DAVID CORWIN

watercolors    past paintings   notes on composition   bio/reviews   contact  

 

 Abstract

Traditional painters seek to achieve compositionally balanced paintings, but the effect they seek has no name or metaphoric description. It is characterized by words to the effect of "that's it." This pictorial effect is described in terms of eye movements and the effect of peripheral vision. and is called visual coherence. It is achieved through tonal balancing of the horizontal and vertical axis and is independent of color. The result of this is that any factor affecting the internal tonal balance of a painting will alter or eliminate fine pictorial coherence. The effect is thought to be caused by a relative increased stimulation of the magnocellular visual pathway in the retina and vision centers. This is the neuropsychological explanation of one aspect of the aesthetic effects of a painting.

 

Key words: Pictorial Composition, Aesthetic Neuropsychological Effect, Visual Coherence,  Luminance Balance, Tonal Balance Magnocellular Pathway, Smooth Pursuit, Eugène Delacroix, Roger de Piles

 

Pictorial Composition: The particular effect observed by some critics and painters: Definition and neuropsychological explanation of its effect.

                                Pictorial composition is a problem that all painters with a traditional view of pictures confront regardless of the subject of their paintings and are often bedeviled by. A painter will ask himself or herself whether a painting "works." This is a typical question in traditional art schools, and in fact, with respect to an abstract work, it is virtually the only question. Other questions such as whether the figures and objects are drawn "correctly" always exist for those who are interested in such matters. Similarly opinions concerning conceptual ideas such as the need for a visually flat effect or the need to illustrate the artistic process may be important but are also irrelevant to this discussion.  Strolling through a traditional art school not dominated by prevailing conceptual ideologies will reveal that students are creating work in every possible style except perhaps Byzantine and Early Gothic.  For these students the question  "does this painting work?" means whether it "works" from a purely visual point of view.

          The problem with the state of  a painting said to "work" is that it changes such that students of  painting may  judge a painting to be wonderful in the evening but dead and lifeless in the morning or marvelous in the studio but not when hung at an exhibition. In the course of painting, the work may appear to be wonderful but "incomplete." It is not infrequent that the attempt to complete the painting somehow destroys it or causes it to no longer "work."  Perhaps a single stroke changes the whole feeling of the painting.

          The earliest known discussion of this special feeling occurs in books by Roger de Piles, a French art critic, an artist of the early 18th century. He described "a particular feeling evoked by a painting  in which the first impression from a distance causes a certain magical effect based on the arrangement of colors, lights, and shadows even before being aware of the subject of the painting." He felt that this purely visual aspect of a painting - the initial effect and its continuation - the  arrangement of forms, colors, light and shade on the canvas  was a more important aspect of painting than the traditional concept of the arrangement of the subject. Good composition as traditionally understood leads the eye through the painting to discover the subject of the painting and  was chiefly concerned with the subject and how it was presented. (1) (2)

          Eugene Delacroix  in the 1830's noted that "there is a kind of emotion which is entirely particular to painting: nothing (in a work of literature) gives an idea of it." He attributed this to the arrangement of colors, lights, shadow and said: "you  find yourself at too great a distance from the painting to know what it represents; and you are often caught by that magic accord."  Delacroix was undoubtedly referring to the same phenomena of the overall visual impression as noted by de Piles. (3

          Neither de Piles  nor Delacroix or others had a name for or gave instructions on how to achieve the effects they observed  while the traditional aspects of composition could be described and rules could be formulated to teach the critics, public and painters how to look at paintings. It is not as if this aesthetic experience disappeared; it has always existed for those who are sensitive to it. There is, however, no metaphor for this visual experience or as Delacroix said: "nothing gives an idea of it." This is important to note because with other complex sensory experiences such as taste there is a whole host of metaphors used  to relate the experience to other gustatory and taste sensations.

          Academic pictorial composition in the 20th century with the introduction of abstract paintings became dissociated from pictorial subject and introduced many conceptual issues. Composition remained connected to the traditional concept of looking at the individual parts of a painting and how they relate to each other and in particular to pictorial balance. The concept of pictorial balance will be discussed in some detail as this has been quite important in relation to what was thought to be good composition.

          In 1903 an artist Henry Rankin Poore wrote an influential book in which he insisted that pictorial balance had to be located at the geometric center and that this was the key to the aesthetic effect of a painting. He maintained that this was a center of mass type of  balance in which every object or space in a picture had a weight proportional to how far it was from the pictorial center (4). These ideas were subsequently incorporated into Rudolf Arnheim's extensive analysis of the aesthetic effects of painting.  Arnheim maintained that there was a center of balance of the "field effects" induced by depicted objects or forms in every painting which did not necessarily correspond with the geometric center but that for a well composed picture, this should be along either the horizontal, vertical or diagonal axis. (5)

          The existence of this center for determining an aesthetic quality of a picture , and its detection was a subject of much work. Locher et al 1998 observed students of industrial design engineering as they created images composed of geometric shapes and noted that the students' compositions were balanced around the center (6).  Locher et al  in a similar study with advanced students in an Academy of Fine Arts observed that they did likewise (7).  As Locher, 2003 summarized the problem: "balance is achieved when the elements of a pictorial field are pitted against each other about a balancing point or center so that the entire composition appears stable and 'visually right' (i.e., 'good')" (8)  Finally  Firstov et al 2007 evaluated 1332 paintings created by "renowned Russian artists" using a mathematical computation of centers of gravity  and determined that these paintings had centers of balance or gravity corresponding closely to the geometric center.  (9)

 

          The results of these and many other studies indicate that, in general, people can agree on the center of balance of a painting using an intuitive notion of a center of gravity,  and that these paintings feel "right." The implication has been that paintings, constructed with a center of gravity approximating the geometric center, are aesthetically successful. However, these images and pictures were not related in any way to the "that's it" idea expressed by painters when a painting appears to be successful.

 

Giving a name to this aesthetic phenomenon: Pictorial Coherence

          A picture in a high state of coherency has a glow that results from being able to see the entire painting immediately from a distance and a feeling of being able to gaze at the whole painting while looking at individual parts. The eye is able to roam freely all over it rather than being attached to isolated elements. For this observer the eye can smoothly slide through path and lines in the paintings. On examining a painting  it might be remarked that a space in the painting seems empty. This is often referred to as the problem of the corners. The cubist were said to have tried to avoid this problem by using oval canvases. If one were to make a painting with the problem corner coherent, this problem  disappears. It is not as if the corner is given some special attention, assuming it was left alone while the painting was transformed. It is that coherency has the effect of freeing the eye to move so that it does not get stuck looking at something of little interest such as that troublesome corner. As will be discussed below, if the painting is moved to a place where it loses coherency or becomes less coherent, the eye would get stuck again. Coherency is a state and not a permanent attribute of a picture for it depends not only on the picture but the lighting and spatial conditions in which it is viewed. 

          As with any sensory perception this visual attribute of a picture has to be learned. In the example given above, the viewer would have to learn to notice that the corners seem empty. Without that the change would not be noticeable.  Like any sensory perception there are people who are particularly sensitive to it or have been taught to see it while others cannot perceive it. Not everyone is capable of analyzing a perfume or a wine and for many  a response of oly like or dislike might be elicited even after attempting to experience a quality that they are told is there. All of this does not mean in any way that the traditional cognitive aspects of looking do not exert a powerful force over how a picture is viewed. The two ways of viewing operate simultaneously with coherency facilitating the careful inspection of the picture. 

          A number of coherent pictures along with less coherent or incoherent variations are illustrated for reference. These images should be seen using the original computer files on a monitor that is color calibrated in a place with limited reflections. A printed image is very unlikely to be at all faithful to the original.

 

        figure 3 Miro 113.5,114,125,124.jpg                      figure 1 Miro UR 102  UL  115  LL 89  LR 110.jpg 

                 figure 1 Miro    Coherent                                                        figure 2 Miro  Not Coherent

                                                                                                             

        The Castle Roche-Guyon ri darker 140, 118,142,121.jpg                                  The Castle Roche-Guyon balanced 121,120,133,130.jpg 

                 Figure 3  NOT COHERENT                                                                             Figure 4   Coherent

                                                      Braque The Castle Roche-Guyon       

 

       Rocks balanced 80,80,90,92.jpg                                   Rocks (variation) 104,84,65,78.jpg 

    fig 5 Cézanne  Coherent                                          fig 6   Cézanne  Not coherent

   

Velasquez Juan de Pareja 33.4,35.6, 42.3,37.4.jpg

figure 7  Coherent

Velasquez 'Juan de Pareja'

 

Figure

State of Coherency

Luminance average

Of Quadrants black=1white=255

ULQ, URQ ,LLQ, LRQ

Deviation of

(CoM) from geometric center and Quadrant location

figure 1

coherent

113.5  114  125, 124                              

2%       in LLQ

figure 2

not coherent

102,115, 89,  110

3%     in URQ

figure 3

not coherent

161,151,141,130

3%      in ULQ

figure 4

coherent

141,141,152,150

2%     right half

figure 5

coherent

80,80,90,92.

5%     in URQ

figure 6

not coherent

103, 96 ,64, 78

6%    in URQ

figure 7

coherent

 52, 56,  62, 58

2%    in  LRQ

  Table A1      

           Physical Attributes of a Picture capable of being seen in a State of Fine Coherence

          Given the example of a painting in a highly coherent state, the conditions for creating it can be determined.  Deviations from these conditions, by for example painting over parts of the painting, lead to a greater or lesser loss of coherency.  Empirically coherency is determined by the average tone or luminance of each quadrant of a picture such that the two upper quadrants should  be equal to each other, and the two lower quadrants should be also equal to each other. The balance of the upper half with respect to the lower half depends on how the picture is viewed. For an average size picture on the wall (or any vertical position) at approximately eye height the upper half must have an average luminance somewhat less than the lower half (be darker). For a picture lying flat before the viewer, the luminance of the upper half should be equal to or somewhat  lighter than the lower half.   Very large pictures and even average size pictures have different requirements depending on how high they are hung as discussed below. Table A one gives the tonal and center of gravity attributes of figures 1 to 6.

          Initially on studying this problem of pictorial composition or the effect called coherence, the author was of the impression (as considered in the different studies by Locher and others) that a center of gravity type of balance was necessary to obtain this state. Prior to using MatLab to performs these calculations, Photoshop's luminance averaging feature was used as an approximation to the center of gravity balance to help modify computer images until they reached a visual state of coherency. Subsequently it was realized that the luminance averaging value of each quadrant was a good fit for a coherent state while the center of gravity did not distinguish between a coherent and an incoherent image. 

          The center of gravity model of pictorial balance in which the center is made to correspond closely to the geometric center or a major axis is correct in so far as the majority of paintings whether good or bad are centrally balanced. It does not explain or create the conditions necessary to arrive at a state of coherence (the "that's it") in an image. The observations of art students creating images balanced around the center and all of Firstov's scanned images show that this is the way all people including artists and initially this writer approach or conceptualize balanced images.

     All of the problems that painters observe with pictures changing from "working" to "not working" relate to this balance of luminance. For example, this writer has noted with chagrin that a perfectly balanced painting lost its glow when a curator placed it very high on a wall. The apparent constancy of size, shape and illumination is maintained with height,  but in reality these attributes are changed due to the change in perspective.   As a painting is raised higher on the wall, the image of the top half being farther away becomes smaller on the retina. A smaller image means less light is reflected or in other words it becomes darker.  As a result the balancing point becomes lower assuming the illumination and reflectance is unchanged. The change in luminance will not be consciously noted because the brain infers that the picture as an object will have the same luminance as it moves in a more or less constant light environment (luminance constancy). This change will only be noticed for a painting moving  from a state of fine coherency to a lesser or worse state. As discussed above concerning the example of the painting with the problem corner, changing the position and therefore changing the luminance will not change the problem corner. There are many different possible types and an infinite number of degrees of deviation from perfection. The subject of how much a painting can deviate from the optimum state while retaining visual characteristics of coherency is quite complicated and will not be discussed in this paper. .

          Viewing a painting from an angle also causes problems for the same perspective reason: the farther side becomes darker. This change also occurs when a painting is flat before the viewer. Watercolors are often created on a table, and photographs are frequently examined that way. A previously coherent image when exhibited on a vertical wall will appear differently.

            A change in lighting or viewing will have a powerful effect on how the picture is perceived as any change in the reflectance of the  surface of the painting or photograph will change the tonal balance. This can also occur on a computer monitor that is mirroring a light source behind it. LED panels also have marked changes of color and contrast with viewing angle. The other effect of illumination is due to the temperature or spectrum of the light. Since the coherent state is created through tonal balance, any color can be interchanged with any other as long as the balance is maintained. However, when a painting, created under a warm incandescent light, is illuminated with a cooler white light, color in the painting will have a profound effect. Incandescent light which is relatively deficient in blue wavelengths will make colors that contain blue darker because there is less blue light to be reflected. When seen under bluer light,  they will appear lighter.


          Other practical considerations are changes when the painting is varnished or becomes dirty or the colors and varnish are altered through time. Modifying the dimensions through additions or cropping create a new painting. With large paintings it is often very difficult to obtain even illumination. Every reproduction is a new image with its own states of coherence. Frames are symmetric and therefore a small edge will not change the top bottom balance significantly. A large frame will change this ratio, but a large frame tends to be seen as a separate physical object and not part of the painting. Not infrequently in museums one encounters the unfortunate situation where the lighting causes the frame to cast a shadow on the upper part of the picture which distorts the luminance balance.

                   Pictures with a lot of little forms of more or less equal value where the forms cannot be visually combined to create larger forms are incoherent. This might be seen as  somewhat similar to the findings of Locher, 1993 that mirror symmetry is most salient in compositions where small objects were formed into larger groups. An example of this phenomena is Kandinsky's "Sky Blue" 1940 figure 8 (10)

 

Kandinsky Sky Blue 1940.jpg

figure 8

158,159,161,160

The average luminance is equal for the top and bottom half. Darkening the upper part somewhat (figure 9) does not change the state of incoherency even though it meets the criteria of bilateral symmetry with the top half somewhat darker. The eye just moves from form to form without being free to see the entire painting.

 

Kandinsky Sky Blue darker upper half.jpg

                                                                 figure 9  

151,153,163,163

When the picture is organized so that the different forms are not so evenly distributed, it becomes more coherent and the eye can slide through the painting. (figure 10)

figure 10 Kandinsky Blue Skies modified 153, 153,161,160.jpg

figure 10

153, 153,161,160

 

Proposed Neuropsychological explanation of findings: Coherent pictures stimulate in a special selective manner the magnocellular visual pathway

     A visually coherent image has several properties. It can be seen at a glance from a distance. Color has no effect on determining coherence other than its effect on luminance. Fine detail and forms are disregarded. The gaze seems to take in the whole picture, and the eye can move around easily. Binocular vision is required for the property of coherence to exist. The viewer cannot experience it when using only one eye.

           Almost all visual information is sent to various vision centers through two different streams: the parvocellular pathway and the magnocellar pathway. The type of visual information processing described above with a coherent picture corresponds to the rapid responses of the magnocellular stream. This pathway is very sensitive to luminance change and not to color. It has low spatial frequency so that forms cannot be delineated precisely, and it is highly sensitivity to motion. All retinal cells send information through both pathways. However, outside the fovea, the density of cone photoreceptors is less but they are more interconnected. They convey local or relative contrast information rather than precise light level information.  Peripheral vision is largely processed through the magnocellular pathway. The information in the parvocellular stream is required for high level object resolution, precise location, texture and color. (11)

 Normally with pictures in a state of more or less incoherence; the eye focuses on the different images in the picture.  When a picture achieves a state of coherence, it would seem that the magnocellular pathway is stimulated in a particular way such that some people who are particularly sensitive to this effect are able to  overriding the need for  precise fixation of the images within the painting. They are able to slide the eye around the painting in a smooth continuous movement and scan rapidly the entire image.

