Geek Articles

Stopping Down Some Bargain Primes and Zooms

In my last post, we looked at how the field curvature of some really excellent 35mm f/1.4 primes affected what happens when we stop down the lens. The takeaway message is that as you stop down, things change a bit differently with the different lenses, but by f/5.6 they were similar. Still slightly different, but pretty similar.

The real purpose of these articles though, was to look at generalization that is often made; ‘all lenses are the same at f/8’. These days f/8 generally means you’re starting to get diffraction softening. That’s not a bad thing and it’s reasonably easy to overcome in post, but it does tend to level the playing field, since diffraction softening don’t care what lens you use.

I thought we should compare a couple of reasonably priced, but pretty good 35mm lenses for the next step. If you’re shooting stopped down, is there any reason to spend the big bucks on an f/1.4 lens? And for those people who still like to say things about ‘this zoom is as good as a prime’ I thought we’d look at a couple of good 24-70mm f2.8 zooms at 35mm for the same reason.

I will mention that you probably should read the last post, since I’m not going to repeat any background on what these MTF vs Field vs Focus plots (AKA field curvature plots) are in this article.

Smaller 35mm Primes

Let’s look at two very popular 35mm lenses; the Canon 35mm f/2 IS and the Tamron 35mm f1.8 SP Di VC. These are both really excellent lenses that, if you don’t need an f/1.4 aperture, offer an excellent alternative to the bigger, more expensive lenses and as a bonus have image stabilization which most f/1.4 primes don’t offer.

Just for a reference, I’m going to repost the graphs of the Tamron 35mm f1.4 SP from the previous article. Notice it starts off with pretty good side-to-side sharpness, although the sagittal and tangential fields are of different shape, so they don’t quite overlap. As it stops down it gets sharper and the increased depth of field results in better overlap., 2019

By f/5.6 it is maximally sharp, although there is some area in the tangential field that’s not quite as sharp as the center. If we average the tangential and sagittal fields, we see at f/5.6 it’s not perfectly sharp edge-to-edge, but it’s close enough for any reasonable purpose., 2019

Now let’s compare it to the smaller 35mm primes to see how they stack up. Since we stopped the f/1.4 lenses to f/5.6, it seems fair that we stop these down through f/8.

Canon 35mm f/2.0 IS

At f/2, it’s apparent the Canon isn’t as sharp away from center as the f/1.4 lenses were at f/1.4., 2018

As we stop down, it becomes apparent that the sagittal field curvature is more pronounced with this lens, too. Still, by f/5.6 and f/8 we can see that the lens is getting good side-to-side sharpness although it looks like the fields still don’t quite overlap.

If we look at the f/8 average, though, things look pretty good. The edges are nearly as good as the Tamron was at f/5.6, although there’s a bit more field curvature. You could pixel-peep the difference and see the Tamron at f/5.6 was a bit better away from the center than the Canon at f/8, but it would take a lot of effort., 2019

Tamron 35mm f/1.8 SP VC

As with the Canon, the Tamron isn’t as sharp wide open as the more expensive lenses were at f/1.4., 2019

While the sagittal field curvature isn’t as pronounced as the Canon was, the tangential plane doesn’t sharpen up quite as well as the Canon. At f/8 our overall field, though, is still pretty good; it would require some pixel peeping to see the difference between the f/1.4 primes and these if they were all shot at f/8., 2019

The more expensive f/1.4 primes, then, are better than the less expensive primes at wide apertures. They’re ‘laboratory’ better even stopped down, but at f/8 the difference is fairly small. If I was shooting stopped down, I’d personally prefer the lighter weight and lower costs. If I was shooting at, say, f/2.8 and wider much of the time, I’d have to seriously consider the better image quality of the f/1.4 lenses.

Now for Some Zooms

Let’s explore the other thing I hear a fair amount; “I don’t need primes because I shoot everything stopped down.” We’ll look at two 24-70mm f/2.8 zooms at 35mm for this comparison, the Sigma 24-70mm f/2.8 DG OS Art and the Canon 24-70mm f/2.8L USM Mk II. The Sigma is a good zoom at a price lower than the Canon or Zeiss 35mm f1.4 primes. The Canon is a really good zoom at a higher price.

Sigma 24-70mm f2.8

A couple of things to note. First, compared to the primes, the field curvature on the tangential field is more dramatic than what we saw for primes. The edges aren’t as sharp wide open. There’s also more field tilt than the primes had. All of this is pretty much typical zoom; we’ll talk about the fields more in the last section of the article., 2019


You really need to be at least f5.6 to get reasonable edge sharpness with this lens, but by f/8 it’s resolving fairly well. The tilt is enough that you’d notice it if you checked carefully, but again, that’s zooms for you. And yes, I know many of you can’t wait to tell me that you have perfect zooms. You don’t but I won’t ruin your fantasy for you just yet. I’ll just say, as we do in the South, “how nice!”, 2019

The bottom line is this zoom, stopped down to f/8 is probably good enough for most people to use for a landscape image. It’s going to have more astigmatism and field curvature out at the edges than the f/1.8 primes did, but it would still be just fine for most people most of the time. I don’t think most people would want to do architectural photos with it, though. The combination of slight edge softness, field curvature, and tilt would be a bit more than most (not all) of you would like.

Canon 24-70 f2.8L Mk II

This gets a little interesting and also demonstrates a bit about why when someone asks if zoom X is better than zoom Y, my usual response is at what focal length and for what purposes? The Canon has better edge sharpness than the Sigma did. And the fields tend to overlap well; there’s not a ton of astigmatism. 2019

However, if you look at the field curvature, that edge sharpness is not in the same zip code as the best center sharpness. You can stop down to f/16 and you’re still not going to compensate for that curvature. (Before you Canon 24-70 owners start screaming about your copy not having field curvature like this, read the last sentence. Or I shall mock you.)

Does it matter? It depends on how you use the lens, what you use it for, and at what focal length. If it’s a landscape with a field of flowers, you might not care if the maximum flower sharpness is in an arc, not a line. If you’re doing architectural photography or a skyline, you would. For other purposes, you might prefer the better subject isolation the field curvature gives. If you take a group portrait, putting the group in an arc would keep everyone in focus, but a straight line of people would result in blurry ones at the edge.

My point is this tool would be quite different to use than the other 35mm lenses we’ve looked at. Great photographers know their tool well and use it to best effect. Many of them would be perfectly happy with this lens for most 35mm work.

One Other Thing About Zooms. Well, Several Things, Really.

A Roger’s Rule for you: All Zooms have multiple personalities.

You’re probably aware that zooms are sharper at some focal lengths than they are at others. This is both on average and for specific copies. For example, the Canon 24-70mm f/2.8 we just looked at is, on average sharpest at 24mm, less sharp at 50mm, and really a bit soft at 70mm. There are a few copies, though that are sharpest at around 50mm. I’ve seen one or two that were sharpest at 70mm.

You may not be aware that the field curvature also changes at different focal lengths. The Canon has a wicked field curve at 35mm we showed you above. At 70mm, though, it’s almost perfectly flat (although less sharp)., 2017

At 24mm the fields no longer match, with a “W” curve in the tangential field, and a sagittal curve that’s slight, but in the opposite direction. So it has significant astigmatism at 24mm, less at 35mm very little at 70mm.

Does this really matter? Only if you want to make best use of your tool. You can frame a shot to make field curvature complement it. But you have to know what the field curvature looks like at different focal lengths to do that. I will add, this is one of the reasons I chose 24-70mm lenses for this comparison, they almost all have significant changes in field curvature from one end to the other. In general, 70-200mm have less dramatic change, zooms greater than 4X more dramatic change, and wide-angle zooms are a bit more random., 2017


The other thing that happens with zooms, almost always, is the tilt varies at different focal lengths in each copy. In the Canon lens above there’s a bit of field tilt at 24m that goes away at 35mm, to be replaced by a bit of decentering at 70mm.

You may remember the Sigma 24-70mm we did above had more significant tilt. Here it is at 3 focal lengths, and I’ve added the line of tilt in an overlay. You can see the tilt increases as you move from 24mm to 70mm., 2019

The main takeaway from this is that zooms are less interchangeable than primes, copy variation is much more noticeable. If you make a general statement about zooms (like I did above for the Canon 24-70s sharpness) be very aware there will be a lot of copies that behave differently.

