It recently came to my attention that lenses rarely perform at it's best when wide open. Is this the case with the Canon 70-200 f/4 L? Will I lose IQ by shooting at f/4 or should I shoot at 7.1 to be sure? (Lens is mounted to a Canon 600D)

  • Kuckel - Crucial question: Is the lens the (1) EF70-200mm f/4L USM (ie non IS) or (2) the IS stabilised EF70-200mm f/4L IS USM ? The two lenses are different. My answer applies specifically to the latter IS version and probably also applies reasonably well but less certainly to the non-IS version (as the DxO tests relate to the IS version. ) – Russell McMahon Sep 21 '16 at 2:46

No lens is perfect. Not at any aperture. Not at any focal length.

Just because a lens may be slightly sharper at f/7.1 than at f/4 is no reason to never use it at f/4.

Some lenses demonstrate a wide variation between wide open and the aperture that gives them their sharpest performance. Other lenses have much less variation in their performance between wide open and the aperture with the absolute maximum sharpness. Although there are exceptions, in general wider angle lenses demonstrate a greater difference between wide open and stopped down than longer focal length lenses when both lenses are made to roughly the same level of optical and manufacturing quality. This is the case not only in terms of acutance, but also in terms of aberrations such as light fall-off (often referred to as vignetting), chromatic aberrations, and geometric distortion.

The EF 70-200mm f/4 L performs very well even at f/4. It does demonstrate slightly better acutance at around f/5.6. The difference is barely measurable using test charts under laboratory conditions. It will probably not be noticeable at all in real world usage. By f/8, depending upon what camera it is mounted, it is already losing acutance to diffraction.

Your Canon 600D has a pixel pitch of 4.3µm which gives it a Diffraction Limited Aperture (DLA) of f/6.8. So at f/7.1 you will already begin to see the effects of diffraction when viewing your images at 100% on your monitor.

There are a number of instances where the utility of using an f/4 aperture will outweigh the very slight loss in acutance you might get with your EF 70-200mm f/4 L. These instances are also applicable to other lenses where the difference between wide open and stopped down are even greater. Some of them are:

  • When you want the Depth of Field (DoF) to be as shallow as possible.
  • When you are shooting in a limited light situation and you need to use the fastest shutter speed that you can to freeze subject motion or reduce blur caused by camera motion.
  • When you desire the absolute highest framerate possible with your camera. Setting the aperture to not stop down, along with setting the lens to "manual focus", the ISO to 100 (less noise means smaller file sizes), and the shutter time to the camera's shortest setting all allow a faster frame rate by reducing or eliminating the time needed for the camera to autofocus, stop down the aperture, take the picture, process the data from the sensor, and write the file to the memory card.
| improve this answer | |

Are you asking about the discontinued non-IS version or the IS one?

With regards to the IS version, it may be better stopped down, but it is one of the lenses that are good enough wide open so that you don't have to worry if you are shooting wide open or not.

The non-IS version was also apparently pretty good, but not as good as the IS one.

| improve this answer | |
  • The name of the "non-IS" version is EF 70-200mm f/4 L. That is the lens specifically referenced in the question. – Michael C Sep 20 '16 at 19:25
  • 1
    @MichaelClark Long experience has shown me, and I assume you also, that unless people are specifically specific about a lense's description then, if there is room for ambiguity, uncertainty should be assumed until full detail is gained. The two lenses that you posted very helpful references to above were (using their titles) the 1999 release EF70-200mm f/4L USM, and the 2006 release EF70-200mm f/4L IS USM. Neither name is an exact match for the OP's "EF70-200mm f/4 L" title. | ebay searching Canon EF 70-200mm f/4 L USM gives 125 hits and adding IS makes that 63. Yet TBD I suspect. Maybe not. – Russell McMahon Sep 21 '16 at 2:43
  • USM is often dropped from the lens description by everyone except Canon. Just because there is now an "EF 70-200mm f/4 L IS" does not mean the "EF 70-200mm f/ L" is no longer the name of the original lens. What would be incorrect is to call the original lens an "EF 70-200mm f/4 L Mark I" or such since Canon has never referenced a lens, camera, flash, or any other accessory using the Mark I moniker. – Michael C Sep 21 '16 at 3:51
  • If you go to any of Canon's support sites you will see with a 100% degree of consistency that the older lenses are listed by their unmodified names they were given when introduced, even if a Mark II, Mark III, Mark IV, or even Mark V version has later been introduced. – Michael C Sep 21 '16 at 3:55
  • @MichaelClark Not important (as far as I can see) but I'm mystified re reference to Mark 1 II III etc. I said nothing about any MKxx versions - I did mention DXO Mark. || He has confirmed that it's the NON-IS version, as you suggested. – Russell McMahon Sep 21 '16 at 10:53

This answer (hopefully) complements Michael's one.
It's liable to be an "eye opener" both for your camera and in general.
The tool discussed below is an invaluable aid to am\nyone wanting an easily assimilated answer to the question you are asking but for their camera / lens combination.