          While the preceding observations postulate a general activation of the magnocellular pathway, there may be an additional more specific visual effect of   a coherent picture. A bilateral tonally balanced circumscribed form with the upper half somewhat darker may be  interpreted as a schematic face. It would be highly desirable to be able to scan a scene and have the attention directed to focus on an embedded face. The gaze using predominantly peripheral vision would be most important for this. Although peripheral vision is fast but blurred it has been shown to have a higher than expected ability for scene gist recognition. (11A) It has been shown that the fast magnocellular stream directly connects to facial recognition centers including centers evaluating the emotional content of facial expressions. The high level evaluation is returned downstream to direct further investigation by lower level vision centers.  What this means is that if a painting were seen on first approximation to be a face, it would induce the feeling that we want to look at that particular picture rather than just move. (12) (13)

          Although slow pursuit eye movements are considered to be impossible without the appearance of a moving object, it is the experience of this writer that a coherent painting permits this type of movement. Eye tracking studies are necessary to determine the nature of eye movements on looking at a picture in a state of coherency as well as fMRI studies to determine what visual centers are stimulated. The activation of the magnocellular pathway is a neuropsychological effect explaining a particular aspect  of  the aesthetic appreciation of pictures.

SUMMARY

Traditional painters have sought to achieve an elusive compositional effect which has had no metaphoric description.  The effect which is very rapidly felt in a pleasurable way, involves seeing the entire picture at a glance and induces the  viewer to gaze at the whole image while the eye moves freely over the entire painting. A painting having this effect is said to be in a state of visual coherence. This effect is achieved through tonal balancing of the upper right and left quadrants and the lower right and left quadrants and with the upper half somewhat tonally darker than the lower half when seen at eye level. This state is totally dependent  on the balance of luminance, and any change either in the position of the painting with respect to the viewer or the quality of illumination will cause a diminishment or elimination of these visual properties. Binocular vision is necessary to experience it. The effect is consistent with a particular stimulation of the magnocellular visual pathway as well as the interpretation of the image as a possible face.


 

REFERENCES

1 Puttfarken 1985 p. 40

2 ibid p. 134

3 ibid p. 121

4  Poore 1903 chapter 3, p.12

5  Arnheim 1974 p. 20

6  Locher et al 1998

7  Locher et al 2001

8  Locher, 2003

9  Firstov et al 2007

10 Locher, Wagemans 1993

11 Purves 2001 p.275

11A  Larson A M, Loschky LC

12 Kveraga 2007 p.13232

13 Tobimatsu 2012

 

BIBLIOGRAPHY

Arnheim  Art and Visual Perception  Berkeley, University of California Press 1974

Firstov V, Firstov V, Voloshinov A, Locher P   2007 "The Colorimetric Barycenter of Paintings" Empirical Studies of the Arts 25(2): 209-217  

Kveraga K, Boshyan J, Bar M 2007 Magnocellular Projections as the Trigger of Top-Down Facilitation in Recognition   The Journal of Neuroscience,  27(48):13232–13240

Larson A M, Loschky LC The contributions of central versus peripheral vision to scene gist recognition

Locher, Wagemans 1993 Effects of element type and spatial grouping on symmetry detection Perception 22:565-587

Locher PJ, Stappers PJ, Overbeeke K (1998) The role of balance as an organizing design principle underlying adults' compositional strategies for creating visual displays  Acta Psychologica 99(2):141–161  

Locher, P., Cornelis, E., Wagemans, J.,&Stappers, P. 2001 Artists’ use of ompositional

balance for creating visual displays. Empirical Studies of the Arts, 19: 213-227.

 

Poore Henry Rankin  1903  Pictorial Composition and the Critical Judgment of Pictures New York

 

Purves D, Augustine GJ, Fitzpatrick D, et al., editors. Neuroscience Third Edition, Sunderland, MA, Sinauer Associates 2004

 

Puttfarken, Thomas  Roger de Piles' Theory of Art, New Haven, Yale University Press 1985

Tobimatsu S. Neural mechanisms of face recognition: an event-related potential study  Brain Nerve.  64(7):717-26.) 2012