This is how zooms are. If you ask me to find a zoom that has no field tilt or centering change throughout the range, I’d tell you it’s not worth me testing a hundred copies in the faint hope of finding one.


I think it’s pretty obvious, at least for lenses shot at 35mm. If you plan on shooting stopped down, a good f/1.8 or f/2 prime is nearly, but not absolutely, as good as the more expensive f/1.4 primes. A good zoom lens, even stopped down, is not as good as the inexpensive prime, at least at the edges of the image, but it will be close enough for most people most of the time. Depending on their copy. The zoom is going to be less predictable and have more variation, so that answer is a bit more copy specific.

That doesn’t mean you can’t take nice landscapes and other images with a zoom; lots of people do. It does mean it’s worth taking a little time to learn the field curvature and stop-down behavior of your zoom at various focal lengths if you want to use it to best effect.

Zooms have other advantages, of course. You don’t have to carry as many lenses or change them as frequently, you can often frame the image better.

A zoom is often the best choice of lens. But it’s never the best lens.

Roger Cicala and Aaron Closz

November, 2019


Addendum: So How Do You Check Your Field Curvature.

Yes, I agree it will be nice if I put out field curvatures for every lens. I may do that for some of the more popular lenses, but I’m not going to do them all, and I’m not going to do all the focal lengths of every zoom. But it is something you could, if you want, check for at home. You can even check more than I can, since I’m just testing at infinity.

My personal favorite way to do this is to find a fairly large area with lots of regular small shapes in it. A field of grass, gravel, mulch, etc. works well. I put something in the center I can focus on and take an image at a shallow angle. For example, if I’m shooting a fair distance (10 or more meters) away I can stand up and take the image of a field of grass. If I’m shooting at close distances (a meter or two) I want the camera close to ground level.

I take the images and run them through something like Photoshop’s Find Edges filter, which will identify the areas that are in best focus and gives me a pretty nice printout of the field curvature.

Roger Cicala, 2016


This article has some more examples of this technique in the last section. As a bonus, if your lens is badly tilted, that will show up when you do this.

Author: Roger Cicala

I’m Roger and I am the founder of Hailed as one of the optic nerds here, I enjoy shooting collimated light through 30X microscope objectives in my spare time. When I do take real pictures I like using something different: a Medium format, or Pentax K1, or a Sony RX1R.

Posted in Geek Articles
  • Have you ever tested the Sigma 30Art, or its later relatives for µ43 and other formats?!

  • As other have pointed out before, these articles have teached me more than reading a lot of books in the subject.

    Without your input I probably would have a lot of zooms, so now I have just long zooms, as I can’t afford long primes, and a 70-200/4.0G, plus primes from 135 down to 14.

    My favorite primes are a diverse mix, from Nikon1 7.9 (a modified 10), via Samyang 14, Voigtlander 20, Sigma 30 Art, Nikon 40 macro, Voigtlander 58, Sigma 105 macro OS HSM, Sigma 135 Art. I have a few more, I do confess.

  • Excellent and informative as always. Thank you for sharing this!

    The mark of all good research is that it raises as many questions as it answers. For me the question is whether a high end prime lens can be designed so that it trades maximum brightness for improved acutance. To rephrase the question: is it a myth that faster lenses are necessarily sharper (once stopped down) than slower ones?

    The reason I ask this is that years ago I remember studying a series of MTF plots by the same person of the Leica 50mm f/0.95 vs f/1.4 vs f/2. Surprisingly, the MTF plots for the f/0.95 were the worst, and, even when stopped down, the f/1.4 and f/2.0 versions performed better. Stopping down the Leica Noctilux didn’t allow the MTF plots to catch up with its slower peers. The moral of the story is that faster lenses are not always sharper and that the Leica 50mm f/0.95 compromised on acutance in exchange for extreme brightness

    Apart from the Leica 50mm f/2.0 lens, there are other high-end prime lens that are a bit slower eg the Sony/Zeiss FE 55mm f/1.8 or the Zeiss 85mm f/1.8. Do these lenses have high-end MTF plots despite their slower maximum apertures?

    Macro lenses are another example of lenses that are relatively slow without necessarily compromising on acutance (though the reduction in minimal focussing distance often does just that) and they may also be designed to perform better stopped down further than a portrait lens.

    The other thing that puzzles me is the way Canon great white lenses appear (at least on DXO Mark….cough…but also on other review sites that measure MTF plots) to reach maximum acutance wide open, with the acutance dropping as soon as you stop down. This isn’t true of the EF 200mm f/2.0 which reaches maximum acutance at f/2.8 before dropping back to the same level of performance at f/4 as it achieved at f/2.0. However, the Canon 300mm f/2.8 seems to attain maximum acutance wide open. So once again, the idea that stopping down necessarily improves acutance seems not to be universally true.

  • Thinkinginpictures

    High MP sensors should give you the ability to overcome some of the corner issues with zooms, no? For instance, the 16-35mm F4 Sony/Zeiss has some astigmatism in the corners throughout the range. From what I’ve seen though it covers about the last 5% of the image if that. Why not shoot wider and crop it out? That leaves out 16mm but that mid range could benefit from such an approach where the corner issues are most notable.

  • SpecialMan

    You have such a generous spirit and are so reasonable and everything you write is so well thought out and researched that you are clearly hiding something from us.
    You are like Dumbledore or Gandalf, too good to be true.

  • FWIW, as a nightscape photographer literally every (fast) lens I have ever owned for more than a year and used for wide-field star photography has changed its plane of focus after extended abuse, from being crammed into the bottom bag of a 4-backpack gear pile on a 100+ mile off-road washboard 4×4 roadtrip, to being bumped around in a crammed-full backpack climbing mountains and hiking long distances…

  • Chris Newman

    Thank you for educating us about the optics or our lenses in such an interesting way, especially with your last two blogs. The downside is that I must accept I’ll never find enough information to decide which brand of two good lenses is likely to be the “best”, even if the tester takes the trouble to test 10 copies of each (which I think is usual for you, but no other testers bother to match).

  • Athanasius Kirchner

    OK, so basically the camera would be “on top” of the charts and not below them, then.

  • Bob Locher

    Thanks Brandon! You are correct of course. Appreciate your comment as well as Roger’s.

  • Out, then back, I believe.

  • No, it’s not possible. But as I said, there is a LOT of copy-to-copy variation in zooms. He may well have had one that was flat at some focal lengths, especially toward the longer end. I’ve never seen one flat at all of them.

  • Lee Wooten

    Thank you for the cultural reminder. I grew up in the South but have lived in colder climates for the last several decades, and I’d forgotten the polite dismissive “how nice!” I’ve grown used to hearing the more straightforward “wrong!” or “you are sooo full of [word not used in polite southern circles].”

  • Brandon Dube

    Of course. It is easier to design a smaller aperture lens. The market constraints (size, price, weight) are stronger on the smaller aperture lenses though.

    Of course, options like the 100/2 MP or 50/2 apo-summicron demonstrate that some of the best lenses available today are smaller in maximum aperture.

  • numbertwo

    Is it possible than in early copies of the canon 24-70 f2.8 II the plane of focus was flatter and in later copies canon changed the optical formula?

    A friend claims he had an old unit that was perfectly flat, but after some time it developed some tilt issues so he bought a new copy that had this wavy field curvature. He replaced it and tested two more copies and all of them were having the issue, but still he’s convinced that his old first version was perfectly flat.

    Maybe it could be interesting to test in case you had an early batch unit…

  • I think it’s way too soon to make that a rule, although it was true for this test group. The really wide-aperture primes (f/1.2 and wider) aren’t going to behave that way. But for the tendency we’ve seen so far, I agree with your suspicion.

  • Athanasius Kirchner

    Excellent article Roger, as always, it’s much appreciated!

    I have one doubt, though – the “W” shape in the fields implies that the plane of focus curves outwards (away from the photographer) or inwards?

  • Bob Locher

    First, thanks for a very interesting report and article.

    I am a bit bothered by the conclusion that a larger aperture prime lens is sharper stopped down than is a slower equivalent prime. I do not dispute the conclusion but rather am puzzled why that would be. My suspicion is this – that the additional glass required to make a sharp larger aperture lens also does a better job of reducing aberrations at smaller apertures.
    Is this correct in your opinion? And if that is the case, would you then think it possible for a lens designer to design smaller aperture lenses but still with improved sharpness either by adding more elements or perhaps using aspherics?


  • I already have in the article I linked to. My only addition was that the curves on the MTF bench are more dramatic than what you actually see in a photo.

  • Ra Ho

    Brilliant, many thanks, this has explained clearly and concisely issues I didn’t even know existed! I have had some expensive but poor lenses from big names in the past, even primes, so I take MTF charts with a pinch of salt. What I have found is that 1.8 primes are the best for my full frame and easier to calibrate for focus on a DSLR than any zoom. Interestingly I also have micro four thirds and the zooms are sometimes better than primes but all of them are excellent.

  • David Bateman

    Thats very promising. The best way to know if you could test it. Would be to select a couple of lenses that you know have specific character from OLAF testing. Then see if the live focus peaking corresponds to what is expected. You could test 5 EF mount lenses on both the Sony A7m4 and Panasonic S1R, using EF adapter. I am assuming the higher megapixels sensors would be more sensitive to the exact focus plane.

  • I agree. I’d also add that the zoom and focus mechanisms either involve rotation of the element (as well as moving forward or back), or in the case of electromagnetic focusing, moving along sliders where force is often applied only on one side.

  • David, I totally agree and I’ve done that a few times and it seemed to work. I can’t say, though, that I’d absolutely trust it since I’m not absolutely certain what algorithms each camera uses in focus peaking. I think in general, though, it should be an effective method.

  • Kers

    Great article ; i always find the optics so complex- these grahics show a lot of information in a easy readable way.

    Field curvature; My most basic problem with lenses especially when used at infinity. ( architecture /landscape)

    I would really like to see the same tests with some 24mm lenses – the more wide angle the more field curvature is my observation. What i notice often with these lenses is a flat curve wide open, but then a very curved field when stopping down that will recover through depth of field by f8…
    nikon 24mm 1.4 the f version, nikon 24mmf3.5 PCE…
    So if you have the time… of coarse i would like to see the latest lenses… s-bayonet nikon , sony etc

  • El Aura

    I would guess the main reason for the larger variability in lens performance with zooms lies in the fact that they have more moving (optical) parts and to a lesser degree that they simply have a (very) large number of lens elements (as some higher end primes also have a very large number of lens elements).

  • David Bateman

    Excellent article. Can you please test this thought. Most mirrorless cameras like Panasonic, Olympus and Sony have in camera focus peaking. Can you compare your find edges in Photoshop, to the live image on a camera in focus peaking that should indicate the best curvature of the image focus?
    I think that make sense. The live focus peaking mode shoulf indicate the same sharpest line of focus.

  • Ok, so now all we have to do is find someone who do the “find the edges” thing for all the lenses and share the results in a convenient website 😀 😀 😀

  • Federico Ferreres

    Fantastic article. it’s like getting to the otherwise elusive soul of the lenses, at least with regards to focus and sharpness, it’s helped me understand them much more. Field curvature and astigmatism become visual things that are easier to understand without error. I think these heat charts are so illustrative, they should win an special award from Mythbuster or some new category of the Nobel prizes.

  • Dave Hachey

    As always, thanks for an excellent article. Although I sort of knew the conclusion at the outset, it validates my decision to focus more on primes for my Sony system.

  • As always informative and entertaining. The results are I suppose what we all sort of expected, these aphorisms are “truth adjacent” but shouldn’t be regarded as dogma.
    Given the results with “good” zooms, it’s fairly safe to say that the kit zooms would only be that much the more so (and why newbies are blown away moving from kit zoom to plastic fantastic primes). I suspect they are less common, but is there anything to be learned from 35/2.8 macro lens thrown in this mix? (Pentax has one, I’m not sure who else) since these should be designed for purposes other than specifically cost or speed and perhaps may be better with the field curvature issue?

Follow on Feedly