This is the Canon RF lens roadmap

Jul 21, 2010
31,089
12,855
The RF 24mm is overpriced for what it is - a lens that REQUIRES correction to work properly.
The RF 14-35/4 L also requires optical correction of the distortion to work properly. When corrected with DxO PhotoLab, I get images at the wide end equivalent to ~13.5mm on my 11-24/4L (which has almost no geometric distortion at 13-14mm), and equivalent corner sharpness. The 14-35/4 is relatively small and light, and takes convenient 77mm front filters (meaning it makes a great travel kit with the 24-105/4L and 100-500L).
 
  • Like
Reactions: 2 users
Upvote 0

SwissFrank

1N 3 1V 1Ds I II III R R5
Dec 9, 2018
503
338
a lens that REQUIRES correction to work properly
You say that like it's a bad thing. It actually doesn't matter. It blurs by at most .5 pixels, which would only be visible on an R5 if you had 130 lp/mm resolution, which no lens does. And with 24mm you're not even going to have most of the shot focused well enough to worry about 1-pixel-wide details.

Even pixel-peeping I doubt you'd be able to show the loss, not even in a staged, unrealistic scenario.

Really, post an example. Even explain textually even a hypothetical, staged example where they thing you're mocking even could possibly be detected much less matter.
 
  • Like
Reactions: 1 user
Upvote 0

Del Paso

M3 Singlestroke
CR Pro
Aug 9, 2018
3,295
4,182
Lol, I've shot Canon for years so your comment made me laugh. Their practices are not cool these days, and yes AF third party lenses are important to have especially when Canon isn't providing good options themselves at a reasonable price. The RF 24mm is overpriced for what it is - a lens that REQUIRES correction to work properly. Canon is cheaping out on us and apologists like you can see no wrong with them. Thankfully this is an open forum so I can have whatever opinion I want!
You dislike optical corrections?
I guess you also dislike Lightroom, DPP , color adjustments, vignetting reduction, CA suppression etc...
What matters is what you obtain with RF L lenses, and that's excellent. No matter how you obtain it!
 
  • Like
  • Love
Reactions: 3 users
Upvote 0
You say that like it's a bad thing. It actually doesn't matter. It blurs by at most .5 pixels, which would only be visible on an R5 if you had 130 lp/mm resolution, which no lens does. And with 24mm you're not even going to have most of the shot focused well enough to worry about 1-pixel-wide details.

Even pixel-peeping I doubt you'd be able to show the loss, not even in a staged, unrealistic scenario.

Really, post an example. Even explain textually even a hypothetical, staged example where they thing you're mocking even could possibly be detected much less matter.
Wait, you're actually defending software correction being required instead of optical? Wow, destroying pixels is apparently a great idea now! Lol, some people...
 
Upvote 0
Jul 21, 2010
31,089
12,855
Wait, you're actually defending software correction being required instead of optical? Wow, destroying pixels is apparently a great idea now! Lol, some people...
Wait, you're actually defending lenses being larger, heavier and more expensive because they need to rely on optical corrections instead of leveraging modern algorithms to correct geometry with IQ loss that is minimal to undetectable? Wow, paying more to carry bigger heavier lenses that deliver similar IQ is apparently a great idea now! Lol, some people...
 
  • Haha
  • Like
Reactions: 2 users
Upvote 0

SwissFrank

1N 3 1V 1Ds I II III R R5
Dec 9, 2018
503
338
Wait, you're actually defending software correction being required instead of optical? Wow, destroying pixels is apparently a great idea now! Lol, some people...
I've asked you for a specific example, doesn't even have to be real-world, where you can notice this "destroying pixels." I've suggested it doesn't even have to be something visible at any realistic use case for the resulting image; it's fine if you want to supply an example where the "destruction" can be observed even looking at individual pixels. Say at 800% mag in Photoshop or what have you.

I've made clear it doesn't even have to be realistic or even a photo as such, just explain textually a case where this "destroying pixels" would be visible.

Do you have such an example image? Or can you think of such an example and explain it to us?
 
Upvote 0
Wait, you're actually defending lenses being larger, heavier and more expensive because they need to rely on optical corrections instead of leveraging modern algorithms to correct geometry with IQ loss that is minimal to undetectable? Wow, paying more to carry bigger heavier lenses that deliver similar IQ is apparently a great idea now! Lol, some people...

This review sums up why its an issue quite well:


Older EF lenses were not overly large and heavy and had optical distortion. Other brands don't seem to have this issue. Canon is being lazy.
 
