Another Perspective on Photography

Perspective -- The effect of viewpoint

Objects appear to shrink in size according to the distance from our eye.

Perspective comes from a Latin word, perspicere, “to look through.”

If you have a grouping of subjects in front of you (a scene) and you draw straight lines from all the points of the objects to a fixed point and then you take a plane surface (tracing paper, glass, etc.) and have those lines pass through that surface, the image formed on that surface is the perspective of that scene. 

In the case of our cameras and lenses, the lines will pass through the lens and be projected reversed left to right and inverted onto a surface (our image sensor or a piece of film). Either way, the projections are exactly the same. What you see are the relationships in sizes between the items in the scene. And this is how I will define perspective for this article. I will also take liberties with the word “distortion” in this article. Technically, distortion is one of the aberrations that can occur in a lens design. You can have barrel distortion where the corners of the image pull inward or pincushion distortion where the corners of the image pull outward. But for the sake of this article, I will use the word distort (distortion, distorted) to describe things that just don’t look quite right in the image. 

 
lens-distortion
 
Perspective projected between the objects and the viewer

Perspective projected between the objects and the viewer

Perspective projected through a lens

Perspective projected through a lens

If all we are to photograph are flat/2D objects on a single plane (such as a painting), we need to set up the camera parallel to the object and evenly centered on the object and we are probably done with the lesson on perspective. But we are often called on to photograph 3D scenes where we want to show the relationship between the objects in the scene. In doing so, we often want to keep parallel lines in the scene parallel in our photographs. So, we will start there. If the camera is held parallel to the lines in the subject they will be parallel on the sensor. But if the camera is tilted up or down (or left or right) from parallel we have the familiar situation of a building that looks like it is falling backward (by tilting the camera up) or we see parallel train tracks appearing to converge in the distance, or rectangular building appears to be a trapezoidal shape with one side being shorter than the other. These effects are also known as “keystoning” because the distortion resembles the shape of the keystone in a stone archway.  The closer you are to your subject(s) the more pronounced the effect will be.

This is because of relative distances. In the case of the building falling backwards, the effect indicates that the photographer is too close to the building and not able to get up high enough to photograph straight on. In order to frame the entire building the photographer points the camera upward to capture the top of the building and while the base of the building might be 10 units away, the top of the building might now be 20 or 30 (or more) units away. Objects closer to the lens (the base of the building) appear larger than objects farther away (the top of the building). So the building takes on a trapezoidal shape and looks like it is falling backwards. It does this when viewed with the naked eye or with a camera. However, with our eye our brain kicks in and compensates because it knows that the building isn’t falling. We expect that the top of the building is farther away and smaller. In a photograph, our brain doesn’t engage that process and we see the distortion.

There are a few solutions to this situation that we can see in this illustration.  

Camera A is in our basic situation. On the ground in front of and near to a building. Even with a wide angle lens, the camera cannot capture the entire height of the building while maintaining a parallel relationship to the building (red lines in the upper diagram). The first solution is to use a tilt-shift or perspective control lens. Using the shift function of the lens the photographer can set up the camera parallel to the building and shift the lens upward to move the image on the sensor to encompass the building (blue lines). Now the perspective on the building is actually wrong in the photo, but appears closer to what we want or expect to see. Personally, I often find these “corrected” photographs to be more disconcerting than the falling building in an uncorrected photograph. It is subtlety “just not right.” And a tilt/shift lens is usually an expensive option and not always readily available.

The next option is to raise the camera (B) up to a higher position, about midpoint on the height of the building so that the base and the top are the same distance from the camera.

The third option is the best, but usually not an option, and that is to move back a few hundred units so that the distance from camera to the base of the building and the distance from the camera to the top of the building are relatively the same. At this longer distance the camera can be parallel to the building or even off a little bit from parallel, and the lines will appear parallel.

The bottom diagram above will help understand what is happening. With camera A the top of the building is twice as far from the camera as the base of the building. So the top will be half as big as the bottom, leading to the “falling backwards” feeling in the photograph. With camera C, the base and the top of the building are virtually the same distance away, so will be the same size in the photograph. 

There is also a fourth option, that is to “correct” the image in post-processing. For this you will need to use a wide angle lens and be back far enough from the building to capture the building top to bottom (with the camera tilted up) so you can stretch out the top of the building to make it similar in size to the base.

Up until the last paragraph I have not mentioned the type of lens (wide angle, normal, telephoto) involved because all of the effects mentioned are determined by the camera to subject distance. When forced to be in close you are forced to use a shorter lens and this causes many people to think that the lens is the cause of the distortion. It is not. All lenses when used from the same position have the same perspective. Camera A or camera C are forced to use a wide angle lens because the field of view of a longer lens will only take in a small part of the building. As you back up to the position of camera C you can use a short, normal, or telephoto lens and the choice of lens will determine the magnification of the building and how much of the surrounding landscape will be included in the frame, but the perspective will be the same with any of the lenses.

This doesn’t only apply to buildings or other tall objects. The same thing happens if you are looking down on a subject. For example, if you are at the top of the grand canyon and photograph down into the canyon from the rim, the bottom of the canyon appears much smaller than the top. But if you were flying over and photographed the canyon from a very high altitude the effect would be much less extreme.

So far, we’ve dealt with one object in the scene and its relationship to itself. Perspective is also about the relationship of various objects in a scene. It can be a landscape with a barn, some trees, and mountain range in the background. Or it could be players on a sports field. Or it can be two wine glasses on a table. 

As with the building, the relationship between the glasses in the following examples is determined by the camera to subject distance. The focal length of the lens just determines what fits into the frame at that distance. The basic set up here is a white seamless paper backdrop with two wine glasses about 46” in front of the backdrop. The glasses are identical in size with one glass six inches in front of the other. There is also a snooted light on the backdrop simulating a spot light. All of that will remain the same and we will examine what happens when the camera is moved closer or farther from the glasses. 

We will start out with the camera in close, about 12” from the front wine glass. To fit the glasses into the frame I have selected a 24mm lens. At this close distance the front glass appears to be almost twice the size of the rear glass, even though they are physically the exact same height. The spotlight on the background is small and tight. The front glass is in focus, but the rear glass is soft and out of the range of the depth of field at the aperture used for the photo (f/10). 

24mm lens 12 inches from the front wine glass

24mm lens 12 inches from the front wine glass

Next, I will keep the 24mm lens on the camera but I am going to back up the camera to be 85 inches from the front glass. Obviously everything in the frame is a lot smaller because we are 7x farther away from the glasses. But now both glasses appear at almost the same size and the spot light on the background has grown larger (even though the light was not moved or changed in any way between shots.

24mm lens at 85 inches from the front wine glass

24mm lens at 85 inches from the front wine glass

Keeping the camera at the new 85 inch distance, let’s change to a 135mm lens and get this image. 

Again the glasses are almost the same size and the background spot is larger than in the first image above. People often call this look “telephoto compression,” but is it? In the next set of images I have taken the photo with the 24mm lens and cropped it to match the framing of the image taken with the 135mm lens. Remember what I said earlier, all lenses have the same perspective from the same camera position.

Here are the images with both lenses taken from the same camera position 85 inches from the glasses and cropped the same compared with the 24mm image that was taken at 12 inches from the glasses.

There are a few things to take notice of here. First, of course, is that even though the photos were taken by radically different focal length lenses (24mm and 135mm), the area of the images common to both images have the exact same perspective. Changing the focal length of the lens did not change perspective. But it did change the field of view and hence the size of the objects in the scene. And what about the spotlight on the background? Why did it change in size between the close in shot with the 24mm lens and the more distant shots with the 24mm and 135mm lenses? It is all back to relative distances. When in close (12”) with the 24mm lens the background is 46” away, about 4 times as far away as the glasses. When the camera is backed up to 85” the background is still 46” behind the glasses, but now the background is only about 1/3 farther away from the camera instead of 4 times as far away, making the background appear larger in the image.

The thing to remember here is that the closer you are to your subject, the smaller items behind the subject will appear. The farther you are from your subject, the larger the items behind your subject will appear. At first this might sound a bit backwards. Shouldn’t items in the photo get smaller as you move back away from them? Yes, they do. But foreground items get smaller much faster than background items as you move back. Then the usual option is to use a longer lens to "zoom" in on the subject and make it larger (along with the background). As you move back the scene compresses. This compression is NOT caused by the lens you use, but it is most noticeable in photos taken with longer lenses because they crop in on the scene and only show you a smaller area than you would see with a shorter lens. The area of the scene that is common to both photos will have the same compressed look. I will come back to this once more a bit later.

Next look at the depth of field in the images. The left image (24mm) has much more depth of field. Both glasses are reasonably in focus, though overall quality of the left image is lower because it has been enlarged so much. In the photo taken with the 135mm lens the front glass is reasonably sharp, but the back glass is beyond the depth of field at f/10 and has gone soft. Does that sound familiar? Go back to the 24mm full frame image above and see that the depth of field is about the same as in the 135mm full frame. Let that sink in. The depth of field with the 24mm and the 135mm lenses is the same when the size of the subject is the same in the frame. The focal length of the lens on its own is not a factor in determining the depth of field in a photograph. It is the magnification of the subject (focal length plus camera to subject distance) that determines the depth of field at any given aperture (f/10 in the case of these examples).

I hope that at this point you are starting to take on the mantra, “it is all about the distance.” Take a look at some of my lighting articles, where you will see a similar situation. The closer your light (camera) is to your subject, the darker (smaller) the background will be. The farther your light (camera) is from your subject, the lighter (larger) your background will be. We can even extend this to depth of field. The closer your camera is to your subject (with the same focal length lens), the less depth of field you have and the further you are from your subject (with the same focal length lens) the more depth of field you have. It is all about the distance.

OK. One more time through this to drive home what happens as you move in closer or move back farther from your subject. This time a landscape scene where we have a house with a mountain range in the distance behind it. Start with the camera 1000 feet away from the house and 5000 feet from the mountains. Now move in closer to be 500 feet from the house. The house is now twice as close and consequently twice as large as it was before. But the mountains in the background went from 5000 feet away to 4500 feet away, hardly any change at all. So while the house doubled in size, the mountains stayed the same size and now appear much smaller in relation to the house than they were originally at 1000 feet from the house.

We can turn that around if that makes it easier to understand. Same house and mountains. We set up the camera 500 feet in front of the house, which is 4500 feet in front of the mountain. Next we back up the camera to be 1000 feet from the house and 5000 feet from the mountain. The house became half the size it was before, but the mountains remained the same size, making the mountains appear that much larger in relation to the house.

Again notice that no mention was made of the focal length of the lens used for these photos. These changes in proportions are independent of the lens used. There is a famous quote from Robert Capa stating that, "if your photos are not good enough, you are not close enough."  There are multiple ways to get closer to your subjects. Your lenses are like a set of tools, you select the right size for the job. The camera to subject distance determines the perspective and then you chose the lens with the field of view that frames the scene the way you want. As a journalist you might be working on an intimate story and want to be in physically close to your subjects. A short/wide angle lens allows you to get in closer, making a stronger impact, giving a feeling that the viewer is right there next to the subject of the photograph. But at other times or in other situations you might want to back away from the subject to increase density, then a longer lens is the choice to bring things closer. It depends on the story you want to tell in your photograph. 

I hope that helps clear up some misunderstandings about lenses and perspective. But there is still more to it. How you view the photographs. But that's fodder for another article about viewing distances and enlargements. For now, please go out and make some meaningful photographs!

Thanks!
JohnC