Truth interrupted
An archaeology of the frame rate
Marek Jancovic Sarphatistraat 81-2 1018EZ Amsterdam The Netherlands 0031-647-169884 marek.jancovic@student.uva.nl Student No. 10233539
Date of Completion: 25 June 2013 Supervisor: Dr. Alexandra Schneider Course: Research MA Media Studies
Department of Media Studies, Faculty of Humanities University of Amsterdam
1. Introduction ... 1
1.1. Methodology ... 6
1.2. Traveling and scheduling: Of maps and calendars ... 10
1.3. Sources ... 14
2. Film – truth interrupted ... 17
2.1. The frame rate as an expressive instrument... 17
2.2. The Magical Number ... 23
2.3. Flicker ... 27
3. Television – truth uninterruptible ... 31
3.1. Of waves, streams and currents: television as a continuous stream... 31
3.2. Interlacing: the image in perpetual motion ... 36
3.3. The television will not be televised ... 40
3.4. Double standards ... 45
3.5. Moving to digital streams... 49
4. High frame rate projection – the new paradigm ... 54
4.1. Staring at reality ... 56
4.2. Before The Hobbit ... 59
4.3. High Frame Rates and Illusion ... 61
4.4. The trouble with truth ... 70
4.5. High Frame Rates and Immersion ... 73
5. Conclusion ... 77
5.1. Research summary and future directions ... 80
5.2. Crisis of History... 82
5.3. An Image Economy of Fear ... 89
Works Cited ... 92
Table of Figures
Figure 1 High speed race recording ... 12Figure 2 Single, double, and triple-blade shutters ... 28
Figure 3 Interlacing artifacts... 39
Figure 4 H.264 motion artifacts ... 50
“There is a cinema beyond the 24 frames per second one, a speed-cinema, which is no longer a time-cinema, a tale. I don’t think that’s bad.”
-Paul Virilio (1991)
Abstract
Frame rates are the operational property of moving images that enables us to see movement. It is the single, invisible component that can on its own distinguish ‘cinema’ from ‘television’ or ‘video,’ infusing moving images with values and positioning them in hierarchies. The frame rate is at the same time a discursive object, a surface along which entities like ‘reality,’ ‘realism’ and ‘truth’ are constructed: ‘realism’ can be expressed in frames per second. Conversely, frame rates also become an epistemological model structuring vision, perception and reality itself: ‘seeing’ and ‘thinking’ also become expressible in frames per second.
In this thesis, I draw an archaeology of frame rates as a complex cultural object composed of practices of seeing, discourses of science and philosophy, technological inventions and visual artifacts that emerge as archetypes of particular media. I explore the displacements and shifts that occur when media change frame rates and travel into other contexts. Why is cinema “truth 24 frames per second,” as Godard had said? Why does a technologically superior digital moving image look ‘fake’ and cinema, with its invisible but perceptible flicker looks ‘real’? Cinema, television and recent ‘high frame rate’ and digital projection practices are analyzed in terms of how they contest the integrity of the ‘frame,’ how they produce standards and aesthetic norms and, ultimately, how they structure our relationship with time, movement and reality.
1. Introduction
It was the night of the Oscars and my friends and I were watching the live broadcast on a big television screen. A reporter was commenting on the celebrities and their festive attires, when one of my companions commented how amazing it was that “this HDTV makes everything look so crisp and real, as if you were there.” In that moment, on the high resolution screen, television’s mission of telepresence had been achieved. The technology had proven
itself. The twist came only minutes later: the live broadcast switched to a trailer of one of the Academy Award nominee films, Beasts of the Southern Wild, directed by Benh Zeitlin. Midway into the trailer, the same friend disgruntledly exclaimed: “these TVs make everything look so fake!”
He himself did not even realize the immediate paradox. I, on the other hand, was astonished by this blatant analytical contradiction of media technology. What he was referring to was the difference in the expected frame rate of the images. Several sophisticated technological processes were taking place which not only provoked his affective response but at the same time made visible (or perceivable) highly complex configurations of media and the expectations that were being posed onto them.
Even a very simplified description should exemplify the technical difficulties behind the moving image that provoked my friend’s enigmatic response. We were witnessing an event happening in real time in Los Angeles being filmed with a television camera with a frame rate of 29.97 interlaced frames per second, one of several high definition television standards currently in use by networks in the United States. In order to be sent to the Dutch TV station we had tuned in to, the signal from Los Angeles was, it can be assumed, first processed by a converter such as those used in transcontinental sporting events into the European digital cable television standard (DVB-C) frame rate of 25 interlaced frames, or 50 Hz, slightly distorting the motion of the footage but yielding a look we are used to seeing on television. The high resolution and fluid motion of our TV set’s image were what made everything so “crisp and real.” When the broadcast zapped to the film trailer, however, something strange happened. The trailer, edited from footage originally shot at 24 frames per second on 16mm film, was first slowed down by an imperceptible fraction to 23.976fps, then processed in order to be shown at the U.S. rate of 29.97, introducing a slight judder into the footage, followed by another the on-the-fly conversion to 25, and finally, our modern television set in a living room 9 time zones away added hundreds of additional frames through interpolation technology in order to show us the product of this process at a staggering speed of 200Hz, or 200 pictures per second. This resulted in the motion of the trailer having a comparable perceived fluidity as the “real” live broadcast. And yet my friend (a film director and scriptwriter) was not impressed. He discarded the trailer section of the transatlantic transmission with the single word ‘fake.’ What was happening?
“La photographie, c'est la vérité et le cinéma, c'est vingt-quatre fois la vérité par seconde,” Bruno Forestier (Michel Subor) remarks while taking photographs of Véronica (Anna Karina). This famous adage of cinema in Godard’s Le Petit Soldat (1960) is significant not only because it echoes what is widely and often uncritically accepted as fact – that cinema records ‘the truth.’ It is also important because it links the truth to the technological process of its mediation: the succession of twenty-four photographic frames per unit of time, the magical number designating the speed with which a camera inscribes the slices of truth onto a strip of 35 millimeter film. If cinema is a succession of twenty-four photographs a second, I wonder if one could echo what Rodin replied when Paul Gsell told him “a photographic instant is truth.” Rodin’s response was: “No, it is false because in reality time does not stop.”1 Is cinema then a recording of truth, or a succession of falsehoods?
The contradictory impressions that my friend reported while watching the Oscars recently very prominently reverberated after the release of the Hollywood blockbuster adaptation of Tolkien’s The Hobbit by Peter Jackson (2012). The film uses a rate twice that of what has been the standard in mainstream cinema for almost 90 years, and is instead close to that of television. This caused a massive outrage from spectators and critics, indicating the uneasy relationships with ‘reality’ that cinema and television operate under. The resulting discussion, addressed in Chapter 4, is the ideal starting point for analyzing what is meant when we address ‘reality’ and ‘truth’ in relation to media. With all our focus on the visual and the indexicality or truthfulness of the moving image, it is useful to keep in mind that cinema has been, until the very recent advent of digital projection, circumscribed precisely by the invisible: the thousands of black gaps we unconsciously unsee every minute. The educational film How you see it by Chevrolet reminds us: “Surprising as it seems, you never see a motion picture while it is actually moving. The only time a movie moves is when the shutter keeps you from seeing the change of picture.” What immediately follows is the self-deceptive contradiction, “Then we see it just the way it looks in real life.” This was in 1936. It is remarkable that anyone would noddingly agree to this claim even as they watch the black-and-white images on a plane surface seemingly moving while the narrator continues his
metallically-tinted voice-over. When we combined these statements with my insightful Oscar evening, it becomes clear that the boundaries of ‘looking just like in real life’ are both surprisingly fluid, and yet very restrictive. And the frame rate is the surface along which this boundary is constantly being redrawn.
A frame rate is not primarily a technological property of a visual medium. It is a mode of seeing. Deleuze’s axiom in Time-Image that cinema makes the invisible perceptible, and Elsaesser’s realization that “[k]ey elements of cinematic perception have become internalized as our modes of cognition and embodied experience, such that the ‘cinema effect’ may be most present where its apparatus and technologies are least perceptible” 2 are both nowhere more valid than in the literally most overlooked component of cinema: the frame rate. Just like when color motion pictures and television changed the (perceived) color of our dreams,3 the fact that a quick Internet search reveals millions of questions asking ‘how many frames do humans see at?’ and its manifold variations such as ‘how many fps are our dreams?’ tells us that the material-technological apparatus of cinema – a progression of still frames – has transformed into an epistemology of seeing. These questions are fervently discussed, answered, proved and disproved with anything from scientific evidence to philosophical pondering to popular misinformation. The effects of visual media on our consciousness, and even more importantly, our immediate sensory perception, have been internalized as cognitive blueprints and guidelines for the recording and projecting standards of the eye and the brain.
The current accepted scientific consensus would generally answer such questions by presenting the argument that our eyes are different from the workings of a camera, that we don’t see in discrete images, but have continuous vision.4 There is thus, despite various popular speculations, no upper frame rate limit for the perception of the eye.
The word ‘continuous’ already raises a few flags: light comes is discrete quantities we call photons, which hit our retina in rapid succession, one by one, not very unlike the frames of a film. But our daily lived visuality immediately puts some more strain on this accepted
2 Elsaesser 2004: 76
3 Murzyn 2008, Schwitzgebel 2001
scientifically produced knowledge: we know that shaking our hand in front of a light source quickly enough will produce a sensation closely akin to seeing several strobing frames with motion blur between them; and we also know that things like airplane propeller blades can move fast enough to dis-appear entirely. The discourse of continuous vision is so prominent that many of the scientists dealing with vision as well as the cinematographers, projectionists, and technological visionaries discussing the pros and cons of this or that imaging technology do not seem to take notice of a remarkably obvious frame rate limit of the eye. I, too, started writing a draft of this thesis in the same vein, criticizing the various theories that lightly equate the film camera with the eye until it was pointed out to me: blinking.
Common knowledge dictates that blinking is a ‘feature’ of the eye that provides it with necessary moisture. The fact is, we blink several times more than is necessary to moisturize the cornea – about 10-15% of our waking hours are spent in darkness.5 It has become known recently that the interruption of vision both precedes the mechanical covering of the eye by the eyelid, and persists after it, and a similar mechanism works also during saccadic eye movements.6 This means two things: first, the closing of the eyelid is only part of the reason why our vision is suppressed during blinking. Second, useful visual information is taken up while the eyes are stationary and not moving, just like a film strip is seen on screen while it stands still in in the projector.
It has been shown that we possess a mechanism that unconsciously times blinking with moments that require less attention to minimize information loss: a team of researchers around Tamami Nakano at Osaka University has shown that we are likely to blink during breaks in speech,7 at punctuation marks during reading and during scenes that require less attention in filmed narrative stories, such as long shots, the conclusion of an action or the absence of characters.8 Attention thus exercises some control over when we blink. But the opposite is also true – in fact, the most recent studies suggest that blinking plays a significant cognitive role: it activates the default mode network (DMN), a set of brain regions that operate during contemplation, daydreaming and memory retrieval, and disengages the parts
5 Nakano et al. 2012, 2009; Johns et al. 2009: 783 6 Johns et al. 2009
7 Nakano et al. 2010 8 Nakano et al. 2009
of the brain involved in attention focusing when involved with the external world.9 In short, our blinking periodically inserts necessary intermittent interruptions into our cognitive processing: “[G]iven its key role in the disengagement of attention, repeated momentary activations in the DMN might be an essential process for any cognitive behavior.”10 This is a significant realization – we need an ‘intermittent flicker’ as a function of the brain! Much like analog projectors needed shutters, our eyes require a shutter to reset and disperse the stream of visual stimuli reaching the eye and brain. What would the consequences of removing these gaps be?
1.1. Methodology
This research bears the subtitle “an archaeology.” What this refers to is the heterogeneous, exciting ‘nomadic discipline’ (to paraphrase Erkki Huhtamo and Jussi Parikka) of media archaeology and its diverse set of methodologies. But in particular, my understanding of the field draws on Foucault’s The Archaeology of Knowledge (1969) as it has been applied by Thomas Elsaesser and Markus Stauff: an analysis of media as discursive practices, practices that operate to enable and produce objects of knowledge, practices that orient the experience of meaning. My primary goal is to analyze how human understanding of and interaction with media shapes the understanding of material reality. How cinema and television (in their past, present, and imagined future forms) constitute different blueprints of visuality: how they become models for perceiving, defining, recording, and analyzing what is believed to be ‘truth.’ The principal point of accessing these objects is the frame rate – the material speed of the succession of images or image elements in projection and broadcasting practices – as a condition of perception.
Although frame rates appear as a set of fixed and clearly delimited technological standards, they are historically a wildly mutable and heterogeneous object. Frame rates are also a perceptual phenomenon analyzed in scientific, artistic and engineering discourses, and a bridge between certain modes of seeing and hierarchical assemblages of media. To put it crudely, the difference in frame rates between film and television is what makes television
9 Nakano et al. 2012 10 Ibid. 3
look ‘cheap’ in comparison. My archaeology of frame rates will also attempt to analyze the conditions of emergence of such discourses that valorize a particular mode of seeing and particular perceptual effects above others.
