Optical Limits reviews RF 24-105 f/4-7.1 STM

AlanF

Stay at home
CR Pro
Aug 16, 2012
8,900
12,214

Klaus Schroiff doesn't mince his words over this one on one of my favourite review sites.

“If you purchased an R5 with this lens, you are nuts. In real life, this lens is used on entry to mid-level cameras with a resolution between 20-30mp. And this is putting much less stress on the lens quality. At 20mp (e.g. the EOS R6) you could mount a coke bottle and the results would still be mostly Ok (sort of). Below are the MTFs (auto-corrected) based on the 30mp sensor of the original EOS R. So as you can see, the results aren't QUITE as embarrassing as they are on the EOS R5. Even the corners are still fairly good. Longer focal lengths are perfectly fine on lower resolution cameras anyway. “

“In absolute terms, it's clearly a sub-standard lens. Its native characteristics range from good to downright terrible. The miserable vignetting and distortions have to be seen to be believed. Conversely, the resolution is actually pretty decent - which is a bit of a pity because you just won't use this lens without auto-correction. “

Mind you, it doesn't look any worse than his testing of the EF 24-105mm f/4 L II at 50 Mpx.

 
Last edited:
  • Like
Reactions: Chaitanya

Del Paso

M3 Singlestroke
Aug 9, 2018
1,321
1,447
Same situation with several new cars: if, for testing, you manipulate the ECU and intentionally switch off the ESP (electronic stability program), you're in for a few bad surprises.
And I don't mean the cheap Korean minis, but some really expensive luxury limousines. Electronics misused to compensate for poor engineering.
I hope for Canon this remains an exception!
 

Sporgon

5% of gear used 95% of the time
CR Pro
Nov 11, 2012
4,302
977
Yorkshire, England
Same situation with several new cars: if, for testing, you manipulate the ECU and intentionally switch off the ESP (electronic stability program), you're in for a few bad surprises.
And I don't mean the cheap Korean minis, but some really expensive luxury limousines. Electronics misused to compensate for poor engineering.
I hope for Canon this remains an exception!
I relate to this; I had a BMW 330i, can't remember the model designation but it's probably about three models back now. Took it round Donnington circuit and had all ESP off. Jeeez, what an experience. It understeered like a mad dog that had just scented a bitch and was only going in a straight line ahead ! Now the Honda S2000 - that was a dream on the track.
 

Del Paso

M3 Singlestroke
Aug 9, 2018
1,321
1,447
I relate to this; I had a BMW 330i, can't remember the model designation but it's probably about three models back now. Took it round Donnington circuit and had all ESP off. Jeeez, what an experience. It understeered like a mad dog that had just scented a bitch and was only going in a straight line ahead ! Now the Honda S2000 - that was a dream on the track.
You just named the brand I was thinking of, same results with its very big brother, which even experienced a rear axle damage according to a German car magazine...
As to the Honda S2000: I totally agree !
 
  • Like
Reactions: Sporgon

SteveC

R5
CR Pro
Sep 3, 2019
2,481
2,346
Same situation with several new cars: if, for testing, you manipulate the ECU and intentionally switch off the ESP (electronic stability program), you're in for a few bad surprises.
And I don't mean the cheap Korean minis, but some really expensive luxury limousines. Electronics misused to compensate for poor engineering.
I hope for Canon this remains an exception!
Have you heard about the huge barrel distortion and vignetting of the 24-240, also corrected in camera?

(This is not an uncommon phenomenon, though in some cases it's not a misuse of electronics. CD players for the car that use a huge memory buffer to compensate for not having a thousand dollars of mechanical stabilization come to mind as a valid use. "Atomic" watches that suddenly lose minutes a day without notice when they don't get their signal are a misuse; an inherently accurate watch, or at the very least one much more accurate than that, is cheap nowadays. [I've won arguments about the correct time with people proud of their atomic watches; I have a 1990's Citizen Navihawk which gains or loses less than a second a month without the aid of an atomic clock.])
 
  • Like
Reactions: Del Paso

Czardoom

EOS RP
Jan 27, 2020
359
770
Have you heard about the huge barrel distortion and vignetting of the 24-240, also corrected in camera?

