hi there! i am a just about 1 week old owner of the Rf 15-35.
In my use, mostly walking about and a few portraits, i find that the lens tends to underexpose by about half or a full stop! is this a behaviour of the lens?
hoping to get some insight!
thank you
Luka
August 9, 2024 4:21 pm
Beautiful photos! Love the detailed review, as always, but the “Canon’s current mirrorless cameras always autofocus at maximum aperture regardless of which aperture value you have selected.”, is not true. R6MkII can be set to preview with aperture, in which case it focuses with aperture set on given value. Cheers!
Christopher
December 7, 2023 10:06 am
Thanks Spencer for another interesting review!
The vignetting figures will be overtaken when you test the Sony 24-105mm f/4 lens, OpticalLimits found it had 5.35EV vignetting at 24mm f/4 www.opticallimits.com/sonya…ss?start=1 That too was with distortion correction disabled, I guess some lenses nowadays are designed only to be used with distortion correction.
Looking at the optical comparison charts, I noticed that the Canon 15-35mm had lower centre sharpness at f/2.8 that the Sony and Nikon zooms, sometimes considerably so. Did this show up in real world shooting, please? From the charts, I must confess I’m slightly puzzled that the conclusion endorsed the sharpness at all apertures, but as I have never used the lens, I presume the difference wasn’t easily apparent in real-world photos.
Thanks for the heads up. On lenses with such extreme vignetting, I think I’ll need to show vignetting both pre- and post-distortion correction in future reviews.
At a certain point, lenses are simply sharp enough that my takeaway will be positive regardless of the competition. That said, the results you describe are why I ultimately gave this lens a 4.5/5 score for sharpness, while giving the Nikon Z 14-24mm f/2.8 a 5.0/5 sharpness score.
Pieter Kers
December 7, 2023 5:38 am
It seems Nikon follows a different path in lensdesign compared to Canon and Sony. Nikons lenses are larger and subsequently have less vignetting. Vignetting over 3 stops & dark corners ! How can you still call it an f/2.8 lens I wonder? Or if only the central area gets light don’t call it 15mm FF but 15mm APS.
Mirrorless lenses also tend to have more distortion than most DSLR lenses. It’s a different design philosophy in general – maximizing sharpness and accepting some consequences in areas that can be corrected in post.
Although, there are limits to everything. Too much vignetting and the corners have more noise; too much distortion and you can’t stretch it back without losing sharpness.
Dmitry
December 6, 2023 10:01 am
The Vignetting value of Canon with some RF lenses Canon RF 28-70mm f/2 USM L – 28mm – 2.93 Canon RF 24-70mm f/2.8 USM L IS – 24 mm – 3 Canon RF 50mm f/1.8 STM – 3.57 Canon RF 16mm f/2.8 STM – 8.32 Canon RF 24-240mm f/4-6.3 USM IS – 24 mm – 12.5 Canon RF 24-105mm f/4-7.1 STM IS – 24 mm – 12.6 And also non-switchable shadow noise reduction in RAW at ISO values from 100 to 800 for R-series cameras.
If our lab testing matches those values, I don’t even know what to say. 12.6 stops of vignetting is beyond just vignetting, in my opinion. It should have been labeled as an APS-C lens at that point.
I set our vignetting charts to have a Y-axis of 4.0, since no previous lens exceeded that value. I’ll need to figure out something to do about that in our upcoming reviews.
Marketing. 24-105 4-7.1 will sell better than 28-70 4-6.3. 16 mm 2.8 in the DX version will sell worse than FF. 28-70 2.8 sounds worse than 28-70 2.0. 24-240 will be more in demand than 50-240. That’s why they constantly change the controls on the R cameras? What prevented them from installing a matrix stabilizer in R or R8? Or why are they blocking all third-party autofocus lens manufacturers? Canon has a dominant position in the market and can dictate its terms. What are they actively using
8-12 stops of vignetting doesn’t pass the smell test for me, but we’ll see when you publish your reviews. I would find it shocking people aren’t complaining about it. 12 stops is 0.02% of the amount of light. That is absurd.
I trust Optical Limits in general. And the situation isn’t as bad as 12.6 stops implies – the extreme vignetting is confined to the very tips of the corners, and just correcting distortion at 24mm is enough to put the lens back into reasonable vignetting territory, according to my reading of the article.
So that basically means the lens was designed to be used with distortion correction enabled. Not everyone will like this design choice, but I agree it is not “this lens has 12 stops of vignetting”-bad.
That makes much more sense. What camera are you testing the Canon (and Sony?) lenses with? Are the chromatic aberration, sharpness, etc. results directly compatible across different brands or should we be looking at each brand in isolation?
Not that it matters since I’m not going to adapt a Canon RF lens to Z mount, but more for general information.
We use the Canon EOS R5 and Sony a7R IIIA, with small correction factors to match our existing Nikon lens tests (which are on the Z7). All of our reviews are fully comparable across brands! That goes for sharpness, distortion, vignetting, CA, and everything else.
It’s actually the whole reason why it took us so long to start reviewing Canon and Sony lenses properly in the first place. We had to do a lot of testing in order to ensure they would be comparable to our existing reviews.
That’s great! For testing methodology, as the camera manufacturers move to higher resolutions (…assuming they do so? I feel like Sony and Canon’s APS-C cameras might be pointing in that direction) what do you intend to do? I’m guessing this is something you’ve dealt with in the past given how long this website has been around.
Good question – we’re planning to stick to these camera bodies for the foreseeable future. 45(ish) megapixels is enough to parse out any meaningful lens differences, and all of our reviews are based on them already. Switching cameras at this point would either mean our results could no longer be compared, or that we would need to redo hundreds of tests.
That “non-switchable noise reduction” has zero impact on image quality, except from tests conducted in laboratories. Those vignetting values are at the very edge of the frame of the “uncorrected” image. For example 24-105mm f4-7.1 is actually wider than 24mm when uncorrected. When you correct it to 24mm, then vignetting is quite normal. I tried it with DxO (RAW image) and before all corrections it is actually about 22-23mm and when corrected vignetting is about 3EV in the “VERY” corners. Similar story with 24-240mm STM (which is a very sharp lens by the way, despite 10x zoom). You really need to understand what is going on before throwing numbers around.
SSM
December 6, 2023 9:34 am
Spencer, thanks for another great review. Your coverage is both comprehensive and fair, and the sample photos are impressive. On a side note: the lens is well damped, not dampened. Dampened means wet or moisturized! Cheers.
Thanks, SSM! Glad you liked the review, and you taught me something new. I wasn’t aware that it was a misuse of “dampen” but you’re absolutely right. I’ll change it as soon as I get back to my computer later tonight.
dampen verb • past tense: dampened; • past participle: dampened.
1. make slightly wet. the fine rain dampened her face
2. make less strong or intense. nothing could dampen her enthusiasm
2a. reduce the amplitude of (a sound source). slider switches on the mixers can dampen the drums
— Oxford Languages
damp verb • past tense: damped; • past participle: damped.
1. make (something) slightly wet. damp a small area with water
2. make a fire burn less strongly by reducing the flow of air to it. he damped down the fire for the night
2a control or restrain a feeling or a situation. she tried to damp down her feelings of despair
3. reduce or stop the vibration of (the strings of a piano or other musical instrument) so as to reduce the volume of sound. the muted notes should be produced by damping the strings at the seventh position
3a. PHYSICS progressively reduce the amplitude of (an oscillation or vibration). concrete structures damp out any vibrations
Who would have a priority taking landscape photos af 15mm f2.8? I think you focus on vignetting too much. The lens is excellent axcept from vignetting at widest angle at widest aperture which is mostly used for creative purposes anyway.
hi there! i am a just about 1 week old owner of the Rf 15-35.
In my use, mostly walking about and a few portraits, i find that the lens tends to underexpose by about half or a full stop! is this a behaviour of the lens?
hoping to get some insight!
thank you
Beautiful photos!
Love the detailed review, as always, but the “Canon’s current mirrorless cameras always autofocus at maximum aperture regardless of which aperture value you have selected.”, is not true.
R6MkII can be set to preview with aperture, in which case it focuses with aperture set on given value.
Cheers!
Thanks Spencer for another interesting review!
The vignetting figures will be overtaken when you test the Sony 24-105mm f/4 lens, OpticalLimits found it had 5.35EV vignetting at 24mm f/4 www.opticallimits.com/sonya…ss?start=1 That too was with distortion correction disabled, I guess some lenses nowadays are designed only to be used with distortion correction.
Looking at the optical comparison charts, I noticed that the Canon 15-35mm had lower centre sharpness at f/2.8 that the Sony and Nikon zooms, sometimes considerably so. Did this show up in real world shooting, please? From the charts, I must confess I’m slightly puzzled that the conclusion endorsed the sharpness at all apertures, but as I have never used the lens, I presume the difference wasn’t easily apparent in real-world photos.
Thanks for the heads up. On lenses with such extreme vignetting, I think I’ll need to show vignetting both pre- and post-distortion correction in future reviews.
At a certain point, lenses are simply sharp enough that my takeaway will be positive regardless of the competition. That said, the results you describe are why I ultimately gave this lens a 4.5/5 score for sharpness, while giving the Nikon Z 14-24mm f/2.8 a 5.0/5 sharpness score.
It seems Nikon follows a different path in lensdesign compared to Canon and Sony.
Nikons lenses are larger and subsequently have less vignetting.
Vignetting over 3 stops & dark corners ! How can you still call it an f/2.8 lens I wonder?
Or if only the central area gets light don’t call it 15mm FF but 15mm APS.
Mirrorless lenses also tend to have more distortion than most DSLR lenses. It’s a different design philosophy in general – maximizing sharpness and accepting some consequences in areas that can be corrected in post.
Although, there are limits to everything. Too much vignetting and the corners have more noise; too much distortion and you can’t stretch it back without losing sharpness.
The Vignetting value of Canon with some RF lenses
Canon RF 28-70mm f/2 USM L – 28mm – 2.93
Canon RF 24-70mm f/2.8 USM L IS – 24 mm – 3
Canon RF 50mm f/1.8 STM – 3.57
Canon RF 16mm f/2.8 STM – 8.32
Canon RF 24-240mm f/4-6.3 USM IS – 24 mm – 12.5
Canon RF 24-105mm f/4-7.1 STM IS – 24 mm – 12.6
And also non-switchable shadow noise reduction in RAW at ISO values from 100 to 800 for R-series cameras.
If our lab testing matches those values, I don’t even know what to say. 12.6 stops of vignetting is beyond just vignetting, in my opinion. It should have been labeled as an APS-C lens at that point.
I set our vignetting charts to have a Y-axis of 4.0, since no previous lens exceeded that value. I’ll need to figure out something to do about that in our upcoming reviews.
Marketing.
24-105 4-7.1 will sell better than 28-70 4-6.3.
16 mm 2.8 in the DX version will sell worse than FF.
28-70 2.8 sounds worse than 28-70 2.0.
24-240 will be more in demand than 50-240.
That’s why they constantly change the controls on the R cameras?
What prevented them from installing a matrix stabilizer in R or R8?
Or why are they blocking all third-party autofocus lens manufacturers?
Canon has a dominant position in the market and can dictate its terms. What are they actively using
8-12 stops of vignetting doesn’t pass the smell test for me, but we’ll see when you publish your reviews. I would find it shocking people aren’t complaining about it. 12 stops is 0.02% of the amount of light. That is absurd.
I agree. A claim such as 8–12 stops of vignetting should be met with a resounding:
Citation needed !
The page I think he is referring to is this – www.opticallimits.com/canon…71?start=1
I trust Optical Limits in general. And the situation isn’t as bad as 12.6 stops implies – the extreme vignetting is confined to the very tips of the corners, and just correcting distortion at 24mm is enough to put the lens back into reasonable vignetting territory, according to my reading of the article.
