Film Restoration

The Culture and Science of Audiovisual Heritage

L. Enticknap

• Publisher: Palgrave Macmillan
• Year: 2013
• ISBN:
  9781137328724
• Topic: Media & Performing Arts
• Subtopic: Film History & Criticism

This is the first monograph-length work intended to enable readers with a humanities background and the general public to understand what the processes and techniques of film restoration do and do not involve, attempting to integrate systematically a discussion about related technological and cultural issues.

1

Why Do Films Need to Be Restored?

Introduction – Articulating the original

Film restoration is a controversial practice. And it isn’t. In May 2010, the British national newspaper the Daily Mail carried a report of the broadcast of an archive-based historical TV documentary series on the Second World War, claiming that ‘Germans have been able to watch the war that changed their world forever in full colour’, and that ‘film footage and photographs restored and colourised using the very latest technology … allow people now to see the war as people then did’.¹ A year earlier and in a similar vein, another press report stated that recently discovered amateur film of Churchill and Eisenhower would be ‘transformed into high definition footage using a state-of-the-art digital film format.’² In both cases, the archival footage being repurposed for contemporary access was processed in such a way as to add significant image information that was not captured at the original point of photography. Nowhere was it implied that this practice was curatorially problematic or ethically debatable as a general rule: it was mentioned towards the end of the former article that the producers of Der Krieg had refrained from colourising footage of Holocaust atrocities on the grounds that doing so would have been ‘tasteless’,³ but the tone of the article was that in overall terms, the practice enhances the film’s perceived authenticity as opposed to compromising it.

This raises a complex series of issues. Firstly, issues of authenticity and provenance in the physically recorded moving image and audio records that comprise a film have never been subject to the same scale of methodological and ethical debates in cinema criticism and scholarship as they are in most other cultural forms based on permanent records. If an art conservator had colourised Leonardo da Vinci’s Virgin and Child with St. Anne and St. John the Baptist, would its mainstream reception have been as uncritical as that of the colourised footage in a television documentary? Why do film critics and historians not routinely refer to the 1958 Vertigo compared to the 1997 Vertigo in a similar way to that in which classicists have conducted extensive debates as to the respective merits of William Melmoth’s translation of Pliny the Younger’s letters compared to that of Betty Radice? And why is there no established vein of scholarship on filmmakers whose works were subject to extensive alterations and revisions in the same way that for the last four decades, musicologists have debated what Deryck Cooke termed the ‘Bruckner problem’?⁴ Almost all the commercially published recordings of Bruckner’s eighth symphony (for example) state prominently which of the five published versions of the score has been used for the performance, and many will be packaged with extensive essays by the conductor or a musical scholar justifying the choice through an analysis of the competing claims to authenticity made for each of them. Yet how many consumer video publications of King Kong (USA, 1933, dir. Merian C. Cooper and Ernest B. Schoedsack) incorporate comparable discussions of the multiple versions that were released in different territories and in different time periods since the initial screenings?

The object of film restoration means different things to different people. As I have argued in the introduction to this book, attempts have been made to define the combination of technical and historical specificity of the moving image medium by which it can be regarded as a discrete document, but these attempts have largely been confined to the archiving profession, with secondary input from historians. Central to this is the notion of originality, and implicit in that is the belief or assumption that there is such a thing as the ‘original version’ of any given film: in other words, a definitive form of the film from which others can be considered deviations and which, if it is not known to exist, it is the restorer’s aim to recreate. The problem is that noone is really clear what ‘original’ means. The International Federation of Film Archives’ (FIAF) code of ethics requires its adherents to ‘not seek to distort or change the nature of the original material’,⁵ while the AMIA charges its members ‘to restore and preserve artefacts without altering original materials, wherever possible’.⁶ Neither of these codes defines what is original and what isn’t, and nor, to any useful level of objectivity, does the only monograph-length work on the subject, Ray Edmondson’s Philosophy of Audiovisual Archiving. The closest Edmondson gets is in proposing two, mutually contradictory, definitions. The first is in a glancing reference in his proposed definition of a record: ‘lasting evidence of transactions, decisions, commitments or process, often in the form of unique original documents’.⁷ The second is ‘a film negative or master recording’.⁸

