Confessions of a film restorer
Confessions of a film restorer
Abstract and Keywords
When restoring a film, the aim is naturally always to provide the definitive version. However, many factors make this an impossible mission. This chapter draws on actual film archival practice and theory, exposing a minefield of obstacles facing any academic study trying to examine film history based on restored works. The focus is on silent cinema restoration, intertitles, and translation issues. Using Mark-Paul Meyer and Paul Read’s categories—from a one-to-one duplication to the creation of an altogether new work—the aim is to give an insight into the complexity of silent film restoration and the practical, and sometimes very unacademic, nature of the actual restoration work. The fact that most film restorations typically concentrate on image quality rather than titles, which are often merely supposed to support the visual action, adds to the complexity of transparency about the provenance of the filmic titles as an object of study.
I HAVE A CONFESSION TO MAKE. I do not think I have ever made a perfect film restoration. Some I still cringe at when specific errors come up during the screening, others have well intended choices which are either not noticed or possibly interpreted inappropriately, and some I am still very happy with.
Drawing on actual film archival practice and theory, here I describe the minefield of obstacles facing any academic study trying to examine film history based on restored works. The focus is on silent cinema restoration, silent cinema intertitles, and translation issues, with the specific aim of pointing out some of the fallacies that film scholars risk falling into, if they are not very aware of the fact that most silent films viewed today have been through many visual and textual alterations before meeting the eyes of the modern audience.
Any film is not just a title, but a collection of versions. For contemporary films these versions are created to fit into different sales windows, often cropped and compressed from the theatrical version, sometimes edited to fit a concrete duration or airline preference. One hopes creative and artistic considerations are taken into account. For historical films many interpretations need to take place, not only as regards extant content from different sources, but also as to the best ways of bringing the look and feel of flammable 35mm cellulose film carriers into the digital formats, and the processes used in restoration and display in the 21st century.
The practical and sometimes very unacademic nature of the restoration work can seem casual to the purist, but the goal here is not to point fingers either at the academics who may be (over-)interpreting restored historical films, or at the film restorers, who sometimes cut corners and forget to document their restorations. Rather I hope to provide some discussion points that might inspire a more nuanced dialogue among academics and film archivists. The fact that most film restorations and restorers typically concentrate on image quality, rather than the titles, which are often merely supposed to support the visual action, adds to the complexity of transparency in regards to the provenance of the filmic titles as an object of study.
Paul Read and Mark-Paul Meyer describe seven different restoration categories from which a film restorer must select when making a restored version of a film:1
1. The film as it is in the restorer’s hands
2. The film as it was seen by its first audiences
3. The film as it was seen by later audiences
4. The film as it was intended by the filmmaker(s)
5. A version that is meant to be seen by a modern audience
6. A new version, a reworking of the original version through a contemporary artist
7. A version for commercial exploitation
This list points to the complexity facing any film restoration and the categories are overlapping rather than clear. Current digital tools have created possibilities that were never available to restorers before, which further complicate the categories by adding potential duplication routes. On top of this, restorations are expensive, and must therefore, directly or indirectly, satisfy an audience or sponsor, thus bringing in a certain contemporary accessibility factor even in less commercial restorations. Most restoration versions are more often than not created to match the original entertainment or audience requirements, hiding the production apparatus, rather than being ‘study versions’. For scholars of language and intertitles, the study of restored versions is further frustrated by the fact that, much as with sound films, most restorations of silent films concentrate on the image, while intertitles are mainly considered supplemental and are typically judged by their ability to not get in the way of the visual narrative.
When restoring a film, the aim is naturally always to provide the definitive and final version. There are, however, many aspects that play into making this an impossible mission. Restoration is always for an immediate purpose and always creates a current version, even if the intended goal is to give access to an authentic and complete historical artefact.
