Double Take: Rotoscoping and the Processing of Performance – Kim Louise Walden

Abstract: In 1915 the Fleischer brothers, Max and Dave, developed a device known as the “rotoscope” which allowed the artist to trace over the original film footage frame by frame to make a more life-like rendition for animation. Rotoscoping’s digital descendent, known as “Rotoshop” was used to style Richard Linklater’s animations Waking Life and A Scanner Darkly, but whilst rotoscoping may originally have been developed to help animators achieve a greater sense of realism, it has never just been about verisimilitude. Drawing on the theories of such diverse commentators as Bertolt Brecht, Roland Barthes and Robert Bresson, the paper will consider two central questions: What have been the consequences of this digital animation technique for screen performance, and what spectatorial pleasures does this means of storytelling afford its audience?

Performing Contexts

During the 1990s the first generation of computer-generated imagery (CGI) animation was in full swing. Pixar produced Toy Story (John Lasseter, 1995), the first full-length computer-generated (CG) feature animation to great acclaim and the marketing which preceded Final Fantasy: the Spirits Within (Hironobu Sakaguchi, 2001) promised a cinematic future populated by “synthespians” or synthetic actors. The associated strategy of deploying well known actors to “voice” computer generated characters proved to be such a popular feature of CGI animations that in productions like Dreamwork’s Sinbad: the Legend of the Seven Seas (Patrick Gilmore, 2003) characters were visualised after vocal performances had been recorded by its star actors Catherine Zeta-Jones, Michelle Pfeiffer and Brad Pitt, rather than before. This enabled animation artists to incorporate some of the most recognisable visual features of the stars in the creation of their animated counterparts. A digital version of rotoscoped animation (whereby live action film footage is traced over line by line to create a life-like effect) used in the films of Richard Linklater went a step further and reinstated the relationship between the actor and their screen representation by re-inscribing the actor’s corporeal presence. In the light of these new kinds of performance, this article will consider two central questions: firstly what performance has come to mean in the context of contemporary animation? And secondly, what spectatorial pleasures are induced by the doubling effects of rotoscoping?

In discussion about film there is often an underlying assumption that performance is the sole concern of the actor. However in the context of animation, a broader definition is needed as the actor’s contribution is distilled through the processes of both digital artistry and technology. James Naremore’s (1988) work in the field of film performance has helped to broaden this definition. Drawing on ideas of E. Goffman, he adopts the conceptual metaphor of a “performance frame” which bestows “special performing significance” on all the individuals who take up designated “roles” contributing to the on-screen theatrical event presented to the audience, from digital artists to actors (Naremore 1988, 107). Recent scholarship in the field of digital animation has articulated a number of ways in which performance can be understood in this enlarged sense and in the next section I will provide an overview of some of these approaches to animation performance, which will serve to stake out the conceptual terrain of the enquiry.

For the first generation of computer-generated animation back in the 1990s, the principle objective seemed to have been the drive to attain an almost “photorealist” look or “performance” by digital means. Commentators observed a preoccupation amongst the digital animation industry with achieving the “reality effect” – a kind of synthetic realism aiming to create the look of real life objects and people through simulating the codes of analogue cinematography (Darley 2000, 83). In a sense it was the technology itself which dominated any consideration of performance as the principle spectacle at the time. Much film commentary was spent marvelling at the meticulous renditions of textures, be they plastic, fur, water or hair, in the films of Pixar and its rival Dreamworks, accompanied by a commensurate growth in interest in how such animations were made.

It was taken for granted that digital animation’s aspiration to achieve photorealism was the obvious goal to aspire to but since then other modes of performance have been identified which are worth noting here (Fore 2001). In her article “All Aboard The Polar Express: A ‘Playful’ Change of Address in the Computer-Generated Blockbuster,” Jessica Aldred (2006) contends that shifts in the broader economic climate coupled with consequent changes in the organisation of the media industries and the advent of multiple media platforms has had further aesthetic consequences for animation performance. Through a close reading of The Polar Express (Robert Zemeckis, 2004) and Final Fantasy: The Spirits Within, she demonstrates the emergence of a “playful” mode of performance which addresses viewers conversant with the operations of video games (Aldred 2006). Aldred identifies a mode of performance in which the viewer adopts the perspective of the CG character just as they would in first person shooter games which provide a sense of “being there” within game space while additionally, the use of multi-perspective performances mimics game play choices (2006). Aldred asserts that these modes of address co-opted from game culture combine with the classical cinematic convention of continuity editing to create a new kind of animation performance.

With the advent of “synthespians”, the implications of technology for performance have become a concern for many filmmakers and commentators in recent years. In her article “We’re OK with Fake: Cybercinematography and the Spectre of Virtual; Actors in S1mØne,” Sidney Eve Matrix considers Andrew Niccol’s cyberpunk film S1mØne (2002). Although the performance in this film is given by a human actor mocked up as a synthespian due to technological limitations at the time which limits its impact (Matrix 2006), nevertheless Niccol’s film raises questions around the issue of authorship of performance and ownership in relation to synthetic performance which bear some consideration (Matrix 2006). She draws the conclusion that performance has always been manufactured to some degree and so the synthespian performance can be understood as part of the spectrum of artifice and simulation which has always surrounded actors and the discourses of stardom (Matrix 2006).

