…it is surprising that phenomena disclosed to a scientist’s
eye are often reported as discoveries. In many cases, those practitioners of
vision, artists, have discovered and manipulated the same phenomena often
centuries before they came under scientific scrutiny, and often with greater
subtlety.
Page 43
Pictures could be paired in a stereoscope, to appear in
depth, or presented in rapid succession, to appear in motion, thereby replacing
the two dimensions missing in static and single pictures – depth and motion. Theories
of spatial vision developed thereafter emphasized its constructive nature, and
placed greater emphasis on the manipulations that could be made of
two-dimensional stimuli. One consequence of this shift from solidity to surface
was the neglect of centuries of surface (pictorial) manipulations made by
artists. Thus, a large body of spatial phenomena was rediscovered in the
confines of the visual laboratory. The perception and representation of motion,
on the other hand, had a different history. Motion had been frozen in pictures
and few novel techniques of alluding to it were developed.
Refs for history of visual experience being captured in
certain ways by artists. Also, Ref for conclusion chapter (How better to depict
close-up double vision with animation techniques).
Page 44
Representation
Pictures are such an integral part of our lives that we tend
to overlook their remarkable peculiarities. We readily accept their relation to
the objects that they represent even though they have few of the properties of
those objects. Indeed, we rarely doubt the equivalence between the pictorial
image and the object represented, and when this equation is questioned
pictorially we are puzzled.
Page 47
Magritte was working within the style of representing
objects in space )linear perspective) that has been adopted since the fifteenth
century. Almost all Western artistic styles prior to the twentieth century
were concerned with representing objects. As such they required a set of rules
for generating the pictorial image – a set of relations that linked the object
with the picture – which could be followed by the artist and understood by the
spectator. Indeed, this was the very function of artistic representation. The
fundamental problem that all pictorial artists must address is the difference
between objects and pictures of them, namely depth. Over the last 40,000
years different solutions to this problem have been adopted, but all contain
some common features that can be linked to optics. Science and art meet in
perspective. Perhaps it should more accurately be said that the optics of
antiquity met the art of the Renaissance in the context of linear
perspective. The technique of perspective painting was rapidly adopted by artists from
the fifteenth century onwards, and many textbooks described its rules. Linear
perspective is now commonplace because photographic images follow the same
optical principals on which perspective is based. However, there must have been
quite different styles of representing objects in space, and the objects
represented can be readily recognized by us. The link between pictorial
art and representation has been examined from the viewpoint of psychology by
many writers.
Page 48
Styles of representing space
The art historian Ernst Gombrich (1909 – 2001) referred to
the different modes of representing space as “the riddle of style”. He posed
the question: “Why is it that different ages and different nations have
represented the visible world in such different ways?” This was so despite the
acceptance that there can have been little genetic change in humankind over the
last few millennia: the cave dwellers would have seen their world in much the
same way as we do ours, although the contents of those worlds differed greatly.
One of the aspects that has not remained constant over
time is how the objects are represented pictorially. Radical changes of style
took place in the way objects were depicted over the last 40,000 years of
recorded art.
Page 49
…Perspective (or more correctly linear perspective) had a
particular significance in the Renaissance because it reflected a return to the
Greek pictorial ideal of attempting to imitate nature. The rules of
perspective were devised in the intellectual cauldron of early
fifteenth-century Florence. Linear perspective was demonstrated by architect
Filippo Bruneleschi and formalised by a contemporary mathematician, Lean
Battista Alberti. Basically it was the application of
Euclid’s visual cone to a glass pane
intersecting it, and this device is now known as Alberti’s window.
Thus, the principals of reducing a three-dimensional scene to a two-dimensional
picture were formulated before the image-forming properties of the eye had been
described. Nonetheless, the difference between looking
at a picture and looking at a scene were clearly appreciated by Leonardo da
Vinci. At the end of the fifteenth century he wrote “A painting, though
conducted with the greatest Art and finished to the last Perfection, both with
regard to its Contours, its Lights, its Shadows and its Colours, can never show
a Relievoequal to that of the Natural Objects”. That’s is, the
allusion to relief or depth in a painting of a scene, no matter how well it is
painted, will be different to the perceived depth separating the actual objects
in the scene.
Page 50
Image Formation
The influence of art was to prove significant because it
frames a recurrent concern for theorists of vision: how can we perceive the
world as three-dimensional when the image cast in the eye is two-dimensional?
Isn’t this precisely the problem that confronts the painter? Artists in the new
perspective style seemed to answer this in practice if not in theory. An
allusion to three-dimensionality could be induced if the rules of linear
perspective were followed; this required a single, fixed viewing point and the
depiction of objects in the scene in accordance with the angle they subtend at
that viewing point.
Alberti described the rules for capturing the image, and a
number of techniques for representing it.
Page 51
Perhaps the simplest means of drawing in perspective was
‘Alberti’s window’. If a single eye is maintained at a fixed position with
respect to a window then the objects in the scene beyond the window can be
represented in accurate perspective simply be tracing their outlined on the
surface of the window.
Page 51-52
Camera obscura
Leonardo da Vinci conducted experiments with a camera
obscura, and drew an analogy between its operation and that of the eye.
Page 52
The camera obscura was used by both artists and
scientists to generate images. The artists realized that the tedium of perspective
projection could be circumvented by its application…
Page 53
Ocular anatomy
The field of view extends through about 208 deg horizontally
for a forward-pointing eye, although light from this region is imaged on the
extreme periphery of the retina, and is blocked by the nose and head on one
side. Vertically, the field of view is about 120 deg.
In its most elongated state, the lens in a normal eye will
bring rays from an object at optical infinity (defined as a distance greater
than 6 m or about 20 feet) to a focus on the retina. This is referred to as the
far point of accommodation. At its most nearly spherical, it will focus rays
from an object at a distance of about 15 cm from the eye (the near point).
Combine with ref to the distance
from the face of Panum’s fusional area?
Page 116
…the situation is a little more complex for vision because
we have two retinas: it is as though we had two similar chessboards,
side-by-side; which one should we use? That is, a single object in space can
stimulate points on each eye, so that there are two origins and two sets of
coordinate values. Despite this, our perceptual experience is not in duplicate:
we see objects singly with two eyes and in a single direction. Moreover, this
direction does not correspond to that from one or other eye but from a point
between them, often termed the egocentric or cyclopean eye.
Page 174
Spatial Representation
Pictures
In the case of humans, significant activities are not
necessarily restricted to the primary demands of survival and reproduction. We
may seek a variety of sources of visual stimulation, many of which are nowadays
cultural prospects. Some of these, like painted pictures, may be almost as
old as the human species itself. Others, such as photographic or digitised
images, are more recent. Writing and printed text are visual stimuli of
particular complexity, because the shapes bear little visual relationship to
the objects or meanings they convey. Thus, human survival seems to depend on
intellectual as well as nutritional sustenance, and in the case of the former
pictures appear to satisfy a universal appetite. The use of created patterns of
visual stimulation as representations of other things is one of the most
characteristic features of modern culture. Of course, pictures may
sometimes be abstract, and appreciated for the shapes they contain, but it is
very hard not to attribute meaning into anything we look at. Pictures can be
both objects in themselves (a surface with marks on it like a painting or
computer screen) and representations of other objects. One consequence of this
is that pictures are likely to be more complex stimuli to process than solid
objects.
Page 267
…we now derive so much of our information about the world
through pictures.
Pictures and Objects
…if we are dealing with pictures of objects we need to
consider the viewpoint from which the pictures are taken, and the orientation
in which the picture itself is presented.
Page 268
Pictures have an advantage over objects in so far as we can
manipulate their characteristics much more readily than we can those of
objects. Therefore experimental studies are much easier to conduct with
pictures, although there are many reasons for concluding that perception of a
picture of an object is not the same as perception of the object itself.
Ref to an earlier poit – da
Vinci’s quote about how pictures cannot replicate visual experience completely.
Page 269
Photographs and drawings
Cameras are designed to produce pictures that are in central perspective; most of the pictures that we see in newspapers and on television have been derived from cameras and so are in perspective. The situation is somewhat different when we consider drawing and other graphic arts. Perspective arrived rather late on the scene in Western art, and cannot be considered in any way as universal. Paintings were made in caves from about 40,000 years ago, and we can recognise the objects (usually animals) that are depicted. They are nearly all represented in the stereotypical view – outline profiles of bison, deer and horses. Thus, from the earliest examples of art outlines have been used to describe and delineate representations of objects. When we draw objects we initially define their boundaries with lines, and often do not proceed any further. Outlines or contours are very informative, and they can be sufficient to establish the relationship between a drawing and the object it is intended to represent.
Page 273
Contours in drawings are abrupt changes in the amount of
light reflected from a surface; they correspond to the boundaries of an object,
but there are rarely such well-defined transitions between light and dark in
the original scene. One of the problems that has beset workers in the area of
computational vision is to define the boundaries of an object in a
well-structured scene. Therefore, an outline drawing of an object is an
abstraction that does not correspond readily to any features of the pattern of
light projected from a scene. We are exposed to such simplified representations
from a very early age, and so we learn how to interpret them. Experiments have
established that we extract information not only from contours but also from
the regions of a drawing where contours change direction or intersect.
Links to Gibson’s theories on
the information taken from line drawings lies in the areas where lined connect
and intersect.
It is clear that humans have
produced pictures for much longer than recorded history, and a variety of
styles can be discerned. Linear perspective is amongst the most contrived
because it requires picture-makers to forgo their perception of space and to
record visual angles from a single station point. Unlike the art
forms that preceded it, linear perspective set out specific rules for
representing the sizes of objects on the picture plane and, more importantly,
the dimensions of texture on the receding ground plane. These were determined
by the distance of the station point from the picture plane and the height of
the station point from the ground. Thus, all the ‘painter’s cues’ to
distance can be enlisted to increase the allusion to distance on the flat
surface of the picture. Artists have to learn to
draw in linear perspective and similarly observers have to learn to interpret
pictures so produced. (Ref in Lit Rev). the
drawing can be assisted by all manner of aides, the simplest of which were
suggested by Alberi and Leonardo in the fifteenth century. If the picture plane
is replaced by a pane of glass then, with a single eye at a fixed position (the
station point), the contours in the scene can be traced on the glass surface:
the drawing will be in accurate linear perspective. Use this to
explain Alberti’s Window.
Page 274-275
Perceiving Pictures
We perceive pictures in the same way that we perceive other
objects in the environment: they are flat, mostly on paper, but occasionally in
frames. There is nothing peculiar about pictures as objects. However, this is
not what is usually meant by picture perception. The term refers to the
interpretation of the marks on the surface of the picture – to the recognition
of the marks as referring to objects that are not actually there. Thus,
pictures have a dual reality: they exist as objects and, in addition, the marks
signify other objects and spaces than those they occupy. In the first sense,
pictures pose the same problems as pipes and people; they have locations and
dimensions that need to be resolved by the perceiver. In the second sense, they
are paradoxical because they allude to another space or to other people.
[trompe l’oeil] painting achieves its aim (i.e., not to
appear distorted) because of the constraints placed on the observer: it can
only be viewed from a fixed station point with one eye, and the viewing angle
is such that the extreme edges of the painted surfaces cannot be seen. Neither
of these conditions apply to the pictures we normally see: we use both eyes,
look at them from many different positions, and we can see the borders that
separate the picture from its surroundings.
Despite the fact that trompe l’oeil pictures have been
produced for hundreds of years, novel variations continue to be produced. Paintings
in perspective produce an allusion to a depth not contained in their flatness –
they are seen as both flat and extended in space.
Page 277
It is…likely that we apply shape constancy to the border of
the picture before processing the contents and interpreting the cues to depth.
If this is the case then processes involved in picture perception must occur
after those for object recognition.
Virtual or simulated realities
Many of the pictures we now look
at are created by computers. Indeed, the technology of computer graphics is
advancing so rapidly that the picture on the screen can mimic the view of a
real scene and change with movements of an observer in the way a view of a real
scene would. Furthermore, artificial environments can be generated by computers
and presented to observers to induce the impression of exploring a
three-dimensional space.
…the issues at the heart of
simulated reality address some ancient questions of visual perception. One of
these is the contrast between monocular and binocular vision, which Leonardo da
Vinci repeatedly struggled with. He made many drawings of viewing with one and
two eyes, and realised that the perception of depth is deficient in a painting
or any other two-dimensional representation. Alberti’s window provides a
monocular match between a picture and a view from a single point, but what
happens when two viewpoints are adopted? It is clear that Leonardo came close
to representing binocular disparities, but he was unable to do so because he
had no adequate theory of image formation in the eye.
Neither Leonardo’s painted
imitation of the world from a single vantage point nor Wheatstone’s paired pictures
entertained the possibility of a moving perceiver.
Page 278
…simulated reality is by no means
a novelty if we consider that pictures provide indirect access to a space other
than the one occupied by the visual display. However, the increasing power of
cameras and computers has added a new dimension to both simulated reality and
remote vision – the visual display of current that can be rapidly updated.
Simulated reality is concerned
with creating an imitation of the visual world. It is considered, wrongly, to
be a modern engineering enterprise, made possible by the power of computers. To
repeat the point made earlier, there has been very little investigation of the
recognition of real, three-dimensional objects, rather than arbitrary symbolic
patterns like line drawings of photographs of objects. Such patterns play an
important role in human activity, in the form of alphanumeric and pictures of
all sorts, but are not encountered in the natural environment.
Summary and Conclusions
Our current theories of
perception have been shaped by ideas in physics, art, physiology, biology,
philosophy, psychology, and more recently computer science.
…pictorial representations of objects could be considered as adequate
experimental substitutes for the objects themselves, and that the dynamic
aspects of vision could be too readily ignored.
Because our eyes are in constant
motion, the retinal image is an abstraction, not actually present in the eye
unless a scene is viewed for a very brief interval, too short for the eyes to
move.
Page 282
No comments:
Post a Comment