Visual Perspective

Visual Perspective concerns how spatial reality becomes a visible appearance—and how that appearance is received, organised and experienced through vision.

It connects physical space, light, optics, the eye, retinal image formation, visual angles, depth information, eye movement and perceptual interpretation. It therefore extends beyond the geometry of graphical perspective and beyond the comparison of the eye with a camera.

Visual Perspective is divided into two related meanings:

Visual Perspective Type 1 encompasses visible appearances, views, images and representations, whether seen directly or produced through natural, graphical, optical, instrument, simulated or digital processes.
Visual Perspective Type 2 concerns perspective as formed and experienced through the human or animal visual system.

All Visual Perspective Type 2 belongs within Visual Perspective Type 1, but Type 1 also includes paintings, photographs, films, displays, projected images and other visible representations that are not themselves human visual perception.

This page concentrates primarily on Visual Perspective Type 2: the optical, retinal, physiological and perceptual processes through which we see and understand spatial reality.

From Spatial Reality to Visual Experience

Visual experience does not arise in a single step. It develops through a sequence of connected perspective processes:

Physical space → Natural Perspective → Optical Perspective → retinal image → Visual Perspective Type 2 → perceived visual space

Objects and scenes possess physical dimensions, positions, orientations, surfaces and material properties. Light illuminates those objects, is absorbed, scattered, reflected or refracted, and travels towards the eye. The optical system of the eye forms retinal images, which are then processed and interpreted by the visual system.

Each stage transforms the information received from the preceding stage.

Physical reality is therefore not transferred unchanged into perception. Its appearance depends upon:

viewpoint and viewing direction;
object distance and orientation;
illumination and atmosphere;
reflection and refraction;
occlusion;
the optical properties of the eye;
retinal resolution;
monocular and binocular information;
eye, head and body movement;
attention, memory and perceptual interpretation.

Visual Perspective Type 2 is the final human visual stage of this chain, but it depends upon all the physical and optical processes that precede it.

Light and Visual Information

Vision requires light emitted by luminous sources or reflected from illuminated objects to reach the eye.

Under ordinary visual conditions, light can often be represented geometrically as travelling along approximately straight paths. This principle explains why objects can hide one another, why shadows possess defined boundaries, why viewing direction can be represented by lines, and why visual and graphical perspective can be analysed using rays, angles and projections.

The term visual ray is used for a representative line connecting an object point with the eye or viewpoint. It should not be understood as the only light ray travelling from that point. Light may radiate or scatter in many directions, while the visual ray identifies the particular direction through which that point is seen.

Visual rays therefore help describe:

the direction of visible object points;
visual angles;
the extent of the visual field;
the visual pyramid;
projected size and shape;
occlusion;
foreshortening;
and the geometrical relationship between scene, viewpoint and image.

The eye receives directional patterns of light, but direction alone does not directly specify the distance of the source. Distance and depth must be inferred from additional visual information.

The Eye and Retinal Image Formation

The eye is a biological optical system. Light first passes through the cornea, which provides much of the eye’s refractive power. It then passes through the pupil, whose diameter is controlled by the iris, and through the crystalline lens, which adjusts its form during accommodation.

Together, the cornea and lens focus light upon the retina.

The retina contains photoreceptors that respond to light:

cones support detailed and colour vision, especially around the central foveal region;
rods are highly sensitive under low illumination and are more numerous away from the centre.

Light from different directions stimulates different retinal locations. In this way, the retinal image preserves important information about the direction, arrangement, colour, brightness and optical structure of the visible scene.

The retinal image is inverted relative to the external scene, but visual experience is not produced by simply turning this image upright. Perception results from the organisation of visual information by the retina, optic nerves and brain.

The retinal image and the perceived visual world must therefore be distinguished:

Retinal image = the optical image formed upon the retina
Perceived visual space = the spatial organisation experienced through vision

The retina registers changing patterns of light; it does not directly measure complete physical distance, scale or three-dimensional structure.

The Visual Pyramid and Visual Angles

The visual pyramid, also called the pyramid of sight, is a geometrical model of the relationship between a viewpoint and a visible object or scene.

Its apex lies at or near the eye, while its outer rays extend towards the visible limits of the object or field. For a circular field this relationship may instead be described as a visual cone.

The visual pyramid is not a complete optical model of the eye. It simplifies many individual bundles of light into representative directions so that visual size, projected shape and spatial relationships can be understood geometrically.

Its most important quantities are visual angles.

A visual angle is the angle formed at the viewpoint by rays extending towards the limits of an object or feature. The apparent extent of an object depends principally upon:

its physical dimensions;
its distance from the viewpoint;
its orientation;
and the direction from which it is viewed.

As distance increases, an object normally subtends a smaller visual angle. Its physical size has not changed, but its retinal and projected image becomes smaller.

Visual perspective must therefore distinguish:

Physical size — the measurable size of an object
Projected or retinal size — the size of its formed image
Apparent size — its experienced visual size

These three quantities are related, but they are not identical.

The Generation of Three-Dimensional Vision

A retinal image does not directly contain physical depth as a separate dimension. The visual system must infer three-dimensional structure from multiple sources of information, commonly called depth cues or generators of the third dimension.

These cues do not operate as isolated rules. They interact, reinforce or sometimes contradict one another.

Monocular Depth Information

Monocular depth information is available to one eye and remains important when viewing both physical scenes and flat representations.

Important monocular cues include:

relative and familiar size;
diminution with distance;
occlusion or superposition;
texture gradients;
perspective of form;
foreshortening;
height and position in the visual field;
atmospheric changes of colour and contrast;
light, shade and cast shadows;
focus and blur;
motion parallax;
emergence and disappearance;
relative movement;
and the convergence or organisation of spatial directions.

A painting, drawing, photograph or ordinary screen can communicate powerful three-dimensional structure without providing separate stereoscopic views. This is because it preserves many monocular relationships that the visual system already uses when interpreting physical reality.

Binocular Vision

Binocular vision combines the partially overlapping fields of two eyes located at different positions.

Each eye receives a slightly different view. The visual system relates corresponding features within these two retinal images and uses their differences—binocular disparities—as information about relative depth.

Important binocular processes include:

vergence, in which the two eyes rotate to fixate upon a common object;
binocular disparity, the positional difference between corresponding image features;
stereopsis, the experience of depth arising from binocular disparity;
and the integration of the two retinal images into one coherent visual world.

Binocular depth information is strongest at comparatively close distances and becomes progressively less effective as distance increases. Beyond the most effective stereoscopic range, monocular information, movement, scale, overlap, context and prior knowledge become increasingly important.

Stereopsis is therefore an important generator of depth, but it is not the sole basis of three-dimensional vision.

Apparent Size, Shape and Scale

The appearance of an object changes with distance, viewing direction, orientation and the resolving capacity of the visual system.

Diminution of Form

As a specified object moves farther from the observer, it subtends a smaller visual angle and normally forms a smaller retinal image.

This is diminution of form.

The object does not physically contract. Rather:

Increased distance → reduced visual angle → reduced retinal or projected size

Relative size can consequently provide information about depth, particularly when the observer knows or assumes that the compared objects possess similar physical dimensions.

Perspective of Form

An object’s apparent shape also changes as its orientation changes.

A circular plane can appear elliptical when viewed obliquely. A long object directed towards the observer appears shortened. Surfaces become more or less visible, outlines change, and parts may overlap or conceal one another.

These changes constitute the perspective of form and include foreshortening, changing aspect and projected-shape transformation.

Scale and Resolution

The amount of visible structure also depends upon scale and resolution.

