Exploratory data analysis and the editing structure of Friday the Thirteenth (1980).
To date, the application of quantitative methods to questions of film form has been overwhelmingly classificatory. This is because quantitative researchers have accepted a limited conception of film form proposed by Barry Salt determined by the frequency with which quantifiable stylistic elements occur:
Questions of style arise when we consider films in relation to other films ... the distributions of these quantities (shot length, etc.) for a particular group of films, say by a particular director, when compared with those for other directors working at the same place and time, give sure indication of the existence of a personal style; in fact this is what formal style is. (Film Style 24)
This definition establishes a strong foundation for the classificatory mode, enabling researchers to distinguish empirically between what is unique and what is commonplace. It has been widely employed to differentiate between groups of films based on their shot length distributions, editing patterns, transitions, and the distribution of shot scales. (1)
It is not, however, a definition that contributes to our understanding of the formal system of a particular film. Salt's definition of film form is too narrow, privileging descriptive stylistics over the functional questions of style that arise when we consider how a film's formal system works to tell a story, to symbolically convey abstract themes to the viewer, to emotionally affect the viewer, or as pure decoration (Bordwell, Figures 34-35; see also Butler 11-15). For example, the frequency of close-ups in two films may be similar but this does not tell us if they use close-ups to create a feeling of intimacy or of threat for the viewer (see Persson), and though we might place these films in the same category, their use of a particular stylistic device may be very different. As Carroll argues:
Stylistic analyses bent on revealing the personal style of the filmmaker, period style, genre style, and school/ movement style are fundamentally taxonomic in nature. Analyses of these sorts may do other things. But they are ultimately committed to categorizing the film. Thus, they are category-relative in a way that guides selectivity. And they may miss the trees for the forest.
As methods, these strategies of stylistic analysis do not promise to account fully for the form of a particular film, given its specific aims and context. They enable us to place the film, and, though that is genuinely informative, they may not explain why the individual film qua individual film possesses the stylistic or formal attributes it does. For that we need a framework for discussing form in the individual film. ("Film form" 390)
The lack of an adequate framework for understanding form in the individual film means quantitative methods have contributed little to the explanatory mode of stylistics.
This essay demonstrates the application of exploratory data analysis (EDA) as a framework that advances the application of quantitative methods to questions of film form beyond the classificatory to the explanatory mode in order to understand how the system of style functions in the individual film. I argue that EDA offers two advantages for the functional analysis of film form. First, adopting EDA as a data-driven framework allows us to identify interesting and unusual features of film style which can then be investigated in order to determine the nature of the processes that produced them. The bottom-up approach of EDA does not require us to pre-specify which elements of film style are relevant to our analysis based on what we interpret the purpose of a film to be. In section two I discuss how the emphasis on pattern discovery differs methodologically from the top-down interpretative approach described by Carroll. Second, the central role of statistical graphics in EDA frees us from the idea that analysis in film studies is a literary activity--"a form of writing which addresses films as potential achievements and wishes to convey their distinctiveness and quality (or lack of it)" (Clayton and Kievan 1)--and allows us to think of film style as dynamic rather than static by representing the formal system of a film in time.
In section three I apply the methods and principles of EDA to understanding the editing structure of Friday the Thirteenth (1980) using statistical graphics. At one level, the choice of this film to illustrate the application of EDA to film style is arbitrary because the method described below is widely applicable and could have been demonstrated using any other film. The choice of a horror film is motivated by the fact that while there are studies of how horror films employ individual stylistic devices to frighten the viewer (such as Baird; Cherry 52-93; Diffrient), to date there have been no detailed studies of editing patterns across entire horror films. My choice of Friday the Thirteenth in particular is to show that in low-budget films such as this one, film style is as deliberate as in the great works of cinema. Tania Modleski writes dismissively about Friday the Thirteenth:
In Friday the 13th [sic] (1980), a group of young people are brought together to staff a summer camp and are randomly murdered whenever they go off to make love. The people in the film are practically interchangeable, since we learn nothing about them as individuals, and there is virtually no building to a climax--only variations on the theme of slashing, creating a pattern that is more or less reversible. (161)
I show that the pattern of the editing in this film is not reversible, that it is logically constructed to frighten the viewer through variations in rhythm and pace, and that it builds to a climax.
