I see a growing number of new lenses hitting the market that rely on lens correction. These lenses have some benefits, but are lens manufacturers taking shortcuts with these lens designs? Will we be seeing more and more lenses with this reliance on lens correction in the near future?
I remember the time when I used my old 15mm fisheye lens. The image was distorted, as one might expect of such a lens. Although it was a rectilinear lens, every vertical and horizontal line became increasingly distorted the farther it was from the center of the frame.
It was possible to correct this lens distortion with software, but at a cost. The image quality degraded at the edges due to the stretching of the image. The distortion of the 15mm fisheye lens was intentional, and I chose that lens for its unique characteristics.
Modern lenses have improved significantly since the release of that fisheye lens. The design of lenses has become increasingly complex to reduce distortion, chromatic aberration, and vignetting. Although physics makes it impossible to achieve the perfect lens, lens technology has come a long way.
Perfect and Even More Perfect
The goal of many lens manufacturers is to produce more elaborate lens designs to achieve an image as close to perfect as possible. New lenses improve with each iteration, mostly in small increments.
We’ve seen innovations in lens coatings, nanotechnology, and floating lens designs, all aimed not only at achieving the best possible image but also at increasing resolution. If the second version of a lens is almost perfect, the third iteration will be even better.
Some may argue that these perfectly designed lenses and their near-perfect images are too clinical and lack personality. It’s perhaps for this reason that vintage lenses, with their unique characteristics, are growing in popularity despite their optical flaws.
Not Perfect, but Made Perfect by Software
Some time ago, I received a new lens for a review—the Sony FE 20-70mm f/4 G—a small zoom lens with a slightly different zoom range compared to its peers. This made the lens unique in some ways.
However, I was surprised by the amount of distortion this lens showed at 20mm and f/4. As expected, it relied on in-camera lens correction software to produce a clean image.
More recently, I received the Sony FE 16-25mm f/2.8 G and FE 24-50mm f/2.8 G for a review. I noticed a similar reliance on image correction. The images were only acceptable after applying lens correction, which removed distortion and vignetting, resulting in good-looking images. Lens correction is even baked into the raw files, making it difficult to remove these corrections if you wish to.
I noticed the same issue with the Canon RF 10-20mm f/4L. Without lens correction, the image at 10mm focal length was heavily distorted, and vignetting was extreme. Now, with the release of the RF 24mm f/1.4L VCM, RF 35mm f/1.4L VCM, and RF 50mm f/1.4L VCM lenses, the necessity of lens correction has become even more common. Without it, the distortion and vignetting are unacceptable.
When Lens Correction Is Not Desirable
Some might say there’s nothing wrong with this trend, and to some extent, I agree. For most situations, lens correction can produce the expected image quality, even if it’s achieved through software.
However, in specialized photography, this reliance on lens correction can become increasingly problematic, even rendering results unusable in some cases. This is especially noticeable in image stacking. Since each lens correction is unique, stacking can result in a significant amount of banding. I’ve seen this issue multiple times with Fujifilm, Sony, Nikon, and Canon lenses when stacking for star trails.
Although I haven’t tested it, I wouldn’t be surprised if this also occurs with focus stacking, especially if a large number of images are used. It would be an interesting experiment to conduct next time I review a lens that relies on image correction.
The reliance on lens correction has other side effects that can be undesirable. Removing vignetting increases noise levels, especially when a three or more stop correction is necessary and high ISO values are used.
Additionally, stretching an image to correct distortion can lead to a loss of detail. I’m aware that modern software can and likely will fill in missing pixels in a smart way, but this may still be noticeable under close examination.
The Benefits of Lenses That Depend on Lens Correction
I recognize the benefits of lenses that rely on lens correction. Since their designs can be simplified, these lenses are often smaller than similar lenses that produce a more accurate image. This can result in reduced weight as well, which is desirable for many photographers.
For instance, Canon’s VCM series—the RF 24mm f/1.4L VCM, RF 35mm f/1.4L VCM, and RF 50mm f/1.4L VCM—are compact and lightweight, making them ideal for a broad range of photographers. This also applies to Sony’s FE 16-25mm f/2.8 G and FE 24-50mm f/2.8 G lenses, which are easy to carry on a trip.
