1.27.2013

Crazy Lens Stuff. Wide Open and Up Close.


I met my friend, Anne, for coffee last week at my favorite coffee shop, Caffe Medici. Out of native silliness I pulled out my camera and made a cursory shot, from a high angle, of my cappuccino and a luscious chocolate croissant. It was a throw away image. Something I might look at for a few seconds before I re-formatted my memory card in preparation for something else.

Anne looked at the image on the back of the camera and then gave me a wry look. "Really?" she asked. "That's all you've got?"

A bit ashamed at being caught in full on hipster photographer mode I sheepishly grabbed the camera again, set the lens at it's widest aperture, focused at the closest focusing distance and pulled out the LCD screen so I could shoot at a much lower angle than I usually do. I snapped a series of three images and then we put the camera away and had a nice chat.

I didn't shoot much else during the day and when I got home I took a look at the images. I was about to erase them but I stopped and took an extra couple of minutes and put them on my desktop.  I took a good look at the images and, like the nerd I am, started blowing them up on the screen to see what a lens really does do when pressed into its worst case scenarios.

And that triggered a whole train of thought for me about why we buy the lenses we do and what misguided metrics we use to make our selections.

When we look at lens tests we look at measurements that are made of a flat target but I rarely find, in real life, that I have a lot of call to photograph flat objects. When I do need to photograph flat objects I can easily reach for a macro lens that was make to photograph flat objects. But most of the stuff I photograph is three dimensional. This cup shot is an example of "real world." No matter how potentially sharp the edges or corners of the Sony 50mm 1.4 might be at its widest aperture, for the most part we'll never see the potential realized because the limited depth of field combined with the three dimensional nature of most scenes negates our ability to see those imagined results.

At the center point of this image, which is the same as the focal point, the image is generally sharp. Outside of this narrow boundary the sharpness and resolution of the outlying areas becomes immaterial.

This is the bane of most lens buyers. We don't have the ability to model accurately the critical factors of a lens in a meaningful way so we trust sites like SLRgear.com and PhotoZone.de and even DXO for their OCD testing of the lenses in which we might be interested. We also treat lenses as totally separate, stand alone tools instead of understanding their real role as integrated parts of complex imaging systems.

I suggest that we actually shoot lenses in the manner we will normally use the lens and make our choices on the merits of the lens in actual use. How does the image look compared to your initial intentions.

So, with all of this in mind I was in a vulnerable state when I made yet another fateful trip to the lens monger here in town, late yesterday afternoon.

Long story shortened, I bought a lens yesterday. It was an interesting buying adventure for me. I've been stalking this particular lens for a while. I am captivated with wide ranging, normal zoom lenses lately. I bought the much vaunted 24-70mm f2.8 Zeiss lens recently, tested it and returned it. Why? Well, the reason to buy the lens, for me, is the performance of the center part of the lens.  The people part. I don't shoot architecture and I'm not so concerned about corners and edges but I do like a sharp, very sharp inner core. I thought the 24-70mm CZ would be a big step up from the "interim" lens I bought back in early December for a PR job. It was the Tamron 28-75mm 2.8 SP lens for Sony. When I shot the two lenses side by side I found that the Tamrom was at least as good in the parameters I found to be important: Core Sharpness and Good Contrast near wide open. At f4 both lenses were good. Equally good. Why drop another $2000 on a lens that was no better than the lens I had in my hands? Not much. Back went the CZ (always keep your boxes until you are certain).

While my friends are much more impressed by the big Zeiss lens my wallet and my financial advisor are much more impressed with my pragmatic assessment of relative quality.

While I liked the idea of the extra reach of the lens I really wished for another 20mm or so on the long side of the Tamron and I really would have welcomed 24mm on the short end. My biggest beef with the CZ 24-70mm is also the very limited range of focal lengths. Shorter on the long end is even worse for me. Something like a 24-85mm makes a lot more sense in actual PR and event shooting. Even more length would be better.

So, back to the lens I've been stalking since I adopted the Sony Alpha system. It's been discontinued since 2008.  It's a rebadged Minolta lens. It's the 24-105mm f3.5 to 4.5. It's much smaller than either of the two lenses I mention above, in fact it's probably smaller than the Sigma 50mm 1.4. In fact, I'm certain it is.

I looked at it and rejected it back in December in favor of the faster Tamron lens but yesterday I looked at it one more time. But why?

At times I can be a sucker for "finder image."  When I put the small zoom on the a99 in the store and carefully set the camera's diopter I found myself liking the image in the finder. Really liking the finder image. I had two lenses in front of me. One was the Carl Zeiss 85mm 1.4 for the Sony and the other one was the 24-105mm 3.5/4.5. I pulled the 85 up to my eye and focused on some boxes about ten feet from me with the lens wide open. The type on the box wasn't particularly out of focus but it sure wasn't sharp. I chalked it up to AF focus error so I switched to MF and magnified the frame to fine focus. And it still wasn't sharp. Not critically sharp.

I started stopping down and checking and right at 2.8 I started to get succinct, sharp type in the center of the frame. I switched lenses and AF'd on the box again from the same position. The zoom rendered the type much more sharply. Wide open. I tried it at various focal lengths. Same results.

Stopping down one stop gave me excellent sharpness.

I tried the test again with the Sigma 50mm 1.4 for the Sony. It could match the sharpness of the zoom, in the center of the frame at around f2.8 as well. I wasn't seeing focus shift as I was focusing stopped down and using MF with magnification.

Since the zoom had been discontinued four years earlier and had been a shop demo for a long time I was able to negotiate an advantageous price and I bought it.

What does all this prove, if anything? That we tend to overlook many good, medium aperture lenses because they lack sex appeal but, in fact, are very, very good. Many people swear by super fast lenses but I find that they don't really deliver incredible performance at their widest apertures. My take is that our penchant for super fast lenses is based on two ideas, one of which is deeply flawed and the other was made irrelevant by our almost ubiquitous dependance on auto focus technologies.

First of all I think that super fast lenses became popular in the 1960's, 1970's through the 1980's because back then people mostly manually focused their lenses in SLRs and the faster aperture meant more focusing accuracy on the screen. The more limited depth of field also helped us get to sharp more quickly. Now that all our cameras are AF these parameters are rendered more or lens meaningless and further impaired in designs that exhibit focus shift upon stopping down.

The second parameter that comes into play is something that Erwin Puts talks about a lot. It is (according to Puts, a Leica expert) eight time harder to accurately make a lens element that is twice as large. Each increase of one stop makes the chances of creating a great optic eight times harder, times the number of elements in a lens. EIGHT times TIMES the number of lens elements.

I first noticed that some lenses that performed very well were slower than the lenses that were most popular or most used. For instance, I would say that in many regards that the Nikkor 55 Micro f3.5 is a sharper lens at f3.5 to f11 than just about any of the much pricier fast lenses on the market. Recently I purchased a Sony 85mm 2.8 lens and my results mirror those I've read in review sites. The lens is very sharp wide open. So sharp it can excite aliasing even wide open. Amazingly sharp. You hear similar stories about the older Contax 85mm 2.8 as well.

In my experiences with 85mm lenses the two Sony 85's tell the story. The big, expensive 1.4 lens is impressive to look at and impressive to look through. But it's hard to get sharp performance wide open due to the vagaries of AF and the added struggle with focus shift as one stops down. By 2.8 the lens is critically sharp (according to most tests) but then it is just on par with its much slower sibling. The 85mm 2.8 matches (to my eye) the performance of the bigger lens from 2.8 up to the diffraction limits of both lenses.

But the real thing a long term photographer should ask about the tools is, "is ultra narrow depth of field something I'll use a lot?" And then weigh that against, well, weight and price. I find most controlled shoots with longer lenses take place between f2.8 and f8.0. With a full frame camera an f1.4 aperture gives one such a narrow slice of sharpness that it becomes schtick.

Given a choice these days I would rather have a lens that's critically sharp but a little slower than a prestige lens that really only delivers when stopped down. In an odd twist the users of smaller formats may have an advantage when it comes to buying smaller, faster lenses (such as the Olympus 75mm f1.8 I talked about last week) because they use smaller diameter lens elements that can be more accurately machined and polished than optics made for larger circles of view...

So, how is the "new to me" little zoom? Where I use it, at f4 to f8, usually making portraits, I find it to be a fine lens and one perfectly suited for events. Is it so sharp it will call attention to its sharpness? No. But it's a nicely rounded sharpness that seems to work well for people. It's better on the a99 than on the a77 and I chalk that up to the interaction between the resolution limits of the lens and the resolution density of the two digital imaging sensors.

I'm heading out the door to shoot with the smaller lens right now. For a little while it will join the Tamron 28-75 and the Minolta 24-85mm in the equipment drawer. I'm putting them all through their paces right now and the winner will get to stay while the other two will be sent packing. Right now it's neck and neck between the Tamron and the Sony. More to come.

added later in the evening: And the early returns are good. The 24-105 has a unique look to it. Sharp but roundy..... I'll post a few and then we'll test later, in depth...












22 comments:

Craig said...

Generally speaking, I'm inclined to agree that faster lenses are not necessarily better. I and others have noticed, for example, that the old manual-focus Nikkor 50mm f/2 lenses tend to perform better (and are more convenient) than the larger, heavier, more expensive f/1.4 and f/1.2 Nikkors. It makes sense, because with an f/2 lens the designers don't have to make additional compromises to get decent performance at f/1.4 because the lens doesn't go there. Instead, they can focus on optimizing for a somewhat narrower range of apertures. This is also, I think, why Nikkors usually don't have insanely short minimum focus distances. I have an old Sigma fisheye that can focus down to six inches, while the equivalent Nikkor only goes to one foot. The Sigma can make some interesting images due to its close focus ability, but the Nikkor, overall, has better IQ. The Sigma is, however, more than adequate for most purposes (well, most purposes for which one would choose a fisheye).

Craig Yuill said...

Kirk, I have noticed that kit lenses these days can produce excellent image quality. I haven't had the desire to rush out and buy faster lenses as I might have in the past, although I might pick up a couple of f/1.8 or so normal/tele primes down the road in order to get faster shutter speeds, when needed, in lower light. I also agree that shooting wide open is not always (seldom?) desirable. I find that some depth of field is needed in most shots, even in a shot I made this morning of a pancake I had just cooked, and shot from directly above.

Sharpness is important, but one characteristic of lenses I have been paying more attention to in recent years is bokeh. Some of my lenses produce very smooth out-of-focus areas, and others not so much. The bokeh in the out-of-focus areas of the photo you posted looks very nice. I know that bokeh is also influenced by the type of background in the shot, but I'm curious if bokeh ever influences your decision when buying a lens.

Peter F. said...

Hi Kirk, You're such a prolific (and excellent) writer. I'm being Off-topic here, but do you type your stuff right up... or, by any chance do you use voice recognition software? Just curious. My wrists would ache after typing what you are able to do... Peter

tomt said...

" It's better on the a99 than on the a99 and I chalk that up to the interaction between the resolution limits of the lens and the resolution density of the two digital imaging sensors."

I'm wondering if this is too deep for me or a typo.

Craig Yuill said...

I think Kirk meant "It's better on the a99 than on the a77 ...". (No?)

Kirk Tuck said...

Hi Peter. This is going to sound strange but I can type faster than I speak and I like to actually look at the words as they come onto the screen. I tried Dragon Dictation recently but I got stage fright, thought what I was dictating sounded dumb and was generally turned off by the process. Even stranger, when I'm out at a coffee shop or on a plane or riding in a car with someone else driving I do all my writing long hand in a larger notebook with a nice, heavy ball point pen. Old habits die hard and new methods sometimes obscure the message. I type with a blue tooth keyboard in my lap. No wrist issues yet and I've been at it for years....

Kirk Tuck said...

I'm attracted to the background rendering of some lenses and not others but I've never rigorously analyzed how it works or how to test for it. I shoot lenses and look at the stuff that comes squirting out. Sometimes the whole package looks good and sometimes not. When I buy new lenses I buy them from a dealer who will allow me ten days to test the lens and see if the lens and I are a good match...

Paddy C said...

Your thoughts on fast lenses are so true it hurts a bit. I have the ZA 85 1.4 and the Sony 85 2.8. The Sony is quite a performer even if you ignore the price and build. I couldn't resist the 1.4 but I don't know that I need it (as in at all).

Still, the Zeiss is acceptably sharp wide-open and very sharp at f2. So it gives a stop advantage there. The bokeh is also better IMO. The Zeiss is substantially better in the corners. This is the big difference and, from what I've seen, is in keeping with the general philosophy of the Zeiss AF lenses for Sony. Which is, a prioritization of edge-to-edge performance over absolute centre performance.

In practice this means that some other lenses will be better in certain areas (such as the Sigma 85 in the centre) but the Zeiss lenses are better when the entire frame is evaluated. And in some cases this overall advantage is there from wide-open and by the time the centres are equal the other lens will never catch up to the Zeiss corners regardless of aperture.

Carlo Santin said...

Sharpness is important but I find most modern lenses are plenty sharp, even the kit lenses that get no love on certain gear forums. I don't do much architecture or landscape photography so corner performance is not critical for me. Bokeh I don't really care about; I must be a bokeh illiterate because I really can't distinguish creamy bokeh from harsh bokeh. If you are spending that much time looking at the bokeh then the image must not be very interesting. I'm not very well versed in the technical aspects of lens design and I'm not a pixel peeper. I think there is much too much fuss about bokeh.

What I look for more than anything would be what I call lens rendering...what does the overall image look like? Do the lens provide pleasing tones and contrast? Does it provide that certain "look" that I find attractive? I know it when I see it. I'm also a sucker for cheap lenses (hey I love an underdog). My old Nikon 50mm 1.8 series E is a great lens, especially on film. The Zuiko 50mm 1.4 has that look that I love...it might be my favorite lens.

That's it, I'm a bit of a simpleton when it comes to this stuff. I like my cameras as uncomplicated as possible, specs don't really get me excited, and lenses I like to get to know over time.

Kirk Tuck said...

You are correct. I've changed the passage to read: " It's better on the a99 than on the a77...." Thanks for letting me know.

Kirk Tuck said...

Love this comment from Carlo: " If you are spending that much time looking at the bokeh then the image must not be very interesting."

Kirk Tuck said...

I found the bokeh on my image above to be prickly and insouciant. With overtones of ripe plum and optically clear cement...

Mike said...

"The second parameter that comes into play is something that Erwin Puts talks about a lot. It is (according to Puts, a Leica expert) eight time harder to accurately make a lens element that is twice as large. Each increase of one stop makes the chances of creating a great optic eight times harder, times the number of elements in a lens. EIGHT times TIMES the number of lens elements."

Just to clarify, if this is correct, it's not eight times the number of lens elements. It's 8 to the power of the number of additional stops.

Let's say we have an f/4 lens.

To make an f/2.8 lens, it's 8 times more difficult to make than the f/4 lens.

To make an f/2 lens, it's 64 (8 to the second power) times more difficult to make than the f/4 lens. (8 times the difficulty of the f/2.8 lens.

And for an f/1.4 lens, it would be 512 (8 to the third power times more difficult to make than an f/4 lens. (8 times the difficulty of the f/2 lens)

In practical terms, I have no idea what any of this means. I like my Canon 50mm f/1.8 lens just fine, and maybe a factor of 8 doesn't amount to much in the real world - I don't know for sure of course.

But for what it's worth, from retailers I've spoken to, the Canon 50mm and 85mm f/1.2 lenses are very commonly returned/exchanged by customers, while the f/1.4/1.8 models are much less so.

Anders C. Madsen said...

Careful now! Next thing we know you will be looking at the "legs" it has down the side of the glass... ;)

Kirk Tuck said...

As photographers we generally love the IDEA of super fast lenses, like the f1 Noctilux and the various 85mm 1.4's and 1.2's it's the day to day reality of them that takes the wind out of the sales... Bigger, heavier and in most uses no better than their slower brethren. I have a friend who saved for over a year to afford the Leica Noctilux f1.0 lens. In practice it was largely unusable and pretty much a one trick pony. It was back to the dealer in less than 72 hours.....

I owned the original Canon 85mm f1.1.2 L lens back in the days of film. I loved the images but it was so slow to focus it was overwhelming. As the price continued to go up and (due to QC constraints) the number of available copies shrank I was happy to divest for a profit after a year of weight lifting....

Etc. Etc.

Jeff said...

Zooms with short telephoto bothers me too. The 18-55 kit lens = 82.5 full frame at 55mm with aps-c always seemed not enough. Perhaps I was used to my Maxxum 7 and 24-105 lens. Yes, on aps-c the effective 36-157 works well, but front heavy on my A33. I got an older 18-70 Konica Minolta to replace the 18-55 that got hit on a corner when I dropped my camera. Definitely better is the 70mm on aps-c = 105 full frame. Even though reviews say 18-55 is better, pictures looked clinical to me and the older 18-70 stuff looks like...photos.

I don't take much in lens reviews. Pop Photo 20 years ago had their metrics and always said "Slides were sharp and contrasy", even for the 28-200 and 28-300's which I disliked for their lack of contrast - dull looking even with just 4x6 prints.

I got a Canon FD to Sony Af converter and put my Vivitar 70-210 on it. Even with the converter glass inbetween lens and sensor, photos still had that "look" the Vivitar gives on my F-1.

Anonymous said...

What you talk about here brings to mind my plea to companies like Cosina. Why oh why do all their new lenses have to be so fast? It makes them big, heavy and expensive. It puts them out of reach for majority of people like myself who have this as a hobby. I happen to like the micro 4:3 cameras. So what does Cosina bring out? Lovely manual focus wide angle and standard lens that are blazing fast, too large and heavy, and completely out of any reasonable price range for all but the very well off. Where are the moderate aperture, reasonably sized and priced lenses for the other 90% of us?

Why do they do this? Is it for bragging rights? After all one of the advantages of m4:3 is the small size bodies. It does not make sense to me to put a 17mm wide angle on these bodies the size of a Coke can. Just my $.02, sorry for the rant.

John Robison

Kirk Tuck said...

The lenses you seek are called the Sigma 19mm and the Sigma 30mm. Slow, sharp, cheap and incredibly well reviewed by people who know. I'm advocating for more like these and less like the Leica Noctilux at nearly $9,000... Vote with your wallet. They'll pay attention.

Kirk Tuck said...

Weird coincidence: I left the comment above and went to visit TheOnlinePhotographer where I found out that Sigma just introduced a new 60mm lens for m4/3 and Nex. It's a 60mm 2.8. Didn't I just ask for one of those...? Hope over to Mike's site and take a look....http://theonlinephotographer.typepad.com/the_online_photographer/2013/01/new-sigma-60mm.html

Anonymous said...

"Lovely manual focus...." and of course real aperture rings.

Call me a crank, but I have an irrational dislike for auto focus lenses. Give me that wonderful silken feel of a smooth, well damped, zero backlash, manual focus lens. A real aperture ring with half click stops, and an engraved DOF and distance scale on the lens. Hey, I admitted to being a crank and I did say it's irrational but the current crop of AF lenses leave me stone cold.

Strangely, Olympus's little plastic bodied 15mm f8 lens cap lens appeals to me. I would actually spend the $50 they want for that thing.

John Robison

Anonymous said...

Well Kirk, after I made my reply I looked up the two Sigma's you referenced and WOW, they are cheap. Someone is thinking about "the rest of us" after all. Thanks for the heads up.

John Robison

lakrids404 said...

I don't know if you buy/use second hand lenses, but the Minolta 35-105 gets good user reviews and is rather cheap.

Personally as an armature I really like (except from some flare problems, that can sometimes can make me seriously thinking about selling it in anger) my (used) Sony 35/1.4, and i had situation with my a900, where I to had used fully open, but normally I prefer it at a least f/2.

It took the the place as my most used lens from the Minolta 85/1.4, which I still use to portraits, nice bokeh ;) and colors. But it's rather big and that can be intimidating to some people.

Good bokeh for me, is on that don't distract from the main motive, which I think some f.eks. a nervous bokeh can do.