EXPLORATORY DATA ANALYSIS AND FILM FORM
Carroll ("Style" 268) argues that the purpose of the formal analysis is "to explain why an individual motion picture is the way it is: why it has the elements of style it does and why they stand in the relations that they do." Thus, the functional approach considers only those instances of film form that will advance our understanding a motion picture as a whole, and where "the formal elements and relations in the motion picture that are pertinent for stylistic analysis are those conducive to the realization of the point(s) and purpose(s) of the motion picture" (275). The form of a motion picture is the "ensemble of choices" intended to achieve those points or purposes. In this respect, the selective nature of the functional mode avoids the generality of descriptive approaches that "duplicates" a film's style, focusing on what is important and thereby illuminates our understanding of film form: functional analysis "does not merely describe the motion picture but explains why it is the way it is, thereby augmenting our comprehension of the work at hand" (276). Carroll does not address how the analyst determines the "point(s) and purpose(s)" of a motion picture in any depth, but recognizes a film is (in some non-defined sense) already interpreted prior to analysis so that a systematic account of the functions of film form is an "augmentation" of the original interpretation and, therefore, proceeds ex hypothesi.
This may be characterized as a top-down approach to analyzing film style in which the semantic field of the critic is mapped onto the film (see Bordwell, Making Meaning 142-145). There is, of course, a place for top-down research on film style and form because what we consider to be relevant to our analysis will often be determined by our knowledge of modes of production, genres, the history of film style, other works by the same filmmakers, and so on. However, Carroll intends the functional approach to be a general method for explaining the form of a film on its own terms that does not depend on such external knowledge. It is a significant drawback of the functional analysis of film form that we are required to make a priori assumptions regarding the point or purpose of a film, as these assumptions will determine those formal features we consider salient in our analyses. This is evident in Carroll's discussion of the functional approach:
A formal or stylistic choice has the intended function to realize the point or purpose of the movie, if the point in question is the intended result of the formal choice and if the formal choice occurs in the work in order to secure the point or purpose of the motion picture.... This, of course, involves having a sense of the point or purpose of the work, a project that often involves interpretation, where the point of the work is to make meaning. ("Style" 277)
In this respect the practice of the functional analysis of film form is no different from other interpretive practices in film studies in that it is "wholly 'finalistic,' based upon an a priori codification of what a film must ultimately mean," substituting the local opinion of the analyst for the Grand Theory of film scholars but being equally blind to potentially interesting formal relations "because the interpretive optic in force has virtually no way to register them" (Bordwell, Making Meaning 260). At worst such approaches produce circular arguments, using the stylistic choices of the filmmaker to justify the interpretation originally stated as a premise for selecting the pertinent relations of film form. A danger of the functional analysis of film style is that we analyze what is important but what is important is what we consider worthy of analysis. The result is a partial account of a film's style that serves the opinion of the critic but which is not an explanation of its form.
As an alternative approach I adopt a bottom-up, data-driven method from statistics. John Tukey proposed exploratory data analysis (EDA) as a "practical philosophy" of data analysis that minimizes the researcher's dependence on prior assumptions and maximizes insight into the phenomenon at hand. Tukey described the role of the researcher as that of a detective:
Exploratory data analysis is detective work--numerical detective work--or counting detective work--or graphical detective work.... [It is] about looking at data to see what it seems to say. It is about simple arithmetic and easy-to-draw pictures. It regards whatever appearances we have recognized as partial descriptions, and tries to look beneath them for new insights. (1)
The goal of EDA is to discover "potentially explicable" patterns in data (Good); with "an emphasis on the substantive understanding of data that address the broad question of 'what is going on here?'" (Behrens 131). This process depends on the researcher adopting a position of skepticism of methods that may obscure informative aspects of data and of openness to unanticipated patterns (Hartwig and Dearing 9). In contrast to Carroll's account of formal analysis in which the analyst specifies the function of a film prior to data collection based on what the critic thinks is important and for whom analysis is primarily confirmatory, an EDA approach aims to arrive at an explanatory model of film form based on our analysis of the data without pre-judging what is or is not important. Adopting EDA as a data driven process to identify "potentially explicable" relations avoids begging the question about the function of film form, and, therefore, places far greater emphasis on the description of film style than Carroll is willing to grant. Certainly, this stage of research should never be characterized as "merely describing" a motion picture. The basis for EDA is abductive and moves from data to hypothesis via a bottom-up process of pattern extraction (Behrens and Yu). From an EDA perspective the descriptive account of a motion picture's form is precisely a guide to what is significant for understanding and appreciating the style of a film, but which avoids description as an end in itself.