At a great distance, a complex object may appear as a simple outline or point. As the observer approaches—or as optical magnification increases—additional structures become visible. Apparent simplicity is therefore partly produced by the limits of the viewing system.

Visual appearance depends not only upon an object’s physical structure, but upon the scale at which it is observed.

The Visual Field

The visual field is the angular extent available to vision at a particular moment and under stated conditions.

It is not an evenly detailed rectangular picture. Sensitivity, colour discrimination and resolution vary greatly across it.

The foveal region supports the sharpest vision, while acuity declines towards the periphery. The surrounding field remains essential for orientation, movement detection and awareness, but it does not possess the same fine resolution as the central field.

Because high-acuity vision occupies only a small region, the eyes continually move. Successive fixations direct the fovea towards different objects and regions, while perception integrates information acquired across time.

The visual field is therefore:

angular rather than simply planar;
unequal in acuity;
altered by eye movement;
extended by head and body movement;
and continuously reconstructed through active observation.

A fixed graphical picture plane can represent a selected frontal field, but it should not be mistaken for the complete organisation of natural vision.

Visual Space

The expression visual space has been used in different and sometimes opposing senses.

It may refer to:

the visual field, perceived world or image space organised through vision; or
the external geometrical or optical world that is being viewed.

To prevent ambiguity, this page uses visual space primarily for the perceived spatial organisation produced through Visual Perspective Type 2.

Visual space includes experienced:

direction;
distance;
depth;
size;
shape;
orientation;
movement;
and relative spatial position.

It is related to physical space, retinal imagery and the visual field, but it is not identical to any one of them.

The Sphere of Vision

Natural vision is not limited to a single forward-facing picture plane.

The observer occupies a position within a world extending above, below, in front, behind and to either side. Through movements of the eyes, head and body, the line of sight can be directed throughout the surrounding environment.

The Sphere of Vision models this observer-centred organisation.

Within this model:

the observer occupies the central station point;
every position upon the surrounding sphere represents a possible viewing direction;
opposite directions occupy opposite parts of the sphere;
the horizon extends around the observer;
and a frontal visual field represents only one region of the complete directional structure.

The Sphere of Vision is therefore associated with looking around:

Sphere of Vision = multiple outward viewing directions around an observer

It provides a basis for understanding panoramic vision, spherical representation, immersive media and the all-direction organisation of visual space.

Looking At, Looking Around and Looking Through

The old page identified several important viewing activities. They can be clarified as follows.

Looking At

Looking at is directed attention towards a particular object or limited region.

When successive viewpoints are formed around an object, the observer or potential station point revolves around it. This produces a Sphere of Revolution:

Sphere of Revolution = multiple views formed around an object

It is an object-centred arrangement.

Looking Around

Looking around directs the line of sight outwards through the surrounding environment.

This is represented by the Sphere of Vision, in which the observer remains the centre of the directional field while viewing forwards, laterally, upwards, downwards and behind.

It is an observer-centred arrangement.

Looking Through

Looking through occurs when vision is mediated by an intervening aperture, window, picture plane, lens, mirror, instrument, screen or display.

Examples include:

looking through a window;
looking through a telescope or microscope;
viewing a camera image;
watching cinema or television;
using a head-mounted display;
viewing augmented or virtual reality.

In these cases, the viewer experiences not only natural visual perspective but also the optical, graphical, instrument or simulated perspective introduced by the intervening system.

Visual Acuity and Optical Resolution

Visual acuity is the capacity to resolve fine detail.

It depends upon:

retinal receptor density;
optical focus;
contrast;
illumination;
object size;
distance;
motion;
viewing duration;
and the retinal location upon which the image falls.

Central vision generally supports finer resolution than peripheral vision. An object that remains physically present may therefore cease to be distinguishable because its image has become too small, faint, blurred or low in contrast.

This produces optical vanishing.

Optical vanishing must be distinguished from geometrical vanishing:

Geometrical vanishing concerns where spatial directions appear to converge or projected forms collapse.
Optical vanishing concerns how far an object or detail remains detectable.

There is no single universal distance at which all objects vanish. Visibility depends upon object size, contrast, illumination, atmospheric conditions, visual acuity and the resolving power of the viewing or imaging system.

Focus, Accommodation and Depth of Field

The eye cannot focus every distance with equal sharpness at the same instant.

Accommodation changes the optical power of the crystalline lens so that objects at different distances can be brought into clearer retinal focus. This process is most effective at relatively close distances and becomes more limited with age.

Objects lying outside the effective focus range may appear blurred. Blur can contribute to depth interpretation, but it also reduces detail and may cause optical disappearance.

The visual experience of focus is dynamic because the observer continually shifts attention and accommodation between different spatial regions. Natural vision is therefore unlike a single photograph possessing one permanently fixed plane of focus.

Eye Movement and the Active Observer

Human seeing is an active process.

The eyes make rapid movements between fixation points. The head turns, the body changes position, and the observer approaches, withdraws from or moves around objects. Each movement produces new projected relationships.

Movement reveals:

previously hidden surfaces;
changes in occlusion;
motion parallax;
changes in apparent size;
alterations in foreshortening;
changing reflections and highlights;
and new relationships between object and background.

Visual perception integrates these successive views across time.

A fixed viewpoint remains useful for analysing a particular visual or graphical projection, but it does not describe the full activity of normal seeing. Visual Perspective Type 2 is normally dynamic, multi-fixational and embodied.

Human Spatial Orientation

The visual system must relate visible information to the observer’s eyes, head and body.

Human spatial orientation can be organised through an egocentric coordinate system centred upon the observer. Important bodily directions include:

above and below;
left and right;
forward and backward;
vertical and horizontal;
and near and far.

Vision also interacts with gravity, balance, posture and bodily movement. The perceived orientation of a room, object or image therefore depends partly upon the relationship between visual directions and the observer’s bodily frame.

When viewing a photograph, painting or screen, the viewer must relate the represented viewpoint and spatial structure to their own position outside the image.

Visual Direction and Object Position

The retina receives information about the direction from which light arrives. The visual system must then interpret this information in relation to eye position, fixation, binocular alignment and movement.

Physical location and apparent location can diverge.

Mirrors produce virtual images. Refraction can displace the apparent position of an object. Lenses can magnify, invert or redirect images. Displays and head-mounted systems can present image content whose apparent location differs from the physical surface producing it.

Object position in visual experience therefore depends upon the complete chain connecting:

physical object → light path → optical transformation → retinal location → perceptual interpretation

This is one reason why visual perspective cannot be reduced to retinal geometry alone.

Visual Perspective and Linear Perspective

Visual Perspective Type 2 and graphical linear perspective are closely related, but they are not identical.

Linear perspective normally assumes:

a fixed station point;
a fixed viewing direction;
a flat picture plane;
rectilinear projection;
and a selected field represented at one moment.

Natural visual perspective involves:

curved retinas;
two moving eyes;
changing fixation;
variable acuity;
binocular and monocular information;
accommodation;
movement of the head and body;
and perceptual integration across time.

Linear perspective is highly effective under appropriate projection and viewing conditions. It can reproduce many important relationships of size, foreshortening, position, convergence and spatial recession.

However, it should not be treated as a complete model of the entire visual field or of conscious visual experience.

The correct conclusion is neither that linear perspective perfectly duplicates vision nor that it is visually invalid:

Linear perspective is a rigorous graphical projection that corresponds closely with selected aspects of visual appearance under defined conditions.

Visual Lateral Distortion

Wide-field viewing and representation raise special problems towards the lateral regions of vision.

The PRC distinguishes several factors that may contribute to Visual Lateral Distortion:

the mapping of spatial directions onto curved retinal surfaces;
the increasing egocentric distance of laterally positioned objects relative to a frontal reference plane;
and changes of orientation and foreshortening as objects are viewed away from the central direction.

Graphical rectilinear perspective introduces a related but different effect. When a very wide field is projected onto a flat plane, rounded and volumetric forms near the margins may appear laterally stretched, even though straight lines remain straight.

Visual and graphical lateral effects should therefore not be treated as identical. One concerns the organisation of natural vision; the other concerns the mapping of a wide angular field onto a flat represented surface.

The Beholder’s Share

Visual experience depends upon both the visible stimulus and the observer.

The eye receives patterns of light, but perception must organise those patterns into objects, surfaces, distances, directions and meaningful spatial relationships. The observer uses context, expectation, memory, attention and prior knowledge when interpreting incomplete or ambiguous information.

This active contribution is sometimes called the beholder’s share.

A partly hidden object may be experienced as complete. An unfamiliar projection may initially appear distorted and later become intelligible. A flat image can evoke a convincing three-dimensional world even though no physical depth exists within its surface.

Visual perspective is therefore both:

an optical process through which visual information is formed
and an interpretive process through which spatial meaning is experienced

Viewing Represented Images

When an observer looks at a drawing, photograph, film or screen, several perspective processes may operate in sequence.

A photograph of a physical scene, for example, may involve:

Natural Perspective → camera optics → Instrument Perspective → image processing → displayed image → optics of the eye → Visual Perspective Type 2

The viewer simultaneously encounters:

the physical surface of the image;
the represented viewpoint;
the apparent spatial world within the image;
and the actual environment surrounding the image.

A represented image therefore produces a layered visual experience. The observer may perceive both a flat material surface and an implied three-dimensional scene.

This dual awareness is fundamental to drawing, painting, photography, cinema, computer graphics, virtual reality and other representational systems.

Why Visual Perspective Matters

Visual Perspective provides the essential connection between physical spatial reality and human experience.

It helps explain:

why objects change in apparent size and shape;
how depth is experienced;
how the eye forms retinal images;
why vision differs from a camera;
how monocular and binocular information interact;
why central and peripheral vision differ;
how movement changes visual information;
why distant objects lose detail;
how optical and graphical images are interpreted;
and why no single projection reproduces every aspect of natural vision.

It is fundamental to art, architecture, photography, cinema, scientific imaging, interface design, computer graphics, robotics, artificial intelligence, virtual reality and visualisation.

Understanding Visual Perspective Type 2 also provides the standard against which artificial image systems are designed, compared and evaluated.

Natural and Visual Perspective — Volume 3

Visual Perspective Type 2 will receive detailed treatment in Volume 3 of The Art and Science of Perspective:

Natural and Visual Perspective

The volume examines the continuous sequence through which physical space, light and viewpoint become natural appearance, retinal imagery and human visual experience.

Its principal subjects include:

Natural, Optical and Visual Perspective;
visual rays and retinal image formation;
monocular and binocular vision;
vergence, disparity and stereopsis;
apparent size, shape, scale and distance;
the visual field and Sphere of Vision;
Looking At, Looking Around and the Sphere of Revolution;
natural scaling and visual lateral distortion;
movement, occlusion, atmosphere and visibility;
geometrical and optical vanishing;
and the relationship between physical, retinal and perceived visual space.

The book develops the foundations introduced in Volume 1 and defined in the Dictionary of Perspective into a sustained study of how spatial reality becomes visual experience.

Natural Perspective — how physical spatial reality changes in appearance with viewpoint, distance, illumination and atmosphere.

Optical Perspective — how light, reflection, refraction and image formation produce optical views and images.

Perspective Category Theory — the classifications connecting Natural, Visual, Optical, Mathematical, Graphical, Instrument, Simulated and New Media Perspective.

Common Errors and Misconceptions — recurrent errors concerning the eye, visual field, depth, linear perspective, optical vanishing and the relationship between vision and representation.