Lisa Bode’s (2006) research in the area has focussed on the spectator’s experience of animation performance and in particular on the critical reception met by The Polar Express and Final Fantasy: The Spirits Within – the first animated release to feature performances by an entire cast of computer generated human characters. She observes how reviewers often drew on the vocabulary of the macabre to describe the much vaunted “synthespians” (Bode 2006). Figures in Final Fantasy were described as “three day old cadavers,” “dolls” and “dummies” due to the unnatural smoothness and waxiness of their skin which was reminiscent of the synthetic qualities of man-made materials like Teflon as well as a certain blankness in their eyes (Bode 2006). She concludes that it is the distance as well as proximity to the human performance that evokes a sense of the “uncanny” for the spectator.

Bode’s discussion of the uncanny draws on scholarship from the fields of 19th century entertainments, psychoanalysis and robotics. She concludes that sensations of the uncanny invariably seem to be associated with uncertainty about the status of an object – whether it is alive or dead. Furthermore this sense of unease is probably not just attributable to the object itself but to experience of the wider historical and cultural situation. In this instance she suggests that the current dominant framework in which audiences understand synthetic performance is arguably cybernetics, cyborgs and that the whole discourse around Artificial Intelligence (AI), arguing that the sense of the uncanny surrounding digital actors in CG animation reflects wider cultural anxieties about what it means to be human in such circumstances.

Bode cites the observations of the Japanese roboticist Masahiro Mori in the 1970s, who noted that when robots were designed to resemble humanoids, they received a warm reception yet when that resemblance became too proximate to human beings, they began to be viewed as “creepy” prompting audiences to identify the differences between these simulated life forms and the real thing (Bode 2006). Mori first coined the phrase “uncanny valley” to describe the impulse of revulsion as well as the need to distinguish artifice from reality and the consequent shortfall between the two (Bode 2006). Vivian Sobchack attributes the unfavourable reception received by the digital actors in Final Fantasy to the ambiguity of their status between live action cinema and animation (2006, 176). She contends that part of the problem lies in the fact that photorealist digital animation actors operate across two modes of representation each with its own set of values and indeed expectations which she terms “indexical” in live action cinema and “emblematic” in animation (Sobchack 2006, 176). Sobchack goes on to assert that each mode generates its own distinct form of realism immersing its audience in a coherent narrative world. However when these modes of performance are brought into juxtaposition with each other, it creates an ontological confusion for the spectator who is invited to scrutinise the shortcomings of digital animation as a realist illusion rather than focussing on its successes (Sobchack 2006, 177).

To illustrate this point Sobchack notes that the marketing rhetoric surrounding Final Fantasy’s promotion promised the animation equivalent of Bazinian “total cinema” – a “totally realist animation” and argues that this hyperbole induced judgmental attention by the film’s audiences (2006, 179). A case in point is the attention received by the character Aki Ross’s hair. Sobchack recalls how one viewer felt that Aki Ross’s hair had become “irritatingly noticeable.” Clearly within Naremore’s expanded notion of performance discussed earlier, Ross’s hair is as much the “point” as any other part of the spectacle. However its dramatic register “upstages” other aspects of the animation and therefore the whole film was doomed to photorealist failure because even though her hair was not the focus of attention in the film, it became the feature which audiences fixated upon (Sobchack 2006, 180). In other words digital animation’s obsession with the “holy grail” of photorealism had generated a particular form of spectatorship in which audiences were distracted by the promised “illusion of life”.

This overview of some of the recent scholarship in the field raises a number of issues about performance in digital animation including questions around realism(s), modes of address to the audience, diverse forms of spectatorship, issues around the identity of the performer and questions of authorship as well as experiences of the “uncanny”. These ideas provide a guiding framework as this paper now turns to the central focus of its enquiry – a form of computer animation called digital rotoscoping – and considers what the implications of this technique might be for performance.

Rotoscoping and Richard Linklater

Rotoscoping is not a new animation technique. With the aim of creating a more accurate rendition of human movement in their drawings, the Fleischer brothers, Max and Dave, developed a device which became known as a “rotoscope” back in 1915 (Crafton 1993, 169). In essence this method of rotoscoping was a forerunner of motion capture technologies. The digital version of the rotoscoping process, known as Rotoshop, was developed by MIT graduate Bob Sabiston during the 1990s and when Texas-based film director Richard Linklater made his first foray into computer animation with Waking Life (2001), it was Sabiston’s software that he licensed to style his films (Ward 2004, 32). What distinguishes Rotoshop from its analogue predecessors is that once the live action film footage has been shot and edited on DV, it can then be converted into QuickTime files and run through the software so that the animator is able to view the footage on a computer screen. Using an interactive pen on a pressure sensitive pad, what appears on the screen can be digitally traced to create the drawn version of the image. The particular feature of this program which frees the animator from the laborious task of having to hand draw every line in every frame is a process called “interpolation”. This allows the animator to draw lines in one frame, use the computer to skip forward a few frames and then trace the same line where it is now, relative to the last line. The programme then draws all the intervening lines and similar to the “tweening” process in Flash, saves the animator a great deal of time (Ball 2006).

