starevich, Pal, and Harryhausen. His way of working was meticulous; it had its own craft history, its own special inventiveness. This technique has limitations, but all films that have to present an unusual vision of the world share the same limitations. Working closely with these films, one gains an understanding of why the filmmakers wanted to make the illusions this way, as opposed to using other methods. As with the Godzilla films, there are conventions of presentation. And the whole body of this model based film-making is set inside the larger enterprise of science fiction cinema in general. The films achieve what the film-makers set out to do in making them. It was hard to do, it was pushing the envelope in a sense, but the special effects support the fairy-tale quality of the film in a very specific ways which come to define in a sense what a science fiction fable is. Honda manufactured worlds in miniature physical form and photographed them, creating illusions of outer space, strange animals and monsters, cities under siege, apocalyptic destruction, the image of desperate people trying to use their tools to understand terrifying and mysterious phenomena, trying to save themselves and the world. Honda used scale, detail, and of course imaginative destruction to achieve the speculative sublime. And Honda`s primary problem in this work is not technical; it is a first of all the question of vision. As much as these films depend on effects, they are also about people, about communities and societies. It`s an interesting approach to filmmaking, fable-like, but very modern at the same time. And it is specifically post-atomic film-making. BATTLE IN OUTER SPACE imagines a human race united by external threat, while MOTHRA recognizes a sense of ecological balance as a corrective to the powers of technology and the greed and corruption of exploiters. Science fiction is never an end in itself for Honda.
For more information on the ICONS OF SCI-FI: TOHO COLLECTION please see the earlier coverage here on SciFi Japan: