Lens Testing - Looking at canon 24-70 2.8L II

[robertDthomas]

I have been reading reviews about this lens and one thing I have noticed is that several reviewers state that they returned their first copy for sharpness or some other reason.  Then the second copy was "tested" and preformed well.

What form of testing is done.  Is there a recommended process, target and software?

When I Google 'lens testing software' there is a plethora of hits but nothing seems to jump out at me.  Any suggestions?

[HSakols]

Better than those lens testing sites is to first use the lens and then polish the front element using a fine grade sanding paper and then to compare again.  Still you might need to return the lens for a better copy.  Keep testing and returning lenses until you find what works for you.

[Ellis Vener]

I use a LensAlign mark II +FocusTune software to first  tune the camera's AF system to the particular characteristics of an individual lens. Why? Because cameras and lenses are like guitars and strings, you  shouldn't expect them to work perfectly together straight out of the boxes - but at least with cameras and lenses, sometimes it  does happen. Once the camera's AF system is tuned I go make pictures  o real world subjects at various apertures, with the camera on a tripod to eliminate that part of user error.

Very funny Hugh!

[Eric Myrvaagnes]

Very funny Hugh!

I think Hugh was talking about the short version. In the longer version you start with coarse sandpaper and work your way down through a minimum of seven grades of sandpaper until the front element of the lens is perfectly flat... 

[Ellis Vener]

I think Hugh was talking about the short version. In the longer version you start with coarse sandpaper and work your way down through a minimum of seven grades of sandpaper until the front element of the lens is perfectly flat... 

for me a sandblaster works best. I can go through more lenses, more efficiently  that way.

[HSakols]

Robert,
I know I need to be careful on the web where everyone has a different perception.  So what I'm saying is I'm sorry if I came off as flippant.  I just came back from actually taking photos in the field and assisting Michael Frye with a workshop.  A lot of technical details were discussed, but the bottom line was great light and composition - period.  I too overly obsess over optics, but after being in the field, I am humbly reminded of what is most important -  light / composition / passion.  Did I say passion?  Still I'd like to do the sandpaper experiment.

[EduPerez]

Better than those lens testing sites is to first use the lens and then polish the front element using a fine grade sanding paper and then to compare again.  Still you might need to return the lens for a better copy.  Keep testing and returning lenses until you find what works for you.

Might be relevant here:
LensRentals - Front Element Scratches.

[Bart_van_der_Wolf]

I have been reading reviews about this lens and one thing I have noticed is that several reviewers state that they returned their first copy for sharpness or some other reason.  Then the second copy was "tested" and preformed well.

What form of testing is done.  Is there a recommended process, target and software?

When I Google 'lens testing software' there is a plethora of hits but nothing seems to jump out at me.  Any suggestions?

Hi Robert,

The required tests depend on the intended use and the type of lens tested, e.g. a macro lens should be tested for close-up photography, perhaps field flatness, and is either expressed as a resolution value or an MTF curve, but resistance to glare could also be an important metric. Therefore, the tests should test those features that are important, and preferably deliver a quantifiable result. A program like Imatest offers a range of tests, and those in the SFR (spatial frequency Response) category are most often used for resolution/MTF testing.

I get very positive feedback on a free resolution test target that I created with several useful features in it, for both visual inspection of the results, and it can also produce a few quantifiable measurement values. One of those values is a parameter that can be used for optimal Capture sharpening.

For resolution tests (which are useful to test for deviant performance) one typically tests the image center and the four corners for resolution. When one of the corners performs significantly worse than the others, and assuming that the focus procedure was very good, then the lens may suffer from decentering, and is a candidate for replacement or repair. I prefer to test by focusing separately on the individual corners, to make sure that they are good, and not allow the field curvature of lenses, or mis-alignment of the sensor, to muddle the optimal score that can be achieved, on non-flat subjects.

This will also produce good reference data for later inspection, e.g. after a drop.

Cheers,
Bart

[NancyP]

There is some item to item variance in all manufacturing processes, and if your camera mount and that particular lens copy's mount vary from specification by a significant amount in the same direction, you may encounter front-focusing or back-focusing. This front-focusing or back-focusing should be relatively rare if quality control is good in the first place.

[Ellis Vener]

There is some item to item variance in all manufacturing processes, and if your camera mount and that particular lens copy's mount vary from specification by a significant amount in the same direction, you may encounter front-focusing or back-focusing. This front-focusing or back-focusing should be relatively rare if quality control is good in the first place.

And that is why higher end Nikon and Canon bodies introduced  since mid 2007 have auto-focus micro-adjust settings. it is rare that a lens or a camera gets out of QC that is so out of spec that those controls do not work. 

[jrsforums]

And that is why higher end Nikon and Canon bodies introduced  since mid 2007 have auto-focus micro-adjust settings. it is rare that a lens or a camera gets out of QC that is so out of spec that those controls do not work. 

AFMA will help some to allow you to get the best out of the lens....while a bit limited to adjusting one point, possibly two (long/wide).  This could be improved.

There is still item to item variance between "identical" lenses which produce differing resolution.  Roger Cicala has shown this often on LenRentals Blog.

[Ellis Vener]

"There is still item to item variance between "identical" lenses which produce differing resolution.  Roger Cicala has shown this often on LenRentals Blog."

This is nothing new. I remember watching Al Satterwhite try out several high end Nikkor lenses one afternoon in the early '80s at Ken Hansen's in NYC, hearing and tales of W. Eugene Smith  and others doing the same with their Leica M series lenses. Also I knew photographers  who had Marty Forscher's group shim the ground glass on SLRs to make sure the ground glass plane was aligned perfectly with the actual film plane on their SLRs. 

[stevesanacore]

"There is still item to item variance between "identical" lenses which produce differing resolution.  Roger Cicala has shown this often on LenRentals Blog."

This is nothing new. I remember watching Al Satterwhite try out several high end Nikkor lenses one afternoon in the early '80s at Ken Hansen's in NYC, hearing and tales of W. Eugene Smith  and others doing the same with their Leica M series lenses. Also I knew photographers  who had Marty Forscher's group shim the ground glass on SLRs to make sure the ground glass plane was aligned perfectly with the actual film plane on their SLRs. 

And it's so much more critical now than ever it ever was with film. I calibrate all my lenses to my bodies when I get them. Most are well within the range of the adjustments available.

[jrsforums]

And it's so much more critical now than ever it ever was with film. I calibrate all my lenses to my bodies when I get them. Most are well within the range of the adjustments available.

Suggested reading....

http://www.lensrentals.com/blog/2011/10/the-limits-of-variation

http://www.lensrentals.com/blog/2012/11/canon-24-70-mk-ii-variation

http://www.lensrentals.com/blog/2012/07/autofocus-reality-part-1-center-point-single-shot-accuracy

[stevesanacore]

Suggested reading....

http://www.lensrentals.com/blog/2011/10/the-limits-of-variation

http://www.lensrentals.com/blog/2012/11/canon-24-70-mk-ii-variation

http://www.lensrentals.com/blog/2012/07/autofocus-reality-part-1-center-point-single-shot-accuracy

I've read most of these in the past. So far i've been lucky and I've had quite a few but that's not saying that a few lenses I've had weren't dogs, but I assumed it was their older design rather than a bad copy. I did get a superb 24-70 2.8 last month but it was a loaner from Canon they wanted me to test against my old version - no comparison. The 24-70 2.8 II they sent was sharper than any of my primes in that range.

I was using a 35 1.4 at f2.8 and we switched to the new 24-70, left it wide open and it was much sharper. Maybe Canon sent me a good one?

It's all fun but also quite important to have a good copy when our livelihoods depend on it.

[robertDthomas]

Thanks to all who have provided reading, links and good suggestions.  (I skipped the sandpaper test).  As a number have said the vast majority of lenses are of high quality but during the initial week or two after a major purchase it is good to verify and I appreciate all the suggestions.  I bought LensAlign and plan to run through the focus testing of my existing lenses before the new 24-70 arrives then I will also test this one for de-centering as described by OP's.