A media archaeologist might interrupt here and ask what exactly it is I mean with ‘succession of still images.’ Rightfully so, as that is anything but a self-evident object, but already presupposes significant knowledge about how audiovisual media produce moving images. At least three important boundaries have to be drawn: First, the frame rate at which a moving image is recorded is not necessarily the same as that at which it is presented. The difference is how slow-motion or high speed effects work. The second distinction is that between the perception of movement, and the (im)perception of flicker: our vision requires different thresholds to splice together a series of stills into movement and to fuse flicker (in other words, experience flashing images as continuously illuminated). This is why analog film projectors use a shutter that flashes every image at least twice: 24 frames per second is enough to represent movement, but at least twice the rate is necessary to hide the on-and-off flashing of the projector while the film is actually moving. These two phenomena have been long known jointly and confused with retinal afterimages under the imprecise term ‘persistence of vision,’ and their workings have been described elsewhere.11 The frame rate is not a matter of a universal number that will satisfy all our visual needs, although it is often discursively reduced this way. Motion and flicker are perceived in a variety of ways, affected both by eye and brain function as well as variables such as the brightness of the environment. The human retina is also more sensitive to motion at the edges than the center, so we consequently perceive flickering more easily in peripheral areas of vision.12 And most importantly, concepts like ‘motion,’ ‘perception’ and ‘vision’ are historically contingent. The exact numerical values of frame rates are secondary for this study. What is important is what role their existence plays in the construction of cinema and television as objects of vision and
11 Alongside Bill Nichols and Susan Lederman, Anderson & Fisher (1978) and Anderson & Anderson (1993)
have made the most prominent attempt to purge the ‘myth of persistence of vision’ from film theory and replace it with knowledge produced in the cognitive sciences. Although two decades old and not without its own problems, especially their short 1993 article gives a good overview of the various phenomena involved in the perception of motion pictures and their changing explanations throughout history.
‘truth machines’: the frame rate, apart from being a technological attribute and perceptual phenomenon, is also method of appropriating media practices for science and of propagating and ordering hierarchies by enabling us to express how ‘realistic’ particular media are in numbers. ‘Frame rate’ in the context of this thesis will therefore refer mostly to the speed of the ‘final product’ as it is presented to the eye from the projector or television screen.
The process of analyzing media-induced changes in the way we perceive the world has a far-reaching tradition in media theory. Among many others, years before the field that is now recognized as media archaeology, there were Kittler, Virilio, or Benjamin writing in very distinct ways and very distinct contexts in Europe or McLuhan and Postman, wildly popular in their time in North America. Before all of them Bergson, in what would later inform Deleuze’s work, described intellect as an immediate correspondence to the cinematographic apparatus.13 There are, in fact, references to all of them throughout this thesis – some brief, others more substantial. A problem with the writings of some of these authors is that they also belong to a tradition of essentialist ontologies that attempt to prescribe what media are, should, and should not be. Television and cinema in particular found themselves fighting the same face-off in this arena over and over, the former often at the losing end. Looking for the essence of a medium assumes coherent, clear-cut borders and stable identities; something both media archaeology and our daily media practices readily disprove. I am taking Markus Stauff’s dissertation on the heterogeneity of television as a superbly valuable methodological guidebook to escape this reductive view.
I do believe that television and film produce and become associated with quite distinct perceptual effects – two different visualities that stand for two different ways of relating ourselves to material reality, to time, to ‘truth.’ But as Jonathan Crary insisted in his genealogy of the observer, “[w]hether perception or vision actually change is irrelevant, for they have no autonomous history. What changes are the plural forces and rules composing the field in which perception occurs.”14 Keeping in mind this along with Stauff’s illuminating approach, an archaeology of the frame rate shows the heterogeneous elements underlying these
13 Bergson 2005[1907]. For Bergson, cinema becomes not only a model for thinking, but a model for the
thinking about thought.
perceptual models. They emerge as products of technical inventions, standards and norms that change throughout history and over time become institutionalized in various ways. These inventions and standards themselves emerge from and influence discourses and previous epistemologies of seeing. The way we speak of media in relation to each other, the way they are taken up and theorized within media theory, the way varying degrees of truth value are assigned to them, the way their sensory effects are scientifically analyzed and publicized, the way they themselves become instruments for producing scientific knowledge and, also the ‘content’ they become associated with – all these are constituent elements that contribute to the construction of a world seen through media. This view precludes any search for ‘essences’ altogether. My understanding of media, then, is that of an apparatus or dispositif as defined by Foucault: “a thoroughly heterogeneous ensemble consisting of discourses, institutions, architectural forms, regulatory decisions, laws, administrative measures, scientific statements, philosophical, moral and philanthropic propositions.”15 The apparatuses of cinema and television emerge from the relations between all these elements. “The diverse surfaces – program structures and technical devices, everyday practices and economic circulation processes – constitute fields of competing strategies and different rationalities, which only bring forth the power and subject effects of television [and cinema] in their interaction.”16 Television and film are therefore names for very heterogeneous, changing, overlapping, interacting and often contradictory practices that organize our experience of the world. What I am ultimately interested in is the intersection between the discourse on film and television, the technological methods they employ to produce images, and cultures of seeing.
The historical delineation of my object unravels from what are currently still the discursive archetypes of television and cinema practice: the period in which shooting and projection rates of 24 frames per second in cinema, and 50 or 60Hz in television broadcasting were the most dominant (if certainly not the only) norms. These norms, or rather their perceptual effects, have become internalized to such a degree that they become the medium itself: with all other formal properties of a moving image such as lighting, shutter speed, grain, resolution, depth of field and so on being equal, a spectator will still be able to tell with ease what looks
15 Foucault 1980[1977]: 194f., c.f. also Stauff 2004 16 Stauff 2004: 14
“like film” and what looks “like television.” On the periphery of these dominant modes of seeing, there are overlaps, technological dead-ends and could-have-beens, or technological topoi that reemerge after decades and centuries and prevent a linear distinction or progression between media. Godard’s definition exemplifies what can be comfortably considered cinema and what supposedly falls outside: early films shot and projected at highly varied speeds or early television broadcasts deviate from the magical number, and are therefore somehow not really ‘cinema’ or ‘television.’ Such hierarchies were nowhere more apparent than in the reactions to The Hobbit, an expensive production with a deviant frame rate – like with the Oscar night broadcast, The Hobbit has been called “kitsch and alienating17”, veiled in a “sickly sheen of fakeness.18”
Centering around this rather long period of unchanging frame rates that can be dated somewhere from the mid-1920s for cinema and mid-1950s for television, both beginning to lose their privileged positions in the mid-1990s, I will branch out both backward and forward in the history of these media in an attempt to document the conventions and discourses around frame rates as a method of accessing reality.
1.2. Traveling and scheduling: Of maps and calendars
There are two particularly fruitful analogies to explain what I mean by referring to frame rates as a condition of perception. I have adopted them as very general guiding points of my analysis and a way of structuring my own thinking of media. Throughout this thesis, I will repeatedly refer to television and film as cartographic and chronometric events.
