How To's

How to Test a Lens

Published November 26, 2010

If you read my articles very often, you may have realized I’m not one to let a lack of knowledge prevent me from writing on a topic. First-hand knowledge I mean. I write a lot of articles that I may spend 6 weeks researching before I feel knowledgeable enough to start writing. But this is one topic that I know well: its what I do for a living. About 8,000 new lenses have come through Lensrentals over the years, each one has to be tested, inspected, and accepted before it goes on the shelf. And every day 150 to 400 lenses return from rental and go through the same process before they ship out again.

I don’t mean to suggest we’re lens reviewers, dissecting and measuring every possible aspect and index of a lens. That takes days, tons of equipment, and a mindset that I don’t have: full-blown obsessive-compulsive nitpicking disorder. I’m glad there are people who do that. I’ll never be one of them.

(Don’t get me started on the whole are-lens-reviews-worthwhile thing. I find it very similar to measuring the height of one third-grader and then saying “all third graders are 4 feet, 7 inches tall and weigh 52 pounds.” Someday someone will start evaluating 20 copies of a given lens, include the copy-to-copy variation in their reviews, and in a short time will own the lens review business. But that’s another topic.)

Why Bother?

When we get that shiny new box home, we expect it to be perfect. And it probably will be. But after opening some 8,000 shiny new lens boxes I can assure not all of them are. Whether its quality control at the factory or getting knocked around in shipping our experience is about 2% of new lenses need to be exchanged. It varies by brand and it varies by lens complexity (an autofocus zoom with image stabilization is more likely to have problems than a manual focus prime). But every lens needs at least a basic checkout when you first get it home. Used lenses, of course, require it even more.

There are several levels of testing and examination one can do. The simplest is go take some pictures and see if they look okay. This works well, and has the advantage of getting to take pictures, but has some shortcomings, too. First, you may not notice a problem until it is too late to return or exchange the lens. Second, if something is wrong, simple photographs don’t usually determine what kind of problem the lens has. This lens is soft is a good article title, but not a good description of a problem.

(BTW – if you send a lens in to factory repair with “This lens is soft” as the only description of the problem, chances are extremely high that it won’t be fixed. Trust me on this. We have 20 lenses a week go in to factory service. We’ve learned.)

At the other end of the spectrum is Extreme Tester Guy (and you are a guy, there are no Extreme Tester females) who will spend 16 hours testing a lens in every way possible including the laser collimator he keeps in the garage. He’ll send back 6 copies to get the perfect lens which he won’t take pictures with because a new lens has just been released and he’s too busy getting a perfect copy of that one. This article will just bore Extreme Tester Guy, he really wants to be a lens reviewer.

But without taking it to extremes, a little basic testing does several things. It will let us know the lens we bought is what we expect it to be, that its assembled properly, functions as it should, and provides a reasonable match for our camera (if you don’t understand what I mean by that last statement, you might want to read the two “This Lens is Soft” articles).

Testing will also show us some important characteristics of the lens that weren’t listed on the B&H blurb or the manufacturer’s website. The marketers are quick to tell you it has +Supermicrosecretformula coating_, 4 aspherical, and 2 ultra-beyond-low reverse distortion elements. BUT do they ever tell you if it focuses using internal, rear or front element groups? If it is parfocal (if you zoom in on the subject is the lens still in focus, or does it need to be refocused)? Is the plane of focus flat or curved? How soft are the corners wide open and at what aperture to they get sharp? Does it have barrel or pincushion distortion? These are things a bit of testing will show us about the lens. And they are good things to know that will let us use the lens more effectively.

129106216-78mwtago
The manufacturer goes to great lengths to show the aspherical and low dispersion elements, but don’t show you which ones zoom or focus.

Step 1: Examining the Lens

But before we start optical testing, we need to do some basic hands-on, touchy-feely examining. Some of this is to make sure the lens is in good shape, some to make sure we know how it is going to work. More new lenses fail my touch-and-feel tests than actually fail optical testing.

The Barrel

If the lens is poorly assembled now, its not going to be better after a year of use. The lens barrel is made of several different cylinders assembled together. There’s usually a joint under the rubber zoom and focusing rings and often another near the end of the lens just under the filter ring. Gently move, rock or extend each segment of the lens and ask yourself the following questions:

  • Are any sections loose from the next?
  • Can you rock the filter ring?
  • Do all the screw holes (including the ones at the lens mount) have screws in them?
  • If there’s an extending internal barrel does it slide back and forth easily when you zoom and focus?
  • Does the front ring accept a filter easily (more important to check in a used lens, but we’ve seen a few new ones with defective threads).

While we’re here, gently shake the lens and make sure there’s not a loose screw rattling around inside. Yes it happens. (If the lens has an image stabilizing/vibration reducing mechanism, this may rattle a bit when you shake the barrel. Don’t freak.)

The Focusing Mechanism

This isn’t for accuracy. That comes later. Set the lens on manual focus and run the ring back and forth a few times (even if you’ll never use manual focus).

  • Does it move smoothly with no catches or gritty sensation? This is also the time to note what part of the lens moves when focusing: front element, rear element, or internal group (neither the front or rear elements move when focusing).
  • If it’s a zoom lens, check the focusing at both extremes of the zoom. On the Canon 70-200 f2.8 IS II lens for example, the rear element doesn’t move during focusing when the lens is at 70mm, but moves a great deal with the lens at 200mm.
  • If the front element moves, does it also rotate (makes it very hard to use polarizing filters) or is it nonrotating?
  • Does the distance scale rotate properly when you move the focus ring?
  • Is there a clutch for manual focus, a switch, or is the manual focus ring always active?

Then switch to autofocus, mount the lens to your camera, and repeat.

  • Does it autofocus smoothly?
  • Does the focus motor sound like your other lenses?
  • Finally, autofocus on something close and then immediately on something at infinity. Then reverse the process. How long did it take?

A slow autofocus system doesn’t matter when you’re shooting Macros or portraits, but it can make a lens useless for sports or street shooting.

(For those of you with too much time on your hands, if you want to quantify focus speed do the near-far autofocus test right next to a microphone plugged into your computer. You can open file in an audio editing program and see exactly how long the lens motor was buzzing.)

Zoom Mechanism

Of course you want to test to see if the zoom mechanism is smooth, both going out and coming back. If the lens zooms by extending the front element (or extending an internal barrel) make sure there are no sharp, sudden catches or areas of high resistance. If there is an extending barrel, make sure the internal barrel is clean with no scratches that might indicate rubbing when zooming. Also check that the barrel isn’t loose when extended, it shouldn’t rock back and forth.

If the front element doesn’t extend, look at the rear element during zooming. If the rear element is part of the zoom mechanism it will move into and out of the barrel during zooming. Why is that important? Rear element zooms tend to have shorter than advertised focal lengths when focusing on near subjects. It may be 300mm at infinity, but only 240mm when focused 9 feet away.

Glass and internals

Look at (not through) the front and rear elements for scratches, coating defects, etc. Front element flaws are of minimal importance, but anything wrong with the rear element (scratch, coating flaw) may have a great impact. Then look through the lens while moving the focus and zoom rings. A bit of internal dust, even straight from the factory, isn’t unusual and is of no consequence. Things like a loose screw, piece of cloth, scrap of metal, or a broken internal element (yes, we’ve seen all of those in brand new lenses) probably are.

Mounting Ring

Mount and dismount the lens several times on the cameras and ask yourself the following questions:

  • Does it go on easily but firmly?
  • When seated is it firm with no looseness?
  • Does the locking pin catch properly and then release easily when you push the camera’s dismount button?

Principles of Optical Testing

There are several principles to testing that are all too often ignored, but each is true (and obvious if you think about it a bit).

  1. If the fixed elements or mount of a lens are out of spec the lens will be bad at all parts of its zoom range and when focusing both near and far distances.
  2. With some lens designs an element 1mm too far forward or back will cause the lens to never focus sharply.
  3. An element tilted as little as 3 degrees from the plane of the lens can cause side-to-side or top-to-bottom issues sharpness differences.
  4. An element just a bit off center can cause chromatic aberration, astigmatism, coma and edge softness.
    Modern lenses may have 15 or 17 elements, each of which could possibly be out of spec.
  5. If the zoom elements are out of spec the lens may be excellent in one portion of its zoom range and horrid at another focal length. (But every zoom is usually slightly better at one end of its range than the other. I’m not talking about that, I’m talking about really bad in one part of the zoom range.)
  6. If the focusing elements are out of spec the lens may work fine on close objects but not at infinity, or vice versa.
  7. If the autofocus electronics and algorithms aren’t accurate or camera-lens communication isn’t good the lens may front or back focus.

The primary point of this is that you must test zoom lenses at least at both ends of the zoom range (and probably in the middle too), and test all lenses focusing on both close and distant targets. And a key point in testing your lens is comparing each quadrant to the other quadrants, not just “is it sharp”. In fact, center sharpness is probably the least important thing to consider when testing a lens.

Step 2: Front- and Back-Focus Testing

We start by testing autofocus accuracy because if the lens is not focusing accurately the other tests are meaningless. Even a manual focus lens should be tested for focus accuracy on an SLR – the viewfinder and sensor may not be exactly calibrated in the camera, and if the lens has “focus confirmation” electronics, that is, in our experience, more likely to be inaccurate than standard autofocus.

Set Up for Depth of Field and Focus Testing

This requires regularly marked surface set up at angles to a focusing object. We use a LensAlign Pro because it’s quick and very accurate, but its not necessary. The marked surface can be a ruler or yardstick set at angles to a focusing target for near testing and a fence (board, chain link, brick, doesn’t matter) with some obvious focusing target for the far testing.

A lot of people use computer generated moir? targets for autofocus microadjustment, and they are fine for that, but they don’t give all the information we need for lens testing so I don’t recommend them for this purpose.

Focus accuracy

Focus accuracy testing is done with a single focus point selected (center point unless you have some very odd reason to do otherwise). We begin with the lens at widest aperture and then repeat with the aperture stopped down a bit.

The process is quite simple: we autofocus the center point at the test target and then check to see where the sharpest focus actually is on the ruler or marked diagonal line. They should be very close. There are several things that need to be taken into consideration:

  • Every lens can focus differently at near and far distances. Test at several distances.
  • A zoom lens usually focuses slightly differently at the minimum and maximum zoom ranges (and sometimes in between too).
    130132661-6lz9czis
    This 24-70 f2.8 focuses perfectly at 24mm (left) but backfocuses slightly at 70mm (right).
  • Some wide aperture prime lenses exhibit focus shift: they may focus accurately wide open, but then backfocus or frontfocus slightly as the lens is stopped down.
    130132659-ctnmool0
    A wide aperture lens with no focus shift: note that the lens is a bit backfocused wide open (f1.2) and remains that way as the aperture is decreased.
  • Autofocus accuracy is an interaction between the camera (every copy of which is slightly different) and the lens (every copy of which is slightly different). Get over the fact that your camera is perfect because it is perfect with some other lens. Below are 5 focus checks done with the same Canon 85 f1.2 on 5 different 5D Mk II bodies, for example (look at the number 4s to see that no two are exactly the same).

130132658-p0hefkgx
The same 85mm f1.2 lens tested on 5 different copies of the same camera body. Notice there are slight differences in front / backfocus with each body.

The summary of all this is we first have to check the lens at widest aperture and at least at both ends of the zoom scale and several different distances from the test target.

If our camera has autofocus microadjustment we then adjust it for the best compromise, getting it perfect where we want to use the lens. If I’m going to use my 24-70 zoom usually near 70mm for portrait shots 15 to 20 feet away, I’ll zero the adjustment there. If I’ll be shooting my 300mm prime lens on distant wildlife, I’ll zero it there.

If the lens is a wide aperture (f1.8 or lower) I’ll also do some sequential shots at various apertures to check for focus shift. When a lens has focus shift, the point of focus changes slightly as the aperture is stopped down from wide open. Eventually the larger depth of field from reducing aperture overcomes the focus shift, almost always by f2.2 or f2.8. This isn’t something I can correct, but its good to know that I can shoot my Sigma 50mm f1.4 either wide open, or at f2.2 or higher, but in between because it won’t focus accurately.

One other thing to check with a wide aperture lens is spherochromatism (also called secondary longitudinal chromatic aberration or longitudinal color shift). Spherochromatism occurs in many wide aperture lenses because while the lens is corrected for for chromatic aberration in the in-focus area, the out-of-focus areas in front of and behind the plane of focus can’t be corrected and will have magenta or green tints. There’s nothing to be done about it, but its good to know how severe it is. In problem lenses you can avoid large white areas if the foreground or background when framing your shot.

129321940-p7nsqvw4
Example of a wide aperture lens with significant spherochromatism. Note the purple tinge in the forground and green tinge in the background.

Once we have the lens focusing as accurately as we can, we move on to Flat Field testing. Of course there is an occasional lens that needs so much autofocus adjustment that you don’t have to test further: if you have to set your camera to +/- 15 or more, it may be better to exchange for another lens. And of course if your camera doesn’t have microadjustment you have to decide what is, or is not, acceptable.

Flat Field Testing

This is the part of testing that will give you the most information about your lens, so its important to set it up correctly. The key to flat-field testing is that everything is flat (duh) and lined up accurately. How flat and accurate depends on depth of field of the lens, which changes with the aperture and the distance from the target. If you’re testing an f/1.4 prime lens at near distance it has to be very accurate. An f/5.6 lens at far distance, not quite so much.

Flat Field Set Up

A fairly large building (windows are great for detail) makes a superb far-distance target. You can assume the building side is perfectly vertical so using a bubble level to set the camera on tripod to vertical is very straightforward.

Any wall can be used for near-distance testing. I recommend downloading and printing some USAF 1951 test charts (they can be found online, including here ) You want one in the center of your image and at least one in each corner. A few pieces of cut up newspaper will work too.

If you want to be more complex, I recommend the simple Resolving Power chart from Edmund Optics that you can buy for less than $20. It may not be worthwhile to test one lens a year, but is certainly not a bad investment for a camera club. There are lots of other test targets available online and most do the job just fine.

130132755-pqicid6j
Edmund Optics Resolving Power Test Chart

Setting up the Near Distance Flat Field

  1. Assuming your wall is vertical and straight make sure your test targets are placed smoothly and tightly on the wall. If they’re lumpy and bumpy, they aren’t accurate.
  2. Place your camera on tripod so the center of the lens is level with the center of the test target. A laser pointer held along the lens barrel can be used for targeting and to confirm centering. A bubble level (built into the tripod or a more accurate hardware level) confirms the lens is lined up vertically.
  3. Make sure the camera is not pointed at an angle right or left: this could make one side of the target out of focus compared to the other. The simplest and fairly accurate way is to measure an appropriate distance from the wall (say 15 feet, but whatever is appropriate for the lens focal length) from either side of the test target. Then draw a chalk line between those two points. That line should be parallel to the wall. You can line up the camera with that line using a yardstick and drop string if you want to be a bit obsessive.

130132755-pqicid6j

But let’s be honest: most of us are going to line up the lens so its centered on the tripod, use the bubble level on our tripod and eyeball judge the rest of it. And that will work just fine 95% of the time. If your results look bad, then you can always go back and measure carefully to make sure it’s the lens with the problem, not your setup.

Flat Field Optical Testing

Most people want to jump in and see how much resolution the lens provides at its center point, but unless the lens is horribly soft that isn’t going to tell us much. Photographing a centered resolution chart gives you almost no information about the lens.

  • Center resolution is pretty good and fairly comparable on most lenses. A more expensive lens usually gets you more aperture or better resolution away from center.
  • A wide angle lens won’t resolve as well as a standard or telephoto.
  • A wide aperture prime doesn’t necessarily resolve better than a good zoom, although most do, at least when you close the aperture a stop or so.
  • By f5.6 most lenses resolve well, by f8 some actually are losing resolution to diffraction(depends on your camera) and by f11 most are losing resolution.
    But I realize everyone wants to know how well their new lens resolves. The AF1951 target you can find online and print for free, or whatever you are shooting, should give you a decent idea of how well the lens resolves.

130137797-02o4jepn
AF1951 Test Chart

If you want more detail, Norman Koren has posted some downloadable test charts along with detailed instructions for using them. Or you can buy detailed (and pricey) mutiple test charts like the Edmund 58940.

130132660-wvwqiggw
Edmund Optics 58940 Test Chart

The only real necessity is to have an idea what you expect from the other lenses you shoot and compare your new lens to them. Be reasonable, though: if you think your new f1.4 prime is going to be as sharp at f1.4 as your old lens is at f5.6 you may be rudely surprised. And that new f2.8 zoom may not be any sharper in the center than your old variable aperture zoom, but it will probably toast the old one on the edges and corners. If all you are interested in is how sharp the center of the lens is then just buy Macro lenses and don’t worry about it anymore.

The bottom line, though, is checking center sharpness doesn’t help too much with testing a lens. What really helps is comparing the quadrants of the image to each other. If they elements are all aligned and centered properly the different quadrants should be similar. If the corners are soft wide open, they should all be equally soft and begin softening a similar distance from center in every direction. Similarly any chromatic aberration should be equal on all sides, not worse on one than the other (assuming your target is equally lighted).

If you look at the 4 corner targets from the lens test below, its obvious that the right hand targets are softer than the left. If you look more carefully you’ll see that the right upper is even softer than the right lower and has a bit more chromatic aberration. This is the kind of thing that will show up if the lens has an element out of center or at a slight angle. The center sharpness on the lens, though, was excellent.

130132758-v6jul4yc
The four corner quadrants of a misassembled lens. Note how much softer the upper right quadrant is than the lower left.

When you see an image like this you want to make sure the effect isn’t caused by the camera not being square to the test target, which could cause some of this effect (probably not the CA). You might try shooting the camera in portrait mode: if its alignment the quadrant in the lower left will now be in the lower right and less sharp, the upper right quadrant will now be in the upper left and more sharp. If the image looks the same in both portrait and landscape mode, its probably the lens.

Once we’ve determined that all 4 quadrants look equivalent on our fist test shots its time to run the lens through a battery of fairly quick tests:

  • Near distance widest aperture. Zooms should be checked at their widest, middle, and longest focal length.
  • If your zoom is variable aperture, be sure to mark down at what distance the aperture changes. You know your f4-5.6 zoom is f4 at the wide end and f5.6 at the long end, but do you know when it changes? Its good to know that shooting at 160mm instead of 170mm will give you f4.
  • At each location the lens should be checked with aperture wide open, and stopped down in incremental (1/2 or 1 stop) steps to f8. You can assume the lens has reached maximal sharpness by f8. Some may sharpen further in the corners at f11, but they’ll usually start sacrificing center sharpness there.
  • The same tests should be repeated at far distance.
  • Testing distances will vary depending on the lens, but I find it logical to test at the distances I’m likely to shoot at. I don’t shoot the 16-35 f2.8 at 4 feet away so near testing may be 10 or 12 feet for it.
  • Unless you’re going to really shoot at the minimum focusing distance, don’t test there. The lens will usually be sharper a few feet away.

Vignetting

Is really part of flat field testing but can be more obvious when the target is a smooth, evenly lit light colored wall. It doesn’t require critical line up like the other parts of flat field testing, and you simply take exposures at various parts of the zoom range and several apertures. Any vignetting should be even in all four quadrants, of course. It is often more severe at one end of the zoom range than the other, and usually less severe when as you decrease the aperture.

Distortion Testing

You can buy accurate optical distortion targets, but they’re pretty expensive ($200-$300) and for all but quantitative work there are plenty of square things you can take pictures of to detect distortion. Squarely laid brick walls work (although most aren’t as square as you’d like), big windows or doors, basically anything that gives you parallel vertical and horizontal lines. You want it to be at your level, though: angling the shot up or down (or side-to-side) will cause its own distortion. You need the camera lined up near the center of the object you’re using to check distortion.

Barrel distortion is most common at the wide end of zooms and pincushion at the long end. More complex moustache distortion may be less evident, and is less common.

129156977-rtcrkmwc

The common types of distortion (above) and an example of barrel distortion (below)

More Advanced Testing

Flare Testing

Flare testing simply requires some bright sunlight. You don’t shoot directly into the sun, but have it shining at an angle from one corner of the field of view. Examine the image for ghosts and flare, and to see how badly contrast is reduced. There are no standards here, but its good to know how the lens performs with the sun in the field.

Star Test

This one requires no more than a tripod and a clear night sky. Basically you simply take a long exposure of the sky in an area with fairly uniform stars. A few second’s exposure should give you fairly pinpoint stars and several minutes to an hour will give you star trails. Observing these you can see very quickly if the lens exhibits coma or severe astigmatism, which will show as blurry stars or star trails that fatten in the corners. With such a long exposure camera sensors and noise will play a part but you can get some simple information you might not otherwise.

Image Stabilization

This is difficult to standardize for a lens reviewer, but it’s a good thing to test with your own hands and shooting style. Simply shoot a test target using different shutter speeds (making sure some of the speeds should be sharp and some blurry when handheld) with the stabilization system on and off. See how many stops YOU get with your hands. Remember IS does not work equally well on every shot so repeat several shots at each range and speed with the IS on.

Don’t try to hold the lens to a made-up standard like “if its 4 stops of stabilization I should be able to hand hold a 1/20 second exposure”. Just see if it helps and how much. Every person gets slightly different results from the same IS system. The important thing is to see how much YOU get.

What Did We Learn Today?

When buying a new or used lens, its worth an hour of time to make sure it’s a good copy and to find out a little about how it works.

My absolute requirements for an acceptable lens are:

  1. Construction and function pass “look and feel” testing.
  2. Front / backfocus can be corrected with microadjustment of 2/3 of the camera’s range or less (leaving a bit of margin of error for future adjustment).
  3. Front / backfocus is similar throughout zoom range and at near and far focusing. How similar is a judgement call, but if it changes less than +/- 3 for a 4x zoom and +/- 5 for a 10x zoom throughout the zoom range, then I’m pretty happy.
  4. All 4 quadrants of the lens are equally sharp, vignette equally, and have similar chromatic aberration at maximum and minimum zoom, near and far focus.
  5. The lens doesn’t have severe loss of sharpness at one end of the zoom range or one end of the focusing range (20% decrease in resolution isn’t too unusual from the sharpest to softest part of a zoom, but 50% certainly suggests something is wrong.)

Most other things are going to be characteristics of the lens, not evidence of a good or bad copy. When you go examine your testing pictures on a monitor (no you can’t use the camera’s LCD) make yourself some notes on that lens. My notes for a made-up zoom might look like this:

  • Maximally sharp at f/5.6 in center at wide end, f8 at long end
  • Corners maximally sharp at f8 throughout the zoom range
  • Wide open sharpness falls off at 1/2 of the distance to the corner in most of the range, but 1/3 at extreme telephoto end.
  • Zoom is sharpest at wide end, begins to loose sharpness at the last 1/4 of its range.
  • There is barrel distortion wide open, gone by 24mm. Slight pincushion at maximum zoom.
  • The lens vignettes mildly wide open at the wide end of the zoom range, but not the long end. It disappears at f5.6.
  • Image stabilization lets me handhold at 2 times the shutter speed expected from 1/focal length X crop.

The notes are specific for each lens. A wide aperture prime lens may have notes about when spherocromatism disappears. With a Macro lens, writing down at what aperture diffraction starts to cause softening may be the most important note I make, since I’m often shooting at high apertures to increase depth of field. If you only have a couple of lenses, you’ll probably throw the notes away after a few sessions with the lens, but if you shoot with a lot of different lenses like I do, its nice to have that note card to refer back to on a shoot.

Good photographers take advantages of these lens characteristics or learn work around them. They also are the fuel that feeds so many online arguments about lenses. For example, I love wide aperture primes and expect to see spherochromatism or soft corners wide open. I use them for center compositions and have learned to compose my shots to large light colored areas out of the foreground and background. But about once a week I get my feelings hurt when a renter tells me my recommendation of that lens stinks because the foreground looks magenta and the corners are soft.

Oh, and one last note. You’ll notice that nowhere in my testing do I say how to calculate “is my copy sharp”. It’s a meaningless calculation unless you have dozens of copies of the same lens like we do and are using some standard measurement of sharpness. I know, for example that on our test charts, at our standard distance of shooting, on the 5D II bodies we test with, that the 70-200 f2.8 IS II Canon lens should resolve about 20 lines/mm. The 16-35 f2.8 will resolve about 16 and the 300 f2.8 about 22 or 24. But these numbers are meaningless to your testing unless you had exactly the same set up as we do.

I also know that one copy of a lens may resolve slightly differently on one camera than another. My only point in all this is instead of worrying about “sharpest possible” just make certain its “sharp enough”. And you do that by examining your prints, not 100% magnification on your monitor. No one ever won a photo contest or award for owning the sharpest possible lens.

The copy is sharp if it resolves the detail you want to see in the prints you make. After you’ve tested the lens and know that its not defective and what its characteristics are, go take some pictures. That will tell you if the lens is sharp enough for you.

Author: Roger Cicala

I’m Roger and I am the founder of Lensrentals.com. Hailed as one of the optic nerds here, I enjoy shooting collimated light through 30X microscope objectives in my spare time. When I do take real pictures I like using something different: a Medium format, or Pentax K1, or a Sony RX1R.

Posted in How To's
  • alsh

    Thank you for the great article. Could you please tell me why there’s some yellowish tint on the left side of the AF1951 Test Chart image? Is it because of light or is it a lens fault?

  • Michael Clark

    The camera focuses on the area of greatest contrast anywhere within the active AF point(s). The active areas for each of those points are larger than the little squares that mark them in your viewfinder, sometimes significantly so. That’s why it is important to have a flat target parallel to the camera’s focal plane to aim the AF at, with the tilted target off to one side. If you aim at the tilted part of the target the camera will almost always focus closer than the part covered by the little square in your viewfinder (unless there is an area of higher contrast further back).

    Here’s a map of the 5D Mark II AF system. Notice the blue areas of sensitivity are several times larger than the black squares that mark them in your viewfinder. Then there are the the red “assist points” used in AI Servo AF that aren’t even marked in the viewfinder at all!

    https://www.flickr.com/photos/33233371@N04/3098752372/in/faves-39420162@N04/

  • Mark

    Thank you for the article. Unfortunately, I’m reading it a few years after you wrote it and a few things have changed.

    1. LensAlign Pro no longer exists. Offered now by that brand is the LensAlign MkII (~$85), a 32″ long ruler add-on for wide-angle as well as telephoto lenses(also ~$85), and a software product called Focus Tune which appears to have few users judging from the general lack of activity on its official product discussion forum. The LensAlight products seem to have generally unfavorable reviews. For less than $65, there is a Datacolor offering called the SpyderLensCal Autofocus Calibration Aid, which appears to be similar in concept and has generally favorable reviews.

    2. The Edmund Resolving Power Chart you recommend and linked to has increased from $20 to $35 since you wrote this. It’s not important, but I thought you might want to know.

    3. You briefly mention the Edmund Optics #58940, which is their smallest ISO 12233 chart. It’s approximately the same size as the Imatest chart that you recommend in http://wordpress.lensrentals.com/2014/02/setting-up-an-optical-testing-station. You seem to recommend using one of these ISO 12233 charts instead of the AF1951 chart, and cited to and example (http://wordpress.lensrentals.com/2013/11/inspecting-an-in-spec-lens) of what can happen using an AF51 chart. Am I correct?

    Thanks again for the great work!

  • Oleg

    I’ve succeeded to replace the lens, tested new one. Sadly, it has uneven corners too (much less pronounced than on first lens). Darkest corner remains bottom-right, brightest now top-left (was bottom-left). I’ve checked camera with my old 24-105 f/4L – all corners appeared even.

    The new lens seems to be usable (not easy to see anything on real shots)…
    I’ve uploaded several test shots and crops into http://goo.gl/mdBWXp
    Would you say this new lens is OK or defective? Yeah, I understand uncertainty of “defective” or “in spec”, but if you received such lens – would you keep it?

    Best regards,
    —oleg

  • Oleg

    Thank you Roger! I truly appreciate your help!

    The camera was square to the target. I’ve positioned the camera using mirror in the center of target – seeing the lens reflection in the viewfinder. I did not use manual focus for these shots, but tested auto-focus before and it was very accurate. I also verified center chart sharpness on every test shot…
    Next time I will follow your advise, use manual focus and live-view, but from educational point of view – do you think some focus inaccuracy might cause unevenly bright corners?

    Best regards,
    —oleg

  • Roger Cicala

    Thank you, Oleg. Assuming you’re squared to those charts (it certainly looks like you are) and those are live-view manual focused images, I think the lens is decentered. It should perform much better than that.

    Regards,

    Roger

  • Oleg

    Here are a few tests – corner crops as well as full size jpeg
    http://goo.gl/10I6hn (that’s folder on Dropbox, the URL shortened by Google)

    Thank you
    —oleg

  • Oleg

    Thank you very much for the response!
    I don’t know what’s with picasa, I just tried the link on another computer – it worked…
    I’ve uploaded the photo to another site (never used before, but seems to be working):
    http://www.use.com/bcaabc57a92d73af2f1e

  • Roger Cicala

    Oleg, I couldn’t get the link to open.

    Roger

  • Oleg

    Hi Roger,
    Thank you very much for extremely useful discussions in your blog, for sharing your knowledge and experience!

    I’ve recently purchased Canon 24-70 f/2.8 II lens and thought some tests would not harm (somewhat influenced by reading here :-)). I was not satisfied with initial results, thus I’ve made better test target and repeated test outdoor in quite perfect lighting conditions. I am afraid the result is not good – I am going to ask Adorama for exchange.
    I would be very grateful to hear your opinion, here is the link to the 100% crop of 4 corners (well, almost corners) as well as center crop.
    https://picasaweb.google.com/lh/photo/wdWDWmVw_YHGXBKgVPkvzl4WxVwkmJASeI9329NgDgo?feat=directlink

    Best regards,
    –oleg

  • Bruce Fairman

    I want to pick your brain. The problem comes with using the darkest ND filters for shooting extended exposure photos. The darker ND filters are so dark you cannot see through them. So, you need to focus first, THEN add the ND filter(s). The problem comes in that as you carefully screw in the ND filters you often move the cylinder which affects the focus – which you set earlier. There is no way of checking focus because the ND filter is mounted. Do you know of any lenses for the Canon EOS system that do the focusing internally and the outside barrel, where you add the filters, is fixed.

    Appreciate your thoughts.

  • Roger Cicala

    David, I would leave well enough alone. I suspect they don’t even routinely check to see if f/22 is accurate – just see if it closes down all the way smoothly. I can’t imagine wanting to actually shoot at f/22 on an NEX.
    Roger

  • David Franklin

    Thank you for this beautiful and helpful discussion! I have a question about aperture mechanics. I recently bought a high-end prime lens, which I’m using on a Sony e-mount camera (aperture setting is controlled through the camera body, not on the lens). The lens is rated f/1.8-22, but I’ve noticed that it doesn’t make any adjustment whatsoever between f/20 and f/22. It’s not a problem with the camera, since my other lenses do make a noticeable adjustment at those settings. I’ve written to the manufacturer, who has suggested that, due to the narrow tolerances of the mechanical aperture, there might be a deviation of +/- 0.3 EV at any given aperture setting, and that these deviations can accumulate at the far end of the scale. In other words, my lens is fully stopped down when it’s set to f/20. Nevertheless, the manufacturer is willing to inspect the lens and service it if necessary.

    My question is, is this a problem that I should even worry about, given that I never intend to shoot at smaller than f/16 or so? I have no idea how common this kind of defect might be, or if it is even really considered a defect. My chief concern is that a small problem with the aperture mechanics now might turn into a bigger problem a couple years down the line.

    I’d be very grateful for advice!

  • AJ

    Roger,

    Thanks for taking the time to put this article together and for sharing the information. I found it quite helpful in understanding how to go about testing lenses.

    Thank you!

  • jil2

    Hi Roger,

    I have a small question according this:
    “The four corner quadrants of a misassembled lens. Note how much softer the upper right quadrant is than the lower left.”

    I’m not sure that if I have different sharpness in the four corner it’s mean that there is a problem with lens assembling.
    How about the field curvature?
    There are two links which shows that sharpens can differ in corners:

    http://slrgear.com/reviews/showproduct.php/product/1082/cat/11

    (take a look on the blur index)

    http://diglloyd.com/articles/UnderstandingOptics/understanding-field-curvature.html

    (the field curvature can be not smooth)

    What do you think?

    Thanks.

    Evgeny.

  • Seb

    Its great to get insights from someone really involved. Especially the part about flawed lenses being rare but nonetheless regular will make me look at my next lens more closely. Thanks for sharing your insights and tips!

  • mike

    Hi Roger,
    Thanks so much for the reply. After reading my post I’m surprised you understood it. .. I have a guitar pick that is very heavy with a pretty sharp point, I lay that on the ruler kind of like a pointer right at the 6″ hash mark and focus at that point. I’m just confused that at 24mm ( especially wide open ) my camera focuses on that 1/8″ thick “end” of my ruler that is 6 inches away from what I’m focusing on, what confuses me more is the fact that if I cover that end with a piece of paper, or cloth, without touching or moving anything, it focuses where it’s supposed to. I don’t expect autofocus to be perfect, I know sometimes the camera doesn’t focus on what I want it to, but usually, if I change the angle slightly or move a bit it will. I will move things further away and try again.
    I know you’re busy and I appreciate your reply. thank you so much for the information….this is too much like work 🙂

    regards,
    Mike

  • Roger Cicala

    Hi Mike,

    If you’re focusing on the ruler itself, sometimes that throws off the autofocus system. You might try putting something vertical at the 6″ mark next to the ruler and focusing on that.

    That being said, working at 3 feet is going to be kind of difficult for the system, it’s so close, especially at wide angles. You might try repeating the test at 24mm but a few feet further away (and probably using several objects rather than a ruler given the focal length).

    All that being said, it is pretty common for a zoom to have different focusing at one end than the other, but not as extreme as what you’re describing.

    Roger

  • Mike Fregeau

    Hi Roger,
    After renting a 24-70L from you guys for my sisters wedding I like it so much I decided to buy one. This is a hobby for me and I want to take pictures not test lenses, but I thought I would try a couple of these when I got the new lens. I have a 450D ( XSi ). I used a 12″ wooden ruler for front and back focus tests and something strange happens only at the 24mm end and I was hoping you might have an answer. I use the center focusing point and focus at the 6″ mark, with the ruler 3 feet from the camera the front end of the ruler is in focus ( the zero point if you will ) when I move the ruler 6 feet away the 12 inch end of the ruler comes out in focus. At 70mm I get a very slight front focus up close and a pretty much right on at 6 feet. The same at about 40mm, this only happens at 24mm irrelevant of the aperture. My camera is on a tripod and I can reproduce this result every time. The ruler is about an 1/8″ thick and if I throw a piece of paper over the end of the ruler, I get a correct focus on the 6″ hash mark. Again, only at 24mm.
    Do you think this is a problem with the lens or the cameras autofocus system ? Since I don’t take pictures of rulers, I’m not overly concerned, I’m just curious, why only 24mm

    These articles are great BTW. Especially for a novice like myself.

    best regards,
    Mike

  • Roger,

    As my T2i does not have microadjust capabilities, is there any other techniques one can use to adjust for front/back focus issues?

    Thanks,

    Tom

  • Thanks Roger – 10x live view autofocus for still objects and center AF otherwise as a starting point.

    I am looking forward to the 24-70 2.8 that I rented to see how it compares.

    Best,

    Tom

  • Roger Cicala

    Hi Tom,

    If you have a still subject, there’s no question that focusing in 10X live view is the most accurate. But when that’s not an option center point AF is generally the best, or at least the most predictable. There is some variation depending on which camera body, though: older or ‘intro level’ bodies are almost always best with center point only, some of the upper end and newer bodies seem to do better with ‘focus groups, etc. The other issue that comes up is ‘focus recomposing’: that may cause more problems, depending on the lens you’re shooting, than anything else.

    Best,

    Roger

  • Sorry about double post…

  • Hi Roger,

    I really enjoyed reading your articles and they have taught me a lot – thanks. I have a T2i and have done a series of front/back AF focusing tests with my lenses. I also tried using the test sheet with manual focus, Liveview AF, Liveview 10x magnification manual, etc. Bottom line is I don’t know what to believe about my camera and lenses. It all started because, even after sending in my Tamron 28-75 for calibration, it seemed blurry and I found myself always preferring to use my 50/1.4 or 100 2.8 L. However when I did the focus test the 28-75 was bad but the 50/1.4 was even worse argh. So, I don’t know if I am over thinking this…

    All analysis aside what, in your opinion, is the best way to get the best focus from my camera? AF single point? AF using multiple AF points? AF in 10x Liveview mode, MF, MF in 10x Liveview mode? Any thoughts?

    Best,

    Tom

  • Hi Roger,

    I really enjoyed reading your articles and they have taught me a lot – thanks. I have a T2i and have done a series of front/back AF focusing tests with my lenses. I also tried using the test sheet with manual focus, Liveview AF, Liveview 10x magnification manual, etc. Bottom line is I don’t know what to believe about my camera and lenses. It all started because, even after sending in my Tamron 28-75 for calibration, it seemed blurry and I found myself always preferring to use my 50/1.4 or 100 2.8 L. However when I did the focus test the 28-75 was bad but the 50/1.4 was even worse argh. So, I don’t know if I am over thinking this…

    All analysis aside what, in your opinion, is the best way to get the best focus from my camera? AF single point? AF using multiple AF points? AF in 10x Liveview mode, MF, MF in 10x Liveview mode? Any thoughts?

    Best,

    Tom

    PS I just rented a 24-70 L from you for the holidays and plan on taking lots of pictures to see how it does in my hands vs the Tamron.

  • Roger Cicala

    Hi Mats,

    There’s a little adjustment screw in the mount that sometimes lets a lens be a little loose. If you can rock it back and forth at all, that is it’s loose enough that it may tilt up or down (especially with the big lenses) it will affect your images, but as long as it doesn’t rock or wiggle it should be no problem.

    Roger

  • Hi Roger,
    Thanks for a great article!

    I just bought the 5D Mark II and also ordered a 70-200 F2.8 Mk II from newEgg. When I mounted the lens i noticed that It has a very slight glitch in how it sits on the camera body. The lens itself is very firm and from what I can see so far (have only used it for a couple of hours) the focusing and speed of AF is amazing. I will test It out a lot more in the next few days, but I am tiny bit concerned about the mounting ring, not fitting properly on my 5D. It doesnt slide in and click and smoothly and “sexy” as my other lenses do, and when i try to rotate it is a tiny bit I can feel a slight looseness. We are talking about half a millimeter of sliding maybe. So its not a HUGE deal.

    So my friend let me test mount it on his older 5D MK I and it was the exact same looseness. THEN, I put his older 35-350L Canon lens (also huge and heavy) on my camera and it has the exact same looseness. So now I am thinking is thing common with these huge lenses?

    I guess I am not sure if I should send it back to NewEgg for a replacement or send it to Canon to have them look at it, or walk into Adorama and have them test it and give me their opinion, or should I buy the 3 year extended warranty and then get it looked that way?

    or….should I ignore it since its the same problem with the other lens and the other 5D MK I no matter how i cross compare.

    Thanks so much, any help is appreciated.
    Mats

  • bart de puij

    Thanks Roger,

    This article about how to test a lens and setting micro adjusting is exactly what I needed. Using good primes, I have often been surprised that although focussing very exactly, with and without AF, the result was out of focus. Yes, focus shift. But I could never nail it out, how to micro- adjust. In this article you now told us to calculate the distance in your are adjusting for. That is very valuable information. Friends have always thought, I’m too technical too have fun with photography, but now I finally know why and when out of focus kicks in. I had started shooting wide open only lately ha ha . Thanks!!!

  • Roger Cicala

    Roberto,
    To some degree these things are true of every lens. Not every lens has focus shift, but many wide aperture primes do. Every lens does focus differently at near and far distances, and at different parts of the zoom range. Not only focus but also maximum sharpness: resolution is different at different parts of the zoom range and at different focusing distances. These things are just reality and part of the many compromises involved in every lens design.

    Roger

  • roberto

    hi, you have said:

    Some wide aperture prime lenses exhibit focus shift

    Every lens can focus differently at near and far distances.

    A zoom lens usually focuses slightly differently at the minimum and maximum zoom ranges (and sometimes in between too).

    Can all this problems be fixed by the Technical Support?
    Would I have to change the lens?
    Those are problems that I would have with every lens I buy?

    Many thanks

Follow on Feedly