4

The lens and perspective

Perception and depth

The eye/brain judges depth by binocular cues (available to two eyes) and monocular cues (available to one eye). Because the eyes are about 6 cm apart the retinas of the eyes receive slightly different images. It is the comparison of these two images and by movement of the head that information about depth is achieved with binocular vision. The short fashion for three-dimensional films in the 1950s replicated stereoscopic vision by having two overlapping images forming the projected image that were separated when viewed by green and red spectacles. Apart from this fashion, a television or film shot is a ‘one-eyed’ system and indications of depth are achieved by:

   relative size of known objects or same size objects;

   linear perspective – parallel lines converge in the distance (e.g., looking along railway lines);

   overlap – any subject that obscures another subject is perceived as being closer to the lens;

   relative brightness – subjects that are clearer and brighter are perceived as being closer to the lens than subjects at a distance;

   motion parallax – as the camera's viewpoint changes, more distant objects will move more slowly than objects close to the lens;

   texture gradient – regular size objects (e.g., blades of grass, fabric weave) will diminish in size as they recede from the lens;

   height in the frame – a subject that is higher in the frame (and smaller) than a similar foreground subject is perceived as being further away.

Depth indicators and their relationship to the lens

The mathematical laws by which objects appear to diminish in size as they recede from us, the way parallel lines appear to converge and vanish at the horizon, were introduced to Western art in the fifteenth century.

Figure 4.1 A detail from The Profanation of the Host (1465), a painting by the Florentine artist Paolo Uccello where he explores the newly discovered linear perspective indicators to represent depth on a two-dimensional surface.

The artist David Hockney suggests that some paintings of the fifteenth century and later were created with the aid of mirrors or lenses. If his theory is correct, artists who worked on a projected image may have faced similar lens problems to those contemporary cameramen are involved with

The Profanation of the Host (Figure 4.1), is a detail of a painting by the Florentine artist Paolo Uccello. As you can see, he has used all the newly discovered linear perspective indicators to represent depth on a two-dimensional surface. They include converging parallel lines such as the timbers in the ceiling, the tiles on the floor and the walls, and a reduction in the size of the tiles as they recede.

So what are the laws of perspective that need to be understood? Unlike the Renaissance artist, the cameraman does not have to puzzle over how to represent a two-dimensional plan of a three-dimensional view before he produces a realistic picture. He does not have to analyse how the eye perceives depth. He simply presses the record button and the camera does the rest. Or does it?

The cameraman has to decide at what distance and with what lens angle he will shoot the scene. That will make a difference to the size relationships within the frame – he will control the perspective of mass.

He has to decide the lens height. Shooting low with a level camera will produce one type of line perspective. Shooting from a high vantage point tilted down will produce another set of line relationships in the frame.

The camera doesn't lie – much. It simply reproduces an image conditioned by one of the four parameters mentioned above – lens height, camera tilt, distance from subject and lens angle.

In identifying how the two-dimensional depiction of space is created, there are a number of lens characteristics that play an important part. In order to control and manipulate these variables in the visual design of a composition, it is necessary to understand how the lens

Figure 4.2 (a) single element lens; (b) multi-element lens

Figure 4.3 Depth-of-field: the three variables that affect depth-of-field are camera distance, ƒno and lens angle

Focal length

When parallel rays of light pass through a convex lens, they will converge to one point on the optical axis (see Figure 4.2). This point is called the focal point of the lens. The focal length of the lens is indicated by the distance from the centre of the lens or the principal point of a compound lens (e.g., a zoom lens) to the focal point. The longer the focal length of a lens, the smaller its angle of view will be; and the shorter the focal length of a lens, the wider its angle of view.

Focal length (lens angle) and the distance of the camera from the subject(s) of a shot play a crucial part in the two-dimensional depiction of space.

Angle of view

The focal length of a prime lens has a specified angle of view. A zoom lens has a variable focal length and therefore a variable angle of view. The approximate horizontal angle of view of a fixed focal length lens can be calculated by using its focal length and the size of the recording format frame of the camera. Because the horizontal angle of view of the lens is proportional to the width of the recording format (e.g., 35 mm film, 2/3″ CCD video, etc.), lenses with the same focal length will produce different angles of view depending on the format in use (see Chapter 12, ‘Composition styles').

For example, a video camera fitted with 2/3″ CCDs the formula would be:

Depth-of-field

The depth-of-field, how much of the scene in shot is in acceptable focus, is another important element in shot composition and in controlling how the viewer responds to the image. Cinemagraphic fashion has alternated between deep focus shots (Greg Toland's work on ‘Citizen Kane’ (1941)), to the use of long focal lenses with a very limited depth-of-field only allowing the principal subject in the frame to be sharp. Changing the ƒno alters the depth-of-field – the portion of the field of view that appears sharply in focus.

This zone extends in front and behind the subject in focus and will increase as the ƒno increases. The greater the distance of the subject in focus from the camera, the greater the depth-of-field. The depth-of-field is greater behind the subject in focus than in front and is dependent on the focal length of the lens and fno. For a correctly exposed picture, the depth-of-field can be adjusted by altering light levels or by the use of neutral density filters that will then require an adjustment to the aperture (fno) to return to a correct exposure.

ƒ no

The ƒno of a lens is a method of indicating how much light can pass through the lens. It is inversely proportional to the focal length of the lens and directly proportional to the diameter of the effective aperture of the lens. For a given focal length, the larger the aperture of the lens the smaller its ƒ no and the brighter the image it produces ƒ nos are arranged in a scale where each increment is multiplied by $$2 (1.414). Each time the ƒ no is increased by one stop (e.g., ƒ 2.8 to ƒ 4), the exposure is decreased by half:

ƒ 1.4   ƒ 2   ƒ 2.8   ƒ 4   ƒ 5.6   ƒ 8   ƒ 11   ƒ 16   ƒ 22

Zoom

The majority of video cameras are fitted with a zoom lens that can alter its focal length and therefore the angle of view over a certain range. This is achieved by moving one part of the lens system (the variator) to change the size of the image, and by automatically gearing another part of the lens system (the compensator) to simultaneously move and maintain focus. This alters the image size and therefore the effective focal length of the lens. To zoom into a subject, the lens must first be fully zoomed in on the subject and focused. Then zoom out to the wider angle. The zoom will now stay in focus for the whole range of its travel.

Readjustment on shot

In live television productions, the zoom lens angle is often altered to trim or adjust the shot to improve the composition when the content of the shot changes. Someone joining a person ‘in shot’ is provided with space in the frame by zooming out. The reverse may happen when they leave shot – the camera zooms in to recompose the original shot. Trimming the shot ‘in vision’ may be unavoidable in the coverage of spontaneous or unknown content but it quickly becomes an irritant if repeatedly used. Fidgeting with the framing by altering the zoom angle should be avoided during a take.

Zoom ratio

A zoom lens can vary its focal length. The ratio of the longest focal length it can achieve (the telephoto end) with the shortest focal length obtainable (its wide-angle end) is its zoom ratio. A broadcast zoom lens will state zoom ratio and the wide-angle focal length in one figure. A popular zoom ratio is a 14 × 8.5. This describes a zoom with a 14:1 ratio starting at 8 5 mm focal length (angle of view = 54° 44′) with the longest focal length of 14 × 8.5 mm = 119 mm (angle of view = 4° 14×).

Extender

A zoom lens can be fitted with an internal extender lens system that allows the zoom to be used on a different set of focal lengths. A 2× extender on the 14 × 8.5 zoom mentioned above would transform the range from 8 5 mm-119 mm to 17 mm-238 mm but it will lose approximately two stops of sensitivity.

Focus

Focusing is the act of adjusting the lens elements to achieve a sharp image at the focal plane. Objects either side of this focus zone may still look reasonably sharp depending on their distance from the lens, the lens aperture and lens angle. The area covering the objects that are in acceptable focus is called the depth-of-field.

The depth-of-field can be considerable if the widest angle of the zoom is selected and, whilst working with a small aperture, a subject is selected for focus at some distance from the lens. When zooming into this subject, the depth-of-field or zone of acceptable sharpness will decrease.

Follow focus

Film and television often have a high proportion of shots of faces and the eyes need to be in sharp focus. Focus on a video camera is determined and adjusted with reference to the viewfinder picture. Sharpest focus can be checked ‘off-shot’ by rocking the focus zone behind and then in front of the eyes. As camera or subject moves there will be a loss of focus that needs to be corrected. The art of focusing is to know which way to focus and not to overshoot. The peaking control (if fitted) on the viewfinder emphasizes the electronic edges and is an aid to focusing and does not affect the recorded image. In setting a film shot, the principal subject distance is measured and any change of focus is calibrated and marked.

Zoom lens and focus

A zoom lens is designed to keep the same focal plane throughout the whole of its range (provided the back focus has been correctly adjusted). Pre-focus whenever possible on the tightest shot of the subject. This is the best way of checking focus because of the small depth-of-field and it also prepares for a zoom-in if required.

Pulling focus

Within a composition, visual attention is directed to the subject in sharpest focus. Attention can be transferred to another part of the frame by throwing focus onto that subject. Match the speed of the focus pull to the motivating action.

If the focus is on a foreground person facing camera with a defocused background figure and the foreground subject turns away from camera, focus can be instantly thrown back to the background. A slower focus pull would be more appropriate in music coverage, for example, moving off the hands of a foreground musician to a background instrumentalist.

Differential focus

Differential focus is deliberately using a narrow depth-of-field to emphasize the principle subject in the frame in sharp focus that is contrasted with a heavily out-of-focus background.

The structural skeleton of a shot

Although a television or film image is viewed as a two-dimensional picture, most shots will contain depth indicators that allows the audience to understand the two-dimensional representation of space that contains the action. Text on a blank background has no depth indicators but the text is still perceptually seen as ‘in front’ of the page.

Figure 4.4 Diagonal arrangements of lines in a composition produce a greater impression of vitality than either vertical or horizontal lines. The square-on shot of a house (a) is visually static because it maximizes the number of horizontal lines in the frame. Angling the camera position to show two sides of a building (b) converts the horizontal lines into diagonals. A line at an angle is perceptually seen as a line that is in motion. Compositions with a strong diagonal element imply movement or vitality (see Chapter 5, ‘Visual design').

Although the subject of the shot remains the same – a house – the structural skeleton of the shot has been rearranged to increase the viewer's perceptual attention independent of their interest in the specific content of the shot

The audience will be looking at the surface of the screen, a two-dimensional plane covered by a series of lines, shapes, brightness points, contrasts, colour, etc., and will respond to any indication of recognizable form and space contained in the shot. They will read into the two-dimensional image an impression of a three-dimensional space.

There are therefore two aspects of the composition. The content – a house, horse or face – and the front surface arrangements of lines, shapes, contrasts, etc., that form the recognizable images. The majority of the audience may only remember the content of the shot – the house, horse or face – but they will also be affected by the series of lines, shapes, brightness points and contrasts, colour, etc., which construct the front surface plane of the image. This ‘abstract’ element of the shot may be crucial to the way the viewer responds to the image.

Each visual element in a shot can therefore serve two functions:

1.  as content – that part of the composition that provides depth indicators and information about the physical subject of the shot;

2.  as form – part of the design that lies on the surface plane of the screen and forms an overall abstract design that can produce a reaction in the viewer independent of any response to the content of the shot.

The reduction of this aspect of the shot, its form, to a simplified diagram of line and shape has been termed the structural skeleton of the image. It reveals the perceptual elements that potentially can hold the viewer's attention over and above the interest in the content of the shot.

The construction of the structural skeleton of the plane of the shot does not simply rely on content. For example, every cameraman knows that a shot of a building can be made more interesting if the camera is moved from a square-on symmetrical viewpoint to an angle of view favouring more than one side or surface, and/or if the height of the lens is varied. Repositioning the camera is altering the structural skeleton for, while the content of the shot remains and is recognizable as ‘a building', converging lines of rooftop, windows, doors, etc., have been altered and restructured to provide a more pleasing ‘front surface’ design (Figures 4.4(a) and (b)).

Figure 4.5 All parallel horizontal lines that recede into the distance appear to converge towards the horizon at one point known as the vanishing point. If the camera is level (i.e., without tilt), the horizon line will bisect the frame at midpoint. A line projected from the camera to the vanishing point will intersect all objects at a height equivalent to the lens

Vanishing point

All parallel horizontal lines that recede into the distance appear to converge towards the horizon line. Receding parallel lines above eye-line slope down, receding lines below eye-level appear to slope up towards the horizon (Figure 4.5) The position of the vanishing point in the frame will control the degree of convergence of any parallel lines to the line of the lens axis.

A single vanishing point in the centre of the frame (e.g., level camera positioned in the centre of a street looking along the street with lens at half house height) will produce a very centralized composition with all the parallel lines of the houses meeting in the centre of frame. It could emphasize, for example, the conformity and rigidity in the planning of a housing estate but the shot may have little or no compositional elements that hold the attention (Figure 4.6(a)).

A very popular compositional device to emphasize the principal subject in the frame by focusing the strongly convergent lines behind the subject is to place the vanishing point just outside the frame.

Tilting the camera

Panning the camera up will move the horizon line down and redistribute the proportion of the converging lines so that the upper set of lines will have a steeper angle than the lower lines. Panning the camera down will move the horizon line up and will have the reverse effect.

A third influence on the degree of convergence will be camera position. Moving the camera back and using a longer focal length lens to keep in frame the original houses will reduce the angle of convergence. Moving the camera forward and using a shorter focal length lens will increase the angle of convergence (Figures 4.7(a) and (b)).

The final influence on the structural skeleton of lines will be to increase or decrease the camera height. Craning up and panning down will produce one set of converging lines. Craning down and panning up will create another set of lines.

These four camera parameters – camera height, tilt, lens-angle and camera position – in combination or singularly, all influence the structural skeleton of the shot without altering content.

Two-point vanishing perspective

A camera positioned at the corner of a building with a lens positioned at half building height will produce a shot with two vanishing points. Depending on the framing and content, the vanishing points may be inside or outside the frame.

Again the four camera parameters listed above will have a significant effect on the convergence of lines. Using a very wide-angle lens combined with a camera position close to the building will create the greatest angles of convergence.

Three-point linear perspective

If a camera is looking at the corner of a very tall building at lens position of eye height and is panned up to include the whole building, a third set of converging vertical lines is added to the two sets of horizontal converging lines. There are now three vanishing points in or out of the frame with the additional flexibility of altering all the angles of convergences with camera height, lens angle, camera distance from building and angle of tilt (Figure 4.8).

Figure 4.6 (a) A single vanishing point in the centre of the frame (e.g., level camera positioned in the centre of a street looking along the street) will produce a very centralized composition, with all the parallel lines of the houses meeting in the centre of the frame. (b) Panning the camera right will push the vanishing point to the left of the frame and produce a different set of converging lines. (c) A common compositional device to emphasize the principal subject in the frame is to place the vanishing point just outside the frame so that the strongly converging lines draw the eye to the main subject.

Panning the camera right will push the vanishing point to the left of the frame and produce a different set of converging lines (b). Continuing to pan the camera right will position the vanishing point outside the frame (c) and progressively reduce the angle of convergence of parallel lines until they become horizontal at the point when the lens axis is at 90° to them

Multiple vanishing points

Any number of vanishing points are created depending on the variety and position of parallel horizontal and vertical lines in relationship to the camera lens. For example, a high camera angle looking down on the rooftops of a village has little control over the structural skeleton of the shot apart from adjustment in framing. This becomes unimportant because the variety and interaction of the lines usually gives the shot sufficient visual interest without the need to control the angle of convergence.

Horizon line and camera height as a compositional device

Our normal perceptual experience of someone of our own size moving on flat ground towards us is that the horizon line will always intersect behind them at eye level.

As was mentioned in Chapter 1, it was a fifteenth-century writer/architect, Leon Alberti, who realized that the controlling design factor when creating a two-dimensional image was the distance of the eye from the scene and its height from the ground. The crucial element in his construct was the horizon line. This illusionary line where the ground plane meets the sky is also the point where all orthogonals, that is, parallel lines running at right angles to the horizon line, meet. This point is called the vanishing point.

We are usually most aware of the horizon line when we are by the sea. If we set up a horizontally level camera at eye level beside the sea it follows from Alberti's reasoning that as only horizontal lines can ever reach the horizon, the horizon line will appear to be at mid-point vertically in the frame. It will bisect the frame at its mid-point because only the centre axis of the vertical lens-angle is horizontal.

If the camera is level, the centre axis of the lens will always be the only horizontal line that meets the horizon therefore increasing or decreasing the lens height has no effect on the position of the horizon line in the frame.

Of course, Alberti's explanation of our normal perception of linear perspective does involve two visual illusions. The first illusion is that the sky meets the sea when it obviously does not; second, that a visual sight line parallel to the sea would eventually meet this illusionary line at what is termed the vanishing point.

Looking from behind the camera, we will see that the horizon line will intersect the camera at lens height. If the lens height is 1.5 m then all 1.5-m objects in front of the lens will be cut at the same point by the horizon line in the frame.

If we tilt the camera down, the horizon line moves up the frame. If we tilt up, the line moves down. But if we crane the camera up, keeping it level, the horizon line follows and continues to bisect the frame.

Figure 4.7 A significant influence on the degree of linear convergence is the distance of the camera from the main subject. Moving the camera back (a) and using a longer focal length lens to keep in frame the foreground chess pieces will reduce the angle of convergence of a projected line at the top and at the base of the other chess pieces. Moving the camera forward (b) and using a shorter focal length lens will increase the angle of convergence

If the camera is level, any object between the lens and the illusionary vanishing point on the horizon will be intersected by the horizon line at the same height as the lens.

American silent film production at the turn of the twentieth century used a convention of a 50 mm lens at eye level and actor movement was restricted to being no closer to the lens than 12 ft. With an actor standing 12 ft from the lens, the bottom of the frame cuts him at knee height. By 1910, the Vitagraph company allowed the actors to play up to 9 ft from the lens and the camera was lowered to chest height.

Lens at eye level

From these static camera positions there developed a Hollywood convention of frequently placing the camera at eye level, which in turn allowed the horizon line to cut the foreground actors at eye level. It is particularly noticeable in exteriors in westerns made before the 1960s. Whether the artistes are standing or sitting, the camera is often positioned at eye height, which places the horizon behind the eyes. This emphasizes the main subject of the frame – the face – and the main area of interest of the face – the eyes (Figure 4.10).

In television production, a more prosaic factor controlling lens height is the need to avoid shooting off the top of studio sets. Keeping the camera at eye level speeds up production as actor movement to camera can be accommodated without panning up and shooting off the top of the set or without the need to relight.

Figure 4.8 If a camera is looking at the corner of a very tall building and is panned up to include the whole building, a third set of converging lines is added to the two sets of horizontal lines

Lens height

The height at which the lens is set will also control the way the audience identifies with the subject. Orson Welles, in the film ‘Touch of Evil’ (1958), played a fat corrupt detective in a Mexican border town. He directed the film and, by using a wide-angle lens and placing the camera at a low height looking up at this character, he created a brooding dominant personality who appeared to be towering over the audience and almost falling forwards on to them. The impression produced by this lens height and angle was of an unstable but powerful figure.

Moving the horizon down below a person makes them more dominant because the viewer is forced to adopt a lower eye line viewpoint. We are in the size relationship of children looking up to adults. Leni Riefenstahl's film ‘Triumph of the Will’ (1934) about the 1934 Nazi Party Nuremberg Rally has frequent low-angle shots of Adolf Hitler. It increases his height and status and is contrasted with the high-angle ‘bird's eye views’ of the massed ranks of his followers.

A low lens height may also de-emphasize floor or ground-level detail because we are looking along at ground level and reducing or eliminating indications of ground space between objects. This concentrates the viewer's interest on the vertical subjects. A high position lens height has the reverse effect. The many planes of the scene are emphasized like a scale model. The viewer is in a ‘God-like’ privileged position observing more than the participants in the scene. We are now adults looking down on children.

Usually it is better to divide the frame into unequal parts by positioning the horizon line above or below the mid-point of the frame. Many cameramen intuitively use the Rule of Thirds (see Chapter 10) to position the horizon. A composition can evoke space by panning up and placing the line low in frame. Placing a high horizon in the frame can balance a darker foreground land mass or subject with the more attention-grabbing detail of a high key sky. It also helps with contrast range and exposure.

Figure 4.9 On a flat surface, the horizon line cuts similar size figures at the same point. The height of that point is the height of the lens. The position of the horizon line is controlled by the degree of camera tilt. (a) Low angle; (b) lens at eye height; (c) high angle

Figure 4.10 If the camera is at eye level, the horizon line (if it is in the frame) will pass behind the foreground actors at eye level. This emphasizes the main subject of the frame – the face, and the main area of interest of the face –the eyes

A lens height of slightly above presenter eye height (whether standing or sitting) is usually kinder to the face, provides a more alert and positive body posture and often improves the lighting on an artiste with deep set eyes, etc.

Camera distance

Our normal perceptual experience of someone of our own size moving on flat ground towards us is that the horizon line will always intersect behind them at eye level. Looking at The Profanation of the Host painting (Figure 4.1) one can observe that Uccello has reduced the size of the tiles as they recede from the observer. The ratio of size change puzzled many painters until Alberti showed the commonsense arithmetic of how the reduction in size is directly proportional to the distance from the eye.

A 1.5-m woman, 2 m from the lens, will produce an image that is twice as large as a 1.5-m woman, 4 m from the camera (Figure 4.11).

Size relationships or the perspective of mass can be confused with the wide angle effect and the narrow angle effect. By this I mean that to increase the size of a background figure to a foreground figure it is common practice to reposition the camera back and zoom in to return to the original framing. The size relationships have now altered. It is not the narrower angle that produced this effect but the increased distance from the camera (Figure 4.12(a)–(c)).

By tracking away from the two figures we have altered the ratio between lens and first figure and lens and second figure. It is a much smaller ratio and therefore the difference in size between the two is now not so great. When we zoom in and revert to the original full frame for the foreground figure we keep the new size relationships that have been formed by camera distance. The two figures appear to be closer in size.

As part of our perception of depth depends on judging size relationships (the further away the subject is, the smaller it appears), our perception of this new size relationship produced by tracking out and zooming in leads us to believe that the distance between equal height figures is not so great as the first framing.

Possibly, the narrow angle and the wide angle effect is misnamed. It should be called the distant viewing effect. Our eyes cannot zoom and therefore we are not so conscious of size relationships changing in normal perception.

The important point to remember is that subject size relationship is a product of camera distance. How the subject fills the frame is a product of lens angle. This, of course, is the crucial distinction between tracking and zooming Tracking the camera towards or away from the subject alters size relationships – the perspective of mass. Zooming the lens preserves the existing size relationships and magnifies or diminishes a portion of the shot (see Chapter 16, ‘Movement).

Figure 4.11 Size relationships are proportional to the distance of the subjects from the camera. The image of a 1.5-m subject 2 m from the lens will be twice as large as the image of a 1.5-m subject 4 m from the camera, whatever lens-angle is used

Lens angle

The choice of lens angle and camera distance from the subject is the controlling factor in the way that depth is depicted in the image. This subject is treated in more detail in ‘Staging’ (Chapter 15) but the ‘internal’ space of a shot often plays a crucial part in setting up the atmosphere of a shot.

A long focal length lens positioned at a distance from a cramped interior will heighten the claustrophobia of the setting. Subject size ratios will be evened out from foreground to background and movement to and away from camera will show no significant change in size and therefore give a subjective impression that no distance has been traversed.

Controlling space with choice of lens angle/camera distance

Both compression of space and the reduction of apparent movement caused problems in the editing of an ‘all action’ film, where a shot of two people struggling on a railway line with a train in the distance was shot with a long focal length lens. The visual impression, because of the compression of space, was that the train was nearly upon them, whereas the narrative required a great deal more action before that point was reached. Secondly, because the train appeared to have little change in size over the duration of the shot, it had the appearance of moving slowly. This negated its threat to the protagonists and reduced the build-up of tension. A wide-angle lens close to the subject will increase space, emphasize movement and, depending on shot content, emphasize convergence of line and contrast of mass.

Placement of vanishing points

Control of convergence becomes important when it is used to focus attention on the main subject of the shot. Converging lines can be used to bring this foreground subject into prominence (see Figure 4.6(c)). The positioning of the vanishing point controls the convergence angles of receding parallel lines. By tilting or panning the camera the vanishing points can be placed within or outside the frame. Where the vanishing point is positioned will have a considerable influence on the composition.

Figure 4.14 (b) shows that Leonardo chose a central vanishing point where all orthogonals (receding parallels perpendicular to the picture plane) converge on the head of Christ. This ‘implosion’ of converging lines is in contrast to the square-on table position that in general tends to reduce the dynamic impact of an image. Placing the vanishing point at the central subject emphasizes the subject as being the centre of the view and therefore psychologically in this viewpoint, the centre of the world. If the vanishing point of these strong converging lines was placed outside the frame or at the edge of the frame, the main subject of the painting could be seen as just another element in the frame.

Many cameramen, in framing up a shot, will seek to maximize the convergence of lines by choice of lens angle and camera height/position. Shooting square-on to a subject usually keeps the vanishing point within the frame and often results in a strong emphasis on symmetry and simple balance, particularly if there are strong horizontal lines in the frame at 90° to the lens axis. By shooting at an oblique angle, the vanishing point is moved out of the frame and there is a greater emphasis of converging orthogonals. These form dominant groups of receding wedge shapes and give a greater dynamic attack to the image compared with strong level horizontal lines. Visual excitement is created by neighbouring parallel lines getting closer and closer together.

Figure 4.12 The distance between the two figures remains unchanged in all three illustrations. The distance between foreground figure and camera has altered. With each camera reposition, lens-angle of zoom is adjusted to keep the foreground figure the same size in the frame. The ‘wide angle’ effect and the ‘narrow angle’ effect is a product of the camera distance from subjects. (a) Mid-range; (b) wide angle; (c) narrow angle

'Normal’ perspective

If an observer looks at a field of view through an empty picture frame held at arms length, he will require the frame to be progressively increased in size the greater the distance the frame is positioned away from him in order that the same field of view is contained within the frame at all times (Figure 4.16).

As we have discussed, the perspective of mass and the perspective of line are created by the distance of the subjects from the observer.

Size relationships and convergence of line in his field of view will depend on their distance from him Therefore if he does not change his position, the perspective appearance of the ‘image’ within the frame will remain unchanged. The frame will simply get larger and larger the further it is from the observer's position.

If a photograph was substituted for the frame and increasingly enlarged to match frame size, the two factors that control the exact replication of perspective characteristics in an image are revealed as image size and the distance of the image from the observer.

No lens produces a ‘wrong’ perspective provided the viewer views the correct size image at the taking distance. A wide-angle shot taken close to the principal subject would require the viewer almost to press their nose to the screen in order to experience the perspective characteristics of the image that they would experience if they had been the camera.

The calculation of which lens-angle provides ‘correct’ perspective (i.e., equivalent to an observer replacing the camera) must include image size of reproduction and the distance the viewer is to the screen. A person sitting in the back row of a cinema may be viewing a screen size that is a tenth of the size the audience in the front is experiencing. There is no lens-angle that can provide both viewing distances with ‘correct’ perspective. The audience in the front row will experience wide-angle shots as having ‘correct’ perspective whilst the audience in the back row may judge narrower angle shots as having more ‘correct’ perspective.

Often a script requires interpretation rather than precise replication of ‘correct’ perspective. Interpretative compositions can therefore be created using perspective characteristics that expand or flatten space.

To visually represent the sensation of vertigo, Alfred Hitchcock, in a famous shot in ‘Vertigo’ (1958), had the camera tracking in matched to a zoom-out to keep the visual elements at the edge of the frame static (see Figure 4.16). Because the camera was moving closer to the subject, the image size relationships and line convergence in the frame changed and gave a greater impression of depth to the shot. The zoom-out compensated by keeping the same size image of the foreground subject producing an effect of space expanding without movement. This zoom effect has been used subsequently by Steven Spielberg in ‘Jaws’ (1975) and Martin Scorsese's ‘GoodFellas’ (1990) and many others.

Figure 4.13

The internal space of a shot

The internal space of a shot is a subtle but important part of the look, mood and atmosphere of the shot. As we have seen, when three-dimensional objects are converted into a flat two-dimensional image, size relationships will be controlled by camera distance to subject and lens angle. A small room can appear large using a wide-angle lens and a large room can appear cramped and condensed using a long focal length lens.

Figure 4.14 (a) Leonardo da Vinci's The Last Supper. Where do the converging ceiling lines meet? See Figure 4.14 (b).

A medium shot of an actor can be achieved using a zoom lens with a lens angle that varies between more than 75° down to less than 5°. The wider angles will produce possible distortion of features or exaggerated body movement but the crucial distinction between using this range of lens angles is that to keep the same size medium shot, the camera will move further and further back from the main subject as the lens angle is decreased. This will alter the size relationship between foreground and background – the internal space of the shot will be altered.

Production style and lens angle

So what lens angle should be selected? This will depend on the mood or feel of the shot and the action that it is to contain. For visual continuity during a scene, or even for the whole production, a limited range of lens angles are often decided upon. For example, one style of production may consistently use a wide-angle lens producing a series of shots that emphasize movement towards or away from the lens giving a great deal of internal space to the shots. This is often accompanied by a low camera height emphasizing ceilings and dynamic converging lines of walls, buildings, etc.

Another ‘internal space’ style is to use long focal length lenses to produce compressed space, extended movement towards and away from the camera and a general mood of claustrophobia. Frequently this style is accompanied by a lack of ‘geography’ shots (e.g., shots that provide information about setting or locale). Shot in tight close-up, the action is followed without revealing the location, resulting in a series of images with swirling backgrounds that generate pace without information. The viewer is sucked into the mystery and teased with a lack of precise visual clues as to the surroundings.

The choice of the lens angle is therefore dependent on how the action is to be staged and the visual style that is required. The narrower the lens angle the more difficult it becomes to achieve smooth and fluid camera development and movement. The camera has to travel further to achieve size change or movement on a narrow lens than on a wide angle; is more prone to movement vibration or unsteadiness on very narrow angles; and requires larger and more precise focus pulls.

Figure 4.14 (b) The ‘implosion’ of the projected converging lines of the ceiling and side panels meeting behind the head of Christ emphasize his central importance in the composition

Figure 4.15 The screen size of the reproduced image will increase proportionally to the viewing distance if the original perspective experienced by the observer at distance Z from the subject is to be duplicated.

The viewing distance Z of screen A is too close to reproduce a ‘natural’ perspective and would simulate a ‘wide angle’ look at that viewing distance.

Screen B would simulate a ‘narrow angle’ viewpoint because the screen size is too small for viewing distance Z

Figure 4.16 Keeping the subject in mid-shot, the camera tracks out while zooming in to keep the subject image the same size. Because of the increasing camera distance and narrower lens angle, a smaller and smaller portion of the background is included in the frame. The visual effect is to freeze the subject in space as the background apparently flows out either side of the frame

Estimating distance

There are a number of perceptual clues that are used to estimate distance or space. The depth indicators include binocular vision, which allows convergent and divergent movement to be estimated by the use of ‘two’ viewpoints. Subjects moving towards or away from the observer alter the size of the image focused on the retina of the eye. This change in size may not be accurately appreciated, as perception often involves deductions from what is known rather than what is seen (see Chapter 3). Objects that overlap and their size relationship, if they are similar sized objects, indicate their relative position in space. Colour change due to atmospheric haze and hazy outline at long distance also aid depth perception. Similar objects moving at different velocities also indicate their spatial relationship.

All these depth indicators can be used in film and television composition not only to replicate normal perceptual experiences but also to create atmosphere or to interpret narrative requirements. The decrease in size of objects as they recede from us is used continuously to check on distance. In can also help to create a false distance. The final scene in ‘Casablanca’ is set inside an aircraft hanger with the doors open, revealing an aeroplane. There was insufficient space in the studio to have an aircraft at the distance required so a scale model was built and ‘casting’ recruited midgets to move about close to the scale model to give it verisimilitude. In ‘Night of the Hunter’ (1955), during the chase sequence in the swamp, a silhouette figure on a horseback crossing the skyline is not Robert Mitchum as implied in the story but a midget on a small pony.

Accentuating depth

To emphasize depth in a shot:

   use overlap of subjects to show foreground and background relationships;

   use camera movement to cover or uncover objects in the frame;

   use camera movement past foreground subjects and a changing background on crabbing movements;

   subjects moving towards the lens or away from the lens will create more depth than subjects moving across the frame;

   lighting treatment can create depth indicators (see Chapter 13 ‘Lighting and composition'),

   exploit depth indicators of relative brightness. By making dark close and bright objects in the background, the eye will be carried into the frame and perceive depth in the shot;

   avoid square-on shots of subject (see Figure 4.4). Show as many planes and sides of the subject as possible to emphasize depth;

   accentuate line convergence by the choice of lens angle and focal length.

Summary

No camera – still, film or video – can record an image without leaving an imprint of the optical properties of its lens, its position in space and some indication of the reasons for selecting that lens position. One of the crucial factors that condition the ‘look’ of the shot concerns perspective.

The form (structural skeleton') of a shot, its dominant lines and shapes, can potentially hold the viewer's attention over and above the interest in the content of the shot. The construction of this ‘structural skeleton’ is dependent on the distance of the camera from the subject and the lens angle, which control the size relationships within the frame – the perspective of mass.

The choice of lens angle and camera distance from the subject is the controlling factor in the way that depth is depicted in the image. Lens height and camera tilt will control line perspective. Shooting low with a level camera will produce one type of line perspective; shooting from a high vantage point tilted down will produce another set of line relationships in the frame.

The important point to remember is that subject size relationship is a product of camera distance. How the subject fills the frame is a product of lens angle. This is the crucial distinction between tracking and zooming