So that basically means the lens was designed to be used with distortion correction enabled. Not everyone will like this design choice, but I agree it is not “this lens has 12 stops of vignetting”-bad.
Exactly. We need to be careful when looking at numbers, and read the text with it.
That makes much more sense. What camera are you testing the Canon (and Sony?) lenses with? Are the chromatic aberration, sharpness, etc. results directly compatible across different brands or should we be looking at each brand in isolation?
Not that it matters since I’m not going to adapt a Canon RF lens to Z mount, but more for general information.
We use the Canon EOS R5 and Sony a7R IIIA, with small correction factors to match our existing Nikon lens tests (which are on the Z7). All of our reviews are fully comparable across brands! That goes for sharpness, distortion, vignetting, CA, and everything else.
It’s actually the whole reason why it took us so long to start reviewing Canon and Sony lenses properly in the first place. We had to do a lot of testing in order to ensure they would be comparable to our existing reviews.
That’s great! For testing methodology, as the camera manufacturers move to higher resolutions (…assuming they do so? I feel like Sony and Canon’s APS-C cameras might be pointing in that direction) what do you intend to do? I’m guessing this is something you’ve dealt with in the past given how long this website has been around.
Good question – we’re planning to stick to these camera bodies for the foreseeable future. 45(ish) megapixels is enough to parse out any meaningful lens differences, and all of our reviews are based on them already. Switching cameras at this point would either mean our results could no longer be compared, or that we would need to redo hundreds of tests.
That “non-switchable noise reduction” has zero impact on image quality, except from tests conducted in laboratories.
Those vignetting values are at the very edge of the frame of the “uncorrected” image. For example 24-105mm f4-7.1 is actually wider than 24mm when uncorrected. When you correct it to 24mm, then vignetting is quite normal. I tried it with DxO (RAW image) and before all corrections it is actually about 22-23mm and when corrected vignetting is about 3EV in the “VERY” corners.
Similar story with 24-240mm STM (which is a very sharp lens by the way, despite 10x zoom).
You really need to understand what is going on before throwing numbers around.
Spencer, thanks for another great review. Your coverage is both comprehensive and fair, and the sample photos are impressive.
On a side note: the lens is well damped, not dampened. Dampened means wet or moisturized! Cheers.
Thanks, SSM! Glad you liked the review, and you taught me something new. I wasn’t aware that it was a misuse of “dampen” but you’re absolutely right. I’ll change it as soon as I get back to my computer later tonight.
I think they’re both correct (though damped seems to be used more in technical contexts and dampened in economics (?)).
dampen verb
• past tense: dampened;
• past participle: dampened.
1. make slightly wet.
the fine rain dampened her face
2. make less strong or intense.
nothing could dampen her enthusiasm
2a. reduce the amplitude of (a sound source).
slider switches on the mixers can dampen the drums
— Oxford Languages
damp verb
• past tense: damped;
• past participle: damped.
1. make (something) slightly wet.
damp a small area with water
2. make a fire burn less strongly by reducing the flow of air to it.
he damped down the fire for the night
2a control or restrain a feeling or a situation.
she tried to damp down her feelings of despair
3. reduce or stop the vibration of (the strings of a piano or other musical instrument) so as to reduce the volume of sound.
the muted notes should be produced by damping the strings at the seventh position
3a. PHYSICS progressively reduce the amplitude of (an oscillation or vibration).
concrete structures damp out any vibrations
— Oxford Languages
a bit scary but first class attention to details!
Hi Lukasz,
My sincerest best wishes to you for the festive season and the New Year.
Kindest regards,
Pete
You need to choose what you’ll sacrifice in lens designs. Having bad vignetting at 15mm f2.8 is actually great if it is the only sacrifice.
It all depends what your priorities are.
Who would have a priority taking landscape photos af 15mm f2.8? I think you focus on vignetting too much. The lens is excellent axcept from vignetting at widest angle at widest aperture which is mostly used for creative purposes anyway.
Everyone interested in shooting the night sky?