One form of photographic perspective distortion is perspective convergence or keystoning. This form of distortion is very common across architectural, street and travel photography. It’s most often seen when tall buildings ‘fall’ or ‘lean’ within a picture. This distortion has become so common that most people have stopped noticing it within their pictures or just think it has something to do with the focal length of their lens.
For a number of photographers, vertical convergence and distortion are unpleasant and unwanted because they don’t conform to Linear Perspective cues where vertical lines remain vertical. This makes the scene look different to how we think it should look.
Why Distortion Happens
It’s actually pretty simple. As mentioned, modern lenses are designed to show straight lines. However, this only works when we the camera is pointed straight at or “in line”, with the object that is being photographed. This is because the distance between the camera and object remains the same. As soon as the camera is titled then the distance changes. For example, if I’m trying to capture a tall building straight on with my camera I can only get the bottom section – vertical lines are straight but I’m missing the rest of the building. Now if I angle the camera higher I can get the whole building into the frame but now the vertical lines, the sides of the building, are converging. This happens because the top of the building is now further away from the lens than the bottom – just like a road receding into the distance.
So how do you take a photograph of a tall building without this happening? There are a few ways. The most obvious is by changing your viewpoint. You need to get higher to shoot more of the building front on, with the mid-point being the best place to capture the maximum amount of the structure. You could shoot out the window of a neighboring building, or if there is nothing around you could spend some money on a crane!
If you can’t physically get higher, then you can change your viewpoint in two other ways. One is with the help of shift lenses the other is through perspective correction software.
The ability to shift the position of the lens has been around since the beginning of photography. The original view cameras (created in the 1840s) used a bellows system which let photographers change the position (shift) the lens in relation to the film.
The shifting mechanism acted like a mini elevator, moving the photographer’s viewpoint higher and lower. Historically, shifting in order to correct perspective convergence was seen as an important part of the capture process, just like focus, shutter speed and aperture.
Shifting was possible because the lens captured a much larger (circular) field of view than the film. The same way that modern shift lenses work. Modern shift lenses are commonly used by architecture photographers to limit vertical convergence on large buildings. However, these lenses can only be used on DSLR or medium format cameras and can also be very expensive (upwards of $1,000).
In recent years, digital technology has been able to replicate this shifting process, artificially changing the photographer’s viewpoint. Software like Photoshop, Lightroom or Gimp reworks the pixels in an image in order to straighten vertical lines in turn reducing distortion. Most of these programs allow you force the image back into a ‘correct’ position through either sliders or by selecting and dragging a corner. One of the downsides of the software is that significant adjustments require resampling of the image, often reducing sharpness in parts of the photograph.