The Realism Continuum, Representation and Perception
Stuart Medley, Edith Cowan University, WA, Australia
Hanadi Haddad, Edith Cowan University, WA, Australia
Abstract: “The realism continuum is a visual model that presents any image as a series of pictures, iteratively reduced in representation from its referrent. A continuum has been used before to gauge the effectiveness of educational instruction (Wileman 1993; Dwyer 1972; Knowlton 1966; Gropper 1963) or to explain the communicative potential of different comics stylings (McCloud 1993). Reference is made in new design theory to less detailed images being easier to scan for pertinent information and generally reducing demand on working memory (Malamed, 2009). None of these theorists explains how it is we can see the less-real-than-real in the first instance, even though we’ve evolved looking only at the real. This paper presentation shows why, psychologically we can see and understand distilled and abstracted pictures and also why our visual systems (the eyes and brain) actually prefer these to photorealistic pictures. The presentation focuses on two majortasks of the visualsystem and how these tasks are facilitated by pictures chosen from deliberate points along the realism continuum. Images of greater realism help to solve the homogeneity problem: distinguishing objects in the same class. That is, telling the difference between Tom, Dick & Harry. Images of reduced or distilled detail facilitate object hypotheses: distinguishing between classes of objects. That is, telling the difference between a person and any other kind of object or thing.”
“Reference is made in new design theory to less detailed images being easier to scan for pertinent information and generally reducing demand on working memory (Malamed, 2009).” with the question of “just how do we see and understand the image that is less-real-than-real in the first place?”
“We propose that images of greater realism help to solve the homogeneity problem: distinguishing objects in the same class; that is, telling the difference between Tom, Dick & Harry. Images of reduced or distilled detail facilitate accurate object hypotheses: distinguishing between classes of objects, for example, telling the difference between a human and any other kind of object or thing.”
The paper proceeds to describe Wilemans work in Visual Communicating (1993) in which he attempts “to cover the whole gamut of image types in terms of their level of realism .. His linear scale runs from ‘concrete’ at the realistic end to ‘abstract’ at the distilled end.”
“There are three major ways to represent objects—as pictorial symbols, graphic symbols, or verbal symbols”
“his work echoed findings in Goldsmith (1984) and Gombrich (2002) that the most realistic image is not the most communicative.”
“McCloud’s model, in Understanding Comics, does not echo Wileman’s approach … McCloud has ‘abstract’ on a separate axis to ‘concrete.’ Instead of “paring away detail” then leaping to a “purely symbolic image” as in Dyers linear model McCloud’s model has a concrete to iconic on an x axis and abstract to real on y axis. Separating “Non-iconic abstraction from iconic Abstraction and Text.”
“iconic abstraction is only one form of abstraction available […] usually the word ‘abstraction’ refers to the non-iconic variety, where no attempt is made to cling to resemblance or meaning […] this is the realm of the art object.” (pp.50-51)
“Lilita Rodman’s (1985) concept that abstraction moves images from the particular to the generic; from a focus on surface to a focus on structure; and from mimetic to symbolic.”
“however we interpret the facts, it remains true that all representations can be somehow arranged along a scale which extends from the schematic to the impressionist” (2002, p.247). Gombrich
“how realistically or otherwise they are depicted, affects how images are received, and therefore the meaning gained from them. ”
“suggests that detail found in realistic images that is not pertinent to the context or the message being delivered may be regarded as noise in a communication (Haddad, 1995). Where source, message and receiver are constant, a channel with noise (distraction) is less effective than one without noise (Bello, 1953).
The authors move on to there next point, “how can we perceive images that are not realistic?” noting that the above studies do not address this question.
“The most realistic image has been demonstrated not to be the most communicative (Malamed, 2009; Gombrich, 2002; Gregory, 1970). Line drawings perform better in this regard than photographs of the same things (Fussel & Haaland, 1978). This may seem surprising. If the human visual system has evolved among the real visual world, it should stand to reason that any means that can replicate that world accurately is the best means to communicate visual information to the reading or viewing audience. The photograph springs most readily to mind: as Susan Sontag explains in On Photography (1977), a photograph is ‘directly stencilled off the real’ (p.154). Yet we can see and understand images that have been abstracted or stylized through drawings of various kinds. This raises the issue of learned versus innate visual understandings; perhaps we must learn to see and understand the non-realistic image.”
The authors then mention studies of infants that suggest that our understanding of abstracted/reduced images may be preexisting. “presented with two dots and a line in a facial arrangement tend to spend more time viewing such an image than they would a ‘non-face’ configuration of the same graphics. This suggests that such an image is understood as representing a face (Morton & Johnson, 1991; Fantz, 1961).”
There are mechanisms in the brain that allow us to identify “less real than real images” found in nature ie. how we can figure out silhouettes and things een in the distance, also see in grayscale (low light situations like night).
“A group of faculties of the visual system, labelled by psychologists, ‘perceptual constancies’, explain that the brain knows what the eye does not. These mental workings override purely visual sensations to prevent the individual from mistaking unique sensations on the retina as unique objects. The visual system is not merely accepting of what presents on the retina, but in fact is measuring that presentation against what the brain knows of the world.”
‘Shape, size and colour constancies are examples of these mental mechanisms (Walsh & Kulikowski, 1998, p.492). Size constancy explains that an object is perceived as having the same size regardless of its distance from us. Knowledge of its size will override its presentation on the retina. Shape constancy explains that an object is seen to have the same shape regardless of orientation. Thus we see things ‘as they really are’ and are not taken in by variations in shape presented to the retina. Colour constancy explains that an object is perceived as having the same colour regardless of changes in light. That is, the brain assumes that an object is less likely to change its own colour than it is to take on different colours as a function of changes to ambient lighting. This connection between the two visual versions of the same thing is what allows us to see the less realistic as having a relationship to the more realistic.
Or rather, the less detailed can stand for the more detailed but in a more general way: the detailed version may be an individual we recognize; the less detailed we may simply regard as ‘a person’. The same would apply for the ideally lit figure and the silhouetted figure respectively. These mental faculties tell us that the real visual presentation of an object upon our retinas must be matched against existing mental information. Implicit in this is that the knowledge already gained of the world exists in some kind of visual form. This form does not precisely match any ‘real’ visual version of such an object but must contain a range of information from different viewpoints and under different lighting conditions.”
Visual constancies could also help explain why the various pictorial representation of objects from different cultures are still recognizable to people not of that culture (provided it is an object that both share).
“Fussel and Haaland (1978) describe how they put visual tests (containing images of “common objects” such as a tree, people, a chicken, etc.) before some 400 Nepalese adults who were unused to pictures. The study was done in order to prepare materials for instruc- tional booklets for illiterate villagers. The study group was shown 10 different things presented in six different styles. These styles, from realistic to distilled, comprised black and white photographs; black and white photographs with background removed from around the subject (‘blockout’); a line drawing with shading and internal detail (a ‘three-tone’ image); the same drawing without shading and with minimal interior detail; a silhouette: and a line drawing. Cumulative correct responses to all 10 of the pictured subjects were as follows: Three-tone, 72%; Blockout, 67%; Line drawing, 62%; Silhouette, 61%; Photograph, 59%; stylised drawing, 49%. The authors conclude that:
“the lessons to be learned from this part of the study would seem to be that the more detailed and realistic a picture is, the more effective it is. The so called ‘simple’ stylised drawings are evidently not simple in anything but appearance, making greater demands on the person trying to interpret them. (p.27)”
So perhaps in the western world, saturated as it is with images, we have pushed and developed out facility for interpreting images because they are everywhere and we need to be able to interpret them quickly and accurately. The other side of this ‘way of seeing’ is that we are now faster at recognizing reduced images (ie. line drawings) than highly realistic one (ie. photographs). The studies above have suggested that this is because the more ‘real’ images contain more information or ‘noise’ that slows us down, not to say that such information isn’t relevant (depending on motivations) we just dont require it to purely recognize an object in a general sense.
However the authors go on to say that the results for the highly real images were only marginally better than the stylized images and that the best results lay in between the two extremes.
“Perhaps the progression along the continuum is problematic because the visual system has more than one task to perform. Psychologists talk of ‘Object hypotheses’ and the ‘homo- geneity problem’ (Rhodes, 1996). What these mean, respectively, are ‘what kind of object am I looking at?’ and ‘which one of those particular objects am I looking at?’. The first is a more coarse problem of differentiating between classes of objects; is that a car or a house? The second is a more fine-tuned question intended to differentiate between objects within the same class; what model of car am I looking at?”
The authors “propose that the coarse problem is more effectively dealt with by communicating with less realistic pictures. And the fine problem more effectively dealt with using pictures higher in detail, more closely matching their real-world referrent.” Also suggest that Fussel and Haaland’s findings happened because the middle classification lay in the “sweet spot for the human visual system as a drawing that can solve both tasks.”
“Training in aspects of the image can change the way one perceives pictures (Noide, et al, 1993, p.219).”
The International Journal of the Image
Volume 1, Number 2, 2011, http://ontheimage.com/journal/, ISSN 2152-7857
© Common Ground, Stuart Medley, Hanadi Haddad, All Rights Reserved, Permissions: firstname.lastname@example.org