Upvote 0
Nov 2, 2020
132
134
I've asked you for a specific example, doesn't even have to be real-world, where you can notice this "destroying pixels." I've suggested it doesn't even have to be something visible at any realistic use case for the resulting image; it's fine if you want to supply an example where the "destruction" can be observed even looking at individual pixels. Say at 800% mag in Photoshop or what have you.

I've made clear it doesn't even have to be realistic or even a photo as such, just explain textually a case where this "destroying pixels" would be visible.

Do you have such an example image? Or can you think of such an example and explain it to us?

I had similar questions, so I attempted to verify pixel destruction as a result of software distortion correction. On a couple of occasions (using the RF 16 and the RF 14-35) I've compared corner detail of a RAW uncorrected image with the same corrected image.

In those cases, I was unable to observe any significant difference between image detail of the corrected and uncorrected images. What I did observe was a slightly smaller corner in the RAW image due to the effects of barrel distortion. In order to compare those corner images at the same approximate size, the RAW image required more magnification to match the size of the corrected image. So, a takeaway was that any apparent softness in the corners of highly software corrected images was possibly a result of the additional magnification required in the software correction process.

And since the detail of the two images appeared much the same when viewed at the same size, the software distortion correction itself wasn't directly reducing image quality; rather it was the need for additional magnification to offset the barrel distortion. So based on my observations, I would say the statement that software distortion correction is destructive to pixels is misleading. Rather, I'd say that the additional magnification required to correct for barrel distortion simply highlights (magnifies) the detail that was lost as a result of the reduced magnification caused by barrel distortion in the corner.

The end result is the same: Substantial barrel distortion contributes to loss of detail in the image corner. I'm just not convinced that software correction (beyond necessary increases in magnification) reduces the image quality any further. Ultimately, it's up to the individual to assess whether the final quality is acceptable, and whether to waste time fretting over how the results were achieved.
 
Last edited:
  • Love
  • Like
Reactions: 2 users
Upvote 0

SwissFrank

1N 3 1V 1Ds I II III R R5
Dec 9, 2018
503
338
And since the detail of the two images appeared much the same when viewed at the same size, the software distortion correction itself wasn't directly reducing image quality; rather it was the need for additional magnification to offset the barrel distortion. So based on my observations, I would say the statement that software distortion correction is destructive to pixels is misleading.
OK, but it's not as if magnification changes by a factor of 2 or something, does it?

And, the alternative to a lens that requires software correction isn't the same lens that doesn't need software correction. Instead, it's a lens that, in order to to keep the geometry perfect, had to have compromises made somewhere else: either more axial CA, say, or more coma, or more size weight price or complexity that weakens it, or less sharpness. This last, especially, means that while the (moderate) extra magnification you describe would certainly lower sharpness over this non-existent mythical lens, it may nonetheless be sharper than the actual alternative lens that could have been designed!

And this is where nunataks makes his error: he is making a confident statement that could only be made if you knew what this actual alternative lens was. Yet no-one knows that except perhaps Canon.

Engineering is the art of tradeoffs. As an engineer I can happily and often trivially improve ANY parameter of my system... at the expense of others. You want less vignetting? I can do it in an hour. If you agree to a 122mm front filter I can eliminate vignetting completely. And the lens will weigh 4x more. Oh, you agree to the filter size but not the weight? No problem! I will keep the weight down to only 50% heavier, and make it out of carbon fiber. It will cost 5x more. Oh, you can't accept the cost increasing more than 50%? No problem! I will just cut the thinness of the components in half, and make all the lenses plastic! It will break if you drop it, and it now has lots of chromatic aberations, but I have now met your vignetting target, AND weight target, AND price point. And it's not even 5PM.
 
Upvote 0
Nov 2, 2020
132
134
Another observation I made from my RAW and corrected comparisons of those two wide angle RF lenses requiring substantial software distortion correction was the difference in corner performance between the two lenses. The RF 16 was somewhat soft in the corners (both RAW and corrected) whereas the RF 14-35 was very good in the corners even at 14mm. The alleged pixel destruction due to software correction of lens geometric distortion was clearly absent with the RF 14-35, where it should have been obvious if it existed.

As you suggested, those claiming software correction of geometric distortion is destructive at the pixel level need to provide convincing evidence that such is true. Especially given evidence to the contrary (RF 14-35 corner performance).
 
  • Love
Reactions: 1 user
Upvote 0
Jul 21, 2010
31,089
12,855
This review sums up why its an issue quite well:

In summary, an ultrawide f/2.8 prime lens that costs a mere $300 has soft corners. I’m severely shocked, surprisingly stunned, and stupendously stupefied.

What, you were expecting the optical performance of a lens that would cost several times as much, but you were expecting not to pay for that? Yeah, I’m sure you were. Do you need those sharp corners to resolve the unicorns at the edges of your images?

Older EF lenses were not overly large and heavy and had optical distortion.
Sure, like the EF 14mm f/2.8L II that is more than double the size, four times the weight, and costs 7 times as much as the 16/2.8? Or the EF 17-40/4L that is more than double the size, three times the weight, costs 2.7 times as much as the 16/2.8…and is just as soft in the corners? (Actually, the prime is a bit sharper if stopped down to f/4.)

Other brands don't seem to have this issue. Canon is being lazy.
Really? Sony, Fuji, Olympus and Panasonic reportedly all force corrections for certain lenses and have been doing so since long before Canon followed suit with the RF 24-240 (Canon does not force corrections for any EF-M lenses). Nikon does so as well (including in some cases where it really doesn’t seem needed, e.g. the Z 28/2.8 that has <3% distortion when not corrected).

You should at least make a modicum of effort to get your facts straight before you post such drivel.
 
  • Like
Reactions: 1 users
Upvote 0

SwissFrank

1N 3 1V 1Ds I II III R R5
Dec 9, 2018
503
338
Another observation I made from my RAW and corrected comparisons of those two wide angle RF lenses
Now if we could only get nunataks to actually investigate this stuff before making categorial statements that there's a problem!

> The RF 16 was somewhat soft in the corners

Sure but it's a $250 lens and the smallest Canon even makes. Compare it to say the EF14mm/2.8, which was nearly the same spec, but 10x the price. The EF14mm was rectilinear, but the corners were soft as oatmeal, despite its size and price. Clearly, Canon made huge sacrifices to keep it rectilinear, as required in the film era: the price was 10x higher, ultimate corner resolution was lower, lens was far bigger and far more likely to break. Heck, the lens CAP for the 14mm cost more more than the entire 16mm.
 
Upvote 0
You guys are a trip...

Remember the fact that the 14-35 and the 16mm have BLACK corners as the lenses don't cover the full frame sensor, and thus they're required to be both cropped and distortion corrected. If you think that's an improvement, then be my guest and buy them. I won't, especially at their inflated prices over the much better quality EF lenses that existed before.
 
Upvote 0
Nov 2, 2020
132
134
Your statements ring somewhat hollow. The overall quality of the corrected RF 14-35 compares very favorably with the EF 16-35, though it does command a steep premium for improvements in size, weight, and IS. As far as the RF 16, there is no EF counterpart, so a fair comparison isn't possible.

Arguments about the need for cropping and distortion correction can be regarded as idealistic if the final image is comparable. Hate those corrections if you wish but results oriented individuals may not care if the final results are essentially the same.
 
  • Like
  • Love
Reactions: 1 users
Upvote 0
Jul 21, 2010
31,089
12,855
You guys are a trip...

Remember the fact that the 14-35 and the 16mm have BLACK corners as the lenses don't cover the full frame sensor, and thus they're required to be both cropped and distortion corrected. If you think that's an improvement, then be my guest and buy them. I won't, especially at their inflated prices over the much better quality EF lenses that existed before.
I see that you decided to ignore my suggestion to try and get your facts straight before you post drivel.

Please tell me, how is it a fact that images from the 14-35/4 and 16/2.8 are required to be both cropped and distortion corrected? That is false, correcting the distortion actually fills in the corners. No cropping is needed. In fact, as I've stated (and shown) previously the 14-35/4 at the wide end, when converted with DxO PhotoLab, delivers images with a field of view equivalent to 13.5mm on the EF 11-24/4L, and the corrected corners are just as sharp (the 11-24 has almost no corner distortion in that comparison because 13mm is the point where the lens crosses from barrel to pincushion distortion, although there is a small amount of mustache distortion in the mid-frame throughout the range). That means that, after correction, the 14-35 at it's wide end (where UWA zoom performance is generally worst) is delivering IQ equivalent to a lens costing double the amount and with excellent optical corrections (and more than double the weight to achieve them), when the latter lens is not at an extreme of its zoom range.

If the above was confusing, the short version is that the RF 14-35/4 delivers IQ every bit as good as the EF 16-35/4L, and does so with an extra 2 mm on the wide end while being smaller, lighter, and still taking the same 77mm filters. Yes, it costs ~$300 more (depending on rebates)...and that doesn't seem unreasonable for a lens that delivers a wider focal range in a smaller package with equivalent optical performance. Maybe the price tag of the 14-35 puts it out of your reach, but that's a you problem.

As for the price of the RF 16/2.8 being 'inflated over the much better quality EF lenses that existed before', that lens costs $300 and there is nothing like it in the EF lineup.
 
  • Like
Reactions: 1 user
Upvote 0

SwissFrank

1N 3 1V 1Ds I II III R R5
Dec 9, 2018
503
338
Remember the fact that the 14-35 and the 16mm have BLACK corners as the lenses don't cover the full frame sensor, and thus they're required to be both cropped and distortion corrected. If you think that's an improvement, then be my guest and buy them.

I don't care if the image uses 100% of the sensor or 1%. I look at the resulting photos to see if they take good photos. I know I should forget the images and just concentrate on trivia, I know, but I guess I'm just clinically insane.

I won't, especially at their inflated prices over the much better quality EF lenses that existed before.
RT16mm/2.8's closes competitor is the EF14mm/2.8. I took a LOT of images with that lens (inc. my first published) but the 16mm is better images, 1/10th the price, 1/5th the volume, and seems likely to be far more resilient. I really cannot fathom why you're so hung up on trivia and don't seem to care about the actual resulting image.

Why not just post some side by sides images and show us how much worse the 16mm is?
 
Upvote 0
I see that you decided to ignore my suggestion to try and get your facts straight before you post drivel.

Please tell me, how is it a fact that images from the 14-35/4 and 16/2.8 are required to be both cropped and distortion corrected? That is false, correcting the distortion actually fills in the corners. No cropping is needed. In fact, as I've stated (and shown) previously the 14-35/4 at the wide end, when converted with DxO PhotoLab, delivers images with a field of view equivalent to 13.5mm on the EF 11-24/4L, and the corrected corners are just as sharp (the 11-24 has almost no corner distortion in that comparison because 13mm is the point where the lens crosses from barrel to pincushion distortion, although there is a small amount of mustache distortion in the mid-frame throughout the range). That means that, after correction, the 14-35 at it's wide end (where UWA zoom performance is generally worst) is delivering IQ equivalent to a lens costing double the amount and with excellent optical corrections (and more than double the weight to achieve them), when the latter lens is not at an extreme of its zoom range.

If the above was confusing, the short version is that the RF 14-35/4 delivers IQ every bit as good as the EF 16-35/4L, and does so with an extra 2 mm on the wide end while being smaller, lighter, and still taking the same 77mm filters. Yes, it costs ~$300 more (depending on rebates)...and that doesn't seem unreasonable for a lens that delivers a wider focal range in a smaller package with equivalent optical performance. Maybe the price tag of the 14-35 puts it out of your reach, but that's a you problem.

As for the price of the RF 16/2.8 being 'inflated over the much better quality EF lenses that existed before', that lens costs $300 and there is nothing like it in the EF lineup.

They literally show the black corners in an uncorrected photo from the 14mm. I can find more examples, but you just want to argue so I'm going to block you instead. Have a nice day!
 
Upvote 0

AlanF

Desperately seeking birds
CR Pro
Aug 16, 2012
12,346
22,520
They literally show the black corners in an uncorrected photo from the 14mm. I can find more examples, but you just want to argue so I'm going to block you instead. Have a nice day!
If you block @neuroanatomist , then you will miss some of the most informative posts here. But, it's your choice if you don't like his direct style. I can't say I learn anything from those who speak only from viewing a few review sites and don't have hands-on experience.
 
  • Like
Reactions: 1 users
Upvote 0

tron

CR Pro
Nov 8, 2011
5,222
1,616
I had a negative attitude towards RF 14-35 f/4L IS but I always liked its small size. When I found an offer in my country about RF lenses (Canon supposedly asks for an EF lens but they actually accept a photo of the EF lens and its serial number and not the lens itself!) of 25% off the price I got it (saving almost 500 euros!). It joined my collection of many RF and EF UWA zooms (and some EF mount fixed UWA lenses) because I still have 5DIV and 5DsR so I kept most of my EF lenses.

What I see at 100% is excellent and the corners good to very good even at f/4 in low light/contrast shots. So no complaints. The end result is very good. The only problem is it will have to compete the 2.8 RF which is my museum selection and my DSLR UWA lenses for astrophotography (starting from next year, this year passed with no such photos unfortunately).

But when I go out for birding I usually take my Nikon for birds and for landscapes I use EOS 200D with EF-S 10-18 and EF-S 15-85 but now I can replace them with R5 with 14-35 and 24-105.
 
Last edited:
  • Like
Reactions: 1 user
Upvote 0