Linklater’s first rotoscoped animation, Waking Life started conventionally enough working with a number of actors from previous Linklater projects with Wiley Wiggins who first appeared in Dazed and Confused (Richard Linklater, 1993) taking the lead role. The live action version was quickly shot and edited in six weeks from start to finish and then the animation began. With no firm preconceptions of exactly how this film would look, this stage in the film’s production process operated like a large-scale experiment. Under the guiding direction of Sabiston who acted as the film’s Art Director, Linklater agreed that artists (rather than exclusively digital animators) should be recruited and given free rein to experiment, which resulted in a film made up of a mixture of drawing styles from coloured crayons to cartoon, primitive to pixellated (Ward 2004).

Waking Life has little that would conventionally be described as plot. It tells the story of a college boy, whose name we never learn, who dozes off to sleep on a train and comes to in a kind of “consciousness” which can only be described as a waking dream. Upon arrival at his destination, the film follows his encounters with a series of individuals who freely share with him their perspectives and points of view on the nature of free will, dreams and life in general. In effect, the film takes place in that indeterminate place between sleeping and waking where the mind combines and recombines experiences from our waking lives into a surreal mix and this premise gives the film its license to string together a series of completely unrelated vignette scenes in which the protagonist comes across real-world people he has encountered in college (there are a lot of philosophy professors in this film), characters reprised from earlier Linklater films namely Julie Delpey and Ethan Hawke from Before Sunrise (Richard Linklater, 1995) and its companion piece Before Sunset (Richard Linklater, 2004) who seemingly continue to exist beyond the diegesis of these films, as well as a sprinkling of bar stool philosophers, random street types and ciphers.

What is interesting from the point of view of this enquiry is the way that from the outset definitions of performance are expanded as performances are duplicated with different modes of graphic presentation overlaid onto live action film. Moreover the notion of performance is extended beyond the parameters of what is seen on the screen to include the animation process itself.

The Animator’s Performance

Conventionally animation houses like Disney run production on factory lines complete with strict deadlines and style manuals (La Franco 2006). By contrast, Linklater set out with a much more experimental, artist-led approach to the production of his first rotoscoped feature, Waking Life. He held auditions and “cast” (his word) artist animators in much the same way as actors were cast in character roles for the live action version of the film (Linklater 2002). Most of the individuals working on the film were not animators and had not used the technology before (Ball 2006). Accounts of the film’s production suggest that artists were invited to experiment with the software and were only “cast” to work on a character once a particular drawing style had been agreed upon (Desowitz 2001).

Acting teacher and writer Ed Hooks has long been an advocate of the importance of acting for animators, asserting that when watching an animated film, you are watching a performance crafted by an animator (Hooks 2003, 6). Academic Paul Wells concurs that animators must also use acting techniques to understand and project the particularities of the character they are crafting (Wells 1998, 104). Drawing on the performance theories of the Russian actor and theatre director Constantine Stanislavski, Wells argues that animators in effect have to engage in a process of identification with a character, much like an actor, and formulate their own modes of performance practice to “embody” movement and emotions graphically. In other words, the animator must develop strategies to graphically signal the motivations of a character (Wells 1998, 105).

Hooks’s guiding strategy is to encourage animators to focus on the principles of motivation through emotions, thinking processes and physical action (Hooks 2003, 20) but the problem with this approach is that it fails to acknowledge that there are fundamental differences between the ways that an actor and an animator approach performance (Veltman 2003). Unlike an actor, the animator usually starts from scratch. There is no body, voice or face – just a blank screen. Character has to be constructed from the ground up with every detail consciously designed (Veltman 2003). Moreover an animator’s working methods are not in essence Stanislavkian – acting in animation is not so much about authenticity, improvisation and chemistry between actors as about a process of mediation: interpretation of the performances of the voice-over soundtrack, translating live action footage of actors into drawing as well as collaborative processes of working with other animators and directors (Veltman 2003).

Hooks acknowledges that actors and animators approach characterisation from opposite directions (Hooks 2003, 10). Generally speaking actors work from the inside out establishing motivation for an action. They establish what they are doing, why they are doing it and then make it visible through movement, gesture and expression. Whereas the opposite is true for the animator who starts with the outside – the storyboard and the voice-over and works backward to establish the inner motivation (Veltman 2003).

Hooks takes a dim view of rotoscoping techniques altogether, which he describes as a “second generation form of performance”, a pale facsimile of the original (qtd. in Veltman 2003). He sees the technique simply as a process of copying a performance and overlooks its potential for expressive interpretation of the actor’s recorded performance, primarily through the very act of doubling itself (Veltman 2003). To illustrate this point it is worth considering one of the most powerful and disturbing scenes in Waking Life (for which it earned its 15 classification) featuring the lone figure of a prisoner incarcerated in a cell. Played by Charles Gunning the prisoner angrily rails against those people he considers responsible for his incarceration. He paces up and down in his cell shouting and fantasizing in gruesome detail his plans for revenge against those who conspired to throw him in jail.

fig 1
Fig. 1: The animator creates the character and so becomes an actor. Source:

A close reading of the scene reveals that whilst the voice-over outlines the type of character we encounter here, it is the animation which really allows the viewer to understand the character using line and colour to dramatise every expression and gesture. The scene takes place in a grey stone-built prison surrounded by barbed wire. The cell walls are grey and even the singlet the prisoner wears is grey. The colouration of the background is unremittingly drab yet juxtaposed with this greyness is the startling puce-red colour of the man’s complexion. His face seems to literally boil with rage. The dark etched lines on this face contort his features as he rants and the jagged white teeth add to his menacing mien. Together these features of line tone and colour express the prisoner’s anger but what we become aware of, as the scene progresses, is that this anger has another purpose. It is the very thing that sustains this man in his captivity. Wells states that the objective of the animator is to make every movement a poignant expression of the nature and implication of the character’s thoughts. Here the animator’s design represents aspects of the character’s personality that cannot be explicitly stated by the voice-over actor. On one level the animator creates the character and so becomes an actor (Wells 1998, 107) but it is the relationship between the animation and the voice-over acting of Gunning that creates the character.

As suggested earlier by Matrix (2006) in her article on “synthespians” and synthetic performance, the authorship of the performance is a collaborative endeavour here. A feature on the 2002 Waking Life DVD called “Animation Scrap Heap” provides us with our first clue. Amongst the discarded clips is an earlier version of animator Mike Layne’s designs for the prison scene, indicating that a process of editorial revision helped to create the final version of the film. Contextual reading reminds us that the animator’s performance is rarely a solitary one but rather it emerges from the intersection of a number of creative contributions. The scene’s script is lifted almost verbatim from a Hubert Selby Jr story called The Room (2002), the voice skills are attributed to actor Charles Gunning and the input of art director Sabiston who also designed the software alongside Linklater. Clearly others could be added to this list and so any definition of performance in animation needs to acknowledge its collaborative nature. It is the synthesis of these ingredients that creates characterisation.

Signification as performance

Many scenes in Waking Life feature Linklater’s family, and friends and people chosen (not cast) because of who they are and what they have to say, rather than for their capabilities as actors. Academics, artists and filmmakers are all invited to express their points of view and so in a sense the film is closer to documentary than fiction. It is in these scenes that the viewer becomes aware of the role rotoscoping plays in transforming “talking head” interviews into something which could be described as a performance. Rotoscoping doesn’t just trace over the live footage, replicating the performed event but operates to construct positions and effects from which the spectator can understand what they are seeing. For example one scene features a university professor explaining transhumanism and biotechnology. He is drawn by the painter Nathan Jenson in a highly expressive manner emphasising the elasticity of his features and gesticulating hands whilst his speech is annotated with Jenson’s sketches and diagrams to clarify the meaning of what he says. Here rotoscoping process is used as a form of graphical sub-editing as well as an artistic tool.

In Camera Lucida (1980/2000, 20), Roland Barthes described photographs as having the qualities of an “animation” – in the sense that certain kinds of image construction “animates” its spectator and indeed the spectator “animates” it. In order to fathom the ways that Waking Life’s rotoscoped images articulate, I found Barthes’ approach to photographs illuminating. To explain how meaning is derived from a photograph, Barthes dissects an image by the Dutch photographer Koen Wessing titled Nicaragua (1979) portraying a rubble-strewn street corner patrolled by two soldiers carrying guns. In the background two nuns cross the street in the opposite direction. Barthes identifies the “co-presence” of two elements in the photographic image: the soldiers and the nuns. He describes the former with the Latin term – studium referring to “the application to a thing, a taste for someone, a general kind of enthusiastic commitment” that we can take to mean the subject of the study – what the picture intends to say in general (22). The nuns he describes as the punctum – quite literally “a sting, a speck, a cut, even a little hole” which I take to mean the point or split which draws the attention of the viewer but doesn’t necessarily uphold the message – indeed, might even run counter to it (26-27). He discerns that it is the juxtaposition of these co-present yet discontinuous elements (the soldiers and the nuns) which generate the “adventure” of the image (22).

A parallel can be found in a scene in Waking Life sub-titled “The Holy Moment” which begins with the protagonist visiting a film theatre and settling into a seat to watch two individuals on screen discussing the nature of film. One is the filmmaker Caveh Zahedi explaining French film theorist Andre Bazin’s ideas about cinema. It is this conversation about Bazin’s proposition that film’s unique quality as an art form lies in its ability to capture and reproduce moments of reality, which forms the ostensive studium of the scene.