When restoring a film, there are always limits. These may be imposed by the surviving elements, the sources available, the restoration supervisor’s goal, the technical duplication process, and/or the skills of the operators. While the restoration takes the original sources as its starting point, the physical process of production is a matter of adapting current creative film-making tools to handling historical elements and film creation processes. For instance, tinting and toning are two different chemical treatments in black-and-white analogue (p.103) film (tinting applies a coloured dye to the surface of the film image, while toning couples a dye to the silver that creates the image), which in current digital processes are translated into digital application of image manipulation.
For restoration of intertitles in silent cinema there are several possible sources for intertitles. One might be lucky and an original print may survive. In this case the restorer might decide to do a one-to-one copy, with some enhancements in duplication, and just pass on the film in as non-interventional a way as possible. But what if the print is of a foreign version, especially if better image material exists from a negative or a different image source? What if the titles in this foreign print are of low quality and cheaply made? What if the titles are of a later date than the original release? Or if they are altered from the original intention because of censorship or territorial and cultural ‘interpretations’?
Intertitle restoration practice
Most silent film prints were made as positive cut prints (also called ‘pos-cut’ prints), since intermediate duplication film did not reach a satisfactory quality level until the mid-to late 1920s. Therefore silent films virtually only exist in their final complete form as prints. Also, the positive cut prints allowed different scenes to be tinted or toned to give visual cues to the time of day, such as blue for night, or elaborate sunsets with combined blue tones and pink tints. The camera negative was edited into reels corresponding to the colour effect and in continuity, with edit numbers etched onto the film. Intertitle numbers were written or edited into the negative to allow the positive cut editors to cut the title with the corresponding number into the print (see Figures 6.1 and 6.2).
Film companies were sometimes referred to as ‘film factories’, an appropriate term considering that many had dozens of positive cutters employed to splice together the hundreds of positive prints, each containing more than 100 intertitle and continuity edits. (For an example of the physical edit between image and intertitle, see Plate 7.) Each print can thus be considered a uniquely crafted version, with local distributors sometimes making their own versions, based on their own preferences or local taste.
The pos-cut process allowed silent cinema to be truly international, since intertitles were in the local language for all audiences. With the introduction of sound, the pos-cut technique was abandoned and cut negatives became the standard. At the end of the 1920s, new ways had to be found for films to reach non-domestic audiences. Sound films were subtitled and/or dubbed into the local tongue, especially for larger markets and languages, where the investment made financial and cultural sense. (p.104)
The restoration type that comes closest to the film as it survives and was seen by its first audiences is where an original print is the only resource available for the restoration. In this case the object of restoration is an original exhibition copy, and the main objective of the restorer is to make as authentic as possible preservation elements and exhibition copies resembling the original with optimum fidelity. For academic and study purposes of the film as ‘text’ this is in many ways the most relevant source for the historical study of cinema. However, most film restorers would prefer to have an original camera negative to work from, since it contains better image quality than a print.
The original print is always an invaluable point of reference for the look and narrative structure of a silent film. When it comes to the look and syntax of intertitles, original prints (in those cases where the titles are preserved as image) offer the only insight into this aspect of cinema study. A common practice in film archives has been to cut down intertitles, especially those in a foreign language, before duplicating onto expensive negative stock. Therefore many intertitles today survive as so-called ‘flash titles’, which consist of one or two frames left in the duplication negative for later reference to intertitle content and placement. In some rare instances the actual title cards used for the production of the film intertitles survive; this is the case for a number of (p.105) Swedish films, where the original paper cards are held in the Swedish Film Institute. The duration of these intertitles and how to make them look ‘analogue’ in digital versions is, however, still a challenge.
