« Forge Ahead! | Main | Quote o' the Day: Doug Thacker »

Friday, 10 February 2017


Interesting conclusions.In my starter days it was a 50 mm, a 28mm and a 135mm lense. That was it. Nowadays a 24-105 (long discontinued by Nikon however freshly rebuilt and doing good service) and a Nikon 70-200
f/4.0 long range zoom live in my camera bag with a Nikon 20mm f2.8 (handy for photographing the inside of railway cars) all keeping my Nikon D750 warm. My need for anything else is naught; mostly photograph trains and their ilk as well as ocean-going ships in Hamilton Harbour.
"If"I could get those same pieces of glass as fixed focal lengths, I'd require many more pieces of glass and that in turn would make my list with camera bag to port even greater.
So for me zoom is it, although I would dump the that 70-200 zoom if I could find a nice 135mm fixed lease; in the Nikon brand.

I'm not that bothered by sharpness; a lens doesn't have to be extremely sharp, just sharp enough. If I was that bothered I would use a tripod, all the time. I don't.

Great to see Douglas Adams quoted; more truth in one number.

And I loved the geeky math jokes incorporated at the end.

Last month you said you "fell in love" with the Panasonic 12-35mm zoom when you had it for testing -- so much so that you were now considering it for OCOLOY. So maybe zooms aren't all that bad?

[It's not perfect of course, but I really do like that lens. --Mike]

Well written and explains why even though I appreciate what a zoom lens may be able to provide (I like them for travel), I have only owned three zooms in thirty years. Now to eating avocados. :)

But, but, but ...DxO said 8-)

The only tests that count are your own tests. Lots of people claiming real world tests, but I've never seen a real world shot that looks like anything that I do.


BTW Why would I carry a 70-200 f/2.8 when all I need is an 85mm f/1.8?

For a couple of years, I used a Nikon D90 with a 16-85mm DX Nikkor zoom. This lens was plenty sharp "most of the time". I would get a soft image from time to time (no focus error, using AF or live view AF or manual focus in live view gave the same result). Especially in portrait orientation, the upper right corner of the frame was sometimes very soft - even noticeable in an A3+ print. Just like the damn lens wanted to remind me that it still existed. Now I don't know if that lens had soul - but for sure it didn't have humor.

Best, Thomas

So, what is the conclusion? Never buy a zoom, or don´t worry about this article, and go out and and take photos with that zoom you love so much.

Thanks for the heads-up on that piece, Mike. Roger's torture-and-test lab certainly does interesting and genuinely informative investigations, answering real questions we ask.

The handiness and generally excellent quality of today's good zoomers vastly outweighs their faults for me. My own rule is simply this: if can control the camera position for a scene and if I have a prime that fits I'll use the prime. Otherwise (and usually) I'll use the best zoom I have. And I have some doozies! My Sony 24-70 F2.8 GM kicks ass. The detail definition I'm getting from it on the A7R2 is the best I've ever seen, even including my Phase One.

In the final consideration, at least for me, the lasting impact and value of most images doesn't rely on optical lens properties.

Gear-heading is a much more expensive hobby than photography. (Mostly.)

Mike, Great read, he writes a good analysis of prime vs zoom lenses. Most of my photography is travel, and thus I tend to gravitate to zoom lens; however when I look at my wall, and think about which lens I used on which image, I find that most of my best work was done with a prime lens. It can be difficult to change lenses when you are traveling in challenging conditions. Also one does not want to burden his back with a lot of glass. So with that said, I just rented the 18-135 Fuji lens from lens rental to see if it will work. I know it will not be in the same league with my 16-50 f 2.8 fuji lens, but it may replace the 55-200 and 16-50 on future trips along with the beautiful 35 f2 fuji lens. I always carry a prime where ever I go, they truly offer better IQ. Thanks again for directing us to the article. Eric

I am, in fact, planning to take my 85mm/1.8 rather than the 70-200/2.8 out this afternoon. But also the 135mm/2. Partly just in the mood, partly wanting the extra stop.

Given the widespread availability of 85/1.4 lenses that aren't that huge, it kind of gripes me that I can't get a 28-85/2. I don't understand zoom lens design (I suspect none of us here really do), but it feels to me that the zoom shouldn't need much more front element than the longest focal length needs for the intended max aperture, hence the 28-85/2 might be heavier or somewhat longer, but the front element shouldn't be particularly larger than the 85/2 needs.

That must be nonsense, right? Since so few fast zooms exist? (f/2.8 is not fast. The Sigma 50-100/1.8, that's fast, but it's for APS-C and they don't make it in M43 mount, and then it's not a "normal" range on APS-C or M43. Still fast, at least, though.)

An interesting read, thanks for the link.

Not much surprising to me in there, although confirmation by objective testing of things I've thought I knew is nice.

I hate lens testing, but did a lot of shooting of test targets when I got my Oly 12-100/4 Pro. What's surprising is how well my one particular copy does. Although I shot at wider apertures as well, I can only pixel peep so much, and only looked closely at f5.6 and f8. (Also, not coincidently, my most used apertures.)

The zoom was, as close as I can see pixel peeping @ 100%, without all that test gear gear, as sharp, center and edge, as the Oly 45/1.8* prime, and darn near as good as the famously sharp 75/1.8. I doubt if anyone could see a difference short of a wall size print.

It's better than the older Panny 20/1.7, not quite as good as the Oly 25/1.8.

I was particularly impressed with how it holds up in the corners.

I know a few empirical things:

1. Pure lens resolution is only one of several factors that determine image sharpness, and not always the most important one.

2. Lens resolution and perceived sharpness are only important factors in some images I make, and only crucial in a minority.

3. I have never been, and will likely never be, content with 2-3, even a handful of primes.

4. I deeply detest changing lenses in the field. I shoot a lot out in woods, fields, swamps, etc. Wind, dust and damp, no place to set things down, not enough hands, and so on and on . . . and the time consumed, yuck. I am happier carrying two bodies with one zoom on each, covering 24-800 mm eq. than changing lenses.

Thus, I am highly delighted that the range and quality of zooms has continued to improve since the first one I used, a so-so Nikkor 43-86 that would look like bottle bottom glass compared to this 12-100.
* Now the Panny 42.5/1.7 has come on the scene. The tests of a handful of samples on the web and one private one show it to better the Oly in the outer zones.

In the 1990's I picked up a zoom lens... a 24-120mm AF Nikkor. It's reputation wasn't the best but the range in focal lengths was ideal and I used for years covering events. As he newer models of this optic came out I would visit my local camera store and shoot some exposures outside with my older lens and the newer model. Each time I did this, my older model 24-120mm lens always would out resolve the newer optics (I did this comparison test quite a few times with always the same outcome).

Another time I needed to digitally capture a series of very finely detailed, circa 1860s maps. I couldn't remove the maps from the State Library but had to photograph them there under guard (they were very valuable). I had to use their copy stand which wasn't all that high and totally negating using either my MF 55mm Micro-Nikkor or my 60mm Micro-Nikkor. Before the shoot, I determined that a 24mm lens would be ideal to use. I borrowed a MF and an AF 24mm Nikkor and ran tests... neither was sharp enough at the reproduction ratio I was shooting at. For the heck of it I tried my 20-35mm AF Nikkor and it blew away the two prime Nikons and to beat that it had an amazingly flat field. Used it to photograph the maps and later to make 100% prints for an educational exhibit.

Over the years have been able to test many samples of the same lenses. Most of the time you couldn't tell the difference, but very often one sample definitely outshone the other.

[I can second that last sentence. --Mike]

Thanks for pointing us to another great article by Roger. You're right that he's the foremost writer on matters of lenses. This article really puts our minds straight in terms of the frequent articles we read about this or that lens being great or crap. Unfortunately, very few people have the ability to test a range of the same lens like Roger does.

Zooms are for tourists. I thought everyone knew that.

Having owned some zoom lenses I didn't like, but others swore on, and some zoom lenses I did like, and others didn't, it's patently clear to me that a) There is variation in quality within lens batches, b) even for equally good zooms, owners' tastes also vary and influence their opinions. So, if you buy a zoom, even if it's a multi-thousand $/€/£ lens, be prepared to return it for another copy. Also, don't judge a zoom (or indeed any lens) in absolute terms based on testing just one copy of it.

And Roger Cicala is always entertaining and informative, though I wonder where he finds the time to get all this fun stuff done.

It's been many years since I've used a zoom, but when I used one I got some great shots, and that includes one of the earlier Tamron 18-200mm lenses, which I used before I knew I wasn't supposed to and which I found to be remarkably high contrast and distortion free even at the edge at 18mm. I know I got shots with that lens I couldn't have gotten with a prime, because I'd have had the wrong prime on my camera.

But it didn't take long to figure out that zooms are - excuse my stating the obvious here - large and heavy and slow, often expensive, and, well, zooms. I much prefer the single focal length for its size and weight and speed, its generally superior optics, and, most of all, for the fact that it gives me a limited structure within which to work. If a zoom is free verse, a prime is a sonnet, or haiku.

In general, structure is key to creativity, because staying within certain confines forces invention, but also because the self-imposed limitations soon become second nature and then fade beyond your awareness. In photography, this is when you start seeing like the lens, which is a good place to be. You don't get there using a zoom.

Addendum: Sorry, that was a Tamron 28-200mm.

Miserere wrote "though I wonder where he finds the time to get all this fun stuff done" but he only does the fun stuff now, see "Roger gets a new job" at https://www.lensrentals.com/blog/2012/05/roger-gets-a-new-job/

Okay. I've read the article, thank you Mike for posting the link. I learned a lot from it.

Reading it raised a question and, not being an optical physicist, it may be a stupid one but here goes. Based on the quoted fact (in the article) that f1.4 primes will be softer than f2.8 primes, is it safe to assume that smaller lens elements will be sharper than larger ones? Does this mean then that, given similar manufacturing tolerances and quality control, a M43 25mm f2.8 will necessarily be sharper (better) than a medium format equivalent (and, to a lesser extent, APSC and FF)? Could this help explain why my Oly kit is now getting more use than my FF Nikon? Is there a tipping point between camera resolution and lens quality (and therefore price)? If so, then this would certainly be another argument for the "invest in lenses and treat cameras a disposable/ replaceable".

BTW, a very good friend of mine insists that the answer to the LTUAE question is 37 and not 42... but he can be a bit eccentric at times.

My favorite field of view is the GR's 28, but I can zoom to 200 in about 16 steps.

Oh, and I enjoyed re-reading "All lens tests are Wrong". But atdmittedly it took me almost a decade until I got to that understanding myself, too :(

And it's a state of heightened awareness, when the mind can concentrate on the world and its appearance aka. subject matter, and eye and brain and camera come to a framing regardless of any mtf or other figures!

So now I live happily with both a fine zoom 24-80mme (the olympus incarnation) and if I want to travel light the 35mme and 90mme primes. This is good enough for most of the days, and when it's not, it is quite often just my lazyness to move.

Zoom lenses are the SUVs of photography -- good at a lot of things, great at none.

Toyota sells a lot of SUVs and Canon sells a lot of zoom lenses.

So: There is (are) no best, only the lens(es) with which you shoot. Enjoy your photography and images, they are what you make it and them.

That is perhaps the best overall article on lens testing ever. I can say that because I read it in its entirety. :) Seriously, most such articles bore me to tears. That doesn't mean they aren't valuable or even important. But if I don't read them because they are anywhere from non-engaging to horribly written, it's not too valuable to me personally.

I prefer prime lenses not only because they are, in most cases, optically better, but because they are both easier to use (for me) and are typically smaller and lighter.

I love reading Roger's articles, they are always infused with his great sense of humor. I also very much like his "polar plots" that show sharpness around the lens at different angles. This so much more useful as a means to evaluate lens performance I don't know why everyone, including the manufacturers, don't use it.

His conclusions align with what I emphasize to my Design for Six Sigma students when I teach them stats: If your measurement system is good enough (and Roger's is), you can almost always demonstrate differences that are "statistically significant". What's important, though, is not necessarily what is "statistically signficant", it's what's practically significant that is important in the real world.

What I read as the key takeaway from Roger's testing (and what I take that he is alluding to) is that while it can be demonstrated statistically that primes on the whole are sharper than (most) zooms, from a practical perspective, many zooms are sharp enough; that is, they are "fit for purpose". From a real-world, practical perspective, "fit for purpose" gets the job done most times in most situations.

This is why so many folks (and pros) can use zooms to get their real work done.

The comments to this entry are closed.