However, during the course of the on-screen conversation, other elements on the screen start to distract the viewer’s attention. Zahedi’s hair is blown by a wind whilst his companion’s hair remains curiously still and unruffled. As if to underline the point a fan is pointedly positioned between the two figures, inert and clearly not the source of the breeze. This incessant “wind” forms the punctum of the scene, visually contradicting Bazin’s proposition. By subjecting the Zahedi figure to an animated “wind”, attention is directed towards the craft of construction and the potential for artifice in the animation. To reinforce this point, at the conclusion of the scene, the animation starts to dissemble – walls fall away, props disappear and the two figures are turned into clouds playfully drawing out some of the contradictions in Bazin’s philosophy on the nature of filmic realism from the perspective of the cinema’s digital age.

fig 2
Fig. 2: Rotoscoping’s “Holy Moment”. Source:

Now, clearly there are material differences between Barthes’ photographs and the digitally animated image. The features of an analogue photograph are “found” referents in the real world framed by the photographer through the camera viewfinder, whereas in digital animation they are consciously designed into the mise en scene of the image. However despite these differences in the way that the visual is produced, this model of signification seems to raise questions about the shifting paradigm of realism and cast a light on the way that rotoscoped images in effect annotate their source material to express the animator’s “take” on the subject.

Rotoscoping may originally have been developed to enable animators to achieve a greater sense of realism in character movement, but it has never just been about achieving verisimilitude. By creating a deliberate tension between the real and the represented, what this digital technique seems to make possible is a form of animation that questions the relationship between notions of truth, verisimilitude and reality. What Linklater’s animation films share is a fascination with mind-altered states of reality whether dream-induced, as in Waking Life, or drug-induced, as in the case of his next animation film A Scanner Darkly (Richard Linklater, 2006). Where they differ is the way rotoscoping is used. In Waking Life, emphasis was placed on exploring rotoscoping’s creative potential for expression, whereas A Scanner Darkly deploys rotoscoping to deliver a much more literal graphic representation of the live action film which has quite different consequences for the actors and their performances.

A Scanner Darkly and Star Performance

Linklater’s second animation film A Scanner Darkly is an adaptation of a short story written by cult science fiction writer Phillip K Dick in 1977. Set in the near future in suburban California, it depicts a world which looks remarkably like the present in which the so-called “war on drugs” is being waged by the narcotics division of the police department with the support of an elaborate surveillance system (the scanner of the title) and undercover agents in animated scramble suits – all-in-one body masks which project onto the wearer an ever changing collection of facial and body features.

In A Scanner Darkly questions of performance circulate more directly around actors and acting than its predecessor, Waking Life. Casting for this film, Linklater capitalised on the currency of the Hollywood star system, employing high-profile stars such as Keanu Reeves in the leading role with a supporting cast that includes Winona Ryder, Woody Harrelson and Robert Downey Jr. Casting well-known stars of the screen has become a standard practice in animation and clearly contributes to the meaning of their performance. At script level, playful references are made to the star cast’s real life notoriety. Winona Ryder’s first line is a complaint about someone stealing – she had a well-publicised conviction for shoplifting. Then there is a reference to the removal of a spleen – Reeve’s ruptured spleen was removed after a motorcycle accident (Esther 2006, 65). Whilst as one film journalist noted, a canny eye for casting has all of Scanner’s supporting performers linked in various ways to drug use (Rowin 2006, 77).

Stars accrue meaning not only from media coverage but from past film roles too. Reeves has had an extensive film career working with a wide range of directors from Bernardo Bertolucci to Kathryn Bigelow which has brought him considerable fame (Giarratana 2002, 69). However it was during the 1990s that Reeves found particular success in the science fiction genre with his deadpan delivery as the eponymous Johnny Mnemonic (Robert Longo, 1995) and in the part of Neo in The Matrix trilogy (Wachowski Bros., 1999, 2003, and 2003). Commentators have noted how Reeve’s “blank” and some would say “vacuous” style of performance was uniquely suited to the cyborg figure and that it became something of a generic feature of performance in science fiction films at this time (Cornea 2003).

When viewed from a historical perspective Reeve’s much remarked upon “blank” style of performance is not particularly unusual. There have always been stars who have offered up performances marked by a facility for little more than an inscrutable facial expression – consider Buster Keaton, Greta Garbo, and Robert Mitchum, who have all based successful careers on such blankness. It could be argued that this is a mode of performance more commonly associated with photography than cinema where the performer is considered to be a “model” rather than a “character actor” much like any other property appearing in the film’s mise en scene. This is a style of performance advocated by the French film director Robert Bresson. He endeavoured to make a distinction between theatrical conventions of performance derived from the Stanislavskian approach and what he considered to be purer cinematic forms of expression. Bresson favoured a form of non-performance (Bresson 1991, 80). Often using individuals who were not actors at all, Bresson required his “models” to repeat multiple takes of each scene until all semblance of presentation and emoting were stripped away from the gesture or action. Bresson asserted that through this non-performance approach what he called the “automatism of real life” could be revealed and the spectator would be able to see truths that were normally concealed beneath the surface of everyday existence. To this end he argued that film actors should aim for performance focused on “being” rather than “seeming” in order to achieve a sense of naturalism (Bresson 1991, 81).