(This is not an uncommon phenomenon, though in some cases it's not a misuse of electronics. CD players for the car that use a huge memory buffer to compensate for not having a thousand dollars of mechanical stabilization come to mind as a valid use. "Atomic" watches that suddenly lose minutes a day without notice when they don't get their signal are a misuse; an inherently accurate watch, or at the very least one much more accurate than that, is cheap nowadays. [I've won arguments about the correct time with people proud of their atomic watches; I have a 1990's Citizen Navihawk which gains or loses less than a second a month without the aid of an atomic clock.])
I was somewhat befuddled when the RF 24-240mm was released and so many reviewers (and users) were dismayed at the fact that you need to auto-correct, and the lens is actually designed so that you must auto-correct. Numerous mirrorless lenses in other systems (Sony, Olympus and probably others) have been made this way for a number of years and no one seemed to mind. The Olympus 12-100mm f/4 is regarded as a really fine lens with fantastic reviews - and most reviewers might barely mention in passing that, "oh by the way, the camera auto-corrects for the large barrel distortion that would happen if you use the lens without auto-correction." Then Canon comes out with a lens designed with auto-correction needed and people went crazy!

So, no, it is not a misuse of electronincs or poor engineering. It is engineering that uses the latest technologies to create something that could not have been created before. Unless, of course, you want a much bigger, heavier and more expensive lens that may not perform as well.
 

neuroanatomist

I post too Much on Here!!
CR Pro
Jul 21, 2010
26,131
4,732
24-105 L-series lenses are compromises between convenience and cost vs. IQ, the non-L is just more of a compromise. I’ve only used L versions (original EF and now RF) and always found them to be adequate. Although I started with the 24-105L on my 5DII, I switched to the 24-70/2.8L when I got the 1D X. Similarly, I started with the RF24-105/4L with my EOS R, at some point around getting the R3 I’ll get the 24-70/2.8 IS or 28-70/2.
 
  • Like
Reactions: Del Paso

JPAZ

If only I knew what I was doing.....
CR Pro
Sep 8, 2012
1,138
521
..... I switched to the 24-70/2.8L when I got the 1D X. Similarly, I started with the RF24-105/4L with my EOS R, at some point around getting the R3 I’ll get the 24-70/2.8 IS or 28-70/2.
Since the R5 has IBIS, the 24-70 f/2.8L MKii might actually be "better" than it was (and it was great) on the 5dIV.
 

Joules

doom
CR Pro
Jul 16, 2017
1,742
2,145
Hamburg, Germany
Critizing a budget lens for requiring auto corrections to be used is like critizing a camera like the R5 for not being able to record unlimited 8K video. If you ignore the specific use case equipment was designed for when evaluating how well it does, what are you even evaluating?

In other words, if the lens is unusable without correction, who really cares about 'native characteristics'? Especially since I guess these include portions of the image that are cropped off anyway when the intended adjustments are applied?
 

Maximilian

The dark side - I've been there
CR Pro
Nov 7, 2013
3,648
2,869
Germany
... At 20mp (e.g. the EOS R6) you could mount a coke bottle and the results would still be mostly Ok (sort of). ...
To me that kind of phrase is just sensational and trying to be cool and to make the audience laugh.
To me this disqualifies the reviewer. And I think Klaus Schroiff doesn't have the need to. Sad.
 

stevelee

FT-QL
CR Pro
Jul 6, 2017
2,180
887
Davidson, NC
24-105 L-series lenses are compromises between convenience and cost vs. IQ, the non-L is just more of a compromise. I’ve only used L versions (original EF and now RF) and always found them to be adequate. Although I started with the 24-105L on my 5DII, I switched to the 24-70/2.8L when I got the 1D X. Similarly, I started with the RF24-105/4L with my EOS R, at some point around getting the R3 I’ll get the 24-70/2.8 IS or 28-70/2.
I got the non-L version as a kit lens with my 6D2. I have been very pleasantly surprised by it. I have used it as my general purpose lens, and haven’t found anything to complain about.
 

Sporgon

5% of gear used 95% of the time
CR Pro
Nov 11, 2012
4,302
977
Yorkshire, England
To me that kind of phrase is just sensational and trying to be cool and to make the audience laugh.
To me this disqualifies the reviewer. And I think Klaus Schroiff doesn't have the need to. Sad.
Agreed. 20 mp on FF is actually pretty high resolution in its own right. I recently compared my 26mp RP to the 50mp 5DS in a ‘landscape’ type scene, using my Tamron 85mm VS, which at f/5.6 is apparently my highest resolving lens. In all honesty there wasn’t much difference other than the larger output of the 5DS.
 
  • Like
Reactions: privatebydesign

InchMetric

Switched from Nikon. Still zooming the wrong way.
CR Pro
Jun 22, 2021
159
187
Summary: This light and inexpensive lens is not for those who insist on turning off the advanced electronic features the lens (or car) was designed for.
 

neuroanatomist

I post too Much on Here!!
CR Pro
Jul 21, 2010
26,131
4,732
Summary: This light and inexpensive lens is not for those who insist on turning off the advanced electronic features the lens (or car) was designed for.
That’s not the point. The point is that those lens corrections are not ‘free’ – you cannot correct distortion without loss of sharpness, you cannot correct vignetting without increased noise, etc.

