“Perspective” properly refers to any of various graphic techniques for depicting volumes and spatial relationships on a flat surface, such as size perspective and atmospheric perspective. The term “perspective,” however, most often brings to mind the drawing system of linear perspective. Linear perspective is a technique for describing three-dimensional volumes and spatial relationships on a two-dimensional surface by means of lines that converge as they recede into the depth of a drawing. While multiview and paraline drawings present views of an objective reality, linear perspective offers scenes of an optical reality. It depicts how a construction or environment might appear to the eye of an observer looking in a specific direction from a particular vantage point in space.
Linear perspective is valid only for monocular vision. A perspective drawing assumes that the observer sees through a single eye. We almost never view anything in this way. Even with the head in a fixed position, we see through both eyes, which are constantly in motion, roving over and around objects and through ever-changing environments. Thus, linear perspective can only approximate the complex way our eyes actually function.
Still, linear perspective provides us with a method for correctly placing three-dimensional objects in pictorial space and illustrating the degree to which their forms appear to diminish in size as they recede into the depth of a drawing. The uniqueness of a linear perspective lies in its ability to provide us with an experiential view of space. This distinct advantage, however, also gives rise to the difficulty often connected with perspective drawing. The challenge in mastering linear perspective is resolving the conflict between our knowledge of the thing itself—how we conceive its objective reality—and the appearance of something—how we perceive its optical reality—as seen through a single eye of the observer.
Perspective projection represents a three-dimensional object by projecting all its points to a picture plane by straight lines converging at a fixed point in space representing a single eye of the observer. This convergence of sight lines differentiates perspective projection from the other two major projection systems—orthographic projection and oblique projection—in which the projectors remain parallel to each other.
The converging nature of sight lines in linear perspective produces certain pictorial effects. Being familiar with these pictorial effects helps us understand how lines, planes, and volumes should appear in linear perspective and how to place objects correctly in the space of a perspective drawing.
Convergence in linear perspective refers to the apparent movement of parallel lines toward a common vanishing point as they recede.
The first rule of convergence is that each set of parallel lines has its own vanishing point. A set of parallel lines consists only of those lines that are parallel to one another. If we look at a cube, for example, we can see that its edges comprise three principal sets of parallel lines, one set of vertical lines parallel to the X-axis, and two sets of horizontal lines, perpendicular to each other and parallel to the Y- and Z-axes.
In order to draw a perspective, we must know how many sets of parallel lines exist in what we see or envision and where each set will appear to converge. The following guidelines for the convergence of parallel lines is based solely on the relationship between the observer's central axis of vision (CAV) and the subject.
In orthographic and oblique projection, the projectors remain parallel to each other. Therefore, the projected size of an element remains the same regardless of its distance from the picture plane. In linear perspective, however, the converging projectors or sight lines alter the apparent size of a line or plane according to its distance from the picture plane.
Perspective drawings also possess other pictorial characteristics found in multiview and paraline drawing systems.
The observer's point of view determines the pictorial effect of a perspective drawing. As this viewpoint changes—as the observer moves up or down, to the left or right, forward or back—the extent and emphasis of what the observer sees also change. In order to achieve the desired view in perspective, we should understand how to adjust the following variables.
The height of the station point (SP) relative to an object determines whether it is seen from above, below, or within its own height.
The distance from the station point (SP) to an object influences the rate of foreshortening of the object's surfaces that occurs in the perspective drawing.
The orientation of the central axis of vision (CAV) and the picture plane (PP) relative to an object determines which faces of the object are visible and the degree to which they are foreshortened in perspective.
In constructing a perspective by hand, we must have experience in setting up the station point and the angle of view to predict and achieve a reasonable outcome. A distinct advantage in using 3D CAD and modeling programs is that once the necessary data is entered for a three-dimensional construction, the software allows us to manipulate the perspective variables and fairly quickly produce a number of perspective views for evaluation. 3D CAD and modeling programs, while following the mathematical principles of perspective, can easily create distorted perspective views. Judgment of what a perspective image conveys, whether produced by hand or with the aid of the computer, remains the responsibility of its author.
Illustrated on this and the facing page are examples of computer-generated perspectives, showing how the various perspective variables affect the resulting images. The differences in the perspective views may be subtle but they do affect our perception of the scale of the spaces and our judgment of the spatial relationships the images convey.
The location of the picture plane (PP) relative to an object affects only the final size of the perspective image. The closer PP is to the station point (SP), the smaller the perspective image. The farther away PP is, the larger the image. Assuming all other variables remain constant, the perspective images are identical in all respects except size.
In any rectilinear object, such as a cube, each of the three principal sets of parallel lines has its own vanishing point. Based on these three major sets of lines, there are three types of linear perspective: one-, two-, and three-point perspectives. What distinguishes each type is simply the observer's angle of view relative to the subject. The subject does not change, only our view of it, but the change of view affects how the sets of parallel lines appear to converge in linear perspective.
If we view a cube with our central axis of vision (CAV) perpendicular to one of its faces, all of the cube's vertical lines are parallel with the picture plane (PP) and remain vertical. The horizontal lines that are parallel with PP and perpendicular to CAV also remain horizontal. The lines that are parallel with CAV, however, will appear to converge at the center of vision (C). This is the one point referred to in one-point perspective.
If we shift our point of view so that we view the same cube obliquely, but keep our central axis of vision (CAV) horizontal, then the cube's vertical lines will remain vertical. The two sets of horizontal lines, however, are now oblique to the picture plane (PP) and will appear to converge, one set to the left and the other to the right. These are the two points referred to in two-point perspective.
If we lift one corner of the cube off the ground plane (GP), or if we tilt our central axis of vision (CAV) to look down or up at the cube, then all three sets of parallel lines will be oblique to the picture plane (PP) and will appear to converge at three different vanishing points. These are the three points referred to in three-point perspective.
Note that each type of perspective does not imply that there are only one, two, or three vanishing points in a perspective. The actual number of vanishing points will depend on our point of view and how many sets of parallel lines there are in the subject being viewed. For example, if we look at a simple gable-roofed form, we can see that there are potentially five vanishing points, since we have one set of vertical lines, two sets of horizontal lines, and two sets of inclined lines.
The one-point perspective system assumes that two of the three principal axes—one vertical and the other horizontal—are parallel to the picture plane. All lines parallel to these axes are also parallel to the picture plane (PP), and therefore retain their true orientation and do not appear to converge. For this reason, one-point perspective is also known as parallel perspective.
The third principal axis is horizontal, perpendicular to PP and parallel with the central axis of vision (CAV). All lines parallel to CAV converge on the horizon line (HL) at the center of vision (C). This is the particular vanishing point referred to in one-point perspective.
The one-point perspective system is particularly effective in depicting the interior of a spatial volume because the display of five bounding faces provides a clear sense of enclosure. For this reason, designers often use one-point perspectives to present experiential views of street scenes, formal gardens, courtyards, colonnades, and interior rooms. We can also use the presence of the central vanishing point to focus the viewer's attention and emphasize axial and symmetrical arrangements in space.
The diagonal point method for constructing a one-point perspective uses the geometry of a 45° right triangle and the principles of convergence to make depth measurements in perspective.
We can use the diagonal point method to construct a one-point perspective grid easily. A perspective grid is a perspective view of a three-dimensional coordinate system. The three-dimensional network of uniformly spaced points and lines enables us to correctly establish the form and dimensions of an interior or exterior space, as well to regulate the position and size of objects within the space.
We can lay a piece of tracing paper over this perspective grid and draw in the major architectural elements of the space. With the same grid, we can also locate the positions and relative sizes of other elements within the space, such as furniture and lighting fixtures.
The section perspective combines the scaled attributes of a section drawing and the pictorial depth of a perspective drawing. It therefore is able to illustrate both the constructional aspects of a design as well as the quality of the spaces formed by the structure.
The two-point perspective system assumes that the observer's central axis of vision (CAV) is horizontal and the picture plane (PP) is vertical. The principal vertical axis is parallel to PP, and all lines parallel to it remain vertical and parallel in the perspective drawing. The two principal horizontal axes, however, are oblique to PP. All lines parallel to these axes therefore appear to converge to two vanishing points on the horizon line (HL), one set to the left and the other to the right. These are the two points referred to in two-point perspective.
Two-point perspective is probably the most widely used of the three types of linear perspective. Unlike one-point perspectives, two-point perspectives tend to be neither symmetrical nor static. A two-point perspective is par-ticularly effective in illustrating the three-dimensional form of objects in space ranging in scale from a chair to the massing of a building.
The following is a method for constructing a two-point perspective grid utilizing measuring points. As with the construction of a one-point perspective, you should first establish the observer's point of view. Determine what you wish to illustrate. Look toward the most significant areas and try to visualize from your plan drawing what will be seen in the foreground, middleground, and background. Review the perspective variables on pages 115–120.
A measuring point (MP) is a vanishing point for a set of parallel lines used to transfer true dimensions along a measuring line (ML) to a line receding in perspective. The diagonal point in one-point perspective is one example of such a measuring point.
In two-point perspective, you can establish two measuring points (MPL and MPR) for transferring dimensions along the ground line (GL) to the two major horizontal baselines that are receding in perspective.
Once constructed, a perspective grid should be saved and reused to draw perspective views of interior and exterior spaces of similar size and scale. Each unit of measurement can represent a foot, four feet, a hundred yards, or even a mile. Rotating and reversing the grid can also vary the point of view. Therefore, you can use the same grid to draw an interior perspective of a room, an exterior perspective of a courtyard, as well as an aerial view of a city block or neighborhood.
These three perspectives use the perspective grid shown on the preceding page. In each case, however, the height of the observer's station point (SP) above the ground plane (GP) has been selected to portray a specific point of view, and the scale of the grid has been altered to suit the scale of the structure.
This interior perspective also uses the grid shown on page 134. Note that the left vanishing point (VPL) lies within the drawing, enabling three sides of the space to be shown and a greater sense of enclosure to be felt. Because VPL lies within the drawing, greater emphasis is placed on the right-hand portion of the space. If the left-hand side of the space is to be emphasized, use a reverse image of the grid.
The combined effects of convergence and diminishing size make it more difficult to establish and draw measurements in linear perspective than in the other two drawing systems. But there are techniques we can use to determine the relative heights, widths, and depths of objects in the pictorial space of a perspective drawing.
Perspective measurements are not a major issue in 3D-modeling programs because the software uses mathematical formulas to process the three-dimensional data we have already entered.
In linear perspective, any line in the picture plane (PP) displays its true direction and true length at the scale of the picture plane. We can therefore use any such line as a measuring line (ML) to scale dimensions in a perspective drawing. While a measuring line may have any orientation in the picture plane, it typically is vertical or horizontal and used to measure true heights or widths. The ground line (GL) is one example of a horizontal measuring line.
Measuring perspective depth is more difficult than gauging heights and widths in linear perspective. Various methods of perspective construction establish depth in different ways. Once we establish an initial depth judgment, however, we can make succeeding depth judgments in proportion to the first.
There are two methods for subdividing depth measurements in linear perspective: the method of diagonals and the method of triangles.
In any projection system, we can subdivide a rectangle into four equal parts by drawing two diagonals.
To subdivide a rectangle into an odd number of equal parts, or to subdivide its receding edges into a series of unequal segments, its forward edge must be parallel to the picture plane (PP) so that it can be used as a measuring line (ML).
Because any line parallel to the picture plane (PP) can be subdivided proportionately to scale, we can use such a parallel line as a measuring line (ML) to subdivide any intersecting line into equal or unequal parts.
If the forward edge of a rectangular plane is parallel to the picture plane (PP), we can extend and duplicate its depth in perspective.
Once we are familiar with how lines parallel to the three principal axes of an object converge in linear perspective, we can use this rectilinear geometry as the basis for drawing perspective views of inclined lines, circles, and irregular shapes.
A more precise method for determining the vanishing point for an inclined set of parallel lines is as follows:
Drawing stairs in perspective is easiest when we can determine the vanishing point for the inclined lines that connect the stair nosings.
The circle is the essential basis for drawing cylindrical objects, arches, and other circular forms.
Reflections occur on the horizontal surfaces of bodies of water, the mirrored surfaces of glass, and the polished surfaces of floors. A reflecting surface presents an inverted or mirror image of the object being reflected. For example, if an object is resting directly on a reflecting surface, the reflected image is a direct, inverted copy of the original. Thus, in a perspective view of the reflection, the reflected image follows the same perspective system of lines already established for the original image.
Any reflecting planar surface parallel to one of the three major axes extends the perspective system of the subject. Therefore, the major sets of parallel lines in the reflection appear to converge to the same vanishing points as do the corresponding sets of lines in the subject.
When drawing a perspective of an interior space having a mirrored surface on one or more of its major planes, we extend the perspective system in the manner described on the previous page.