The silent period: Incorporating exhibition

Defining ‘original’ as a combination of content-based and temporal (i.e. immediately following a film’s initial production), to varying levels of specificity, is problematic. Early and mature silent cinema presents a particular challenge because, to adapt Edmondson’s phrase, a major part of the ‘transactions, decisions, commitments or process’ this form embodies – the audio – is not contained within the ‘unique original documents’; or at least, not in the same one and in the same way as the picture. That uniqueness is not self-contained and textual in the sense that it is embodied entirely within the audio-visual artefact itself. It was a common practice in early cinema exhibition in multi-purpose entertainment venues, notably fairgrounds and music halls, for a showman or lecturer to accompany the projection of a film, often explaining a filmic narrative with the aid of notes or a crib sheet supplied with the film print itself. During the period in which the techniques of continuity editing were still undergoing basic development and evolution, these notes provided elements of the narrative that were not understandable from the actual moving images, as in the case of A Daring Daylight Burglary (UK, 1905, dir. Frank Mottershaw), during the projection of which the audience was told that a telephone call had been made to arrange for a criminal’s arrest.⁹ The film makes no sense without this information, yet when it is screened today it is usually with a musical accompaniment and without any commentary. A claim could possibly be made to originality in respect of the sequence of film images in isolation, but not to the authenticity of the viewing experience.

Between approximately 1908 and 1915, a primitive version of what might now be called a music video emerged, in which films were made of a performer lip-synching to a pre-existing, commercially published music recording, and then exhibited using a proprietary method (there were several, each with their own branding, e.g. Cinephone, Vivaphone) of interlocking and synchronising an acoustic gramophone to a film projector.¹⁰ In many cases the films survive but the audio recordings do not (or vice-versa), and today the films are sometimes screened with contemporary live musical performances of surviving scores. To what extent can this be regarded as original or authentic, especially if the situation is further complicated by the fact that upon such a film’s initial release, it may well have been screened with either recorded or live music, depending on the equipment and personnel available in the theatre?

After the combination of live musical performance and mechanically generated sound effects became the dominant form of accompaniment for mature silent cinema in North America and other countries that were significantly influenced by Hollywood’s business and cultural practices (approximately from 1915 until the late 1920s), a number of variables informed how this took place. In some cases studios would provide scores, to varying degrees of specificity (i.e. ranging from a complete score to a loose framework that enabled significant improvisation by theatre musicians) and complexity, from a piano score to full-scale orchestral parts.¹¹ One school of thought among the archiving profession believes that ‘a tour of important silent films without the original score [ … ] carries the message that half a restoration is good enough. No-one would accept this in the art world, [sic] why should it be acceptable in the world of early film?’¹² But ‘the original score’, if defined as an authenticated form of accompaniment dating from the film’s initial screenings, meant something very different in a large, city centre theatre than it did in a small, rural one; a fact that needs to be borne in mind when considering the response to Anderson’s injunction, offered by restorers such as Kevin Brownlow and Carl Davis, of synchronising high-quality recordings of full-scale symphony orchestras to the final release versions of their projects.

And furthermore, surviving evidence of the culture of film exhibition during this period is that it was a far more volatile and anarchic environment. For example, a comic record published in Britain in 1926, Mrs.’iggins at the Picture House, portrays a noisy projector, an out-of-tune harmonium, a customer who noisily reads out the film’s intertitles, disinfectant being sprayed in the auditorium and a myriad of other distractions that created an exhibition context very different from the one in which the end result of archival restoration projects are typically seen today.¹³

The sound period: Multiple texts

Defining originality in the case of a sound film (and a silent one if the scope of coverage is restricted to the moving image component) would superficially appear to be easier, as there was usually a lot less inherent variation in the historical reception context. However, the definitions of originality commonly used by archivists have the potential to be every bit as problematic as they do with silent films, both in evidential terms and in terms of the specificity of the film medium. Unique documents are often not ‘original’ in the sense of being an unmediated record of their content, e.g. if they are the end result of censorship, re-editing for commercial reasons or other alterations; and a film negative can often be two generations removed from the one exposed in a camera and incorporate substantial modification to the camera original’s content during the post-production process. It may be the nearest (in generational terms) surviving element to that camera original, but be missing content that is present in other copies that are removed by further generations. In other words, a given physical element of a given film can be ‘original’ according to some definitions and in some contexts, but not others. As the historiographer Arthur Marwick points out, ‘certain materials do not fit neatly into the categorisation as primary and secondary sources. Some are primary sources from one point of view, but secondary from another.’¹⁴ The same health warning applies to the notion of originality in film. For example, the 1938 re-release of King Kong, in which scenes believed by censors to portray unacceptable levels of sexual violence were cut,¹⁵ is ‘original’ in the sense of being a record of the effects of the Production Code half a decade after its introduction, but not of King Kong as it was initially distributed in most of the USA shortly after its production was completed in 1933. Or the British Film Institute’s ‘restoration’ of A Colour Box (UK, 1935, dir. Len Lye) is original in the sense of revealing the aesthetic detail of its maker’s hand painting onto film, but not of the subjective impression of seeing the film in the 1930s, as the two versions used entirely different colour reproduction systems and the theatre projectors in widespread use now use a different light source, with a slightly lower colour temperature, to those of the 1930s.¹⁶

