Of the great strands of human curiosity and experimentation that led to the development of photography, few inventions were more important than the camera obscura. A simple device, the camera obscura is a dark room or box with a hole on one side. The scene outside the camera obscura is projected onto a wall opposite the hole.
Starting in the early Renaissance, mathematicians and artists, such as Filippo Brunelleschi, rediscovered the principles of single-point perspective, which had been known to the Romans but lost in the Dark Ages.
As time went by, artists interested in painting realistic scenes began using portable camera obscura devices to project single-point perspective renderings so this could be incorporated into their work. For example, Johannes Vermeer enthusiastically used a portable camera obscura in creating the perspective projections in his paintings.
Before the invention of the camera obscura and widespread adoption of single-point perspective, most art was rendered as basically flat or following a cartoon-like map structure. Occasional elements, such as tables or chairs, were rendered in perspective, but there was no consistent method for achieving this.
Single-point perspective was so incredibly effective as a way to present the world that it has basically taken over our sense of visual reality. It has changed the way we collectively see the world.
Paintings with a single vanishing point appear absolutely real, and we sense them as accurate representations of three-dimensional reality. But, of course, this is an illusion. A painting, like a photograph, is a two-dimensional object that can appear to present three-dimensional reality based on a visual convention—but it can never be three-dimensional reality itself. It’s odd—and ultimately delusional—that we have come to automatically accept single-point perspective as reality itself.
When it comes to photography, we have one-upped the camera obscura. There’s no need to project the lines of perspective onto a wall for tracing and transfer. The vanishing point and related projections are captured optically using the pinhole effect, and can be saved for future use, either chemically or digitally. Perspective is already baked into the mechanics of photographic rendition, as well as into our consciousness.
Nonetheless, single-point perspective with a vanishing point remains a convention—a method of mapping three dimensions onto a two-dimensional space—and only one way of depicting three-dimensional reality. As a matter of photographic composition, one can use the absolute certainty of the viewer’s belief in the veracity of perspective to shift the way the viewer sees reality.
If you are using single-point perspective in your image without considering how it is being used and, most importantly, what the alternatives are, it’s time to change things up a bit. The results can be startling!
Working with the Vanishing Point
A good working definition of a vanishing point is the point at which receding parallel lines appear to meet when rendered in linear perspective. There is more to this definition than meets the eye. For one thing, the parallel lines in question only “appear,” or seem, to meet. In many perspective renditions, the vanishing point is implied and does not actually appear anywhere in the image. The viewer’s brain assigns a position to the vanishing point, even if it isn’t seen, by following the perspective rendition.
As I’ve mentioned, one of the progenitors of today’s camera is the camera obscura. The projective device of the pinhole in the camera obscura has been improved and enhanced using optical science and lenses. The projection of the camera obscura on the back wall of the box or room is now captured using a digital sensor (or film).
This literal one-to-one correspondence between the camera obscura and today’s cameras has led artists and critics astray! The mistake that is made is to assume that because today’s cameras have camera-obscura functionality, they can only render perspective the way a camera obscura does. David Hockney’s quip that photography is like a “paralysed Cyclops”—or is seeing things with only one, fixed eye—comes from this misapprehension.
True, the easiest and default mode when photographing the kind of subject that responds well to perspective—such as landscapes and architecture—is to fall back on basic single-point perspective. But even within the realm of single-point perspective, the photographer has choices as to where the vanishing point is and how to apply that perspective. As I’ll also discuss later in this chapter, multi-point perspective is possible photographically. Once refraction and curvature enter the picture, the vanishing point becomes a complex issue, and multiple vanishing points are possible. In addition, the Photoshop darkroom provides a number of very effective ways to alter the “rules” of perspective.
The simplest relationship to the vanishing point and single-point perspective is to have one vanishing point pretty much in the middle of the image. This central vanishing point is likely to be raised a bit above the midline of the image, just as the horizon line tends to be raised when looking across the landscape, and in some cases tends toward the top of the image.
There’s certainly no rule that says that even a single vanishing point needs to be in the middle of the image, and often it isn’t.
Move your camera, and use your feet to “scootch” around and try different positions to experiment with changing the vanishing point (see page 61).
As with a number of photographic issues, very small changes in camera position can have a big impact on the final image. Even if the vanishing point doesn’t itself change, its location in relationship to the boundary frame of the image will move as you shift your camera’s position.
If the camera’s angle in relationship to the subject is shifted to the left, the vanishing point will probably move to the right.
If the camera’s angle in relationship to the subject is shifted to the right, the vanishing point will probably move to the left.
If you move the angle of the camera up in relation to the subject, most likely the vanishing point will go downward.
If you move the angle of the camera down in relation to the subject, most likely the vanishing point will go upward.
It’s worth taking the time to get a feeling for how the camera angle in relationship to the subject can impact the location of the vanishing point. Even in the simplest images, location of the vanishing point plays a big role in how your viewers will perceive the “reality” that your image has captured.
The vast majority of images—particularly landscape and architectural studies—show a world with a single vanishing point. Just because you only have one doesn’t mean you are out of options!
Understand the importance of the vanishing point to your image, and to the way your viewers visually interpret the relationship of your image to the world, and to their overall sense of the reality of your work. Keeping this importance in mind, it is important to work with your camera when photographing to find compositions where the location of the vanishing point best complements your compositional intent.
Multi-Point Perspective
If one vanishing point is good, then two should be better, and multiple vanishing points should be great—at least, so you’d think.
We have been trained since early childhood via almost all the media that we see—televisions, movie screens, books, our mobile phones—to view the world in a single-point rendition. Any image with more than one vanishing point automatically seems weird and feels uncomfortable.
If you look for them, out in the world there are many scenes that do have more than one vanishing point. The most common scenario is when something splits our vision. This could be a building, a fence, maybe a post, a rock, or even a car. The key issue is that we see something differently on either side of the object that splits our vision. Any time we see something like this, visual parts of our brains tend to work overtime, and it can be like decoding a visual puzzle.
To make an image that captures real-world, multi-point perspective, most likely you’ll want to use a wide-angle lens. The best approach is to use a fairly wide-angle lens in the 21mm–28mm range. This moderate to extreme wide-angle lens will often work better than a wide-angle lens with distortion, such as a fisheye lens.
The problem with using a fisheye lens, and with curvature appearing at the edges of your image due to an extreme wide-angle lens, is that the viewer’s eye notices this distortion immediately and assumes that the multiple vanishing points are related to this distortion rather than occurring naturally.
The idea of an image with more than one vanishing point is to create a sense of being off balance because the multiple perspective renditions contrast with the normality of the photo. The distortion inherent in a fisheye lens goes against this because the viewer expects things to be distorted, and they know right off the bat that they are not looking at a “normal” photo.
As with many aspects of composition, small movements in the camera position make all the difference with multi-point perspective images (for more about moving the camera to enhance composition, turn to pages 61 and 116).
For example, the photo on page 167 shows a street corner where a fence meets in a V-shape. This was photographed as two-point perspective as I pointed the camera directly at the area a little above the level of the fencing where the fence meets at the corner post. If I had moved the camera even a tiny bit in any direction, the image wouldn’t have worked visually.
Wherever I am out and about in the world, I look for vistas that have multiple vanishing points. I believe that having more than one vanishing point adds interest and intrigue to an image. If there’s an effective way to integrate multiple vanishing points into the purpose of a composition, then I know I am well on my way to making a compelling image.
Refraction, Curvature, and the Perspective Model
It’s a fundamental principle of photography that you cannot actually photograph an object. All you can photograph is the light reflected or emitted from the object. This implies that any model of perspective is using reflected or emitted light rather than the actual scene itself.
With it understood that one is capturing light, then it is clear that you can alter the perspective of almost anything if you can curve or bend the light. The most common way to play with light is to use reflection and refraction. A reflection mirrors existing light, and thus will create multiple perspective points if the camera is correctly positioned.
A refraction alters light by bending it. For example, when light passes through a bottle filled with liquid, the light within the bottle bends and any objects in the bottle, such as a spoon, will appear distorted. A curtain blowing in the breeze with shadows projected onto it follows normal perspective for the curtain, but no longer follows normal perspective for the shadows because they will appear curved (for an example, see pages 86–87).
Another example is a reflection in a curved surface, such as a building with large, curved panes of mirrored glass. The reflections in this building no longer follow normal rules of perspective and may show other buildings and scenery that wiggles across the glass. This mirrored scene in no way follows what we think of as normal perspective. The response is not to understand that this is an alternative projection, but rather to see it as a weird reflection.
When working in my studio, I often try to see what weird effects I can come up with using mirroring, curvature, and reflection. The odder the visual effect, the more impact the image has. This is true out in the wider world as well as in the studio.
Changing Perspective in Post-Production
Last but not least, it is possible to change perspective using Photoshop or other postproduction tools. The advantage of changing perspective in post-production is essentially that the sky is the limit. You can modify all aspects of the perspective in your image as much as you want.
The disadvantages of modifying perspective in post-production are that it is surprisingly difficult to pull off with élan, and the results are all too likely to look artificial, faked, and “Photoshopped.” It’s often best to look for situations where what you are doing is enhancing or modifying reality, rather than creating from scratch.
One of my favorite techniques for modifying perspective in post-production involves duplicating an image, horizontally or vertically mirroring the duplicate, and then using portions of the mirrored duplicate to composite with the original image.
In addition, many photo-manipulation software tools allow you to alter perspective. For example, Photoshop’s Edit ►Transform menu provides a number of tools that can be used to alter the perspective in an image. These tools include Distort, Perspective, Skew, and Warp.
Photo-manipulation tools, such as the transformation tools in Photoshop, allow the image manipulator to alter perspective in a variety of ways either locally, for a portion of an image, or globally, for an entire image.