I have described a silent film restoration from original negatives and title books in detail in ‘Restoring a Danish Silent Film—Nedbrudte Nerver’.2 Many of the feature films produced in the 1920s by Nordisk Films Kompagni, the dominant Danish film company at the time, are preserved from the original camera negatives. This means that there are no intertitles in the preservation elements, but different numbers and sometimes written text to indicate double exposures or alternative endings. Nordisk used a relatively simple tinting scheme in this period, consisting mainly of ‘Amber’ (see Plate 7) and ‘Blue’ (for night scenes). Restoring continuity therefore mainly consists of inserting the (blue) outdoor night scenes, typically only a small separate reel of footage, into the film. Since the starting point is the camera negative, this type of restoration has the greatest potential for maintaining image quality.
However, it was common for feature films in the 1920s to have several hundred intertitles. Nordisk kept its ‘title books’ spanning 1908–25, which survive to this day as a unique source for the company’s intertitles. (See Plate 8.)
Even if the original wording of the intertitles survives in these title books, there are a number of possible pitfalls. One of the practical decisions to make is related to the fact that Danish spelling changed in 1948, so that the spelling of a number of words is now different from the silent era. Also, after 1948 the first letters of nouns were no longer capitalised, as in German, but styled in small letters. This is in line with the other Scandinavian languages, which the change tried to create stronger proximity to. Since the title books contain the intertitles in their original syntax, and the films would have been created using the original Danish, the choice to not update the spelling in the restoration is opted for as the right decision.
However, in films where footage is missing and new explanatory titles are created, new Danish is used as a ‘marker’ that these titles are not original in their content. This was the case for the restoration of Carl Theodor Dreyer’s Once Upon a Time (Der var engang, 1922), where one third of the film is missing. Here explanatory titles in modern Danish are inserted. Also, for Dreyer’s Love One Another (Die Gezeichneten, 1922), a few explanatory titles were inserted using modern Danish, to distinguish them as restoration additions. It is quite possible that this distinction between original titles in pre-1948 Danish and newly created titles in post-1948 Danish is lost on the general spectator,
(p.107) but I find that using it discreetly is the correct approach, to allow the film narration itself take the fore. (See Figure 6.3.)
When working on the negative/title list restorations, other choices are made that bring these restorations into the more critical restoration categories. In order to cater to a broader audience, I have opted to create double-language versions, with combined Danish and English titles in the same intertitle. Both of the Dreyer films mentioned above, as well as Himmelskibet (Holger-Madsen, 1918), Glomdalsbruden (Carl Th. Dreyer, 1926), and more have been restored with new combined intertitles. While this was common in some territories (Finland, Belgium, Switzerland, etc.), a Danish/English combination is a new creation solely to enable easier exploitation and access in the western (academic) hemisphere. To further problematise this, the Danish titles aim for historical authenticity, while the English titles are typically modern translations of the Danish titles, since very few original English title lists survive for the Danish films as they were shown in English-speaking markets. The resulting version is thus in several ways a bastard version.
While restoring Dreyer’s Love One Another in collaboration with Dreyer expert Casper Tybjerg from the University of Copenhagen, I had a number of sources available, but none that had the exact titles we needed for the combined Danish and English titles. The film survived only in a somewhat shortened Russian titled print. Dreyer’s Danish screenplay, based on a novel by Aage Madelung, survived in the Danish Film Institute collection, and the Swedish intertitles were on record from the Swedish censor. In this case we created new intertitles in modern Danish and a translation in English. In many cases the Danish was close to the Danish sources, but as the exact wording was not available, modern Danish was used to underpin the fact that the titles were a new creation for the restoration.
To say documentation is not always perfect would be a gross understatement. Some restorers do produce documentation of their work in the form of edit decision lists, titles lists, and translations of different sorts. However, aside from the occasional article, the documentation of restorations remains hidden in the institution that performed the restoration, or even just on the personal hard drive of the restorer.
One must come to terms with the fact that not making a decision is also a decision. No one-to-one restoration or duplication will be the same as ‘the original’. There are so many factors that play in, such as light source, film transport, and colour systems, that one cannot replicate the original experience without showing an original film on an original projector, an option only very rarely achievable.
In a 2013 article entitled ‘There is No Such Thing as Digital Restoration’, David Walsh has explored a way of assessing the potential offered by duplication technology and the danger zones relating to trying to improve on the film (p.108) being restored.3 Some very grey areas open up in relation to digital scanning of negatives. Since digital does not have the graceful ‘degradation’ in resolution offered by film-to-film duplication, the digital element from a negative scan may be in higher resolution than any version of the film to have been screened before. This can sometimes be a very real challenge, especially when the filmmakers and cinematographers anticipated an analogue drop-off to cover up coarse make-up, or the fishing lines holding up model airplanes. The recent restoration of Dr. Strangelove (Stanley Kubrick, 1964) is probably the best-known example of this. Most of my colleagues try to balance the tools and possibilities in an objective manner, but there cannot really be such a thing as ‘neutral intervention’.
In film restoration the terminology and categorisation of the specific restoration job is somewhat loose. One might use the term ‘restoration’ to describe the lower numbers in Read and Meyer’s list of categories, while the term ‘remastering’ could be applied to the higher numbers, since more compromises and guesswork would have been applied. In a sense it is a range, where the lower numbers aspire to an objective reconstruction, while the higher numbers introduce a higher degree of new creative approaches and repurposing of the original material.
It is in my opinion dangerous to use a more casual approach to techniques in order to please modern audiences. For example, some silent film restorations use tints, even when there is no direct evidence of a specific colour scheme. This is done since the current standard in silent film restoration is to retain the tints and tones (whereas until the end of the 20th century many films were preserved in black and white because of technology, cost, and sometimes preference), so a new restoration does not always feel done, until it has been coloured.
Another remastering technique that is better avoided is the use of old-fashioned fonts and decorative intertitle frames. Since these techniques were also used originally, it can be very difficult to distinguish the real thing from the fake. One of the worst tools in the digital toolbox is the application of the plug-in ‘old film look’, which adds scratches and blemishes to film that otherwise looks fine. It is one thing to use it in modern productions, another to apply it to historical footage! Of course, you might see it applied to new inter-titles or freeze-frame titles, since the startle from authentically wavering live footage to rock-steady intertitle can be very disturbing. You may never know!
One of the most frustrating experiences of film restoration is when previous restoration efforts have closed the door to further investigation. This happens when the original elements were used as new production elements. A number of early restorations performed at the Danish Film Museum edited the original negatives into continuity order, from the reels of film arranged in order for tinting. This was done to enable easier printing, trimming out colour indications, since cinema was dominantly black and white in the period 1927–60. If there is a first commandment of film restoration it must be: never edit the originals! Later generations of restorers and scholars should be allowed to retrace the elements and sources, enabling them to make their own ‘final’ versions and interpretations informed by the knowledge and scholarship of their time.
Back in 2000, when DVD was the medium that gave a revival to classical and silent films, a scholar once asked me why there were no extras on a DVD I had produced. My response was that the extra was that the films were now available. My excuse for not making perfect restorations is much the same. Now the films are accessible, and I try to convince myself that at least I have not improved too much on the sacrificial altar of providing new access. (p.110)
(1) Paul Read and Mark-Paul Meyer (eds), Restoration of Motion Picture Film (Oxford, Butterworth-Heinemann, 2000), p. 71.
(2) Thomas Christensen, ‘Restoring a Danish Silent Film—Nedbrudte Nerver’, in Dan Nissen, Lisbeth Richter Larsen, Thomas C. Christensen, and Jesper Stub Johnsen (eds), Preserve Then Show (Copenhagen, Danish Film Institute, 2002), pp. 138–45.
(3) David Walsh, ‘There is No Such Thing as Digital Restoration’, in Kerstin Parth, Oliver Hanley, and Thomas Ballhausen (eds), Work/s in Progress: Digital Film Restoration within Archives (Vienna, SYNEMA, 2013), pp. 30–42.