Camera Lenses Testing - Vignetting
Part 4 - Vignetting check (uneven illumination, dark corners)
Vignetting is the darkening of the corners of an image due to the optical or physical design of a lens. It's a defect in the sense that it's something you don't want, but even a perfectly well constructed and designed lens can show vignetting, so it's not a defect in the sense that it's not something which is due to poor workmanship or design, or which can be corrected by a repair.
Vignetting comes from two sources. One is a physical obstruction effect due to the mechanical construction of the lens and the size of the optical elements. There are always compromises to be made in this area between lens size and cost and the degree of vignetting. If you make a lens big enough, with large enough diameter elements, you can in principle pretty much eliminate this source of vignetting.
A second source of natural vignetting comes from what is known as Cos^4 (Cosine to the 4th power) falloff. The basis of this is that the light from the lens has to travel further to get to the corners of the image than it does to get to the center, so it spreads out more and the brightness per unit area decreases. This is a particular problem with very wide angle lenses. It can be compensated for to some extent by lens design (e.g. retrofocus or telecentric designs), but it's very hard to eliminate.
There can also be vignetting caused by mechanical obstruction of the lens aperture by using a filter that is too thick on a wideangle lens, or by using a badly designed lens hood that is too small, but these aren't really associated with the lens itself.
You can test lenses for vignetting by shooting images of a perfectly evenly illuminated white or grey card. Getting even illumination isn't always as simple as it sounds since this is a pretty sensitive test. Sunlight is a good light source. Indoors using artifical light it's much harder to get perfectly even illumination.
The image of a grey card below was taken using a Canon EF-S 18-55/3.5-5.6 lens wide open at 55mm. You can see some vignetting present since the corners are slightly darker than the center. This is normal and to be expected. If you want to measure the amount of vignetting, it's quite easy to do. Just take a series of shots at 0. -1/3, -2/3, -1, -1 1/3, -1 2/3 and -2 exposure levels. Then measure the brightness of the center and corners of the shot taken with no exposure compensation. It may be easier to do this if you convert the image from color to B&W since then you'll just get a single number. Most image editors allow you to see the brightness value of the pixels.
Let's say the center of the 0 EC image measures at a brightness level of 100 and the corners measure 62. Let's say the center of the -1/3 EC image measures 89, the -2/3 image measures 73 and the -1 stop image measures 62. That means that in your 0 EC image, the corners are 1 stop darker than the center. In the image above, the corners are about 2/3 stop darker than the center, which is reasonable for an f5.6 lens shot wide open with the lens (EF-S) matched to the format size (APS-C). Vignetting due to limited element size and lens construction is reduced by stopping down, though cos^4 falloff isn't.
For a fast lens (e.g f2 or faster) it would not be unusual to see 2-3 stops of corner vignetting. Stopped down to f5.6 you might expect to see 0.5 to 1 stop. These numbers apply to lenses designed for the format in use, i.e. an EF series lens on a full frame camera or and EF-S series lens on an APS-C format camera. If you use a full frame lens on an APS-C camera, you'll see much less vignetting than you would with the same lens on a full frame camera. Vignetting will change with aperture, the focal length setting of a zoom lens, and it may also change slightly with focus.
If you want to see the vignetting pattern more clearly you can stretch the histogram so that the brightest areas are moved to a level of 255 and the darkest areas are moved to a level of zero. The result is shown below:
There is a little more vignetting at the top than at the bottom, but it's not too bad. I'd say that this lens is in acceptable alignment.
There's nothing you can do about intrinsic lens vignetting, it comes from the lens design. The only thing to watch for is that it's more or less symmetrical. If you see a LOT more vignetting in one corner of the image than the others, it's probably an indication of misalignment of the lens optics. If you see a significant difference in vignetting between the corners of an image, do the test again using a different target under different lighting and see if you get the same result. After stretching the histogram, you're seeing very small differences in illumination magnified. If you test target wasn't perfectly evenly illuminated, it will affect the observed intensity distribution in the image.
