Photographically, vignetting means the darkening of the corners relative to the centre of the image. At low levels it is not noticeable to the average person, and not objectionable photographically, and at its worst it destroys a picture.

All 'normal' photographic lenses vignette. The only exception that comes to mind at this time is the 10mm OP fisheye from Nikon, which might still be available on special order. This lens was designed specifically to produce no light fall-off over the whole image field. See the topic on fisheye lenses.

A simple lens obeys an optical law called the cos-4th law (cosine of the angle of incidence off the lens axis to the 4th power). The cos-4th thing comes from a number of factors, including that the light gets spread out over a larger area, and that the corners of the image are further away from the center of the lens than the on-axis part of the image. This means that for a lens with an angle of view of 20 degrees, where the maximum off-axis angle of incidence is 10 degrees, the corners receive cos(10 degrees) to the 4th power as much light as the center. Cos (10 degrees is 0.9848, so that to the 4th power is 0.9406. This means the corners get 94 percent as much light as the center, a generally imperceptible difference. A 20 degree angle of view for a 35mm camera equates to a 123mm lens. A 50mm lens has a 47 degree angle of view, and the very corners are 23.4 degrees off axis. Cos 23.4 degrees is 0.9178, and that to the fourth is 0.71. This is about 1/2 a stop worth of light, and is barely noticeable, if one looks for it. For a 28mm lens, with a 75 degree angle of view, cos4th comes to 0.39. So now the corners only get 39 percent of the light that the center gets. For a 20mm lens this goes down to 21 percent, and for a 15 this goes to 10.5 percent. Obviously the last is unworkable; this means the corners are more than 3 stops down from the center. Something has to be done.

Today few lenses are simple; usually the only simple lenses we use for photography are cheapy close-up lenses, which usually give cheapy results, and our glasses and contact lenses and eye lenses that we judge our pictures through before, during (on direct view cameras) and after taking the picture. How we manage to make do with such a poor optical system for our own vision is a topic for another discussion.

Cameras use complex optical systems. For long focal lengths we use 'telephoto' design lenses, and for wideangles we use 'retrofocus' design lenses. The result of these designs is that we screw up the cos-4th law. Due to the mirror in SLR cameras, and for the need for some space between the back of the lens and the shutter curtain in TTL metering rangefinder cameras, the part of the lens closest to the film has to get mo ved further out. Due to these shenanigans, the angle of incidence of light from the object to the lens is not the same as the angle of exit (through the center of the lens) as it would be in a 'simple' lens. The angle of exit is really the one that finally determines how harsh the cos-4th law is, and for retrofocus lenses it is not nearly as harsh. So, for SLR's, the whole issue becomes a moot point, as all lenses of less than 40mm are retrofocus designs, and the shorter the focal length, the more retrofocus they are. End of (severe) problem. The light falloff stays within acceptable bounds. Unfortunately (here we go again), other problems, like distortion, arise. But that's life.

Very wide angle lenses that are not retrofocus in design, which are mainly found in view-camera lenses, get fixed in a different manner. When the angle of view is very great, and is used to its fullest, the light falloff due to cos-4th would be objectionable, so a filter is put over the lens which is a graduated neutral density, being clear at the edges and darker at the center, so that the illumination on the film turns out even. In 35mm, the Hologon lens for the Contax uses a Center-filter.

Telephoto lenses are designed as such so that an 800mm lens can be constructed so that th front of it is less than 600mm from the film, at infinity focus, instead of 850mm or so. Extreme telephoto designs have allowed one 300mm lens to be only about 100mm long. Makes lens handling (and carrying) extremely convenient. Here, the result for the cos-4th law is that the exit angle is much larger than the entrance angle, and there is more cos-4th falloff than one would expect with a lens as long as a 300, or 600, or whatever. In any case, this doesn't matter, as the cos-4th falloff is just not a concern where the exit angle is less than that of a 50mm lens. Again, with the telephoto design, other problems such as distortion crop up.

All the above deals with the type of light falloff that may be termed 'optical vignetting'. That is, this falloff occurs because of intrinsic optical construction and design.

Then there is 'physical vignetting'. This is the reduction of light in the corners due to the lens mount, or filter or lens hood or finger getting in the way and not allowing as much light in the corners of the image compared with the center as the optical laws, including the cos-4th law, would have us believe. Again, almost every photographic lens exhibits physical vignetting. This is due to a number of factors. Firstly, it is extremely difficult to design a complex lens with a number of elements so that no vignetting takes place. The front elements and the rearmost elements would generally have to be huge, especially for fast wideangles. This in turn would make the lens much heavier and more expensive. Secondly, some aberrations can be controlled by not allowing some of the edge rays to exist, in effect, stopping down the lens at the edges. Again, this benefits the fasters, wider lenses particularly.

In any event, if you mount your lens of concern on your camera, open the camera back and lock you shutter open with a cable release or whatever, hold it up to a brightly lit surface such as a window, you can look through the front of the lens and see the film aperture through it. Look for the corners, and stop your lens down. If you have a 50/1.4 lens on your camera, you probably have to stop the lens down 4 stops to f/5.6 or so before it is the aperture blades that are cutting off the light coming from all directions, rather than part of the lens barrel. A 105 macro lens with an aperture of f/4 might need to be stopped down only one stop, or also to f/5.6 before the same condition applies. Only at this aperture or smaller does physical vignetting cease, and the only vignetting that still applies is optical vignetting. (There are actually some very minor other losses such as glass and surface losses, but they can generally be ignored.)

What does this mean? Well, it means that very fast lenses, with apertures like f/1, f/1.2 etc., often image a central area with an effective f stop as you would believe from the advertised number, but the corners often are 2 to 3 stops darker when using the lens wide open. This is only a defect if you choose to believe that, and depending on the type of pictures you take. Since most images have the important items in the central region, especially in very low light situations, you generally don't suffer. You just have to know your lens, and use it accordingly. There is more of a problem with some zooms, especially wide range zooms which are inexpensive. Here you might find that physical vignetting is not eliminated until you stop down to f/11, and I have seen a few where it never goes away, and is very severe at wide apertures. It is a problem with these lenses because they are often used outdoors in normal bright light shooting situations, and then the sky gets these dark areas in the corners. This can be annoying.

Please note that physical vignetting is affected by the focussed distance of the lens, and by the focal length of zooms. If you want to check your lens, check it set at various distances and at various focal lengths. Note also that you can check this from the back of the camera as well as from the front, and a complete test includes both. Some new, smart electronic cameras don't like having the back open, or no film in the camera before a time exposure is started. For these, the easiest method to do these tests is to insert an old outdated film that you don't want to shoot with anyway, wind it to 1, start the time exposure and lock it, and then open the camera back and take out the cassette and leave the bit of film hanging from the take-up spool. After doing your tests, you might have to pull out the film, close the back and start the rewind, even though there is no film in the camera, just to get the camera to think everything is as it should be. Smart camera mostly, but stupid in some ways. Note that checking for vignetting by looking through an SLR's lens from the front and trying to see the corners of the viewfinder is neither very easy nor accurate. The finders generally don't show 100 percent of the image, so you might get vignetting on your negs in the extreme corners that your viewfinder doesn't care about.

Another way that vignetting can crop up is by doing it yourself. Put the wrong lenshood on, screw 5 filters onto your wideangle, or curl your finger in front of your lens, and you have your own, personal vignetting. Testing the view through the lens as described above will quickly let you know how many filters are too many, or whether you can stack a lenshood and filter on a particular lens. The finger has to be dealt with more severely. If, when looking through the lens, you can't see the corners of the frame due to the lenshood or filters, the corners of your pictures will obviously stay as black as the film and printing allow. If the light to the corners is obsructed partly by the filters or lenshood, you will have some vignetting. Using this test method on wideangles will show you whether you can use a regular polarizer safely, or whether you have to invest in a special wideangle polarizer such as Nikon's (which are all wideangle polarizers), or special ones from B&W. Note that this is a more definite test of vignetting than taking picures, a lot quicker, and a lot cheaper. Taking pictures tells you whether the vignetting you have is bothersome to you, but does not tell you whether or not you have vignetting, unless you can examine your negs with a densitometer.

Finally, be aware that vignetting will also be introduced during the printing stage, and with negatives, the vignetting of the printing equipment counteracts the vignetting of the taking equipment. Balance is unlikely, though, and printing equipment usually is limited to optical vignetting of the lens and light condensing system, which is generally quite low. Projection systems produce more vignetting. Against all this come our aesthetic concepts, which, in general, much prefer corners which are darker than the center, rather than the other way around. So in printing we often artificially introduce darkening of the corners.

Therefore, in the end, we come again to the main concept of photography. If you like the picture that's produced, it's probably all OK. But better than that is if you are in control of the whole process, know what you want to produce, and can achieve just the amount of vignetting that you desire.