Arriving at a definition is in fact easier said than done, the principal reason being that (with the arguable exception of home movies shot on reversal stock), audiences never in fact see an ‘original’ film, if one defines original as being the physical artefact that is exposed in the camera, i.e. on which the recording is actually created. Like the mass distribution of the written word, moving image media depend entirely on their ability to be copied. Indeed, it was the ability to create multiple physical copies from a single film element exposed in the camera that made the cinema a viable commercial enterprise: it is always worth remembering that the Lumières’ Cinématogràphe, one of the earliest mass-produced piece of film industry hardware, was a machine that combined three functions in one: a camera, a printer and a projector. In other words, it embodied the three sectors of what would later form the vertically integrated structure that enabled Hollywood’s economic dominance: production, distribution and exhibition. The creation of multiple copies can be done at comparatively little cost (relative to that of the initial production, or the ‘negative cost’ in film industry parlance) and with little if any perceived loss of technical image quality by the lay viewer.

And furthermore, that copying is almost never a content-neutral process and is often integral to the creative one. In 2008, the inventor of a film scanner designed for archival use published a press release claiming that the machine offered ‘non-judgmental preservation’, by which he meant that it extracted image and audio information from the element being scanned and represented it in digital form without altering its subjectively perceived aesthetic or sonic properties.¹⁷ The implication, of course, is that conventional film copying processes do incorporate some form of ‘judgement’ as to what the duplicate should look and sound like, which isn’t necessarily the same as the source. The change can be deliberate, for example optical printing to incorporate dissolves, mattes or other special visual effects, or the incorporation of computer generated imagery (CGI) within a digital post-production workflow. In such cases, a ‘non-judgmental’ copy of the original camera negative will bear little if any resemblance to the finished product the filmmaker intended audiences to see on the screen, and in some restoration scenarios it may be deemed necessary to, in effect, recreate original post-production steps. A very simple example would be the practice of tinting and toning using coloured dyes. These colours were not captured at the point of photography, yet were routinely added in the processing of the final release prints. Some restorations incorporate colour tints and tones, either produced by authentic methods or simulations using contemporary technologies, whilst others have not, with the restored film being screened in black-and-white. Claims to originality could be made for either output.

Exhibition technologies in use following a film’s initial release are likely to have been very different from the ones in mainstream use when it is subsequently restored, thereby raising further questions as to the nature of authenticity and originality. Technicolor is an obvious case in point: throughout the period in which it was a mainstream technology, the dye transfer printing method used to produce the copies shown in theatres was incapable of reproducing the definition or tonal range captured by the beam-splitting camera; so much so that many release prints dating from a three-strip production’s initial release ‘do not provide a true idea of the fidelity that the system was capable of [recording] at the time’.¹⁸ The dye-transfer printing system was notoriously resistant to quality control, producing release prints that varied from those that rival modern chromogenic ones to copies in which colour fringing, poor saturation and other defects resulted in them being treated as ‘fly-over state’ (USA) or ‘north of Watford’ (UK) prints, i.e. distributed to second-run theatres that charged lower admission prices. In some cases, Technicolor’s technical characteristics were deliberately manipulated for artistic effect, effects that cannot easily or accurately be reproduced with alternative technologies in current use.¹⁹ The final aesthetic variable that needs to be thrown into this mix is that during the period in which Technicolor dye transfer printing was a mainstream technology, the light source in most theatrical cinema projectors was the carbon arc lamp, which produces a colour temperature of about 5,000 kelvin. Between the late 1950s and early 1970s, carbon arcs were largely superseded by the xenon arc lamp, with a colour temperature of 4,000 to 4,500 kelvin (i.e. more towards the red end of the visible spectrum). Viewing a 1930s or 1940s Technicolor release print using modern equipment, therefore, will give a significantly misleading impression of what the cinematographer and grading technicians were trying to achieve.

A restored version can often reveal detail recorded at the initial point of photography that, due to shortcomings in post-production, printing and/or projection technologies used thereafter, were lost from the film’s early screenings. Surviving release prints can provide evidence of this (subject to the health warning on illumination sources), thereby presenting the restorer with an ethical dilemma: does he try to recreate the ‘original’ exhibition experience, retaining what are now perceived as technical drawbacks, or does...

Table of contents

• Cover
• Title Page
• Copyright
• Contents
• List of Figures
• Acknowledgements
• Introduction – Why Restoration Matters
• 1 Why Do Films Need to Be Restored?
• 2 Where Are Films Restored, Where Do They Come From and Who Restores Them?
• 3 The Technique of Film Restoration
• 4 The Presentation of Film Restoration
• 5 Conclusion: The Ethics and Study of Film Restoration
• Appendix: Technical Glossary
• Notes
• Bibliography
• Index of Film Titles
• Subject Index