Graphical displays are the primary tools of exploratory data analysis (see Ellison). As a film typically comprises several hundred (if not thousands) of shots describing its style clearly and concisely can be challenging; and it is often simpler and more informative to use graphical representations as an efficient method of representing large and complicated datasets. Description is not separate from analysis, and far from merely reproducing a film's style, such displays make it possible to see the formal relationships across a film and to extract interesting and unusual patterns.
Effective visual presentations highlight interesting and unusual aspects of quantitative information under investigation. This encourages the researcher to pursue these features to identify their sources and implications for understanding the processes that are generating the data in the first place. (Jacoby, Statistical Graphics 7)
Graphical displays are not a substitute for the films we wish to study: we use graphs to identify those features that could be interesting and then we go back to the film to see what those features are and to understand how they function in the film's formal scheme. The model of the film produced is an account of these patterns and their causes, but avoids the need for a priori assumptions regarding what is important and what to select because visual representations of data "forces us to notice what we never expected to see" (Tukey vi). For example, we might observe in a graph of the editing structure of a film a handful of shots that are of much longer duration than others and wonder why this should be the case and, though the graph has brought this feature to our attention, it is only when we go back to a film itself that we can begin to understand what discourse elements these features are associated with and how they shape our experience of a film. "Graphics reveal data" (Tufte 13, original emphasis), and provide an unparalleled method for open-mindedly discovering pertinent relations of film form.
David Bordwell describes film style as a system because it "mobilizes components--particular instantiations of film techniques--according to principles of organization" (Narration 50). In light of this definition, I emphasize that style is a dynamic system organized and evolving over the duration of a motion picture and that the dynamics of film style are an essential part of any formal analysis. However, analyzing the evolution of just a single element of style over time is a challenging undertaking given the complexity of the system of style. Andrew Kievan (74) writes that "film--visual, aural and moving--is a particularly slippery art form" (original emphasis) that "sets up peculiar problems for analysis and description because it is tantalizingly present and yet always escaping." Consequently, film studies tends to treat style as a static phenomenon, either abstracting individual scenes from the flow of the larger work or ignoring the temporal dimension altogether. The failure to use graphical methods in the study of film is one explanation for the lack of research on the temporal evolution of style in motion pictures and the use of time-ordered graphical displays allows us to overcome these problems to look at the dynamic structure of style in a simple, quick, and informative manner. This encourages us to interact with the data so that formal relations across a whole film may be perceived whilst also identifying relevant features at smaller scales. They bring to our attention subtle dynamic variations that otherwise may pass unnoticed and afford a degree of precision in discovering when differences in the pace and tempo of a motion picture occur and in characterizing the size and nature of those differences not attainable by other methods.
THE EDITING STRUCTURE OF FRIDAY THE THIRTEENTH
I collected shot length data from Friday the Thirteenth (1980, dir. Sean S. Cunningham, ed. Bill Freda), loading the film into a non-linear editing package and analyzing it frame-by-frame. A shot is defined as a continuous sequence of frames, and an edit as the transition between two shots, either as hard cut or a gradual transition (wipe, fade, dissolve, etc.). When the edit is gradual, the edit between two shots is located at the approximate mid-way point of the transition. As a PAL DVD is used a correction factor of 1.0416 is applied to the duration of each shot. (2) The data includes titles establishing time, date, and location, but does not include the opening and closing credits.
A time series is a sequence of observations ordered in time, and the goal of time series analysis is to reveal the temporal structure of a data set so that we may come to understand the underlying dynamic process. I use the ranks of the shot lengths rather than the raw data, with shots ranked from shortest (r = 1) to longest (r = N) and average ranks assigned to tied values. When analyzing the dynamic editing structure of a film, it is important to remember that time (i.e., the duration of a shot is the time between two cuts) is the variable of interest, and by using a rank-based method the fact that time is not the independent variable is not a concern as I assume only that the data are ordered in time. Fitting a trendline to the ranks of the shot length data using locally estimated (LOESS) regression reduces noise in the ranked data making it easier to identify the underlying trends in the editing of the film (see Jacoby "Loess"). These methods allow us to identify clusters of longer or shorter takes, the presence of intermittent cyclical patterns, and to locate change points in the editing of a film. Once identified, these features can then be analyzed in more detail by going back to the film to see the narrative events, characters, and other elements of film style with which they are associated. I make no assumptions about the functions of editing in this film, though I take it for granted that film form is related to function and that the objective of formal analysis is to explain why a film is the way we experience it to be.