The connection between cinematic, televisual and cartographic practices is not far to seek – after all, the organizing principle of all of them is a form of projection. There are various methods of projecting the physical surface of the Earth onto a map, some of them more conventional than others, none of them technically entirely accurate and all of them socially and historically prejudiced. Cartographic projections are events because they emerge and decline, as they respond to a certain need in a given historical moment. Most importantly, television and film, like maps, order our knowledge of the physical world: a given frame rate is
17 Macnab 2012: unpagin. 18 Collin 2012: unpagin.
like a conventionalized projection of the world that becomes the world, in the sense that it is an epistemological model which can, through discursive and non-discursive means, acquire the status of truth. Ptolemy’s world map would be now considered less accurate, less truthful to the material world than, perhaps, Mercator’s projection, which again is perhaps less ‘accurate’ than a satellite image. The satellite image as the representation that is currently believed most accurate becomes the world.19 Regardless of its feasibility, this teleology of truth, of being as real as possible, is an important (if often vague and conflicting) driving force behind many inventions that would become what we call television and cinema. And as the following three chapters will show, there is a difference between the truth value associated with these media. Television is, through various discursive practices, positioned as an accurate representation of the real, yet paradoxically possesses lower prestige.
If my media-as-cartography occupies the spatial, then chronometry is its temporal equivalent. Frame rates as a time-keeping practice are an even more evident analogy than the map: the frame rate is a measure of time: in editing software, time is measured in minutes, seconds and frames, and one can sometimes hear filmmakers speak of something in reality that happened ‘so fast it only lasted a few frames.’ 1931 and 1932 saw a splendid example of how the cinematic apparatus had become a method of both measuring time and recording reality that was more accurate than a stopwatch. Several sporting events, most notably the 1932 Los Angeles Summer Olympics, deployed a high-speed 16-mm motion picture camera operating at 128 frames per second, which was modified to simultaneously record track runners at the start and finish lines and a clock dial (Figure 1). The film strip literally became “a new way of splitting seconds.”20 The mechanism celebrated a string of successes: It reversed three decisions during the Olympics, made the International Amateur Athetic Federation suggest hundredth-second timing as a standard, was invited to the following Olympics in Berlin, 1936 and recommended by Amelia Earhart for timing airplane races, which it was later indeed used for.21
19 Lisa Parks was perhaps the first to dedicate her career to an extensive engagement with satellites and
satellite images as knowledge-producing media objects, most recently in Down to Earth: Satellite
Technologies, Industries and Cultures (2012), co-edited with James Schwoch.
20 Fetter 1933 21 Ibid.: 339f.
Figure 1 One frame from a recording of a race in Berkeley,
California on July 1, 1932. Image taken from Fetter 1933, SMPE Journal.
But my analogy extends beyond the microtemporal event; I envision media-as-chronometry primarily in reference to calendars. We are accustomed to the comfort of a neat division of time. Our lives, our mode of production depend on a segmentation of time as predictable, regular, and automatized as the successive frames of film,22 with our sleep patterns perhaps being the black gaps in-between which (perceived) movement and progress takes place. And yet we are reminded, not all too often but persistently enough, that our efforts to control time, our time-keeping practices, are inaccurate.
Because the frame rate of our material world is not exactly 24 hours per day, every once in a while we are forced to mess with the flow of time and add a day to our otherwise so regular calendars, those nicely intellectable tables of time perception. We call this event an intercalation, the interposition of a leap day into the calendar. It is no coincidence that Gregory Flaxman used the same word to describe “the modern project of cinema”23 in his own contribution in The Brain is the Screen. Where Flaxman sees cinema as disruption, I, on the other hand, recognize regularity. The projection of a motion picture (or at least the current archetype thereof) requires and relies on a frame rate as stable as the typing of a seasoned stenographer. The disturbances, on the other hand, occur when the calendar is adjusted to fit
22 A related argument was given by Crary 1996[1990]: 112 23 Flaxman 104
back into its slightly mismatched mold: the intercalation happens when frame rates shift – from cinema standards to television or otherwise, – when the incompatible models for perceiving time are forcefully aligned. The same film viewed in various regions of the world in different contexts will differ considerably: when switching between the existing frame rate standards, films get sped up, their duration changes, audio shifts, the voices change in pitch, motion becomes jittery.24 A two-hour movie displaced from film to a DVD can change in length by five minutes. The effects of such intercalations (we could call them ‘leap frames’) are slight, but are they insignificant enough to be ignored completely by theories of cinema and television? Is it still the same film? A media practice as mundane as watching a movie on DVD carries with it the weight of the conflict between cinema and television, and over a century of technological inventions and limitations, scientific knowledge, economic circumstances and power relationships. The various strategies devised for changing frame rates and the material artifacts they introduce are discussed in more detail in Section 3.4. If maps are an instrument of displacement, so are the heterogeneous power relations between audiovisual media embodied in frame rate conventions. After Bazin’s ontological inquiry, media archaeology has added the temporal dimension and asked “when is cinema?”25 Thinking of frame rate changes as displacements can help us contribute the next logical increment: “Where is cinema?” This points us not only towards the ‘interface’ between the eye and the brain addressed earlier, but also to the locus of the moving image, the spaces in which spectators are located when they absorb it. Media archaeologists have been productively exploring this aspect for some time now. Hagener (2008) had asked the same question in an issue of Cinéma&Cie dedicated, fittingly, to ‘relocation.’ However, when he claimed that “the cinema has lost much of its material, textual, economic, and cultural stability,”26 Hagener falls into the trap of opposing an apparently stable past to the multitude of heterogeneous functions and properties of current cinema practices (loosely associated with the organizing term ‘digital’27). There is no such thing – cinema has never had one stable material form (not after the institutionalization of ‘the cinema’ as a specialized architectural
24 cf. Fox 1983: 206. 25 E.g. Elsaesser 2004 26 Hagener 2008:18
27 Stauff: 15 calls digitalization a ‚symbolic place of refuge […] which allows a plausible and manageable
space, and especially not in its earlier forms before that) and was always performed in very diverse economic and cultural contexts. In the same issue of the journal, Schneider (2008) also asked “Where is Cinema?” and showed precisely this on the example of non-theatrical moving images. From its earliest days, cinema was a mobile medium, finding itself in homes, schools, elevators, on mobile phones and amusement rides.28 And indeed, with the frame rate as the map and calendar, it becomes possible to draw new topologies and chronologies of cinema and television in which the movie theater and the living room are only two of the many possible spaces where the practice of media takes place: The Pathé Baby amateur movies ran at 14fps,29 8mm home film cameras at 16fps,30 Super 8 ran at around 18. The large format Todd-AO from the 1950s briefly experimented with 30 fps, theme park attractions often project at 60, whereas early mobile phones with video recording capabilities had variable frame rates starting as low as 5fps. If we follow the frame rate, even the slightest change can tell us much about when and where particular moving images are situated and under what economic circumstances, artistic intentions and imagined functions the are being put to use.
1.3. Sources
I have a certain ‘agnosticism’ in regards to my sources, meaning that I believe a comment on a YouTube video can reveal just as much as a scientific journal article. My research is not structured by a teleological desire to create a history of frame rates or to show that media always the goal of greater, better, more accurate representation of ‘reality.’ Instead, my motivation is to comprehensively include diffuse and even contradictory knowledge. It is certainly true that ‘realism’ is one of the major motivating forces behind technological shifts and inventions in media, but ‘realism’ is itself a volatile organizing principle which different social groups attach different meanings to.31 Nevertheless, there is a level of editorial unity in the sources and data I use enforced by what was accessible to me. First, there is a problem of
28 For an even earlier overview of cinema’s mobile exhibition, see also Rossell 1999: 68-71 29 Herbert 1998: 70
30 Brownlow 1980
31 This view is loosely based on Deac Rossell’s use of Wiebe Bijker and Trevor Pinch’s concept of
technological frames, which suggests that technological artifacts have interpretive flexibility in what they
language: the vast majority of my sources are written in English. This, naturally, introduces significant local and cultural biases. Especially when it comes to the primary literature on early projection and broadcasting technology, I have heavily relied on technical journals, trade papers and magazines digitized and available from the online Media History Digital Library32 and the Society of Motion Picture and Television Engineers, as well as English-language patents and news reports. More specifically, I have reviewed for relevant information issues of The Journal of the SMPE from 1916 to roughly 1935, issues of Projection Engineering from the late 1920s until the early 1930s and various other isolated sources such as Optical Lantern and Cinematograph Journal and various periodicals from around the turn of the century. The bulk of the information thus invariably focuses on early cinema developments in the United States. As for sources on contemporary technology, they are comprised similarly of industry papers and advertisements, interviews with engineers and conference records (some of them in video form). Some of the most insightful input came from informal online forums such as Film-tech.com, an online point of exchange for professionals from the post-production and projection fields (mostly in North America and Western Europe) active since 1999. The forum is a treasure trove for minute technical details on projection devices and contraptions and various obscure projection standards, many of which are otherwise poorly or inaccessibly documented. Since the forum’s foundation, a number of continuously contributing members have been discussing the various standard and high frame rate formats in existence; many have had direct viewing experience and noted down their impressions and predictions. As such, their publicly viewable discussions over the past 15 years are a valuable form of ‘recorded oral history’. Similar forums that have been a priceless resource for information otherwise not available are, for example, the Adobe Community,33 CreativeCOW.net or doom.org. The first two are also frequented by professionals in the film industry but involved rather in film production, and often deal with daily problems in the production and distribution workflow: standards conversion, compression codecs, frame rate changes, and so on. Doom.org was originally founded to address DVD conversion, but has over many years expanded into an informal library specializing in video capture and encoding, broadcasting, DVD and Blu-ray authoring and so on. Its members are mainly ‘amateurs’ – a very misleading
32 http://mediahistoryproject.org/ 33 http://forums.adobe.com
term for highly skilled experts not participating professionally in formal industries of cinema and television.
Members of various other forums are also engaged in critical knowledge-production and archival activities that are addressed neither by professional literature, film studies research nor industry papers. An example is the documentation and categorization of Blu-ray players, projectors or television sets based on their support of various frame rates. Such information is often fragmented over dozens of hyperlinks and threads, but at the same time readily accessible, searchable and (because of the informal peer-review of other forum participants) most often highly reliable.
A fraction of the sources relating to early television technology consists of German patents and industry reports. Similarly for high definition television but also for medical papers regarding vision and the function of blinking, a number of Japanese-language publications and scientific papers have been used. All of these are instrumental not only in the ‘hardware’ portions of this thesis as an archive of technological history, but are necessarily included in the discourse-analytical ‘software’ part in order to understand and describe the governing epistemologies of media. I have coupled these technical sources (from ‘insiders’ of technology) with a range of literature dealing with the use and reception of the media – in the case of television and cinema, this meant engaging with part of the canon of media studies in which prominent thinkers and theorists have, as has been mentioned above, in significant ways constructed and shaped these objects both within the discipline and in popular parlance. In the case of high frame rate projection, as it yet lacks a comprehensive media theoretical inquiry, I have attended to film reviews of The Hobbit which because of its perceived and factual technological ruptures with previous projection technology has sparked a debate confronting cinema, television, seeing and reality – precisely those objects that I intend to analyze. These sources will allow me to approach how conflicting media practices are imagined, what their properties, functions or essences are believed to be, how their material properties and the discourses about them affect how human vision is performed.
Although I used the word ‘insiders’ to technology, I certainly do not believe that engineers or inventors have a privileged access to knowledge, or that they precede discourses on media by the ‘outsiders’ – spectators, philosophers, film critics, and so on. Instead, media technologies
are understood here always already as embedded in these various modes of reception that necessarily imply institutions, power relations and hierarchies.34
The thesis is structured into three main chapters dealing with film, television, and high frame rate projection, in the order that I believe our dominant visuality is shifting. I am aware that such a categorization might somewhat undermine the argument against media essentialism, as it presents three media practices as if they were clearly delineable. Nevertheless, the discussion within the actual chapters will hopefully blur the boundaries as I draw out the multiple overlapping, internally contradicting and thoroughly heterogeneous relationships of these media.
All three chapters discuss the media practices as fields of contesting technologies operating within discourses of science, economies and epistemologies of seeing. Out of these technologies, some have emerged as dominant forms that have, reversely, come to shape, structure and leave their marks and artifacts on discourse and relationships with reality.
2. Film – truth interrupted
2.1. The frame rate as an expressive instrument
The camera, whether photographic or filmic, is an instrument that enables the manipulation and production of time. Victorian-age post-mortem portraiture exemplifies the desire projected onto photography to stop the flow of time and preserve a fleeting instant. Muybridge’s high-speed photographic experiments, already anticipating motion pictures, proved to be a brilliant example of the ‘production’ of reality by dilating time through the photographic eye. The control over time was profoundly tactile: cameras, photographs, film strips and optical toys could be – had to be – touched and felt. Although Mary Ann Doane was referring to pocket watches, the same logic applies to these time-manipulating devices: “Time was indeed felt.”35 Many optical contraptions of the time like the Kineograph, now more commonly known as a flipbook (whose cinematic roots are still evidenced by its German name “Daumenkino”)featured the same kind of interactivity and afforded its spectators/operators
34 cf. Rossell 1999 35 Doane 2002: 4
time-bending abilities, allowing them to slow down, stop, or speed up the passage of time with the movement of their hands.36 Time was reversible. The fourth generation of the Optigraph projector sold by Sears Roebuck in 1907 advertised as one of its main selling points the ability to be easily operated both forward and backward,37 enabling projectionists to turn around time at will (and double the exhibition opportunities of their one-reelers, which audiences now could enjoy backwards). The brothers Lumière had already demonstrated this in Démolition d'un mur in 1896.38
What the camera, uncritically assumed to be the ultimate instrument for the recording of “truth,” created was a curious conception of reality: a reality consisting of a succession of separable, self-enclosed and integral moments “quantified into lifeless units of time and movement”39 that could be re-animated at will. Racing horses everywhere were faced with a profound rupture: after Muybridge’s experiments, they began galloping a new, documentable kind of motion. A motion that implied no movement at all, but a stasis that could be archived and revised, subjected to scientific analysis or prolonged reflection with a claim to ‘objective’ visual proof.40 A strip of film, the product of the photographic and filmic representational authority, granted us control over a reality that is undeniable and manageable, a reality segmented into frames. This was the reality cinema emerged in.
It is important to remember here again the distinction between the speed of the camera and the speed of projection: the ability of cinema to manipulate temporality is not an inherent capability of the camera alone, but arises only from the discrepancy between recording and viewing speed. Godard’s magical number of 24 frames per second refers us to a certain cinema tradition after the institutionalization of sound film recording and projection norms in the late 1920s, where the recording and projection speed were mostly the same. But before that (and on its periphery), difference was the ‘norm.’ In earlier cinema practice, recording and projection belonged to entirely separate trades which rarely harmonized. In his fascinating autobiographical account of the birth of the Vitascope, Thomas Armat recalls his
36 cf. Huhtamo 2005:9, Strauven 2011 37 Malkames 1967 [1957]: 101. 38 cf. Ceram 1965: 151
39 Crary 2001: 144
40 Aumont traced this “new confidence in seeing as an instrument of knowledge, even of science” (1997:
and Charles Francis Jenkins’ early attempts to construct a film projector. In 1895, he constructed what he claimed “the first projection machine ever made that embodied an intermittent movement.”41 Armat and Jenkins used Edison’s Kinetoscope films, filmed at approximately 40 frames per second, but their machine would not run at more than half the speed.42 The projector was a “complete failure” in Armat’s own words. Apart from its impracticable weight and loud noise, it produced a slow-motion effect as well as, presumably, strong flicker. These shortcomings (as they were clearly perceived) stood at the beginning of a 30-year-long period in which separate recording and projection speeds were the field in which cinema’s realism was shaped. Some saw differences between the recording and exhibition speeds as a problem. Already in 1903 the Optical Lantern and Cinematograph Journal called for the establishment of common norms: “unless the projecting of the subject upon the lantern screen is conducted precisely at a speed corresponding to that at which the negative picture was taken, false representations of nature will result.”43 The dictum of a truthful representation of reality is governing here, something André Bazin would approve of, but the prevailing view was quite different. Although producers often marked their films with a preferred speed, exhibition speed during the silent film era was at the discretion of projectionists and theater directors. They adjusted projection scene-by-scene depending on the action, and generally preferred to screen movies faster that the recording speed in order to give motion pictures energy and liveliness greater than that of real life.44 In 1920, even F. H. Richardson, the most vocal advocate of increased attention to projection harmonization and relentless critic of over-speeding film admitted that “there are occasional exceptions where a scene may actually be improved by moderate overspeeding of projection.45” For my later discussion, it is important to point out that both the discourse and projection practice present
41 Armat 1935: 243
42 ibid. 245. In Richard Abel’s Encyclopedia of Early Cinema, Charles Musser specifies the Kinetoscope speed
as “shot and exhibited at approximately 30-36 frames per second” (Musser 2005: 516). A discussion of the
Kinetoscope speeds is continued in Section 4.3.
43 Optical Lantern and Cinematograph Journal 1903: 11
44 Lenning 2004: 138. There are other reasons as well. Fox (1983: 208) lists flicker as the motive behind
faster projection speeds: even with a three-blade shutter, a film projected at 16fps would still appear “jerky and unnatural.” Possibly the most important incentive was that by rushing the presentations as much as they could, theater owners were able to squeeze in more shows.
45 Richardson, F.H., “The Various Effects of Over-Speeding Projection”, SMPE X (1920), quoted from Wurtzler
the original frame rate as something possibly not realistic enough – ‘realism’ for early cinema audiences meant a moving image that had to slightly improve on reality. This is in sharp distinction to current high frame rate projection, but more on this will follow in Chapter 4. Overspeeding was deliberately counted upon also during production: Griffith’s The Birth of a Nation (1915) included under-cranked action scenes filmed only at around 12fps in order to heighten drama when projected at higher speeds.46 The practice of over-cranking was, at the same time, a game of catch between producers and projectionists. Knowing that their films were being rushed, producers also ordered cameramen to film faster47 – which, of course, did nothing to prevent the theaters from rushing even more. Because of this (rather than any perceptual reasons), frame rates kept steadily increasing during the 1910s and 20s.
James Card’s (1955) and Kevin Brownlow’s (1980) comprehensive analyses of cue sheets, interviews, reports, and journals revealed the diversity of filming and projection speeds: anything from 12 to 26fps even well after the spread of sound film. Frame rate choices were a complex combination of aesthetic, technical, and economic factors: dynamism but also increasing film length and narrative complexity that began bursting out of the confines of the one-reeler, underexposure, fire hazards and cost of film, among others.48 By altering the speed, cinematographers and projector operators of this age held – in a very material sense – control over the passage of time divided into frames. As late as 1926, producer Richard Rowland expressed his doubts about any norm for the projection of film:
[I]n the making of motion picture productions the tempo is one of the most powerful tools in the hands of the director for obtaining the desired effect. Some films are made, therefore, to be run slowly and some at greater speed.49
This was the same year 24 fps became a standard on the Movietone sound-on-film system – a standard whose perceptual effects would become so dominant that it would be called “the magic of cinema” almost 90 years later, after the premiere of The Hobbit. So dominant that Kodak, Sony and Panavision would receive a Technology & Engineering Emmy award in 2001, when they finally succeeded in displacing cinema onto video recorders by developing a video
46 Lennig 2004: 137 47 Brownlow; Card
48 Another thorough analysis of the discussion on projection over-speeding can be found in Chapter 2 of
Richard Koszarski’s An evening entertainment: the age of the silent feature picture 1915-1928 (1990).
camera that could record in 24 frames per second.50 The visually much inferior film standard, entirely inadequate in representing motion in comparison to video, had become such an epistemic reference point that other media regardless of their scientifically confirmed ‘better’ access to reality had to try to emulate it.
Standardizing the frame rate meant the final eradication of the last remnant of an interactive “player mode” of pre-cinematic optical toys51 which still lived on in cinema – one which had already retreated from the Mutoscope parlors to the hands of projectionists, and would eventually be lost to mechanical precision. Such a breach between touch and vision has been identified by Crary52 as a trait of the transition from classical to modern vision. Although cinema is nominally situated already in the paradigm of the modern observer, its own technological transformations and the automation of frame rates followed a similar transition. Many decades later, the same process would reappear in an entirely different cinematic optical toy: the mobile phone. I still have most of the video files I shot nearly a decade ago with the first mobile phone I owned that could record moving images. ‘Moving’ is almost an exaggeration with a frame rate barely reaching 5fps, but if I had to choose whether to preserve my little ‘home documentaries’ or some masterpiece by Méliès, the choice would be quite clear. Although I did not turn a crank, when my phone was recording these videos, it was balancing the same variables as silent films did. There are technical limits on how fast the tiny camera and processor can record image sequences. The frame rate changes depending on an economy of the image – the available data rate. And it also changes when the sensor balances light exposure. The shutter speed and frame rate are decreased to counteract the loss of brightness if the surrounding light drops, clearly a (pre-programmed) aesthetic decision. Unlike skilled cinematographers of the silent era who prided themselves in their own idiosyncratic cranking speeds, I, the amateur operator, have no real operative control besides moving the phone and stopping the recording. Other decisions are confided to the recording algorithms and compression codecs, but the mechanics of the process remains not very different from early cinema production and exhibition practices.
50 Perazzi 2002; SONY 2004 51 Strauven 2011: 152 52 1996[1990]: 19
As with early films, the frame rate of a phone video varies between recordings, but will also be fluctuating within the same ‘film’ or file. Despite varying frame rates, cell phone media formats have found other ways of roughly synchronizing audio and video, but they inscribe a different map, a different dispositif of cinema, and many difficulties arise if we try to displace the images produced this way. Even the Apple iPhone, so highly praised for its camera quality, records at variable speed. According to users’ experiences documented online,53 the rate between multiple recordings can range from 15fps up to the NTSC standard of 29.97 depending on light conditions, but will fluctuate mildly around these values also within a single file. It is striking that such information is documented largely by users trying to recontextualize these moving images into cinema production practices, and in the process become confronted with frame rates as markers of cultural and economic value. Professional editing software such as Adobe Premiere Pro, Apple’s Final Cut Pro, or Avid make it exceedingly difficult (sometimes impossible) to work with variable frame rate files. Synchronization issues, motion disturbances and incompatibilities haunt those who attempt to work with videos shot on mobile phones in a context reserved for institutionalized cinema.54 The whole lifespan of a movie, the whole cinema industry technologically locks out the possibility of variable frame rates – from the ‘professional’ camera to the theatre projector.
Although irritating, such experiences also illuminate how technology is never passive, never neutral, never not part of diverse and contradicting desires, economies and practices, and that it participates in constructing contexts, infrastructures and potentialities for both itself and other technologies. It is impossible to include a phone video in a theatrically released movie not because the technology would not allow it or because the ‘quality’ would be insufficient, but because these two technologies emerged in a media epistemology in which it was impossible to imagine something like that happening. iPhone videos are incompatible with professional editing suites because their envisioned purpose was entirely outside of that possibility – the recordings are assumed to be viewed privately as personal memories or shared with others online; the file size needs to remain as small as possible while retaining a
53 This information is taken from various discussions on online forums such as Apple’s own
http://discussions.apple.com, and others described in Section 1.3.
reasonable image quality. A mobile phone recording simply is not ‘cinema.’ But it is precisely this knowledge of what cinema is or is not supposed to be that mobile phone cinematographers are actively redefining when they discuss solutions to combining variable frame rate footage with professional editing software, the mobile phone with the film industry. They are producing new knowledge of media, imagining new possibilities of the moving image, and opening up new uses and practices.
2.2. The Magical Number
How, then, did cinema’s rigid standard of 24fps come into existence?
One of the established myths about the history of cinema (most often found in popular or industry literature) is that 24 fps was chosen as the lowest possible speed at which sound could be reproduced in a satisfactory manner.55 This gives the impression of a planned economic consideration – the cost of film in conjunction with sound quality. Recently, SMPTE fellow Mark Schubin has begun debunking this widespread myth based on Scott Eyman’s history The Speed of Sound, in which Western Electric engineer Stanley Watkins recalls how the speed was chosen in 1924:
According to strict laboratory procedures, we should have made exhaustive tests and calculations and six months later come up with the correct answer. […] What happened was that we got together with Warners’ chief projectionist and asked him how fast they ran the [silent] film in theaters. He told us it went at eighty to ninety feet per minute [21.3-24fps] in the best first-run houses and in the small ones anything from one hundred feet [26.6 fps] up, according to how many shows they wanted to get in during the day. After a little thought, we settled on ninety feet a minute as a reasonable compromise.56
The magical number at which cinema records truth was a largely arbitrary decision. Moreover, the speed had nothing to do with sound at all: Watkins was talking about Western Electric’s Vitaphone, one of the many sound-on-disc systems used at the time, for instance for A Plantation Act (1926), The Jazz Singer (1927) and several shorts before it. Because it was not tied to the film strip, the film could have run at any speed. The only requirement was that it
55 e.g. in Read&Meyer 2000: 26; Christie Digital Systems 2012: 1; and countless blogs dealing with high
frame rate cinema.
would run at a constant one, because of the sensitivity of the human ear to slight frequency changes.57 The Vitaphone standard would eventually migrate onto Movietone, a sound-on-film system developed by Theodore Case and Earl Sponable. Movietone originally ran at 22.6 fps but was switched to the faster rate after the technology was acquired by Fox Film, which, due to cross-licensing agreements between Western Electric, Warner Bros. and Fox allowed the engineers to use motors already developed for Vitaphone.58 From there, it was adopted rapidly. In 1928, Richardson was satisfied that sound films had finally synchronized shooting and projection speeds at a reasonably high and flickerless speed and sparked new interest in projection quality – sound film was “a marvelous progress.”59 By 1929, the Society of Motion Picture Engineers (which had advocated 21.3fps as a universal standard just five years prior60), without much fanfare, accepted 24 frames per second as the official norm of both taking and projection speed of standard 35mm film.61
But standardization was not entirely the ‘fault’ of either sound-on-disc or sound-on-film systems. Despite the relative autonomy to direct the exhibition that projectionists enjoyed, projection speeds even before sound films were not completely arbitrary. As Altman has noted, the musical accompaniment of silent films also necessitated a stable and predictable running time – in fact, cue sheets with an indication of the desired projection speed are sometimes the only way to establish what the speed of a silent movie was.62 Long before ‘sound speed,’ sound had already begun standardizing the frame rate.
57 Edison had already realized during his experiments with the phonograph in 1880. Altman (2004) deals in
some detail with the difficulties involved in keeping constant speed for sync-sound systems during the so-called ‘silent era.’ An early, technically very elaborate description of early systems and their difficulties is also given in “A Review of the Quest for Constant Speed” by E. W. Kellogg, published in Journal of SMPE 1937, 28.
58 Eyman: 122. Sponable’s (1947) journal recounts in detail the relations and license deals between the
various involved companies and individuals.
59 Richardson 1928: 872. 60 Brownlow: 166
61 SMPE 1929: 37. It is an interesting numerical coincidence of history that the number 24 is recurrently
deemed appropriate to record reality: For his Animal Locomotion photographs, Muybridge used at first 12, then 24 cameras (Lawrence 2003, Musser 1994: 49), and François Willème’s photo-sculpture studio also deployed the same number of cameras arranged in a circular wall to create a 360° image of a model positioned in the center (Wedel 2009). Anschütz’s first Electrical Tachyscope had 24 photographic images (Ceram: 132; Hendricks: 12). As a matter of fact, Musser (1994: 48) reports that Muybridge’s exposures “were separated by approximately 1/25 of a second. Had they been animated, this would have yielded a “frame rate” close to sound film – or PAL television.
It is The Hobbit that made the calcification of this standard of visuality visible by introducing a technology that is at the same time radically new and yet nothing more than a return to a time before there was anything magical about 24 frames per second. There is an admittedly magical power in the standard in that it made things disappear: the force with which it penetrated projection practice worldwide meant that nearly every cinema in the world for the greater part of its existence could only exhibit movies at one stable rate. The viewing conditions of what media archaeologists like to call early cinema truly became a thing of the distant (technological) past: impossible to simulate or recapture without distortions, inaccuracies, technical problems, monetary expenses. 24fps is the film archivist and historian’s nightmare: it is the institutionalized irrecoverability of the past. We are seeing many early films being released on DVD or Blu-ray or screened at film festivals and museums. Such objects and institutions derive a significant amount of legitimacy from their function as gatekeepers of film history,63 enabling access to the most ‘authentic,’ most ‘complete,’ ‘digitally remastered’ or ‘restored’ and ‘authoritative’ versions of early cinema’s preserved artifacts.
After the Berlinale 2010, the TCM Classic Film Festival 2010 launched the North American premiere of Lang’s Metropolis expanded with the recently found footage. “This nearly complete copy of Lang’s 204-minute original now stands as the authoritative version of the film,” the press release said.64 This language implies a fidelity to an ‘original’ (regardless of what that might be or why some of the many versions are more ‘original’ than others), but disguises the displacements and intercalations that must inevitably occur when a film created before frame rates were mechanically stabilized is presented with equipment and technology that does not allow such flexibility. Already expressing the exact duration in minutes is possible only in hindsight. Regardless of the intention of its creators (and how do we establish that, too?), no screening of a hand-cranked movie would ever last the same amount of time. The past is not being restored here, but actively produced.
63 Marijke de Valck has written on festivals and their function as gatekeepers of cultural capital, for instance
in “'And the winner is...': what happens behind the scenes of film festival competitions”. International
Journal of Cultural Studies. 2010. 13(3).
64 TCM 2009. Such events are, needless to say, greatly profitable as well: passes for the 4-day festival were
Cinema has, in fact, not one but two magical numbers. Brownlow showed how projection equipment has helped construct contemporary perceptions about early cinema: that silent films were jerky, badly shot or had flicker. These are not artifacts of the early cinematic image, but marks of the inadequacy of current technology to retrieve the past. Analog projectors often feature a ‘silent’ switch, which slows down the projection rate from 24 to 16fps, the speed of spring-wound clockwork cameras.65 Even current film restoration manuals give these as “the two basic speeds employed throughout film making history.”66 How and when projector manufacturers turned this number into a synecdoche for all silent films is less important than the fact that as with the Metropolis ‘original,’ a projector equipped only with the options ‘silent’ and ‘sound’ is also a knowledge-producing machine that orders and displaces the past multiplicities of silent cinema into a homogenous history constructed retrospectively from the present. Brownlow criticized how unbearably long silent movie projections become when projected at ‘silent’ speed, but James Card, film preservationist and what I consider a brilliant proto-media archaeologist, had already pointed out the same problem in 1955:
In presenting a silent film to a group of spectators, the program director should be sure of his purpose. Does he wish (as he often claims) to show the film as it was seen originally? Or does he wish to present the film as its maker intended it to be seen? […] [H]e should realize that Douglas Fairbanks' Robin Hood, for example, might have been shown in two hours and a half during slack periods of the day or in a little less than two hours during the evening, to squeeze in an extra show. If he wants to show Robin Hood at the speed originally specified in 1922, he will run it at 12 minutes per reel which is very close to sound speed. (The film will then last two hours and eight minutes.) If it is decided to show Robin Hood at the arbitrary 16 frames per second, the film will last exactly three hours! And this is the way poor Robin Hood is usually shown to Film Society audiences, painfully limping through the forests at a rate that gets him through his adventures a full 52 minutes later than Mr. Fairbanks intended.67
Which of the possibilities, then, do we choose as the ‘authoritative original’ that deserves to be shown at film festivals and released on a Blu-ray with fancy packaging? Where does it derive its authority from? Is the authority based solely on the longest possible accumulation of ‘lost footage’? These questions also present the issue as if we were able to play the films at
65 Brownlow 1980: 164 66 Read & Meyer: 24 67 Card 1955: 55
arbitrary speeds and pick the most appropriate one, but the truth is that it is exceedingly difficult to displace a silent film from the past accurately into any other context.68 A portion of cinemas and film museums have analog projectors equipped with “Microverters” and potentiometers that allow varying the frame rates, but they also need another ‘device’: experienced projectionists with historical knowledge of silent cinema. Not an easy find. Very recent digital projection standards (addressed in Chapter 4) appear to be moving towards somewhat greater liberty when it comes to frame rates, but they will certainly produce their own myths and epistemologies of silent cinema.
2.3. Flicker
More problematized than over-speeding in ‘silent’ years was the issue of flicker, the notorious ‘defect’ of cinema responsible for one of its many names. Contemporaneous sources suggest that the total elimination of flicker – the black gap in-between every frame – was posited as the great telos of the time. In their evaluations, film critics at the turn of the century included flicker as a point of judgment: reporting from a motion picture showing, it was customary to praise the projection if the flicker was little or not noticeable. The Optical Lantern issue mentioned previously presented several experimental methods of dealing with flicker, among them various irregular and composite shutter types, continuous movement, 69 simultaneous double projection of two positives and others.70 Deac Rossell has written on the success of mechanical intermittent projection over optical solutions to flicker (using revolving mirrors and lenses), tracing it to the mechanical dissolving effects of the magic lantern which formed
68 On the technical difficulties also cf. Brownlow 1980. The situation has changed little even now: Patrick
Loughney has briefly mentioned the issues in presenting for the first time Edison’s Dickson Experimental
Sound Film, currently recognized as the earliest sync-sound film recording in existence, in 1998. Among
them were projectors that could not run the strip at its original high speed, then believed to be 46fps (Loughney 1999). What emphasizes the irretrievability of the past even more, is that not even this speed is certain: Editor and sound designer Walter Murch who synchronized the footage for the Library of Congress later identified the speed of recording to be 40fps instead, as he wrote in the online forum of Cinema Audio
Society on 6 June 2000. (A copy of his report is currently available at FilmSound.org/murch/Dickson.htm).
69 Television pioneer C. F. Jenkins in particular remained an active experimenter with continuous projection,
as his numerous articles in the publications of the time show.
70 “The Flickerless Projection of Motion Pictures” , Optical Lantern and Cinematograph Journal 1903. The
brothers Skladanowski used a double film loop to decrease flicker on the Bioscop in 1895. More on this in Section 3.3.