However, these compact lenses aren’t for everyone. If you need to disable lens correction for your type of photography, the results may be disappointing. In that case, it’s best to choose a different lens.
What If?
Imagine if lens manufacturers decide to produce even more lenses that rely on lens correction software. These lenses may become increasingly popular due to their size, weight, and reduced cost, which is likely more financially appealing for manufacturers.
If this trend continues, new lenses that produce the best possible image quality without correction may become harder to find, if they’re even available at all. The difference in distortion between the old Canon EF 35mm f/1.4L II USM and the new RF 35mm f/1.4L VCM is a good example. What if Canon decides to stick solely with the VCM series and its reliance on image correction?
I hope this doesn’t happen, and that lenses with the best possible image quality and minimal distortion continue to be released for years to come. I believe there’s a market for both, and I hope lens manufacturers won’t overlook that.
What do you think about the increase in new lenses that rely on in-camera image correction? Do you think it’s a growing problem or not? Please share your thoughts in the comments below.
So the idea is to turn off Lens correction for the shooting, do the stacking and add lens correction to the final image ?
But Lens correction do not apply to the RAW image, right.
so problem solved ? no?
Sometimes they are baked into the RAW images.
Do you have an example (brand/model) of pro camera with such things?
RAW is without lens correction for me, but willing to discover that I'm wrong.
I noticed this with the Fujifilm GFX50s and GFX100. I have seen it with Sony cameras as well
"I remember the time when I used my old 15mm fisheye lens. The image was distorted, as one might expect of such a lens. Although it was a rectilinear lens, every vertical and horizontal line became increasingly distorted the farther it was from the center of the frame."
A rectilinear lens is by definition not a rectilinear lens. Fisheye lenses allow curved lines, rectilinear lenses don't.
Both the rectilinear and the fisheye lenses show distortion, though the fisheye has arguably less. Human vision has a relatively narrow angle of view. Imagine standing in front of an infinitely long straight wall, maybe ten feet away, and you can't perceive the whole thing from end to end. Look at the part of the wall right in front of you and the top and bottom are straight and parallel. Look to the left and you'll see the wall disappearing to a vanishing point. Look to the right and you'll see the wall disappearing to another vanishing point. Nowhere can your eyes detect a curve in the straight wall.
Now take a piece of paper, and draw a rectangle to represent the frame of a photograph. Inside the rectangle draw the whole wall, including the parallel lines in front of you and the vanishing points on both ends.
What you have isn't the distortion of optics, it's the distortion of condensing a wide field of view into a narrow field of view.
You are right. That fisheye lens is not rectilinear. But it produces an image that covers the full frame sensor.
Some fisheye lenses cover the full frame, some have a round image within the full frame, some even cover medium format up to 6x7 cm. But coverage has nothing to do with a lens being rectilinear.
We've seen what it takes to make an "optically perfect": the Nikon NIKKOR Z 58mm f/0.95 S Noct! And that thing is huge, heavy, and crazy expensive. As you note: lens design is a constant evaluation of tradeoffs. And usually, when we say "additional optical correction" that translates to "additional glass elements." That's why the Nikon NIKKOR Z 58mm f/0.95 S Noct, a prime lens, has 17 elements in 10 groups.
I personally prefer to save the weight, size, and cost where it makes sense. I don't think I have to enumerate why less weight, smaller sizes, and cheaper costs are good. I'm not here to defend bad image quality: if it looks bad, then it looks bad! But if the end image looks good, what does it matter if that was achieved with digital lens correction?
For regular photography it doesn't matter that much. I agree.The point is, what if all future lenses rely only on software to produce a usable image?
I'm not sure I understand the thinking here.
Prior to in-camera lens correction the addition of lens correction algorithms in image processing programs, eg Photoshop, was of great value and a necessary step for optimal processing first of scanned film and then of optical dSLR originals. The Adobe converter still contains options for a number of lens corrections.
Doing the lens correction/optimization in camera is convenient and improves the apparent optical performance of lenses.
Win win.
Perhaps the demise of written content with objective bench testing of lenses has led many to forget that without the built in image correction of EVF dSLR systems eyebrows might rise even further at the prices and "raw" performance of modern lenses.
Regardless of price no lens is optically perfect, no is possible.
Niche processes like stacking of any ilk require software futzing to work at all, lens correction algorithms or not. I don't see a problem.