Figure 1 (a) presents the distribution of shot lengths in five-number summary for the shot length data of Friday the Thirteenth; and Figure 1 (b) presents the time series of the ranks of the shot length data. Exploratory data analysis of the time series using the time series of the shot ranks reveals this film to be comprised of six different narrative segments, each characterized by a different editing style. The beginning and endpoints of these segments are indicated by the shading in Figure 1 (b) and are lettered "a" through "f." Furthermore, transitions between these segments are associated with changes in the dominant mood of the film.
The first section (indicated as "a" in Figure 1 (b)) of the film is the originating event of the murder of two counsellors at Camp Crystal lake in 1958 (shots 1-17, [summation] = 294.1s, median = 6.7s, interquartile range (IQR) = 24.0s). This sequence includes several tracking shots from the point-of-view of the unseen killer, with six shots running to more than 19 seconds. The second section ("b") of the film is set in the present, introducing the main characters and establishing the location of the narrative (shots 18-144, [summation] = 965.5s, median = 5.5s, IQR = 7.9s). Within this sequence there are two narrative threads--Annie hitchhiking to the camp (shots 18-74) and the arrival of the new counsellors at the camp (shots 75-122). There is a tendency to higher ranked shots (i.e., longer takes) in the latter as this part of the sequence is organized around a series of group shots compared to the shot-reverse shot pattern of singles that characterize Annie's interactions with the locals as they warn her about the past of the camp. This sequence ends with the murder of Annie by an unknown assailant, including a series of very short takes at shots 139-144 in Figure 1 (b) as the violence of this assault reaches its peak. From shot 145 shot lengths become increasingly longer until shot 262 ([summation] = 1222.1s, median = 5.5s, IQR = 8.8s). This can be seen in Figure 1 (b) in the trend towards higher ranked shots indicated in segment "c." This third segment does not feature any violence and focuses on the creation of sense of foreboding through the near drowning and the snake in the cabin, the warnings of the police officer and "Crazy Ralph," and the appearance of a mysterious figure in an oil slicker. This is achieved formally as the film builds tension by progressively lengthening the duration of takes so that over the course of this sequence the pace of the film slows down.
Changes in editing style between these first three segments of the film occur when there are large differences in mood between consecutive scenes. Thus, the transition between the first and second segments of the film occurs with the shift from the violence of the original murders at the camp to the sunny optimism of the present day as Annie sets off to be a counsellor at the camp. (This transition is also separated by the opening titles, which are not included in the data set). Similarly, the shift between the second section ending with Annie's murder and the next section with the other counsellors swimming at the camp is based on a shift between extreme violence and the counsellors relaxing at the lake (see Fig. 2). These shifts are also marked by the use of white flash transitions that do not appear at any other point in the film, and which are used to cover shifts in time (past/present) and location (Annie/the camp). The use of white flash transitions to cover shifts in location occurs only at the transition between different narrative segments and the onset of a new editing pattern, and they do not feature within narrative sections even though multiple locations are used (e.g., the shifts between the counsellors at the camp and Annie in the second part of the film).
In the later parts of the film, transitions between editing regimes occur within scenes and are not associated with any particular optical effect as narrative space and time are continuous. The fourth sequence ("d") is the "stalk-and-slash" segment of the film (shots 263-399, [summation] = 2095.0s, median = 8.4s, IQR = 15.9s), and includes the murders of Jack, Marcie, Brenda, and Steve. The sequence begins in the middle of a scene with the sudden murder of Jack, and from this point longer takes dominate as the emotional tone built up in the preceding segment is sustained. The use of long takes emphasizes the anticipation of violence and creates for the viewer a continuing sense of dread that round every corner lurks a new terror. This is the slowest segment of the film, but also contains two clusters of very short takes at shots 273-280 and shots 318-323 associated with the murder of Marcie (see Fig. 3) and the terrorizing of Brenda, respectively. These sudden outbursts of shocking violence realize the fears of the viewer and represent a more aggressive form of terror in counterpoint to the dominant mood of the sequence. However, no such clusters are associated with the murders of Jack or Steve which are shot as long takes, and there appears to be a clear difference in the way in which violent scenes are edited according to the gender of the victim.
The shift from the stalk-and-slash section to the final girl sequence also occurs within a single scene when Mrs. Voorhees's demeanor suddenly changes while remembering Jason's death; and again a change in the dominant mood of the film is accompanied by a change in style. The final girl sequence ("e") runs from shot 400 to shot 534 and is edited much more quickly than other parts of the film ([summation] = 647.7s, median = 2.8s, IQR = 2.9s). This sequence is characterized by the heightened emotional intensity of Mrs. Voorhees's psychosis, creating a mood of aggressive tensity that continues through the sustained violence she inflicts on the panicked Alice. This segment breaks Alice's fight for her life into three rapidly edited violent scenes divided by slower cut clusters as the final girl runs and hides from the killer; and this alteration is evident in the peaks and troughs in the ranked data in this section in Figure 1 (b). The final sequence of the film ("f") comprises shots 535 to 559 ([summation] = 269.8s, median = 9.3s, IQR = 12.9s), and includes Alice's dream of encountering Jason as she floats across the lake in a canoe and waking in the hospital. This sequence begins with a shot of the moon over the camp recalling the opening shot of the film, and like the opening sequence is edited slowly. This coda reintroduces the idea that a threat remains out there, and returns--emotionally and stylistically--to the pervading sense of unease and disquiet of the earlier narrative sequences.
Brett Adams, Chita Dorai, and Svetha Venkatesh observed that in Hollywood narrative cinema, large changes of pace occur at the boundaries of story segments, while smaller changes in pace are identified with local narrative events of high dramatic import (Adams, Dorai, and Venkatesh; Dorai and Venkatesh). The example of Friday the Thirteenth shows that while change points in the time series may coincide with the end of one narrative segment and the beginning of the next, this is not always the case; and that changes in the dominant editing patterns are consistently associated with large differences in the mood of the film. This suggests the editing structure of the film is organized at different scales with a particular editing pattern dominating the macro-structure across whole acts or sequences (the building of tensions, the stalk-and-slash sequence), and localized clusters linked to specific narrative events (the attacks on the campers).
The different editing styles are associated with different types of horror and their emotional impact on the viewer. This supports the findings of earlier research on editing and pace in motion pictures and their relationship to affective content. Alan Hanjalic and Xu Li-Qun used variations in the duration of shots as a means of describing the affective content of video, with shorter shots creating stressed, accented moments and longer takes used to de-accentuate an action. In slasher films, the slower edited sequences function to create a sense of suspense for the viewer by contrasting a fearful context with neutral visual scenes. For example, in Friday the Thirteenth we watch as the group at the camp perform a series of mundane tasks (repairing the cabins at the camp site, unpacking boxes, setting up the archery targets) while the killer--unidentified at this time--is out there in the woods. Research on the neuroscience of fear has shown that the apparent neutrality of such scenes is transformed by the viewer's knowledge of a threat even though the visual information is not fearful in its own right (Willems, Clevis, and Hagoort). This effect depends on the disparity of knowledge between the viewer, who knows the killer is "out there," and the characters whoare unaware of the threat that surrounds them. The use of shots of increasing duration at this point in the film de-accentuates the action so that the threat to Alice is created in the minds of the viewer rather than presented to them. For a film that features eight murders (including one that takes place off screen), there is remarkably little violence in Friday the Thirteenth.
J. P. Telotte points out that "simply knowing that horrors do, in fact, exist 'out there' is insufficient, however; the full consequences of this knowledge also have to be thrust home" (125). The slasher film must deliver to the viewer a series of brutal murders as the psychological horror of the suspenseful gives way to the physical violence of body horror, and in these moments adopts a different editing style. Hang-Bong Kang identified rapid editing as characteristic of scenes in which the dominant emotion was fear; and Wang Hee Lin and Cheong Loong-Fah noted that the excitement level of a scene increases as shot lengths decrease, so that action scenes are typically edited more quickly than dialogue scenes. These associations are certainly evident in Friday the Thirteenth, and the most emotionally intense and physically violent moments of these films are associated with rapid editing. Additionally, there is no new narrative information to be processed in these scenes and therefore no need for the viewer to store information for later reference. The use of shorter shot lengths does not compromise the forward flow of the narrative, and allows these films to present extreme events to the viewer without loss of coherence.
Film critics tend to speak of the style of a film in singular terms as though it definitely has one--and only one--mode of expression, but this analysis indicates it would make more sense to talk of the styles of Friday the Thirteenth with different editing patterns used to create different types of horror. However, these different styles should not be regarded as separate but as operating in conjunction. Emotional priming is an extensively studied phenomenon in which a subject's responses to stimuli are modulated by their active motivational state, and empirical research has shown the amplitude of the startle reflex to be augmented when a subject is already in a state of fear and is inhibited if a subject is primed with pleasant stimuli (Lang and Davis; Lang, Davis, and Ohman). The viewer's responses to the abrupt onset of violence in Marcie's murder or the final girl sequence in Friday the Thirteenth are intensified because he/she is already in a state of fear created during the slowly edited stalk-and-slash segments of these films. Arousal is also a factor in modulating the startle response with the greatest reflex associated with the most unpleasant images, including mutilations, animal attacks, and human attacks (Bradley, Codispoti, Cuthbert, and Lang). Such images are the standard fare of the slasher film and also intensify the level of our reaction to the onset of violent sequences. An interesting feature of the emotional modulation of the startle reflex is that it does not seem to depend on novelty, and persists even when a subject is presented with the same picture stimuli (Bradley, Lang, and Cuthbert; Lang, Bradley, and Cuthbert). This may account for the fast-slow-fast editing pattern of the final girl sequence, enabling a film to repeatedly generate a startle response in the viewer as the killer rises to attack again and again without loss of effect. Thus in Friday the Thirteenth, Mrs. Voorhees can attack Alice three times in a short space of time and scare the viewer on each occasion because the intervening slow cut segments prime the viewer by re-creating a sense of suspense prior to the onset of the next attack.
Exploratory data analysis provides a quantitative framework for studying form in a single film that removes the need for prejudging the significance of the elements of film style on the part of the analyst. In this essay, I used the time series of the ranks of the shots as a simple and fast method for exploring the editing structure of motion pictures, and applying this method to Friday the Thirteenth I was able to locate change points in their editing patterns and to identify clusters of shots as potentially explicable relations of film style. Manovich writes of computer-based approaches to visualizing film style that
A computer does not have the same insights into the meaning and structure of a film as its director and editor. However, it can help us notice subtle patterns in editing, composition, movement, and other aspects of cinematography and narrative that maybe hard to see otherwise.
This is the case in the example demonstrated in this article. Though it is immediately apparent to the casual viewer that some moments in the film are edited rapidly more rapidly than others, it is far from obvious to even the careful viewer that the editing pace slows over the course of a 20-minute section of the film as it does in the third part of Friday the Thirteenth. Without the economy and the detail of the time series graph, this aspect of the film's editing is easily overlooked, demonstrating the value of the bottom-up data-driven approach of EDA. It is also clear from visualizing the editing pattern of Friday the Thirteenth that Modleski's dismissive assessment of the film is not justified. Linking features in the time series plot to narrative events and to characters, I found this film to be characterized by multiple regimes of editing associated with different types of horror but which nonetheless operate in unison, that there is a close relationship between the large scale narrative structure and editing structure while localized relations of film style are associated with specific narrative events.
(1) On the comparison of shot length distributions see the articles by Redfern ("Shot length," "Impact," "Comparing"). On editing patterns see the articles by Cutting, DeLong, and Nothelfer, and O'Brien. Cutting discusses transitions; while Baxter and Salt ("Shape") look at shot scales.
(2) PAL DVDs have a frame rate of 25 fps whereas film uses 24 fps, resulting in a 4.16% difference in running time that is corrected by multiplying the duration of each shot by 25/24 = 1.0416.
Adams, Brett, Chitra Dorai, and Svetha Venketesh. "Role of Shot Length in Characterizing Tempo and Dramatic Story Sections in Motion Pictures." Proceedings of the First IEEE Pacific Rim Conference on Multimedia, 13-15 December 2000, Sydney, Australia. Eds. G. Lu and R. L. K. Liu. Sydney: University of Sydney, 2000: 54-57. Print.
Baird, Robert. "The Startle Effect: Implications for Spectator Cognition." Film Quarterly 53.3 (2000): 12-24. Print.
Baxter, Mike. "Picturing the Pictures: Hitchcock, Statistics and Film." Significance 9.5 (2012): 5-9. Print.
Behrens, John T. "Principles and Practices of Exploratory Data Analysis." Psychological Methods 2.2 (1997): 131-160. Print.
Behrens, John T., and Yu Chong-Ho. "Exploratory Data Analysis." Handbook of Psychology: Volume 2--Research Methods in Psychology. Eds. Jofin A. Schinka and Wayne F. Velicer. Hoboken, NJ: John Wiley & Sons, 2003: 33-64. Print.
Bordwell, David. Narration in the Fiction Film. London: Methuen, 1985. Print.
--. Making Meaning: Inference and Rhetoric in the Interpretation of Cinema. Cambridge, MA: Harvard UP, 1989. Print.
--. Figures Traced In Light: On Cinematic Staging. Berkeley, CA: U of California P, 2005. Print.
Bradley, Margaret M., Maurizio Codispoti, Bruce N. Cuthbert, and Peter J. Lang. "Emotion and Motivation I: Defensive and Appetitive Reactions in Picture Processing." Emotion 1.3 (2001): 276-298. Print.
Bradley, Margaret M, Peter J. Lang, and Bradley N. Cuthbert. "Emotion, Novelty, and the Startle Reflex: Habituation in Humans." Behavioural Neuroscience 107.6 (1993): 970-980. Print.
Butler, Jeremy G. Television Style. New York: Routledge, 2010. Print.
Carroll, Noel. "Film Form: An Argument for a Functional Theory of Style in the Individual Film." Style 32.3 (1998): 385-401. Print.
--. "Style." The Routledge Companion to Philosophy and Film. Eds. Paisley Livingstone and Carl Plantinga. London: Routledge, 2009: 268-278. Print.
Cherry, Brigid. Horror. Abingdon: Routledge, 2009. Print.
Clayton, Alex and Andrew Kievan. "Introduction: The Language and Style of Film Criticism." The Language and Style of Film Criticism. Eds. Alex Clayton and Andrew Kievan. Abingdon: Routledge, 2011:1-26. Print.
Cutting, James E., Kaitlin L. Brunick, and Jordan E. DeLong. "The Changing Poetics of the Dissolve in Hollywood Film." Empirical Studies in the Arts 26.2 (2011): 149-169. Print.
Cutting, James E., Jordan E. DeLong, and Christine E. Nothelfer. "Attention and the Evolution of Hollywood Film." Psychological Science 21.3 (2010): 432-439. Print.
Diffrient, David Scott. "A Film is Being Beaten: Notes on the Shock Cut and the Material Violence of Horror." Horror Film: Creating and Marketing Fear. Ed. Steffen Hantke. Jackson, MS: U of Mississippi P, 2004: 52-81. Print.
Dorai, Chitra, and Svetha Venketesh. "Bridging the Semantic Gap in Content Management Systems: Computational Media Aesthetics." Proceedings of the First International Conference on Computational Semiotics in Games and New Media, 10-12 September 2001, Amsterdam, The Netherlands. Eds. A. Clarke, C. Fencott, C. Lindley, G. Mitchell, and F. Nack. Amsterdam: CWI, 2001: 94-99. Print.
Ellison, Aaron M. "Exploratory Data Analysis and Graphic Display." Design and Analysis of Ecological Experiments. Eds. Samuel M. Scheiner and Jessica Gurevitch. New York: Chapman & Hall, 1993:14-45. Print.
Friday the Thirteenth. Dir. Sean S. Cunningham. Georgetown Productions Inc., 1980. DVD.
Good, Irving J. "The Philosophy of Exploratory Data Analysis." Philosophy of Science 50 (1983): 283-295. Print.
Hanjalic, Alan, and Xu Li-Qun. "Affective Video Content and Representation Modelling." IEEE Transactions on Multimedia 7.1 (2005): 143-154. Print.
Hartwig, Frederick, and Brian E. Dearing. Exploratory Data Analysis. Newbury Park, CA: Sage, 1979. Print.
Jacoby, William G. Statistical Graphics for Univariate and Bivariate Data. Thousand Oaks, CA: Sage, 1997. Print.
--. "Loess: A Nonparametric, Graphical Tool for Depicting Relationships Between Variables." Electoral Studies 19.4 (2000): 577-613. Print.
Kang, Hang-Bong. "Affective Contents Retrieval from Video with Relevance Feedback." Digital Libraries: Technology and Management of Indigenous Knowledge for Global Access. Eds. T. M. T. Sembok, H. B. Zaman, H. Chen, S. Urs, and S. H. Myaeng. Berlin: Springer, 2003: 243-252. Print.
Kievan, Andrew. "Description." The Language and Style of Film Criticism. Eds. Alex Clayton and Andrew Kievan. Abingdon: Routledge, 2011: 70-86. Print.
Lang, Peter J., Margaret M. Bradley, and Bruce N. Cuthbert. "Emotion, Motivation, and Anxiety: Brain Mechanisms and Psychophysiology." Biological Psychiatry 44.12 (1998): 1248-1263. Print.
Lang, Peter J., and Michael Davis. "Emotion, Motivation, and the Brain: Reflex Foundations in Animal and Human Research." Understanding Emotions. Eds. S. Anders, G. Ende, M. Junghofer, J. Kissler, and D. Wildgruber. Progress in Brain Research, Vol. 156: Amsterdam: Elsevier, 2006: 3-29. Print.
Lang, Peter J., Michael Davis, and Arne Ohman. "Fear and Anxiety: Animal Models and Human Cognitive Psychophysiology." Journal of Affective Disorders 61.3 (2000): 137-159. Print.
Manovich, Lev. "Visualizing Vertov." Russian Journal of Communication 5.1 (2013): 44-55. Web. 12 June 2016. http://softwarestudies.com/cultural_analytics/ Manovich.Visualizing_Vertov.2013.pdf.
Modleski, Tania. "The Terror of Pleasure: The Contemporary Horror Film and Postmodern Theory." Studies in Entertainment: Critical Approaches to Mass Culture. Ed. Tania Modleski. Bloomington: Indiana UP, 1986. 155-166. Print.
O'Brien, Charles. "Sous les toits de Paris and Transnational Film Style: An Analysis of Film Editing Statistics." Studies in French Cinema 9.2 (2009): 111-125. Print.
Persson, Per. "Towards a Psychological Theory of Close-ups: Experiencing Intimacy and Threat." Kinema 9 (1998): 24-42. Print.
Redfern, Nick. "Shot Length Distributions in the Short Films of Laurel and Hardy, 1927 to 1933." Cine Forum 14 (2012): 35-71. Print.
--. "The Impact of Sound Technology on the Distribution of Shot Lengths in Hollywood Cinema, 1920 to 1933." CINEJ 2.1 (2012): 77-84.
--. "Comparing the Shot Length Distributions of Motion Pictures Using Dominance Statistics." Empirical Studies of the Arts 32.2 (2014): 257-273. Print.
Salt, Barry. Film Style and Technology: History and Analysis, 2nd ed. London: Starword, 1992. Print.
--. "The Shape of 1999." New Reviezo of Film and Television Studies 2.1 (2004): 61-85. Print.
Telotte, J.-P. "Through a Pumpkin's Eye: The Reflexive Nature of Horror." American Horrors: Essays on the Modern American Horror Film. Ed. Gregory A. Waller. Urbana, IL: U of Illinois P, 1987: 114-128. Print.
Tufte, Edward R. The Visual Display of Quantitative Information, 2nd ed. Cheshire, CT: Graphics Press, 2001. Print.
Tukey, John W. Exploratory Data Analysis. Reading, MA: Addison-Wesley, 1977. Print.
Wang, Hee Lin and Cheong Loong-Fah. "Affective Understanding in Film." IEEE Transactions on Circuit Systems for Video Technology 16.6 (2006): 689-704. Print.
Willems, Roel M., Krien Clevis, and Peter Hagoort. "Add a Picture for Suspense: Neural Correlates of the Interaction Between Language and Visual Information in the Perception of Fear." Social Cognition and Affective Neuroscience 6.4 (2011): 404-416. Print.
Caption: Figure 1: Shot length data in Friday the Thirteenth (1980) (a) Adjusted box plot of the shot length distribution. The axis labels are the five-number summary of each distribution, (b) Time series of ranked shot lengths with fitted LOESS trendline. The six sections of the time series discussed in the text are indicated a to f. The y-axis has been inverted for a more intuitive reading of the ranked data in which higher ranks indicate longer takes.
Caption: Figure 2: The transition between (top) the violence at the end of segment "b" and (bottom) the tranquility at the start of segment "c," separated by the use of a white flash (middle).
Caption: Figure 3: Shots 275-280 from the murder of Marcie.
Please note: Illustration(s) are not available due to copyright restrictions.
|Printer friendly Cite/link Email Feedback|
|Date:||Jan 1, 2015|
|Previous Article:||The difference between crows and blackbirds: Alfred Hitchcock and the treason of images.|
|Next Article:||"Like an Animal": figures on the move in The Godfather.|