5
\$\begingroup\$

I'm using the EF 100mm f/2.0 USM lens on an EOS 5D Mark IV body. There seems to be a specific focusing distance (about 1.5 m) below which all photos come out at least one stop underexposed when relying on the camera's metering system.

The following two photos were both taken through the viewfinder in aperture priority mode using evaluative metering. The subject distance is about 1.5 m.

100 mm, f/2, 1/80 s, ISO 640

100 mm, f/2, 1/100 s, ISO 1600

As you can see, the camera's metering system chose a much darker exposure for the first photo, with the only difference being that the camera was slightly further away from the target in the second photo.

This happens in both AF and MF and seems to depend solely on the focusing distance, rather than the actual subject distance. When changing focus in MF mode, I can easily see the jump the metering system does when crossing the focusing distance at which the effect starts. I can provide more examples if needed.

This does not happen in live view. It seems to be specific to mounting the EF 100mm f/2.0 USM lens on the EOS 5D Mark IV body, as it does not happen with any of my other lenses on the EOS 5D Mark IV body and it also does not happen with the 100mm f/2.0 USM lens on my EOS 750D body. Both bodies and the lens in question run the most recent firmware at the time of writing this.

What is causing this effect? I will probably return the lens since this makes shooting with it very difficult and I didn't like it that much overall anyways, but it would be nice to know what is wrong so I can avoid this problem in the future.

I do not think that my question is a duplicate of this question. The jump in exposure is very sharp and happens at a specific focusing distance, rather than continuously changing with focusing distance as I would expect with light loss at high magnification factors. Furthermore, the magnification is not particularly high as the subject distance is about 1.5 m. I am fairly certain that the issue is with the metering. Indeed, when using manual mode to circumvent all metering or live view shooting to use a different metering system, everything works as expected.

\$\endgroup\$
5
  • \$\begingroup\$ Possible duplicate of Does maximum aperture change with focus distance with the Canon 60mm f/2.8 macro lens? The answer applies to most macro lenses. \$\endgroup\$
    – D. Jurcau
    Jan 12, 2017 at 17:20
  • 1
    \$\begingroup\$ What happens if you set a fixed ISO and take your shots in aperture priority? Auto ISO seems to be fiddling with the sensitivity. \$\endgroup\$
    – BobT
    Jan 12, 2017 at 18:31
  • \$\begingroup\$ @BobT In that case, it still sets the same exposure, except it uses only the shutter speed to do so. For instance, I fixed my ISO at 3200 and it switched form 1/80 s to 1/30 s when transitioning across the 1.5 m focusing distance value (different example scene than in the question). \$\endgroup\$
    – Jules
    Jan 12, 2017 at 18:57
  • \$\begingroup\$ Imgur has stripped the exif data, is it possible you can upload the photos to another website that preserves the exif data? It might give us more information about what the camera is doing. \$\endgroup\$
    – Rob
    Jan 27, 2017 at 11:32
  • \$\begingroup\$ @Rob Here you go: drive.google.com/file/d/0B-gY7tKF6PAtelgwU2k1WmRNVE0/… drive.google.com/file/d/0B-gY7tKF6PAtdEpra0d1V1B5ZzQ/… \$\endgroup\$
    – Jules
    Jan 27, 2017 at 22:53

1 Answer 1

2
\$\begingroup\$

The modern camera’s metering system measures exposing light thru-the-lens. We expect the results to be accurate, even when close focusing.

You should be aware that the f-numbers associated with a camera lens are based on imaging an object at infinity (symbol ∞). In other words, the math is performed when imaging a point source, like a star. When we image objects that are closer than ∞, the f-numbers become invalidated. The f-number error is super tiny when imaging objects 3 feet (1 meter) or further. When imaging object at about 1 meter, or closer, the error becomes significant.

This phenomenon is called “Bellows Factor” (BF). In the past, most camera makers stopped close focusing when the error approaches 1/3 f-stop. They did this because general purpose camera lenses are optimized for distance thus they are compromised when tasked to work close. Additionally the f-number error (BF) will likely result in an under-exposure unless a correction is applied. Modern cameras that read the exposure through the lens are generally immune as the metering system likely applies the needed correction.

On the other hand, a macro lens keeps the f-numbers valid when close focusing and they are optimized to image close-up., They are slightly challenged when tasked to image objects at a far distance.

Bellows Factor math: BF=(m+1)^2. M = magnification. If you close focus an achieve “unity” (life size or magnification 1), then m=1. At unity BF=(1+1)^2 (magnification + 1 squared). Solving for BF at unity: BF =( 1+1)^2 = 4 The BF is handled just like a filter factor; we multiply the exposure time by the BF to compute a revised shutter speed. If the factor is 4 the correction is 2 f-stops compensation (you open up the aperture since each f-stop is a 2X change). At unity the camera set to f/5.6 is actually functioning two f=stops more stopped down at f/11. Applying the BF to shutter speed: The shutter is set to 1/60 of a second, we multiply 1/60 by 4 = 1/15 (the revised shower shutter speed).

Why is your close-up images unexposed? Perhaps the camera’s logic does not fully compensate for BF. Perhaps the mundane background and it’s color / hue is fooling the metering logic. I suggest aperture priority thus allowing the camera logic to select the shutter speed. You can compensate in manual mode and open up the aperture or slow the shutter.

\$\endgroup\$
10
  • \$\begingroup\$ Thank you for your answer. However, I do not think that it has anything to do with bellows factor. The jump in exposure is very sharp and happens at a specific focusing distance, rather than continuously changing with focusing distance. Furthermore, the magnification is not particularly high as the subject distance is about 1.5 m. I am already using aperture priority. I am fairly certain that the issue is with the metering. Indeed, when using manual mode to circumvent all metering, everything works as expected. \$\endgroup\$
    – Jules
    Jan 12, 2017 at 17:32
  • \$\begingroup\$ I think it is the color and hue of the background -- try a background that is not uniform (mundane). Also try a uniform medium gray. \$\endgroup\$ Jan 12, 2017 at 17:40
  • \$\begingroup\$ I have already tried many different scenes, it seems to happen independently of the subject. I chose this simple scene as an example since changing the subject distance does not change the contents of the photo significantly, but it happens with all sorts of other subjects as well. \$\endgroup\$
    – Jules
    Jan 12, 2017 at 17:44
  • \$\begingroup\$ Try it with an 18% grey card in room with only the subject light on it, I.E. a light pointed at the subject and no other room lights on. I think this a metering issue not a lens issue. take series of photos at different lengths all in FULL manual mode and change nothing but the focal distance to insure that any changes are indeed the lens. \$\endgroup\$
    – Alaska Man
    Jan 12, 2017 at 19:38
  • 1
    \$\begingroup\$ @Jules "full manual circumvents all metering" That is incorrect. Full manual circumvents the camera from making changes base on the meter reading. The meter still takes a reading. In full manual YOU need to set the aperture and shutter speed according to the meter reading for the iso you have set. The point is are the photos you have taken at different focal lengths exposed exactly the same - same fstop, shutter, speed iso. if so and you are still getting photos that are not exposed the same then you have eliminated the all the variables and can then consider that it the lens. \$\endgroup\$
    – Alaska Man
    Jan 14, 2017 at 3:38

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge that you have read and understand our privacy policy and code of conduct.

Not the answer you're looking for? Browse other questions tagged or ask your own question.