This notion of “being” rather than “seeming” as a mode of performance provides a clue to Reeve’s most notable quality as a film star. In contradistinction to Ed Hook’s account of actorly processes which entail the actor establishing a rationale for their character’s actions and then visualising this through movement, gesture and expression, Reeves’s performance lies not so much with his ability to act as with the way he looks (Rutsky 2001, 189). Filmmakers have tended to capitalise on Reeves’s singular ability to “perform” his appearance with long lingering close-ups and it has been argued that the very presentation of his appearance seems to achieve the status of a dramatic act in itself (Giarratana 2002, 69-70). For Reeves, his very star presence substitutes for performance, all of which has bearing on the way he is presented in animation.

In A Scanner Darkly we first encounter the character of Bob Arctor, played by Reeves in the exposition scene at the start of the film. During the course of the scene the star is revealed to the viewer incrementally through three different registers of appearance. The first time we see him from the outside in his “scramble suit” disguise. Next we see the world from his perspective with “point of view” shots reminiscent of the cyborg repertoire with which Reeves has been so closely associated, and then a third perspective is introduced which is the extreme close-up rotoscoped version of Reeves in character (see below).

The rotoscoped face of Reeves as Bob Arctor is a curious image made up of layers: the animated character and the star, the rotoscoped and the real. Each of these layers brings with it their own language and lying one on top of the other builds up a kind of composite figure. For the viewer, this is a “double take” as the translucence of the rotoscoped figure means the viewer is never able to forget the recognisable face of Reeves lurking underneath the animation and hence the layers are never completely able to coalesce into a single unified figure. Taken together these layers of representation exceed the conventions of animation, which has the effect of making the figure “spill out” over the outlines of the character.

fig 3
Fig. 3: Keanu Reeves: the rotoscoped and the real. Source:

In her work on the early rotoscoping of the Fleischer brother’s cartoon character Betty Boop, Joanne Bouldin (2004) has also noted the intriguing spectatorial experience that rotoscope produces and makes observations which have some relevance here. Bouldin asserts that the rotoscoped figure seems to possess a “weightier” presence than conventional animated figures as a result of its material connection with the original. The figure literally borrows substance from the star body and that connection “fleshes out” the animated body, bringing it into a closer approximation to the real – creating what she describes as an “ontological duality” to the rotoscoped image (Bouldin 2004, 13).

Translucence is not rotoscopings only aesthetic quality. Writing on Waking Life, Ward identifies the way in which rotoscoping has the effect of simplifying what it represents (Ward 2005). In a manner reminiscent of Andy Warhol’s screenprints of such figures as the film star Marilyn Monroe, rotoscoping reduces live action recordings to a combination of line, texture and colour. The rotoscoping process translates the indexicality of the photorealistic filmic image to its essential emblematic components which has the effect of reducing extraneous visual information and, in some ways, making its meaning clearer. Like Warhol’s image of Marilyn which picked out the peroxide blond hair, half closed eyes and lipstick coated, half open lips as the quintessential features of the glamorous star, the rotoscoped image of Reeves focuses on contours and planes of his familiar, beautiful face together with the characteristic goatee beard by which the star is known.

Here we see the “doubling” effect of rotoscoping, which in turn evokes a digital reprise of Warhol’s critique of celebrity culture in which star images are commodified and endlessly replicated across media platforms in the contemporary media landscape. The viewer has a sense of the individual underneath the animation who is obscured by its rotoscoped version, and so takes on a ghost-like, almost spectral, quality (Ward 2005, 169). It seems that rotoscoping has paradoxical qualities which on the one hand serve to capitalise on an audiences’ recognition of star personas and live performance but on the other hand obscure its subject constructing an animated iteration of the star (Ward 2005).

To further complicate this paradox, it is interesting to consider this imaging process in the light of animation performance theories. As noted earlier, in his discussion of the animator’s practice, Wells argues that the animator must learn how to act and drawing on the performance theories of Stanislavski, he describes how an animator must empathise with the character in order to render them in such a way that the audience is able to believe in them (Wells 1998, 104).

In the case of rotoscoping however it could be argued that the opposite is true. Rotoscoped animation is more akin to Bertolt Brecht’s “alienating effect” than Stanislaskian empathy, as the viewer is always aware of the layering effect and so are prevented from being immersed in the actors’ performances. Brecht advocated an approach to theatrical performance that sought to create a gap between the performance and the audience in order to distance the audience and thereby prompt the spectator to question the very nature of theatrical construction (Willett 1964, 201). I would assert that the processing of performance by rotoscoping technologies provides the required disruption to create a distance between text and spectator even though it is generated by the technologies of animation rather than the performers themselves. The rotoscoping of performance is the opposite of “naturalistic acting”. Just as we saw with Barthes’ punctum earlier in the article, rotoscoping creates a deliberate tension between the real and representation, which brings into contention questions about the configuration and composition of star identities and the constructed nature of stardom. When a reporter asked Reeves in interview how he coped with stardom and the attention from fans, he replied: “ I’m Mickey [mouse]. They don’t know who’s inside the suit”. “But you’re a movie star”, replied the reporter. Keanu laughed, “So’s Mickey” (Giarratana 2002, 75). As if to prove a point here Reeves aligns himself with an animation character and acknowledged the synthetic quality of contemporary stardom as well as the shifting role of the actor in digital cinema.

The “cartooning” of the actor’s performance

The following, final example continues to mine the questions of performance which circulate around the actor, acting and the presentation of character but illustrates the tendency alluded to by Reeves of how actors adjust their performance in anticipation of the shift in the process of production from live action to animation.

In interviews, some of the supporting actors indicated that they had altered their acting in view of the fact that their performances were to be rotoscoped (Linklater 2006). Rory Cochrane talked about caricaturing certain of his facial expressions in the knowledge that the performance was to be animated. Similarly Woody Harrelson admitted his performance was at times “over the top”. Knowing that the film was to be rotoscoped he felt this gave him “license to go a little more nutty” than usual (Klein 1993).

In a series of interlocking scenes which punctuate the main plot, the supporting cast of Arctor’s feckless friends (including the paranoid drug addict Freck played by Rory Cochrane, hapless-hippy Luckman played by Woody Harrelson and garrulous Barris played by Robert Downey Jr) perform their roles as if they are cartoon characters in a series of comic sketches more familiar to conventions of cartoon animation. After their vehicle breaks down on the road, Barris, Luckman and Bob attempt to mend the car. At the start of the scene Barris blinds his companions with technical explanations for the mechanical failure of the car but his companions refuse to take what he says seriously. Frustrated, Barris starts to “toon” his talk accusing Luckman and Bob and pointing to each one in turn saying: “You are just heckle tweak but you’re a bug-bite squared”. Luckman played by Woody Harrelson picks up the cue, rhyming his retort: “ Step back Freck & Frack, Ernie’s on the attack.” The two characters circle around one another, posturing in an increasingly aggressive manner but the scene is framed as a gag. Barris signals the play-within-a-play: “OK, I’m the technician and you’re the interloper” – and we see the characters fall into what animation commentator Norman Klein has identified as standard cartoon gag roles (Klein 1993, 38).

The Reeve’s character, Bob, adopts the role of the “controller” – the figure who remains aloof from the fray throughout – while Luckman takes on the role of the “nuisance” provoking Barris into a fight. Barris assumes the function of the “over-reactor” and disappears into the house returning with a hammer as his weapon of choice. In the role of the “nuisance” Luckman maintains the gag by playing along. He picks up a rock and taunts his assailant jeering and calling him “Hammerhead”. The cinematography underlines the gag too circling with the assailants and when Barris reappears from the house, the camera shot is taken from behind him at hip height in silhouette to draw attention to the fact that he is now “armed” in a clichéd shot reminiscent of Clint Eastwood spaghetti westerns and punctuated by an Ennio Morricone-style soundtrack to hyphenate the filmic “quotation”.

fig 4
Fig. 4: The gag- “OK, I’m the technician and you’re the interloper”. Source:…/AScannerDarkly07.jpg

In each of these scenes we see live actors directed to explicitly “cartoon” their performances for comic effect in order to bring their performances into a more graphical mode. Costume also signifies the “cartooning” of the scene. Barris sports a costume of a car mechanic for the part: flack jacket complete with numerous tool pockets as well as plastic gloves, all juxtaposed with a pair of incongruous outsized pink slippers reminiscent of Mickey Mouse’s oversized boots. Framed by the costumes, props, sound effects and camera shots, the mode of performance clearly indicates to the audience how it should “read” these episodes. Here the film explicitly shifts register from the rest of the drama and opens up alternative lines of viewing these comic visual and aural modes of performance. By the end of the scene Barris seems to have lost the power of speech, no longer expressing himself in words but in Warner Brothers-style cartoon voice effects calling after Freck as he jumps into his car. Clearly here we are seeing a kind of osmosis between the live action and the animated version of the film as rotoscoping processes the performances of its actors.


The starting point for this article was a sense that technical advances in digital animation have had aesthetic consequences not just for the quality of imagery but also more generally for the nature of performance within animation. Linklater’s use of digital rotoscoping in Waking Life and A Scanner Darkly seems to retain a particularly close association between its live action origins and their animated counterparts, providing a novel form of doubling to examine in the light of performance theories.

In different ways all of the animator and actor’s various modes of performance encountered in this investigation have been predicated on rotoscope’s particular facility for articulating the tension between the real and the represented. What began at the start of the 20th century as an animator’s tool for tracing over live action to achieve a more realistic result has evolved into a rich graphic language of performance which has opened up fresh new ways of seeing for the animation audience. These new forms of spectatorship have arisen from ongoing questions around authorship and performer identity in the wake of the “synthespian”, experiences of the “uncanny”, and the layering or doubling effects enabled via digital rotoscoping processing.


