Gianluca Colla: Reasons To Calibrate Autofocus (Part 2)
Oct 2012 09



*Note: This is second piece of a two part article, read part 1 here.

We’ve learned that despite brand, cameras and lenses, most of the time our gear need to be calibrated to obtain a razor sharp focus. Let’s go in depth and learn how to do it!

Now that we have seen in this article why we need to calibrate the autofocus of our camera/lens combination (CLC) let’s see how to do it.

In a nut-shell, we point the camera at the SpyderLENSCAL, we focus on the vertical target, take a picture, and by zooming in the picture (on the camera display or, better, on a computer screen) we can verify on the ruler whether the front or rear section is sharper (ideally should be the zero, but….)

In case the numbers in the ruler’s front area are sharper than zero, then the lens is ‘front focusing’, in case the numbers on the rear are sharper, then the lens is back focusing and we can use the in-camera menu to fine tune the focus.

That’s the quick explanation, let’s go through the in depth process. To make things easier, I’ll split the process in 6 parts

  1. Alignment
  2. Camera – SpyderLENSCAL Distance
  3. Camera Settings
  4. Lens Settings
  5. Test Image Evaluation
  6. Corrections
  7. Periodicity

1) Alignment

It’s important to get things lined up properly: we need to be square on to the target and it is worth to invest a little bit of time to do it correctly. First of all, we make sure that camera and SpyderLENSCAL during the entire process are placed at the same distance: it can be a achieved using tripods, light-stands (SpyderLENSCAL has a tripod mount) a flat surface like a table, a desk, and so on.

We need to make sure our CLC (camera/lens combination) and SpyderLENSCAL are aligned on the vertical and on the horizontal axis. The SpyderLENSCAL has a built in bubble spirit level to ensure it gets positioned correctly. A level on the camera, thought not mandatory, will definitely make the process easier: you can use an external spirit level mounted on the camera hot shoe (see Image 1) or a level on the tripod plate, or an electronic level in your smartphone (see Image 2) and do not forget that many of recent cameras includes electronic levels (Image 3).

Image 1

Image 2

Image 3

Look at Image 4A and Image 4B to see the correct positioning of the SpyderLENSCAL and the camera. If your camera has LiveView, it can be very helpful in this procedure.

Image 4A

Image 4B

When aiming the CLC at the target, the central focusing point of your camera should overlap the small bullseye at the right edge of the vertical target, but make sure to not overlap the ruler with the center point, or it will result in wrong focusing (see Image 5.)

Image 5

When looking through the viewfinder, the right side of the vertical target and the left side of the ruler should be parallel (see Image 6). If there is some space between the two (see Image 7), you are positioned too much on the right and if you don’t see the entire scale of the ruler, and it’s “cropped” in the upper part (see Image 8) you are positioned too much on the left.

Image 6

Image 7

Image 8

2) Distance

This is probably the question that all of the SpyderLENSCAL users have: what should be the distance between the camera and the SpyderLENSCAL? Some producer like Canon suggest the autofocus regulation should be done at 50 times the focal length. In other world, if I am using a 50 mm, the distance should be 50mm x 50 times = 2500mm or 2.5 meters, if I am using a 100mm lens the distance should be 100mm x 50times = 5000mm or 5 meters, and so on…

Now, while I do trust Canon and I do believe as a general rule 50 times the focal length gives a good overall result, I tend to calibrate the autofocus at my most used shooting distance for that particular lens (if known). A 100mm lens, for example, is a lens I use mostly for portraits, and most of the time I am shooting between 1 and 3 meters, rarely at 5 meters as suggested by Canon, so I’d rather have the peak performance of my CLC in that distance range. If I do not have a specific or most used shooting distance, than the 50 times rules apply. Several “geek” I know prefer to use a distance of 30 times the focal length, because it’s a bit easier see the focusing error (if any).

Needless to say, you should NEVER change the distance between the camera and the SpyderLENSCAL during the calibration!

3) Camera Settings

To evaluate the focusing performance of our CLC, it’s important to shoot whit the following settings:

  • Camera in Aperture Priority Mode (Av on Canon, A on Nikon)
  • Widest aperture: we want the smallest depth of field to see the real focusing point
  • Fixed white balance. From frame to frame, depending on the background, the camera might choose a different white balance, which makes the evaluation of the images a bit more difficult. Setting the white balance with an appropriate value, guarantee all the images have the same colors.
  • Low ISO setting, generally 100 ISO or 200 ISO, gives a better image quality and makes the judgement easier.
  • Raw mode: in-camera jpegs have often some sort of sharpening applied which can alter the judgement of the images, so it’s much better to have “neutral” raws.
  • Central focusing point selected
  • Self Timer and Mirror Lock up: low ISO might result in long shutter time and to avoid problems caused by camera shaking or long exposures, setting the self timer (generally 2 seconds will be enough) and the mirror lock up function is a good practice to avoid issues
  • Before taking every picture, manually set the lens focus to infinity. This ensure your lens is actually searching for the focus, so it fiscally moves the elements.

4) Lens Settings

Of course if you calibrate a prime lens the question doesn’t even bother, but what if I calibrate a zoom lens? should I calibrate at the widest or the longest focal length? Normally you can only calibrate at one focal length and by nature, if there is a need for calibration, it will be more visible at the longest focal length, due to the shallower depth of field. I suggest that you do a test at both focal length, to verify which length is more in need of correction.
With newest cameras like the Canon 5DMkIII you can calibrate the autofocus at the widest focal length and at the longest, resulting in a much improved precision and control in focusing (see Image 9)

Image 9

5) Test Image Evaluation

When calibrating a good quality fast prime lens, like a 50mm 1.4 or a 85mm 1.4, it’s quite immediate to see where the real focus is on the SpyderLENSCAL ruler. Things are a bit more difficult when the quality of the lens is not so good or when the lens is not so fast.

In the first case, with a mediocre lens, we might have an overall fuzziness that makes spotting the real focus quite hard, since everything really looks non in-focus. In the second case, i.e. shooting with a 18-55mm f 5.6, the relatively short focal distance and the big depth of field given by the f5.6 aperture make everything look in-focus, therefore we are unable to find where the focus really is.

Two things can help in this scenario: most of the lenses (even the good ones) have some sort of chromatic aberration in the out of focus areas (see Image 10, click to enlarge it) and generally show some magenta on the out of focus area in the front section of the ruler and some green in the rear section of the ruler. If you start from magenta, you notice that going towards the back of the ruler the color tend to vanish, till some green show up. The point where there is no color aberration, or better, where there is a transition from magenta to green, is your sharpest point. Zoom in at 100% view to see exactly where the aberration shift from magenta to green. You can try increasing the saturation or the vibrance in your raw converter, If this color shift is not clearly visible.

Image 10

Regarding the big depth of field given by some slow and wide lenses, which translates in everything in focus, the only real thing you can do is to evaluate if the DOF is spread 1/3 in the front and 2/3 in the back, as it should. If not, you can correct accordingly. Let’s say your depth of field is about 1/2 in the front and 1/2 in the back, than you have some front focusing and you should “move” it toward the back.

A clarification is now required: if you shoot several images of the target and, as suggested earlier, you manually “defocus” before every shot, you’ll eventually find that the focus is slightly different from image to image. For this reason I suggest you take at least 3 frames of the first test image, and 3 frames for every other image with corrections applied, so you can compar and ensure you actually get consistent focusing from picture to picture.

6) Corrections

Once you spotted the real focus point and the sharpest area on the ruler, is time to fine tune the focus from the in-camera menu. This function is normally located in the camera custom functions menu, in image 11 is an example of Canon cameras and in image 12 an example of Nikon cameras. Please refer to your camera’s user manual to find this function. In case your CLC presents front-focusing, you have to move the focus backward, in case it presents back-focusing you have to move the focus frontward.

Image 11

Image 12

Some cameras allow lens per lens correction (normally you can store up to 20 lenses per body) and a overall correction that apply every lens mounted on that body. You DO NOT want to use this second possibility, and you ALWAYS want to correct lens by lens.

But the question is: how much correction do I apply?

With some (very few) producers the correction can be made in millimeter, which makes life lightning fast: your CLC has a back focus of 15 millimeters? than you just insert 15 millimeters of front focus and.. that’s it!

But with cameras like Canon and Nikon, you can only adjust the focus in “steps”. On top of this, the steps are different from lens to lens. If for example a 20 millimeters front focus on a 50 mm lens is corrected with +7 steps in the camera menu, 20 millimeters of front focus on a 135mm lens might require a +12 steps correction.

In other words using cameras like Canon or Nikon, there is no specific correction to a given amount of front or back focus, and we must take a few images to find out how many steps to apply. We start with a correction, than we take another image, we check if the correction is enough, and in case is not, we proceed until we find the right value.

As a general rule, if your CLC present a severe focusing error, in the range of 5 centimeters or more (front or back) I would start with 8 steps, if the error is in the range of 2.5 centimeters, I would proceed in steps of 4, and if there is a minor focusing error, in the range of plus or minus 1 centimeter, I would correct in steps of 2.

Of course when you get close to correct the problem, you want to proceed one step at the time, to make sure you arrive “gradually” at the sharp zero on the ruler.

7) Periodicity

Like in the case of monitor calibration, there is not a rule regarding how often you should calibrate a lens. Of course a brand new lens is less prone to loose precision than an old beaten lens used every day for the last 15 year. Also, temperature and usage can influence the behavior of a given CLC. While it takes many words to explain in detail the calibration procedure, in real life it takes no more than 5 minutes per lens, while a short autofocus check with the lens cal takes just the time of aligning and taking a picture. For an average user, I would say a quarterly check would be enough, for professional photographer shooting every day news and reportage, with camera always hanging on the shoulders, a monthly check would be ideal. It really depends, but every time you feel you’re not getting the perfect focus you should, a quick check with SpyderLENSCAL can clarify every doubt.

In conclusion, this was a long explanation that translates in few minutes in real life. Just remember that this procedure will not make your lens sharper than it is, it will make your CLC more precise and working at its peak performance. So if you have a standard kit zoom lens, do not expect it to perform as a top notch prime lens after a calibration with SpyderLENSCAL. But sure enough, a calibrated cheap lens will return sharper images than a high end prime lens with a severe focusing error…

Now, just calibrate all you lenses and start enjoying razor sharp photography…

An in depth knowledge of photography has led Gianluca to travel to many diverse destinations around the world, from the Arctic Polar Circle to Africa’s deserts, from the Far East to the Amazon. Gianluca’s photographs have appeared in various publications including National Geographic Magazine, New York times, Newsweek, Los Angeles Times, CondeNaste, Bloomberg News.

He photographed advertising campaign for clients as Canon, UBS, Saatchi & Saatchi, and his images are represented by National Geographic Image Collection.

Gianluca has an intense lecturing and teaching activity counting various seminars and workshops, including TPW, Canon Academy and Cruise Photo Factory.

When not on assignment Gianluca spends his time in Switzerland and in Italy. If you want to see more of Gianluca’s adventures through his photographs, visit his website and