Danger, Sex, and Everything Else
A Comparison of Camera Angle and Camera Distance Effects Across Pictures of Varied Emotional Content
Abstract
Abstract. This study tests the effects of camera distance and camera angle on emotional response across four categories of pictures covering a large emotional range (positive and negative miscellanea, erotica, and threat), using the International Affective Picture System (IAPS) –a large database of emotionally evocative photographs. We content analyzed 722 images for the content category and camera framing (distance and angle), employing these as independent factors in analyses, and used the IAPS’ pre-existing normative average ratings of emotional valence, arousal, and dominance as dependent variables. As hypothesized, affective responses were generally increased by closer framing and high and low angles (compared to straight angles), but the content of the picture played an important role in determining effect strength and direction. In particular, closeness increased arousal for all picture groups but had the opposite effect on positive miscellaneous pictures, straight angles decreased the emotional response for the two miscellanea groups, and low angles increased the emotional response for threatening pictures. This study is the first to show that previously found camera framing effects apply to pictures of high emotional intensity (e.g., erotica and threat). We suggest that future work should consider formal manipulations alongside message content.
Photographs are ubiquitous in all types of media, ranging from billboards and newspapers to dating apps. Pictures are a powerful means of communication because they convey a massive amount of information that is automatically processed in fractions of a second, thanks to their optical similarity to natural, non-mediated visual perception (Hochberg, 2007). Although people’s psychological responses to pictures mostly depend on the objects and events depicted, there is a layer of meaning that is determined by how the camera frames the content of the picture. Camera framing encompasses the distance, angle, height, level, and focus of the camera (Bordwell et al., 2016, p. 187), all of which are structural aspects present across all visual media, unlike content aspects that differ from message to message (Detenber & Lang, 2010). Camera distance and camera angle have received the most attention in media research, with a growing body of experimental studies backing up theoretical claims from media theorists on distance and angle effects. However, most of the empirical research on camera framing uses few messages or single-message designs, and stimuli that tend to be of low emotional intensity. The use of such homogenous stimuli brings up two concerns: first, it leaves unstudied any potential form and content interactions, and second, it limits the generalizability of previous findings, because they are based on stimuli that do not accurately represent the wide variety of emotional content found in visual communication. As it has been pointed elsewhere (cf. Reeves et al., 2016), the lack of stimuli variability is a pervasive problem in media psychology that limits the usefulness of published findings. Accordingly, this study aims to address this shortcoming by assessing camera framing effects across a large dataset of pictures of varying emotional intensity and types of content.
Literature Review
Camera distance, also called shot type, is typically defined in relation to a human subject, and ranges from extreme close-up to close-up, medium close-up, medium shot, medium-long shot, long shot, and extreme long shot (Bordwell et al., 2016). Experimental research has shown that closer shots facilitate elaboration about the mental life of the subject (Bálint et al., 2020) and polarize affective perception, emphasizing either the negativity or the positivity of the impression (Mutz, 2007; Reeves et al., 1992). However, a closer look at the literature suggests complex effects when message content is considered. For instance, Mutz (2007) found that close-ups of uncivil speakers in a televised debate led to more intense evaluations and emotional responses than medium shots, as expected, but not when speakers exhibited civility. Additionally, Canini et al. (2011) found the largest percentage of close-ups in film scenes containing calm events, while long shots were the most prevalent in arousing scenes, which seemingly contradicts an increase in emotion as a function of closeness.
Unlike distance, camera angle is not subject-dependent but instead uses the ground as the point of reference (Giannetti, 1999, p. 14). Three broad categories describe differences in camera angle: low angle (when the camera is aimed upwards), straight on (when the camera is parallel to the ground), and high angle (when the camera is aimed downward). As conceptualized by film theorists, camera angle has a linear effect according to which straight angles are neutral, low angles emphasize power, and high angles indicate weakness (Giannetti, 1999). Such linearity has been supported by multiple experimental studies with high and low angles exhibiting opposing effects on power (Giessner et al., 2011; Kraft, 1987; Mandell & Shaw, 1973; Sevenants & d’Ydewalle, 2006), as well as other dimensions such as credibility (Avery & Long, 1976; Tiemens, 1970) and favorability (Meyers-Levy & Peracchio, 1992). However, there is also evidence to the contrary, with some studies have found that both high and low angles decrease perceived trustworthiness (Baranowski & Hecht, 2018), and that the pattern is rather disparate and non-linear when considering the degree of angle (Kepplinger & Donsbach, 1987).
Thus, there is cumulative evidence suggesting that camera framing effects may in fact be more complex and possibly more dependent on the content of the message than previously thought. However, the question of how form interacts with content remains largely unexplored because the stimuli employed in previous experiments have been of limited variability. This is particularly true for camera angle studies, where stimuli range from monologues, lectures, or newscasts (Avery & Long, 1976; Baranowski & Hecht, 2018; Mandell & Shaw, 1973; Reeves et al., 1992; Tiemens, 1970), to representations of everyday events (Kraft, 1987; Sevenants & d’Ydewalle, 2006), and images of objects (Meyers-Levy & Peracchio, 1992). Although some studies on camera distance have used more emotionally complex stimuli (Bálint et al., 2020; Canini et al., 2011; Mutz, 2007), they have typically used few or single message designs, with the exception of Canini et al. (2011), who did include messages from a larger emotional spectrum but found a reversed relationship between closeness and emotional arousal. Although previous research has used emotionally neutral stimuli, there is no theoretical reason to doubt that camera distance and camera angle effects should also occur for pictures covering a large emotional and content range, while it is possible that some differences could manifest.
As posited by a dimensional approach to emotion, the two dimensions that explain the most variance in affective responses are arousal (i.e., intensity) and valence (i.e., direction, positive or negative) (Bradley et al., 2001), and dominance is often added as a third dimension that taps into how in control one feels while experiencing an emotion (Osgood et al., 1957). We will look at the effects of camera framing on these three dimensions of emotion using four groups of pictures. The first two groups include a wide variety of pictures ranging from low to high arousal, depicting either positive or negative miscellaneous scenes. The third group contains erotic pictures, and the fourth group contains pictures of threatening content. Indicated by their affective ratings in the dataset, these last two groups score highly in arousal: one extremely positive (erotica), and one extremely negative (threat).
We expect that camera framing can intensify the emotional response of the viewer. For positive pictures, a more intense response would entail high arousal, positivity, and dominance ratings, while for negative pictures it would entail high arousal, high negativity, and low dominance ratings. Specifically, and in line with previous findings, for camera distance, we hypothesize that closer compared to further framing intensifies the emotional response. Thus:
During positive picture viewing, closer framing will increase (a) arousal, (b) positivity, and (c) dominance.
During negative picture viewing, closer framing will increase (a) arousal, but decrease (b) positivity and (c) dominance.
We also expect that camera angle can affect viewers’ emotional responses. Since prior evidence suggests that straight angles are emotionally neutral, we predict that they will diminish emotional response compared to both high and low angles:
During positive picture viewing, straight angles will lower (a) arousal, (b) positivity, and (c) dominance ratings compared to either high or low angles.
During negative picture viewing, straight angles will lower (a) arousal and (b) negativity, and (c) increase dominance ratings compared to either high or low angles.
To address any potential differences between high and low angles, we ask:
Do high and low angles differ in their impact on emotional responses to pictures?
Lastly, to determine whether the type of content influences responses, we will compare the findings of the two novel categories, erotica and threat, to those of the positive and negative picture groups:
Do pictures of threatening and erotic content show (a) camera distance and (b) camera angle effects?
Methods
Stimuli and Dependent Variables
We conducted a secondary analysis of pictures available in the International Affective Picture System (IAPS; Lang et al., 2008) that had not been digitally altered and contained people (N = 722). IAPS is a standardized database of pictures used in studies of emotion and was constructed to include images that vary across the full emotional spectrum. In the IAPS technical manual, each picture has three normalized ratings – arousal (i.e., intensity of emotion), valence (i.e., direction of emotion, positive or negative), and feelings of dominance. These ratings, which are the dependent variables used in our analyses, were obtained by Lang et al. (2008) by averaging the evaluations of groups of 100 participants using the Self-Assessment Mannequin (SAM; Bradley & Lang, 1994), a 9-point pictorial scale ranging from 1 = negative, calm, or not dominant to 9 = positive, arousing, or dominant. A more detailed description of the rating procedure used for IAPS is included in the Electronic Supplementary Material, ESM 1. The pictures analyzed include a wide range of content, including portraits, landscapes (with human beings), depictions of violence, body mutilation, nudes, sexual imagery, and sporting events. Because IAPS pictures have been selected to cover the emotional space, the sample includes a wide range of values for self-reported arousal (range = 2.41–7.35), valence (range = 1.31–8.22), and dominance (range = 2.15–7.71).
Content Analysis Procedure
A codebook was developed to measure the camera distance and camera angle of pictures. Four coders (only one of whom was aware of hypotheses) were trained and intercoder reliability was assessed using Krippendorff’s α. While coding, coders were not aware of the pictures’ emotional ratings. One hundred ten images were used during training (15.4% of the sample), and 70 images were used for reliability assessment (9.8% of the sample). During training, discrepancies were corrected using consensus agreement. All coders used a 20.25 × 12.25″ monitor, positioned the monitor directly at eye level for each coding session, and viewed all images in full-screen mode. For pictures with multiple people (N = 342) coders used the person closest to the camera whose face was visible as the point of reference. If multiple people were equally close, the leftmost person was selected. Each image was categorized for camera distance (α = .81) and camera angle (α = .82) (see ESM 1 for details). Using consensus coding, two coders determined each image’s content category, which included threat, erotica, positive miscellanea, and negative miscellanea (see also ESM 1).
Design and Analysis
This study employed a 4 (category: positive miscellanea, negative miscellanea, erotica, and threat) × 2 (camera framing: distance or angle) × 3 (levels of framing: close/medium/far distance or high/straight/low angle) design. Multivariate analysis of variance (MANOVA) was used to analyze the data for all three dependent variables. The final MANOVA model excluded two higher-order nonsignificant effects: a Category × Angle × Distance 3-way interaction (p = .50), and a Distance × Angle 2-way interaction (p = .72). Significant multivariate effects were investigated using univariate analyses of variance (ANOVAs). Significant univariate findings were further explored using Bonferroni-corrected pairwise comparisons.
Results
The MANOVA revealed significant main effects of distance, angle, and category on valence, arousal, and dominance (see Table 1). There were also two significant interactions: Distance × Category and Angle × Category. To understand these interactions and to test our hypotheses, these significant results were followed-up with separate univariate ANOVAs for each dependent measure (arousal, valence, and dominance). MANOVA statistics are shown in Table 1, and findings for the univariate ANOVAs are shown in Table 2. Table 3 includes the means and standard deviations for camera distance and angle, as well as significant results from the post hoc comparisons.
Camera Distance
We predicted that closer shots would elicit stronger emotional responses than farther shots for positive (H1) and negative (H2) content. The interaction of Distance × Category was significant for arousal and valence, but not for dominance (see Table 2). Post hoc analyses showed that camera distance had the expected effect on arousal for three of the four content categories. Close framing was more arousing than far framing for erotica (p = .001) and negative miscellanea (p = .036). Additionally, close framing was more arousing than a medium for threat (p = .008), and medium more arousing than far for erotica (p = .021). Surprisingly, positive miscellanea showed the opposite effect, with farther framing increasing arousal from close to medium (p = .02), and from medium to far (p = .003). The only significant effect of camera distance on valence was for erotic pictures, with close framing being more positive than both medium (p = .011) and far (p = .001). Considered together, these results suggest that closer framing primarily influences arousal, with modest effects on valence. The exception being positive miscellanea, where far framing increased arousal.
Camera Angle
We predicted that high and low camera angles, compared to straight angles, would elicit stronger emotional responses for positive (H3) and negative (H4) content. There were significant interactions of picture category and angle for the three dependent variables (see Table 2). Post hoc analyses showed that angle only significantly affected arousal in a manner consistent with the hypothesis for positive miscellanea, where straight angles were significantly less arousing than both high (p = .003) and low (p = .012). Also, in line with the hypothesis, the two negative content groups showed significant effects of angle on dominance. High angles were lower in dominance than straight for negative miscellanea (p = .003), and low angles were lower in dominance than both high (p = .005) and straight (p = .006) for threatening pictures. Regarding RQ1, there was only one instance where high and low angles significantly differed from each other: in the threat group, low angles were more arousing (p = .015) and lower in dominance (p = .005) than high angles. Overall, these results support the prediction that straight angles are the least emotionally intense, but which angle intensifies (low, high, or both) what dimension (arousal or dominance) depends on image content.
Comparing Threat and Erotica to Miscellanea
Regarding RQ2, we found that emotional evaluations of the two high emotional intensity categories (i.e., erotica and threat) were influenced by camera framing in a manner consistent with the other categories (i.e., positive and negative miscellanea). This supports the possible generalizability of camera framing effects across the spectrum of emotional intensity, with some content-related caveats.
Discussion
This study aimed to assess the generalizability of camera angle and distance effects by testing them on the normed emotional ratings of a large dataset of pictures with great emotional variability, extending prior work that has largely neglected the potential role of content. Overall, our findings indicate that arousing pictures containing cues for sex or danger have camera framing effects comparable to those of emotionally milder stimuli. Camera distance and angle primarily influenced arousal, while the effects on the other emotional dimensions (i.e., valence, and dominance) varied across groups. Based on our findings, the camera angle might have stronger effects on negative images while camera distance could have equivalent effects between positive and negative images. Although our findings support the notion that camera framing modulates emotion, they also suggest that framing interacts with the content of the image.
Closer framing increased arousal for three of the picture categories (i.e., erotica, negative miscellanea, and threat), extending previous research by confirming effects across valence (positive and negative) and in stimuli of high emotional intensity. However, we found the opposite for positive miscellaneous pictures, possibly due to the range of image contents. For example, Canini et al. (2011) suggest that some positive events (such as sports) require further framing to capture the meaning of the scene. Indeed, we examined a subset of the positive miscellaneous group depicting single-person portraits and found that closeness increased arousal (F(2, 84) = 4.207, p = .018), supporting this interpretation. Additionally, closer framing significantly increased positivity for erotic pictures, but not for the negative groups.
Conversely, camera angle had more significant effects in negative compared to positive images. There was only one effect for the positive miscellanea group (straight angles being less arousing than low or high), and no significant effects for erotica. For the negative groups, straight angles increased dominance compared to high angles for negative miscellanea and compared to low angles for threatening pictures. These findings suggest that both high and low angles are more emotionally intense than straight angles, in contrast with findings from previous research where high and low angles exhibited opposite effects.
An exception is the threat group, where low angles decreased dominance and increased arousal compared to high. This finding would support a naturalistic interpretation of camera framing as a surrogate of our every day, non-mediated perception. In real life, the distance and the angle from a scene determine the actions available to the observer, which in turn are likely to affect their emotional response. For example, having something looming over you makes it more difficult to engage in defensive or aggressive behaviors (e.g., fleeing or punching) compared to being eye-level or above, so low angles should lead to stronger emotions when presented with a threat than in another negative, non-threatening contexts. More research is needed to understand if formal effects are indeed rooted in non-mediated perception, as suggested by a naturalistic approach to media effects (cf. Anderson, 1998; Lang, 2014).
This study comes with some limitations. The affective responses in IAPS were obtained from a relatively homogenous demographic group (i.e., US college students, mostly white), so future studies are needed to confirm these findings across more demographically representative samples. Another limitation comes from the statistical strategy we adopted. Due to the aggregate nature of IAPS’s ratings, we could only do a by-stimulus analysis, potentially raising two issues: statistical dependence and inflated degrees of freedom. Although the data might seem dependent, this issue is mitigated because IAPS ratings were obtained from 20 groups of approximately 100 people instead of a single individual. It is unlikely that these groups would significantly differ from each other in a way that would bias camera framing effects. Additionally, although our by-stimuli analysis yields more degrees of freedom than strict by-participant analysis, it nonetheless has a smaller sample size than a full repeated measures design. Lastly, these are limitations that apply to any study that uses the affective ratings of IAPS as dependent measures. Future research could benefit from using pictorial datasets that permit methods such as repeated measures ANOVA or multilevel modeling, which distinguish between participant and subject variance.
Based on our findings we have some practical recommendations. Researchers using images to study emotional responses to specific stimuli (e.g., phobia research) might want to control camera framing to avoid conflating the effects of form and content. For example, a closely-framed mild stimulus (e.g., a ladybug) might be experienced as more arousing than a far-framed scary stimulus (e.g., a spider), wrongly leading to the conclusion that ladybugs are scarier than spiders. Moreover, since we only found modest effects on valence and no effect on dominance for camera distance, we also recommend that researchers experimenting with camera distance prioritize measures that directly tap intensity of motivational activation (e.g., self-reported arousal or electrodermal activity). Our results also suggest that researchers interested in camera angle effects would benefit from using measures of arousal and of power/control (e.g., self-reported dominance), the latter especially for negative stimuli.
We advocate for future researchers to continue the work of understanding how content and emotion combine to modulate the effects of camera framing and angle on viewers’ effective responses. We know that the use of single-message designs and homogeneous stimuli limits the generalizability of media research (Reeves et al., 2016) and prevents us from discovering potential interactions between structural features and specific types of content. Thus, we hope that future studies of structural features use larger pools of stimuli with different types of content varying in emotional direction and intensity, furthering our understanding of how formal manipulations interact with the content of pictorial messages.
Electronic Supplementary Materials
The electronic supplementary material is available with the online version of the article at https://doi.org/10.1027/1864-1105/a000295
We thank M. M. Bradley, P. J. Lang, and the Center for the Study of Emotion and Attention at the University of Florida for letting us use the IAPS database and normed ratings. We thank Harry Yaojun Yan and Kenneth Rosenberg for their help in the coding process.
Lucía Cores-Sarría is pursuing a joint degree in Media and Cognitive Science at Indiana University with the support of a Fulbright Scholarship. Lucía received her joint BA degree in Journalism and Audiovisual Communication from Carlos III University of Madrid, and her MA degree in Cognitive Semiotics from Aarhus University. Her research looks at how cinematography, such as the distance, the movement, or the angle of the camera affects the emotional and cognitive responses to moving images.
Brent J. Hale (PhD, Indiana University) is an Assistant Professor of Communication Studies at the University of Southern Mississippi. His research focus includes interlocutory and psychological responses to social media content through commenting, including messages related to health and politics.
Annie Lang (PhD, University of Wisconsin) is a Distinguished Professor at Indiana University. Her research focuses on understanding communication and media from a dynamic human-centered complex systems perspective. She is particularly interested in rethinking communication theories as changes in qualitative states occurring over time rather than as reductive, additive, linear, processes. Recent work focuses on the dynamics of persuasion, the role of the body in attitude change, destabilizing real-world attractors through game play, and on how to direct perception bypasses the mind to affect our motivational responses.
References
(1998). The reality of illusion: An ecological approach to cognitive film theory. SIU Press.
(1976, November). The effect of vertical camera angle on the perceived credibility of a televised speaker. Paper presented at annual convention of Western Speech Communication Association, San Francisco, CA.
(2020). Shot scale matters: The effect of close-up frequency on mental state attribution in film viewers. Poetics, 83, Article 101480. https://doi.org/10.1016/j.poetic.2020.101480
(2018). Effect of camera angle on perception of trust and attractiveness. Empirical Studies of the Arts, 36(1), 90–100. https://doi.org/10.1177/0276237417710762
(2016). Film art: An introduction (11th ed.). McGraw-Hill.
(1994). Measuring emotion: The self-assessment manikin and the semantic differential. Journal of Behavior Therapy and Experimental Psychiatry, 25(1), 49–59. https://doi.org/10.1016/0005-7916(94)90063-9
(2001). Emotion and motivation I: Defensive and appetitive reactions in picture processing. Emotion, 1(3), 276–298. https://doi.org/10.1037/1528-3542.1.3.276
(2011, September).
Affective analysis on patterns of shot types in movies . In IEEEEd., 2011 7th International Symposium on Image and Signal Processing and Analysis (ISPA) (pp. 253–258). IEEE.(2010).
The influence of form and presentation attributes of media on emotion . In K. DovelingC. von ScheveE. A. KonijnEds., The Routledge handbook of emotions and mass media (pp. 275–293). Routledge. https://doi.org/10.4324/9780203885390(1999). Understanding movies (9th ed.). Prentice Hall.
(2011). The power of pictures: Vertical picture angles in power pictures. Media Psychology, 14(4), 442–464. https://doi.org/10.1080/15213269.2011.620541
(2007).
The Psychophysics of Pictorial Perception . In M. A. PetersonH. A. SedgwickEds., In the mind’s eye: Julian Hochberg on the perception of pictures, films, and the world (pp. 30–56). Oxford University Press.(1987).
The influence of camera perspectives on the perception of a politician by supporters, opponents, and neutral viewers . In D. L. PaletzEd., Political communication research: Approaches, studies, assessments (pp. 62–72). Ablex Publishing Corporation.(1987). The influence of camera angle on comprehension and retention of pictorial events. Memory & Cognition, 15(4), 291–307. https://doi.org/10.3758/BF03197032
(2014). Dynamic human-centered communication systems theory. The Information Society, 30(1), 60–70. https://doi.org/10.1080/01972243.2013.856364
(2008). International affective picture system (IAPS): Affective ratings of pictures and instruction manual. NIMH, Center for the Study of Emotion & Attention.
(1973). Judging people in the news – unconsciously: Effect of camera angle and bodily activity. Journal of Broadcasting & Electronic Media, 17(3), 353–362.
(1992). Getting an angle in advertising: The effect of camera angle on product evaluations. Journal of Marketing Research, 29(4), 454–461. https://doi.org/10.2307/3172711
(2007). Effects of “in-your-face” television discourse on perceptions of a legitimate opposition. American Political Science Review, 101(4), 621–635. https://doi.org/10.1017/S000305540707044X
(1957). The measurement of meaning (No. 47). University of Illinois Press.
(1992, May). Faces on the screen: Pictures or natural experience. Paper presented to the annual meeting of the International Communication Association, Miami, FL.
(2016). The use of media in media psychology. Media Psychology, 19(1), 49–71. https://doi.org/10.1080/15213269.2015.1030083
(2006). Does it matter to be pictured from below? Psychologica Belgica, 46(3), 199–210. https://doi.org/10.5334/pb-46-3-199
(1970). Some relationships of camera angle to communicator credibility. Journal of Broadcasting & Electronic Media, 14(4), 483–490.
