Misconceptions in photographic optics


Here is my top ten of misunderstandings, misinformation and misnomers in relation to the photographic lens. Some of the statements are encountered so often in photo magazines, popular text books and internet resources that they have become routine knowledge of many photographers. I used to believe in some of the below statements just because they are encountered so frequently. Empirical findings, serious literature in the field of photographic optics, or just going through the physics and mathematics should eventually make a disbeliever of the believer.

1. Long-focus lenses come with a narrow depth of field.

Whether a telephoto lens offers less DOF than a wideangle lens depends on the conditions of the comparison. However, it is a common mistake to judge the depth of field by the degree of background blur. Yes, a telephoto lens may give rise to a large (absolute) blur of the background, but this is not a matter of DOF. From the definition of the depth of field it is clear that DOF should should be judged by the in-focus parts of the image, now matter the amount of blur in the out-of-focus parts. (more)

2. The rear DOF is twice the front DOF.

This is just a very silly statement. In the macro regime the DOF distribution (front:rear) is 1:1, in the portrait regime it is still close to 1:1, and a landscape captured at hyperfocus will have a distribution of 1:infinity. Somewhere in between there will be a scenario that gives 1:2, but you will have to look for it. Perhaps the misconception is due to the observation that a 1:2 distribution represents some kind of average over many photographic situations. So much is clear, every scenario should be considered individually. (more)

3. The depth of field only depends on the image magnification and the F-number.

For a given sharpness criterion (circle of confusion), the depth of field quite generally depends on the focal length, the F-number, and the object distance. The statement in white is a very good approximation under certain conditions, but it is no universal truth. The approximation becomes better the smaller the object distance, the larger the aperture (smaller F-number) and the smaller the COC. In particular when the object distance is not small compared with the hyperfocal distance, a short-focus lens will offer more DOF than a long-focus lens—at the same F-number and image magnification. At a higher level of sophistication we may consider asymmetrical lenses and introduce the pupil magnification into the theory of DOF, to find that the claim is once again incorrect— especially in the macro regime. A telephoto lens used at unit magnification gives more depth of field than a retrofocus wideangle lens at the same magnification and F-number. (more)

4. Tele lenses have a flat perspective and wideangle lenses a steep perspective.

A lens does not offer a perspective. The perspective is determined solely by the viewpoint adopted by the photographer, not by the lens choice. In optical terms the viewpoint corresponds to the center of the lens entrance pupil. For a fixed camera position, different lenses may yield small changes in perspective as their entrance pupils will be at slightly different locations, an object-space telecentric lens being an extreme case. Photographers may deliberately look for a 'dramatic wideangle effect' by choosing a close viewpoint to exaggerate the size of foreground details. Still this effect is strictly due to the deliberate close viewpoint. Much of the confusion is also due to incorrect viewing of the print or slide projection. A wideangle photograph must be viewed from close, a telephoto photograph from a larger distance. In practice people adopt a fixed viewing distance, more or less appropriate for normal lenses, which will lead to apparent perspective effects for photographs taken with short or long-focus lenses. (more)

Concerning rectilinear lenses. 5. Wideangle lenses give rise to converging verticals.

Converging verticals only occur when the camera is tilted. In practice, when a wideangle lens is used the photographer finds himself relatively close to a building and tilts the camera more than he would if he were to photograph the same building from a distance with a normal or tele lens. In all cases the convergence is not due to the lens, but to the tilt. (more)

6. With panoramic stitching, where the camera and lens are rotated as a unit, the rotation axis should pass through the lens nodal point.

Many sources that write this don't specify whether it is the front or rear nodal point, but either way it is incorrect. To preserve the perspective during rotation, which prevents a displacement of nearby objects relative to the background (parallax), the axis of rotation should pass through the entrance pupil, which is the center of perspective of a lens. No doubt instructions to experimentally find the position of the rotation axis will lead to the correct point, it's just that the designation nodal point is wrong. The good news is that, unlike a nodal point, the position of the entrance pupil can be approximately found by visual inspection. When you look into a lens from the front, the entrance pupil is the image of the diaphragm opening. (more)

7. When a lens is put on a camera with a smaller sensor, the focal length increases.

With the advent of digital cameras with miniature sensor chips the "increase of the focal length" seems to have become a standard phrase in photo magazines and internet resources. However, the focal length is a property intrinsic to the lens and does not depend in the slightest way on the dimensions of the film frame or CCD. The relevant quantity that is influenced by the sensor size is the field of view, not the focal length.

8. In a lens distortion graph, negative values correspond to barrel distortion and positive values to pincushion distortion.

Strictly speaking barrel distortion corresponds to a negative slope of a distortion(%) curve and pincushion distortion to a positive slope. When there is a local minimum in the curve, which happens for certain wideangle designs, there will be a transition from barrel to pincushion distortion when going from the image center to the periphery—even though the distortion has negative values everywhere. (more)

9. A high-quality UV filter will never degrade the image quality.

Under normal circumstances a UV (or other) filter will be safe to use, with no perceptible image degradation, but that does not take away the fact that the filter does add two reflecting surfaces to the optical system. No matter how small the reflections, they are not zero and flare may appear when the contrast in the photographed scene is high enough. Owing to the flat filter faces, the flare most notably presents itself as ghost images of highlights such as the sun in daylight photography or street lights in nighttime photography. (more)

10. When a lens is used at a small aperture, only a small part of the front lens is used.

When a lens is set to a small aperture, each image point utilizes a small part of the front element. That part has an area that approximately equals that of the lens entrance pupil. However, different image points depend on different parts of the front lens. The image center uses a small part around the optical axis, other areas of the image use portions that are displaced from the axis. In total a substantial area of the front lens is needed to illuminate the entire image. It is true that the very margins of the front element are often sidelined at small apertures, the more so as the focal length increases. (more)

Available light is any damn light that is available! — W. Eugene Smith