This light and inexpensive lens is not for those who insist on the best possible image quality.
 

Czardoom

EOS RP
Jan 27, 2020
359
770
That’s not the point. The point is that those lens corrections are not ‘free’ – you cannot correct distortion without loss of sharpness, you cannot correct vignetting without increased noise, etc.

This light and inexpensive lens is not for those who insist on the best possible image quality.
I'm guessing that most every light and inexpensive lens is not giving you the best possible image quality. Those that are designed with no auto-corrections would likely be even less sharp in the corners and most photographers would be post processing to reduce the vignetting with the same (or even worse) results in terms of noise. So, yes, expensive "L" lenses will give you the best possible image quality. But every lens design will have some compromises. Even "L" lenses have some distortion, vignetting and loss of sharpness away from the center of the lens. So one could argue that no lens is really designed for the best possible image quality, unless it makes no compromises in terms of weight, size and cost. Those lenses are probably few and far between.

While Canon has designed a couple lenses with auto-correction that are lower priced, non L, consumer lenses, other brands such as Olympus have done so with higher end "pro" lenses such as their 12-100mm f/4. The reviews for this lens are very positive and having owned it for many years, I would say that the quality equals any of the Canon L lenses I have owned. No doubt the auto-corrections for distortion reduce sharpness, but do you really think that if Olympus had designed a bigger, heavier lens without the auto-correction, that it would be sharper in the corners than this lens? Or is sharpening a part of the auto-correction? In my opinion, it is still possible to design lenses to be auto-corrected and still deliver very high image quality.

Now, in the case of the Canon RF 24-105 non-L that is under discussion, I have owned this lens and found the image quality a little disappointing, but not at all terrible. At full screen size (just a standard 1800 dot computer screen) I had to pixel peep a little to see the reduction in corner sharpness compared to the EF 24-70 f/4 L. I've owned two samples of the lens and the biggest issue I had was that both were noticeably de-centered, which had nothing to do with the auto-correction.

Just curious for those that insist on the best image quality...if you are purely a Canon shooter and found that the new Nikon Z lenses gave you better image quality, would you switch systems? Or would your Canon lenses still be "good enough." How about compared to medium format? Or is "best possible image quality" suddenly a relative term after all?
 
  • Like
Reactions: Nemorino

neuroanatomist

I post too Much on Here!!
CR Pro
Jul 21, 2010
26,131
4,732
But every lens design will have some compromises. Even "L" lenses have some distortion, vignetting and loss of sharpness away from the center of the lens.
All lenses are compromises, mainly among image quality (which is somewhat subjective), convenience, size/weight, and cost. Primes have much less distortion than zooms, but zooms are more convenient. A Zeiss Otus delivers excellent IQ but manual focus is inconvenient. My 600/4 delivers much better IQ than a 100-400/500, but at a much larger size/weight and a much higher cost. My EOS M system is very portable and inexpensive, the IQ is decent but not great. The wonderful thing is that we have so many choices.
 
  • Like
Reactions: JPAZ and stevelee

InchMetric

Switched from Nikon. Still zooming the wrong way.
CR Pro
Jun 22, 2021
159
187
That’s not the point. The point is that those lens corrections are not ‘free’ – you cannot correct distortion without loss of sharpness, you cannot correct vignetting without increased noise, etc.

This light and inexpensive lens is not for those who insist on the best possible image quality.
Actually, that was my point. You appear to have a different point that no one could disagree with.

Mine has taken some fine images of my kids sledding in the snow when I didn’t want one of my heavier lenses on hand. Even those who insist upon the best have varying needs. I also own a 50 1.8 along side the 1.2 for similar reasons. For tue cost of sales tax on the 1.2.
 

Jethro

EOS R
CR Pro
Jul 14, 2018
589
492
To me that kind of phrase is just sensational and trying to be cool and to make the audience laugh.
To me this disqualifies the reviewer. And I think Klaus Schroiff doesn't have the need to. Sad.
A bit unfair on Klaus - he is making the point that a lens like this will be 'better' on a lower MP body. He's one of the few reviewers who often tests on different bodies to show the practical effect of the difference in MP. Of course, R5 is more of a 'torture test' for a non-L kit lens like this, but that's the point of scientific evaluation and comparison. He also tests with and without auto-correction, so you get to see exactly what you're getting in all cases. He did seem genuinely shocked at the non-corrected version though ....
 
Last edited: