So does the 16-35 f/2.8L III have a vignetting problem or not?

[ahsanford]

At three different sites, we've seen a consistent read on 4+ stops corner darkening at 16mm f/2.8 with the new 16-35 f/2.8L III:

TDP: https://goo.gl/pO4uJi (Greater than 4 stops vignetting)
"At 16mm wide open at f/2.8, this lens has very strong vignetting with the corners darkened by over 4 stops, impacting the benefit that the f/2.8 aperture provides."

PZ: https://goo.gl/6dwafV (4.6 stops vignetting)
"A major weakness is the high amount of light falloff in the image corners at 16mm @ f/2.8. A vignetting of 4.6 EV (f-stops) is WAY beyond our usual scale for full format lenses. Unless you have to go for f/2.8 or simply like the effect, you should stop down to at least f/5.6 in order to reduce the light falloff to an moderate degree."

LensTip: https://goo.gl/e2sWwR (Greater than 4 stops vignetting)
"It would be difficult to call this situation other than dramatic. Such high vignetting we haven’t seen so far in our tests. At 16 mm and by f/2.8 in the frame corners disappears 75% of light (−4.07 EV)"

YET, the much pilloried-with-how-they-review-lenses DXO dropped this peach on the world recently:
https://goo.gl/dLKYFt

Can't link it exactly, so either go to Measurements --> Vignetting --> Profiles or just look at the graphic below. They are claiming the vignetting on the III effectively being the same as the II at 16mm f/2.8.

And DPR, partner to DXO (I believe) went so far as to confirm this (more manageable) level of vignetting -- from the DPR author of the lens review, responding to a forum question about it:

"The DXO data reports a 2 and 1/3 stop decrease in the corners when shot at 16mm and an aperture of F2.8; that matches what we saw in the field. Check out our real world image comparison widget and see for yourself, but our observations agree with the DXO data."

So... Does anyone have this new lens yet and can tell us where this really landed?

- A

 

[neuroanatomist]

Would that be the same 16-35/2.8 II that DxOMark 'showed' had similar corner sharpness when stopped down to f/8 as the 17-40/4 they tested? I suppose it's remotely possible that was the case, if they borrowed a diety's personal copy of the 17-40...

 

[ahsanford]

Would that be the same 16-35/2.8 II that DxOMark 'showed' had similar corner sharpness when stopped down to f/8 as the 17-40/4 they tested? I suppose it's remotely possible that was the case, if they borrowed a diety's personal copy of the 17-40...

I'm just wondering if there is something in their workflow that would drive right past this, like leaving peripheral illumination on or something.

- A

 

[privatebydesign]

Some people 'test' lenses by analyzing an image. Some test lenses on a bench without a 'camera'.

Now we might well need a camera to realize any image from a lens, but if that camera induces some of the aberrations then different cameras will get different results.

The guys at Lens Rentals use a bench. Most testers TDP included, look at an image.

We all know that UWA lenses have steeply angled lightwaves that can be impacted by the sensor stack.

I took a sharp intake of breath when I saw the first tests of the new 2.8 lens until I realized I am actually using one with similar vignetting characteristics already, the 11-24. The results from that lenses vignetting cause me zero issues either when I want that vignetting naturally or when I want to remove it in post.

 

[ahsanford]

Some people 'test' lenses by analyzing an image. Some test lenses on a bench without a 'camera'.

Now we might well need a camera to realize any image from a lens, but if that camera induces some of the aberrations then different cameras will get different results.

The guys at Lens Rentals use a bench. Most testers TDP included, look at an image.

We all know that UWA lenses have steeply angled lightwaves that can be impacted by the sensor stack.

I took a sharp intake of breath when I saw the first tests of the new 2.8 lens until I realized I am actually using one with similar vignetting characteristics already, the 11-24. The results from that lenses vignetting cause me zero issues either when I want that vignetting naturally or when I want to remove it in post.

Agree, but LR hasn't sounded off on vignetting. So every review to date that has published vignetting information has done it with images: TDP, PZ, LensTip and now DXO.

I asked Uncle Rog at LR about it but got no reply, unfortunately.

- A

 

[rfdesigner]

I discovered, discussed in a previous thread (if I find it I'll add a link here) that vignetting can vary with focus position.

If DxO weren't careful about focus position when doing vignetting then things could look VERY different. On both my 28f1.8 & 50STM infinity vs min focus would change vignetting on a 2:1 basis.

Althought DxOs results look like a simple mixup and are showing the MkII instead.

 

[Eldar]

I have used my vIII lens a few times now and I do not find the vignetting to be as problematic as I expected. And then I just read this from DPReveiew;

"In terms of vignetting the lens sees a two and one-third stop decrease in the corners at 16mm when shot wide open. This improves by F5.6, but is never entirely eliminated. As you move through the focal range you do see a decrease in overall vignetting and it all be disappears one you stop the lens down to F5.6 at focal lengths beyond 24mm."

If I had not read the numbers from previous reviews, with more than 4 stops vignetting, I probably would have guessed on about 2,5 stops, which is in line with DPReview´s findings.

 

[IglooEater]

That wouldn't be the first time dxo messed up which data goes with which lens. Also wouldn't be the first time their "testing" produced results that were flat out wrong.

 

[LordofTackle]

I have used my vIII lens a few times now and I do not find the vignetting to be as problematic as I expected. And then I just read this from DPReveiew;

If I had not read the numbers from previous reviews, with more than 4 stops vignetting, I probably would have guessed on about 2,5 stops, which is in line with DPReview´s findings.

I agree. I have used my lens now for several shots and also compared it to the mark II and I found it has more vignette than the mark II, but not much. I think Eldar's guess is pretty much spot on, maybe 2.5-3 (just guessed from actual pictures, I have done no scientific tests). Combined with the 1DXII it is a superb combination and the pictures are extremely sharp, especially compared to the mark II lens, and even compared to the excellent 24-70 2.8.

@ Eldar: have you already tried it on the 5DSR? How is the quality on that high res rig?

That wouldn't be the first time dxo messed up which data goes with which lens. Also wouldn't be the first time their "testing" produced results that were flat out wrong.

Well, it wasn't only dxo, as summarized by ahsahnford, almost all reviewers reported this high vignette, including reviewers that I trust, such as TDP and Dustin.

My lesson from this is (and it's not the first time I have to learn it :-[), just try the gear for yourself and YOUR usage. Reviews will give you always only a general hint on the performance of a lens or camera. Also, they are almost always biased in some way or another (and that's not meant negatively, just by preferences or methodology of the testers).

-Sebastian

 

[rrcphoto]

At three different sites, we've seen a consistent read on 4+ stops corner darkening at 16mm f/2.8 with the new 16-35 f/2.8L III:

vignetting depends on your contrast curves if you are looking at your data off a JPG derived from the camera.

TDP and PZ use the default jpg, not sure what DXO does, but it probably looks at a raw file pre contrast curve applied.

PZ has noted before that canon has a relatively strong default jpg contrast curve so vignetting numbers may look worse than say, Nikon.

all the sample images at dpreview for their lens review, show it's not a "problem" unless you want to make it out to be one.

 

[tr573]

That wouldn't be the first time dxo messed up which data goes with which lens. Also wouldn't be the first time their "testing" produced results that were flat out wrong.

Likewise it wouldn't be the first time internet commenters blew a quirk of a piece of gear entirely out of proportion to its actual effect on real world usage.

In 2 years 'is the 16-35 vignette too high for me?' Will be the new 'can I live with the 24-70 focus shift?'

 

[rfdesigner]

I have used my vIII lens a few times now and I do not find the vignetting to be as problematic as I expected. And then I just read this from DPReveiew;

"In terms of vignetting the lens sees a two and one-third stop decrease in the corners at 16mm when shot wide open. This improves by F5.6, but is never entirely eliminated. As you move through the focal range you do see a decrease in overall vignetting and it all be disappears one you stop the lens down to F5.6 at focal lengths beyond 24mm."

If I had not read the numbers from previous reviews, with more than 4 stops vignetting, I probably would have guessed on about 2,5 stops, which is in line with DPReview´s findings.

For anyone that has this lens in their hands, could you do us a favour?

put the lens to infinity and take an image of a flat white evenly lit object (like you ceiling), then do the same with the lens at minimum focus distance, the images of course will be out of focus, this is good.. and post the results?.

A "small-raw" will be more than good enough, but RAW is what I'm after.

 

[Eldar]

@ Eldar: have you already tried it on the 5DSR? How is the quality on that high res rig?

I don´t have images I can share at this time, but I have done some testing and enough pixel peeping to conclude that this lens works very well on the 5DSR.

 

[ahsanford]

Likewise it wouldn't be the first time internet commenters blew a quirk of a piece of gear entirely out of proportion to its actual effect on real world usage.

In 2 years 'is the 16-35 vignette too high for me?' Will be the new 'can I live with the 24-70 focus shift?'

Of course, use the gear and see for yourself, but when three four (forgot about Dustin Abbott, sorry!) trusted reviewers who are fairly transparent with their methods come to a similar conclusion, I thought we could accept that finding as more real than anomaly. So I don't see this like the 24-70 f/4L IS focus shift. I think it's a legitimately confirmed issue, but it may not necessarily punish you, depending on what you shoot:

1) If you are a dedicated landscaper -- I still don't get why you'd pony up the money and carry additional weight of the f/2.8L III lens when at landscape apertures the f/4L IS would the wiser call -- but I see no harm to the vignetting given that you'll be shooting around f/8 - f/11. Non-issue --> buy the lens.

2) If you are buying this lens for why most folks I presume buy a 16-35 f/2.8 lens -- sports, events, concerts, reportage, etc. -- it probably lives and dies set to f/2.8 and then I think it's a judgment call. Some of those arenas of work will live at lower ISO where you might not worry about making so large a correction, but others (possibly concert photography) might not want to push an ISO 3200 or 6400 shot so far.

3) If you are buying it for astro, the coma is good but not great, and again, do you want to boost your corners that much when you are already shooting high ISO for such long exposures?

- A

 

[ahsanford]

If DxO weren't careful about focus position when doing vignetting then things could look VERY different. On both my 28f1.8 & 50STM infinity vs min focus would change vignetting on a 2:1 basis.

Wow. Had no idea about focus distance affecting this. From a link I just read on this that TDP pointed me to:

"All types of vignetting are at their worst with the lens focussed at infinity. At close focus the field of view decreases and the size of the image circle increases. The vignetted area is pushed outward with the image circle and when the focus is close enough the optical or mechanical vignetting will be outside the film frame (or digital sensor). As a matter of fact, it is optical vignetting that determines the size of the image circle in the first place. Natural vignetting improves too, since the exit pupil moves away from the film (at least, with most lens designs)."

#learning

If also true with the 16-35 f/2.8L III, that would mean landscape/astro folks and events/sports folks would have a very different experience. This could very well explain the differences between DXO/DPR and the rest of the reviewers.

- A

 

[YuengLinger]

Likewise it wouldn't be the first time internet commenters blew a quirk of a piece of gear entirely out of proportion to its actual effect on real world usage.

In 2 years 'is the 16-35 vignette too high for me?' Will be the new 'can I live with the 24-70 focus shift?'

Of course, use the gear and see for yourself, but when three four (forgot about Dustin Abbott, sorry!) trusted reviewers who are fairly transparent with their methods come to a similar conclusion, I thought we could accept that finding as more real than anomaly. So I don't see this like the 24-70 f/4L IS focus shift. I think it's a legitimately confirmed issue, but it may not necessarily punish you, depending on what you shoot:

1) If you are a dedicated landscaper -- I still don't get why you'd pony up the money and carry additional weight of the f/2.8L III lens when at landscape apertures the f/4L IS would the wiser call -- but I see no harm to the vignetting given that you'll be shooting around f/8 - f/11. Non-issue --> buy the lens.

2) If you are buying this lens for why most folks I presume buy a 16-35 f/2.8 lens -- sports, events, concerts, reportage, etc. -- it probably lives and dies set to f/2.8 and then I think it's a judgment call. Some of those arenas of work will live at lower ISO where you might not worry about making so large a correction, but others (possibly concert photography) might not want to push an ISO 3200 or 6400 shot so far.

3) If you are buying it for astro, the coma is good but not great, and again, do you want to boost your corners that much when you are already shooting high ISO for such long exposures?

- A

+1 (Nice process! ;D )

 

[tr573]

Likewise it wouldn't be the first time internet commenters blew a quirk of a piece of gear entirely out of proportion to its actual effect on real world usage.

In 2 years 'is the 16-35 vignette too high for me?' Will be the new 'can I live with the 24-70 focus shift?'