It seems you can look at raw images in DxO without lens correction and then use theirs, not sure of any other way for most users to see uncorrected images from modern lenses.
What if all lenses you will purchase in the future heavily rely on software to produce an acceptable result. It's like photographer who says: "I'll fix it in post" That's not a good photographer. This is similar in my opinion.
Remember when lens reviews told you how much pincushion or barrel distortion they had?
The distortion on the Canon RF 10-20mm f/4L is insane for a $2300 lens.
It is like they charged a premium but cut corners and made a lens that does not cover the full frame sensor. This is especially the case given that many older lenses in that range did not have the completely dark corners.
The main times when issues like that would be seen with older lenses, was when someone used Crop sensor lens on a full frame camera and did not enable the crop mode.
Much of these issues remind me of this article. https://photographylife.com/using-nikon-dx-lenses-on-fx-cameras
Simply Canon is robbing its customers. The distortion on the 50 mm 1.4, without the mandatory profile corrections is incredible, in the saddest meaning of the word. A few years back Canon rumors had posted canons patents for smaller image circle and using profile corrections to make up the difference along with plastic elements etc. These are the results of those patents. Basically cheaper design for the same or more money. Bleh.
I've waited a decade and more to be able to buy an updated 50 mm to 1.4. I'm not buying this one from canon. How telling is it that you can't even turn off profile corrections. And the bokeh isFINE at best. Definitely not as smooth as you would expect for a $1,400 Canon.
What I will be buying is a Panasonic S52 and grab a sigma or one of panasonics 50 mm. The only reason I'm not running two systems right now is because of the time current financial outlook. Don't get me wrong I love my R5 all of my old EF lenses that I adapt but Canon is really running a racket here. I guess I won't be shooting 50 mm at one point for on my Canon for years to come. What a damn shame. I refuse to buy these compromised lenses with mandatory profiles at hyperinflated prices. Nope nope nope.
+ Every month is so called photographers complaining that nobody needs ibis, or it will add to the price of lens. Well that didn't stop canons smaller than full frame image circle, heavily distorted, $1,400 lens price for a 50mm that does not include the awesome blue goo from the last ef 35 version. Cmon.
I agree.
However, I think there's a market for both types of lenses: the ones that require heavy correction and are small and 'cheaper', and the optically as perfect as possible ones. I'm waiting for the latter before I upgrade from my old trusty but as perfect as possible EF 35mm f/1.4L II
Optically correcting something like barrel distortion doesn't just make lenses larger and heavier - it's always a trade-off that creates other aberrations, like increasing astigmatism.
If your choice in lens design is more barrel distortion that has to be corrected digitally and less astigmatism, and it results in a better finished image, or less barrel distortion but more astigmatism that has to be corrected digitally, and that results in a worse finished image, which would you choose?
The lenses I have been using for years have a perfect trade-off. But now the balance is going towards the lenses that rely completely on software correction.
I've wondered about pano merge and focus stacking with frames from heavily corrected lenses but so far have not encountered difficulties. Anyone with experience and comments?
As long as there is enough overlap between the separate images, I don;t think this will be a problem with panoramas. Regarding focus stacking, this relies already on a lot of minor image distortions to get everything perfectly lined up.
I don't think this is an issue.
Thank you for a thoughtful, balanced discussion of the issue. I have zero nostalgia for the days of film, but the recent increasing dependency on digital correction of lenses has me also wondering how much lens craftsmanship we will be asked to sacrifice for the sake of cheaper, lighter lenses. And this has led me to be more sympathetic to photographers who long for more analog.
Canon didn't cut many corners with its first prime-lens releases, though the lenses were painfully expensive. And since then the global-market has been roiled by inflation, so I understand why Canon is going the cheaper route.
As a long-time Canon customer, if I'd like to see them go no further with L-lens corner-cutting.
My kids have gotten into photography, but they are still at the very beginning stages. They use an R50 for now, and they get mad at me when I refuse to fix their extremely blurry images, remove crazy distractions, or add friends to group shots with Photoshop. Will they ever think getting something right within the frame matters in our digital dreamworld?
Again, thanks!
they say if one wants a lens which does not rely much on software correction that one should look for good DSLR lenses as the imperfections are visible to the naked eye through the viewfinder whereas the imperfections of mirrorless lenses can be concealed from the photographer