A vitally important point that soon becomes obvious is that for a given lens the answer changes with the camera used and not always in a manner that is intuitive.


Using the tool discussed below you can say "at a glance" that the lens and camera combination is best across the range at f/5.6 with a slight fall off at the top end, a tad worse at f/4 and f/6.3 getting rather poor by f/8 and best not used above that aperture.


To start with a completely different camera and lens.
My experience is that on a Sony A77 a Sigma 70-200mm, f/2.8 produces essentially maximum sharpness at f/2.8 across muh of the zoom range. I'm not complaining. On other cameras the result with the same lens may be different.

This page provides summary test results by DXOMARK for the Canon EF 70-200mm f/4L IS USM lens.

As presented the results are for when used with the EOS 5DS R.
The camera used can be selected.
With the 5DS R the DXOMARK score is a very good 27, the effective Mp is 24 and they note that the lens scores highest (all their factors considered) at 70mm, f/4.

I recommend duplicating the tab or opening an identical window or similar for reasons noted below.

The release sequence was 550 - 600 - 650. They do not have results for the 600D, or 650 but do have them for the 500D, 550D, 700 ..... The 550 would seem the closest for comparison. (CMOS 22.3mm x 14.9mm, 5184 x 3486 = 18 Mp effective).

In the 2nd window change the camera to EOS 550 D.
Results 5DS R / 550D are

27 / 14 - DXO mark score 70mm f4 / 135mm f4 <- best at 24 / 9 - Mp sharpness effective 0.3% / 0.1% - Distortion
-1.1 / -0.4 - EV vignetting

You need to pore over the DXO mark definitions to make sense of those figues (or such sense as may be made) BUT it is clear that

  • there is a vast apparent difference in performance between the two combinations, even though the lens is the same in each case

  • The FF camera performs best overall at 70mm 'wide open", while the APSC camera is best at 135 mm also "wide open".

  • While the FF camera is far superior in terms of sharpness, the APSC camera is somewhat better in vignetting and distortion.

ALL of these are relatively blunt measures with lots of hidden detail due to the summary form of the results.

Now, on each of the two pages (5DS R & 550D) click the "measurents" tab giving 5DS R and 550D.
These are 3D "maps" for DXOMark score with Y axis = aperture, X axis = focal length and Z axis differentiated in colour showing gradations of "score".
Rather than analysing these graphs, click the "sharpness" tab in each case giving 5DS R sharpness & 550D sharpness

Wow. The map at left is simultaneously astoundinly impressive and really stupid. In this case, for comparison purposes, having identical X & Y scales means the two maps can be directly compared, but given that we have seen that the lens is rated at 24 Mp when used with the 5DS R, having the perceptual Mp scale clipped at 12 Mp for the FF camera, means that the nuances and disasters of performance that are clearly visible for the APSC camera, are utterly lost for the FF 5DS R. We can see that the sharpness is reducing somewhat at increasing aperture, but is "rather good" anywhere up to f11, and that its worst at around 200mm and > f/22 (no surprise), but the solid dark green at the bottom lumps together all PMp scores of 12-24. A shame.

Returning to the 550D (hopefully close enough to your 600) you can see that the lens is probably best at 70mm, f/5.6 (while the text result said 135mm, f/4) but that it's about 'as good as it gets' at any focal length, not vastly worse anywhere from f/4 to f/6.3 at all focal lengths, is fading somewhat at f/8 and above f/11, well ... . IF you believe the DXO Mark results.
As this more or less matches what Michael says, you probably do.

enter image description here

| improve this answer | |
  • 2
    DxO doesn't even have any data posted on the EF 70-200mm f/4 L. The EF 70-200mm f/4 L IS is a different lens with a different optical formula. Compare the two block diagrams at the Canon camera museum: global.canon/en/c-museum/product/ef391.html and global.canon/en/c-museum/product/ef356.html – Michael C Sep 20 '16 at 10:57
  • 2
    Further, the disparity in DxO Marks scores with regard to the EF 70-200mm f/2.8L IS II compared to every other reputable online reviewers results casts some serious doubts about DxO Mark's entire methodology. – Michael C Sep 20 '16 at 11:01
  • 2
    Their test results are only useful when it is remembered that they are the results of a single example of a given lens on a single example of a given camera body. The overall conclusion that larger sensors with higher resolution will give better results for the same lens is expected when one understands that the "print" scores are all normalized after downsizing to 8MP. – Michael C Sep 20 '16 at 11:05
  • 1
    Most DxO Mark lens scores are fairly close to what most other reputable testers that post results online find for most lenses. For whatever reason, the Canon EF 70-200mm f/2.8 L IS II has suffered poor scores at DxO Mark while performing much better in other tester's labs. It's been a fairly well documented and discussed issue, especially with regard to the EF 70-200mm f/2.8 L IS II. – Michael C Sep 20 '16 at 12:18

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

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