Of course, use the gear and see for yourself, but when three four (forgot about Dustin Abbott, sorry!) trusted reviewers who are fairly transparent with their methods come to a similar conclusion, I thought we could accept that finding as more real than anomaly. So I don't see this like the 24-70 f/4L IS focus shift. I think it's a legitimately confirmed issue, but it may not necessarily punish you, depending on what you shoot:

1) If you are a dedicated landscaper -- I still don't get why you'd pony up the money and carry additional weight of the f/2.8L III lens when at landscape apertures the f/4L IS would the wiser call -- but I see no harm to the vignetting given that you'll be shooting around f/8 - f/11. Non-issue --> buy the lens.

2) If you are buying this lens for why most folks I presume buy a 16-35 f/2.8 lens -- sports, events, concerts, reportage, etc. -- it probably lives and dies set to f/2.8 and then I think it's a judgment call. Some of those arenas of work will live at lower ISO where you might not worry about making so large a correction, but others (possibly concert photography) might not want to push an ISO 3200 or 6400 shot so far.

3) If you are buying it for astro, the coma is good but not great, and again, do you want to boost your corners that much when you are already shooting high ISO for such long exposures?

- A

The focus shift is real too - but it all depends on whether or not it will affect your shooting if it's a big deal or not.

I'm saying this more in a "the sky isn't falling" way - the lens has a lot of vignetting in the *extreme* corners wide open @ 16mm. I'm not accusing your OP of this, you're just looking for info, but there is a lot of nonsense going around about how canon dropped the ball totally or the lens is useless and blah de blah on the usual locations (a la dpr)

So I foresee several years of this coming up over and over again when people read these reviews and for some reason can't make up their own mind about whether the specific conditions the quirk exists in matter to and/or effect them or not.

 

[tr573]

If DxO weren't careful about focus position when doing vignetting then things could look VERY different. On both my 28f1.8 & 50STM infinity vs min focus would change vignetting on a 2:1 basis.

Wow. Had no idea about focus distance affecting this. From a link I just read on this that TDP pointed me to:

"All types of vignetting are at their worst with the lens focussed at infinity. At close focus the field of view decreases and the size of the image circle increases. The vignetted area is pushed outward with the image circle and when the focus is close enough the optical or mechanical vignetting will be outside the film frame (or digital sensor). As a matter of fact, it is optical vignetting that determines the size of the image circle in the first place. Natural vignetting improves too, since the exit pupil moves away from the film (at least, with most lens designs)."

#learning

If also true with the 16-35 f/2.8L III, that would mean landscape/astro folks and events/sports folks would have a very different experience. This could very well explain the differences between DXO/DPR and the rest of the reviewers.

- A

It would indicate that the only purpose the lens is really not appropriate for is ultra wide field astro

 

[ahsanford]

Here's how DXO measures it:

Test setup: https://www.dxomark.com/About/In-depth-measurements/DxOMark-testing-protocols/Distortion-LCA-and-vignetting

How they report it out: https://www.dxomark.com/About/In-depth-measurements/Measurements/Vignetting

They do not state how the focus distance is set for the vignetting test.

(They do, however, state that they set focus at infinity for the purposes of the true lens FL measurement that follows all the vignetting shots. So apparently they care about focus distance for that metric.)

- A

 

[TWI by Dustin Abbott]

A lot of wide angle lenses have a heavy vignette, but here's the reasons I made a big deal out of it:

1) I was surprised to see that using Canon's own profile correction in DPP (the lens wasn't supported yet by Adobe when I reviewed it) resulted in visible noise being created along the edges...and that was on the 5D Mark IV, which has the best shadow lift of any Canon body. This was also at ISO 100.

2) Canon really priced this lens high. When a lens is expensive it means that I will personally be more demanding in my expectations than of, say, a $125 50mm f/1.8 STM. The 35L II was also priced really high, but when I got through reviewing it I couldn't really gainsay it. The lens delivered (and incidentally I'm expecting a delivery of my own copy any day now). I wasn't quite as wowed by the 16-35L III, and when it comes in at double+ the cost of the already good 16-35 f/4L IS or Tamron 15-30 VC, well...

I don't think this is all that people should fixate on, as it may not really be a big deal in your real world shooting, but it is a real optical drawback. It's a great a lens in so many ways, but that price...uggg!