Aldred, J. 2006. All Aboard The Polar Express: A “Playful” Change of Address in the Computer-Generated Blockbuster. Animation: an Interdisciplinary Journal 1, no. 2, (accessed August 22, 2008).

Ball, R. 2006. Poetic Scansion. Animation Magazine 20, no. 8,;hwwilsonid=A (accessed April 16, 2008).

Barthes, R. 2000. Camera Lucida. (1980). London: Vintage.

Bode, L. 2006. From Shadow Citizens to Teflon Stars: Reception of the Transfiguring Effects of New Moving Image technologies. Animation: an Interdisciplinary Journal 1, no. 2, 2/bode (accessed April 8, 2008).

Bouldin, J. 2004. Cadaver of the Real: Animation, Rotoscoping and the Politics of the Body. Animation Journal 12: 7-31.

Butler, J. 1991. Star Texts: Image and Performance in Film and Television. Detroit: Wayne State University Press.

Cornea, C. 2003. David Cronenberg’s Crash and Performing Cyborgs. The Velvet Light Trap 52 (Fall): 4-14.

Crafton, D. 1993. Before Mickey: The Animated Film 1898-1928. (1982). Chicago: University of Chicago Press.

Darley, A. 2000. Visual Digital Culture: Surface Play and Spectacle in New Media Genres. London: Routledge.

Desowitz, B. 2001.Waking Life Frees Animators with User-Friendly Software and Indie Approach. Animation Magazine 15 (10): 4-5.

Esther, J. 2006. The Transparency of Things: An Interview with Richard Linklater. Cineaste 31 (4): 64-65.

Fore, S. 2001. Romancing the Rotoscope: Self Reflexivity and the Reality Effect in the Animations of Jeff Scher. Animation: an Interdisciplinary Journal 2, no. 2, 2/fore (accessed April 8, 2008).

Hooks, E. 2003. Acting For Animators. Portsmouth: Heinemann.

King, G. 2005. The Spectacle of the Real: From Hollywood to Reality TV and Beyond. Fishponds: Intellect Books.

Klein, N.M. 1993. 7 Minutes: The Life and Death of the American Animated Cartoon. London: Verso.

Korkis, J. 2003. If at first you don’t succeed…call Peter Jackson. Jim Hill Media 25 June,

La Franco, R. 2006. Trouble in Toontown. Wired 14, no. 3, (accessed April 10, 2008)..

Langer, M. 2004. From the Guest Editor. Animation Journal 12: 3-6.

Linklater, R. 2002. “Commentary”. Waking Life (2001), DVD. Directed by Richard Linklater. Beverley Hills, CA: Twentieth Century Fox Home Entertainment.

Matrix, S.E. 2006. We’re OK with Fake: Cybercinematography and the Spectre of Virtual Actors in S1mØne. Animation: an Interdisciplinary Journal 1, no. 2, (accessed August 22, 2008).

Naremore, J. 1988. Acting in the Cinema. Berkeley, LA and London: University of California Press

Ndalianis, A. and C. Henry, eds. 2002. Stars in Our Eyes: The Star Phenomenon in the Contemporary Era. Praeger: Westport, USA.

Rowin, M.J. 2006. A Scanner Darkly. Cineaste 31, no. 4.

Rutsky, R.L. 2002. Being Keanu. In The End of Cinema as We Know It, ed. Jon Lewis, 185-194. London: Pluto Press.

Selby Jr, H. 2002. The Room. London: Marion Boyars.

Sobchack, V. 2006. Final Fantasies: Computer Graphic Animation and the (dis)illusion of Life. In Animated ‘Worlds’, ed. Suzanne Buchan, 171-182. Eastleigh: John Libbey.

Veltman, C. 2003. Pulling Faces. Chloe Veltman. (accessed September 3, 2008).

Ward, P. 2004. Rotoshop in Context: Computer Rotoscoping and Animation Aesthetics. Animation Journal 12: 32-52.

Ward P. 2005. ‘I was dreaming I was awake and then I woke up and found myself asleep’: Dreaming Spectacle and Reality in Waking Life. In The Spectacle of the Real: From Hollywood to ‘Reality TV and beyond, ed. Geoff King, 161-171. Fishponds: Intellect Books.

Wells, P. 1998. Understanding Animation. London: Routledge.

Willett, J. ed. and trans. 1964. Brecht on Theatre: The Development of an Aesthetic. London: Methuen.


1. With thanks to my journal reviewer for raising this perspective in feedback comments.

Author Bio
Kim Louise Walden teaches digital culture and discourse in the School of Film, Music and Media, in the Faculty for the Creative and Cultural Industries at the University of Hertfordshire, UK. Her research interests revolve around the impact of new media on film and in recent years she has presented conference papers and published articles on film and video games, film web sites and the first generation of films made for mobile phones.

Email Contact: