A77 - Nearly here

[David Kilpatrick]

The photozone.de test is just plain odd - I suggest you read their tests for a few other lenses, and then you'll see that the NEX 16mm significantly outperforms many very expensive lenses of the same coverage, but is then subject to a poor 'verdict'. It's actually a test from which the results can be pulled (distortion, vignetting) to make a case for the lens and not otherwise.

I am sure the corner result from the NEX at 1250 ISO with the 16mm shows more detail than a KM5D shot at ISO 400 dead centre. Personally, I don't care about old versus new, FF vs APS-C, mirrorless versus true SLR. I will find out whether I care about the SLT design shortly; it may be as I fear, something which reduces contrast and sharpness and suffers from ghosting; it may overheat, who knows? I was not blown away by trying the camera at photokina, the viewfinder flickers when you pan and does not (to my eye) look as 'clean' as the Panasonic finder. But if I find something which works, I'll use it.

I suspect the test reports on the 16mm are partly down to the so-called 'prototype' variation, but more down to tester/reviewer variation. Anyone using a standard test suite like Imatest must either invest in a small warehouse to test wide-angle lenses, or accept that you can't test wide angle lenses properly.

As for the field flatness issue, it's almost certainly impossible to track the NEX 16mm into the extreme corners at focus chart distance. To get a table lamp in the corner of my room dead sharp by placing the focus point in the extreme corner of the shot, at about 13ft, the lens ends up with centre of the field focused around 5ft. It's that extreme. But - at f/2.8 I can get a surprisingly clean image right into the extreme corner. It's a sudden falloff, and it's due to a sudden focus distance shelf.

David

[David Kilpatrick]

OK, here are the sort of unscientific but very meaningful tests I do with lenses! I used to use test charts and stuff from 1974 to the mid-80s. Eventually I learned that practical tests and careful study of images can tell you much more than a flat chart at 27X the focal length (etc).

Second - the extreme top corner, 100%, of a NEX 16mm shot taken wide open, first focusing on the lamp with the focus point at the field centre, then recomposing to place the lamp in the corner, without changing focus. Sorry, the order is odd but I'm having trouble with the attachments system on PHPBB3 - must investigate.

First - focused by placing the focus point in the top right corner, directly over the lamp and shooting. I have left a little more on the centre-focus+recompose shot so you can see how suddenly the focus field recedes. But the second shot shows that photozone's result is purely an artefact of flat field close up chart testing, and the lens does not actually turn top complete mush in the corner. These are normal raw conversions - first one is ISO 1000, second ISO 500, due to exposure differences created by the focus point position.

[Greg Beetham]

Did you do those from a tripod David? I'd imagine at wide open you wouldn't need to move very far when recomposing too nearly invalidate the experiment.
Greg

[David Kilpatrick]

Irrelevant Greg - no tripod. A couple of inches, or even six inches, in thirteen feet is not significant (you'd be amazed at the depth of field even at f/2.. I have been using the lens for hundreds of shots in many conditions, including many pictures wide open. I know how it performs. This is not a 'test', just a quick demonstration of something I know about the lens.

If there's one thing I know about it's lenses. I can't begin to count the lenses I've tested for magazines over the last 40 years, and I did indeed start on this as a teenager.

I really don't have time to do this while trying to work through my photokina reports and pictures, but I feel so strongly about the negative propaganda put out in respect of the 16mm E. I am not even entirely sure there were any faulty prototype or pre-production lenses - just inept users, I suspect. I have been cautious myself and assumed the worst.

One of the most annoying aspects of this bad review status is the opinion that the 16mm E does not match comparable 'primes'. Well, there are NO comparable primes. There has never been a single 16mm f/2.8 rectilinear wide angle made for APS-C, or even anything slightly wider such as a 15mm or 14mm. Or slightly narrower like an 18mm. It took Leica 50 years to finally create a 24mm f/2.8 and no maker has released a new generation 24mm f/2.8 for full frame yet - only high speed superlenses.

If you check almost any reviews of zooms, you'll find the 16mm f/2.8 does compete.

This is such a great lens to use that it frustrates me to find people being put off by bad reports, when what Sony did was create something entirely new and unexpected. Do you have any idea just revolutionary it is to make a lens with this angle and aperture using only 5 elements? The last innovation on this level was Minolta's 28mm f/3.5 G for the TC-1.

David

[Greg Beetham]

I'm not overly attracted to wide lenses myself but they are very useful when you DO need one, and when you do it needs to do it's job. I saw some pix early on from the E16 and I thought they were very good, I don't know why it got a bad review here and there when it seemed to me to be as good as any wide I'd seen before...and better than quite a few actually.
Someone, agorabasta I think it was took some pics over in Russia when the NEX first came out and I was amazed at the quality from that 16mm lens in some of those photos.
Greg

[pakodominguez]

I don't have a NEX (nor will I ever) so I cannot test anything (and yes I did handle one). If the lens is so great why did photozone thump it then? (oh I hear the cries of samples variation..sure I'll buy it) Still looks zoom quality to me and not prime. And we're talking around a 24mm equivalent, when you get to UWA (below that) things get much more demanding edges wise.
NEX isn't for me that is very obvious

You handled one -but did you ever "use" one?
whatever.
I can respect that "NEX isn't for me", but it is annoying reading you trashing the NEX system just because it is not for you. NEX is a wonderful system for some people and WE are happy with it's limitations -the 16mm f2.8 is not one of those limitations, au contraire!
You can keep waiting for a system tailored for your needs, in the meanwhile be coherent: "try" the stuff before you start "reviewing" (trashing...) it.

[agorabasta]

That 16mm is quite capable of producing great results, just as it is quite capable of delivering a total abomination.

A busy scene with a lot of detail that is neither quite blurred nor quite sharp would definitely look plenty ugly with this lens. It's exactly the mid-blur that it simply can't render well.

E.g. using f/2.8 while focusing not farther than 1m away would be often quite fine. Then tighter than f/8 while focusing over 3m away is also quite often very nice. But focusing at 1.5m away at f/5.6 or so would almost guarantee a perfect abomination. It surely depends on the scene, though.

[catalytic]

That 16mm is quite capable of producing great results, just as it is quite capable of delivering a total abomination.

A busy scene with a lot of detail that is neither quite blurred nor quite sharp would definitely look plenty ugly with this lens. It's exactly the mid-blur that it simply can't render well.

E.g. using f/2.8 while focusing not farther than 1m away would be often quite fine. Then tighter than f/8 while focusing over 3m away is also quite often very nice. But focusing at 1.5m away at f/5.6 or so would almost guarantee a perfect abomination. It surely depends on the scene, though.

So the worst performance is neither at MFD nor at infinity, but around 1.5m? Really? This is the first time i have heard of that. Not that i am an expert here, but it just seems odd, as most lenses would be tuned for infinity, and some lenses are tuned for 2-4m (portrait distance), but to have a "bi-modal" tuning for MFD and infinity would be unusual, no?

[catalytic]

The photozone.de test is just plain odd - I suggest you read their tests for a few other lenses, and then you'll see that the NEX 16mm significantly outperforms many very expensive lenses of the same coverage, but is then subject to a poor 'verdict'. It's actually a test from which the results can be pulled (distortion, vignetting) to make a case for the lens and not otherwise. [...]

I suspect the test reports on the 16mm are partly down to the so-called 'prototype' variation, but more down to tester/reviewer variation. Anyone using a standard test suite like Imatest must either invest in a small warehouse to test wide-angle lenses, or accept that you can't test wide angle lenses properly.

As for the field flatness issue, it's almost certainly impossible to track the NEX 16mm into the extreme corners at focus chart distance. To get a table lamp in the corner of my room dead sharp by placing the focus point in the extreme corner of the shot, at about 13ft, the lens ends up with centre of the field focused around 5ft. It's that extreme. [...]

I have often wondered why some relatively expensive wide-angle lenses show surprisingly poor corner performance on photozone's tests, but when asking Klaus about it, i have always gotten the impression that he is careful and thoughtful about things, and actually brought up caveats like field curvature and testing distance himself, which means he considered those factors in his testing methodology. However, your explanation does make a lot of sense, and it would make sense that he cannot optimize his testing setup for wide-angles as much as he would like given the resource constraints.

I will find out whether I care about the SLT design shortly; it may be as I fear, something which reduces contrast and sharpness and suffers from ghosting; it may overheat, who knows? I was not blown away by trying the camera at photokina, the viewfinder flickers when you pan and does not (to my eye) look as 'clean' as the Panasonic finder. But if I find something which works, I'll use it.

I am anxiously awaiting your evaluation of the SLT system, as i am very seriously considering the purchase on an A33/55. First thoughts: would you go for an A33 or A55 based on the specs, and why? I am not sure the A55 is worth the extra $100, as i don't need the 16mp. But i wonder if the 3x deeper buffer would be useful in the field... not sure as i don't know how usable the continuous shooting is on the SLT with Live View blackout / delay and flush times (which have been tested at 32-50 seconds by I-R and DPR, respectively, using a SanDisk Extreme III card, and by others at 10-12 seconds using the fastest card available).

[catalytic]

[...] If there's one thing I know about it's lenses. I can't begin to count the lenses I've tested for magazines over the last 40 years, and I did indeed start on this as a teenager.

I really don't have time to do this while trying to work through my photokina reports and pictures, but I feel so strongly about the negative propaganda put out in respect of the 16mm E. I am not even entirely sure there were any faulty prototype or pre-production lenses - just inept users, I suspect. I have been cautious myself and assumed the worst.

One of the most annoying aspects of this bad review status is the opinion that the 16mm E does not match comparable 'primes'. Well, there are NO comparable primes. There has never been a single 16mm f/2.8 rectilinear wide angle made for APS-C, or even anything slightly wider such as a 15mm or 14mm. Or slightly narrower like an 18mm. It took Leica 50 years to finally create a 24mm f/2.8 and no maker has released a new generation 24mm f/2.8 for full frame yet - only high speed superlenses.

If you check almost any reviews of zooms, you'll find the 16mm f/2.8 does compete.

This is such a great lens to use that it frustrates me to find people being put off by bad reports, when what Sony did was create something entirely new and unexpected. Do you have any idea just revolutionary it is to make a lens with this angle and aperture using only 5 elements? The last innovation on this level was Minolta's 28mm f/3.5 G for the TC-1.

David

And i, for one, am very grateful that you not only patiently shared your knowledge and experience with us, but took the time and effort to do some testing that illustrates this for us.

I am curious to know why / how Sony are pulling off good performing lens designs with so few elements and why they wouldn't want a couple more elements in there to control aberrations and the focus field better insulated from the level of precision and QC required when the design calls for such high contribution of power from each element? If they are doing such revolutionary things, why are they the only design team that's capable of it, when presumably, most lens design teams should have the same NASA-level computational abilities and ability to optimize designs these days?

BTW, please don't mistake my tone for one that is questioning or challenging your assertions; i am just very curious and am fascinated about such things!

[David Kilpatrick]

We have to remember that 24mm is not as extreme as you think - Fox Talbot's origjnal 'mousetrap' had a lens with exactly the field of view of a 24mm (but only around f/16!). The minimal element thing is the Holy Grail of lens designers, from Thomas Sutton (one element - water in a glass skin!) to the Wray Unilite (the first 5-element f/2 standard lens) and on through the 1960s where Leitz were able to turn seven-element designs into six-element using new glasses. I was always surprised that Fuji's experiments with 5-element fast standards and EBC coating, in the late 1970s, were never noted as a future direction.

What's unusual about the 16mm is that it is also using aspherics/airspaces as part of the design (a concept seen in the TC-1's G-lens, and the very simple 3X zooms from the Riva era) and is 5/5 - 5 elements in 5 groups, no cemented groups. The geometric correction is provided by one aspheric, and that alone removes the need for two further elements. The coatings appear to be moderately efficient but not wonderful; inherent contrast is very low (which may be why some reviewers have produced low figures, as low contrast drops the apparent MTF). But they don't mess up the colour balance and they manage, despite the air-glass surfaces, to allow light sources in the picture without much ghosting.

As agorabasta observes, the lens lacks one thing which would improve it - internal focusing/floating element design in which the elements all move separately. If it had this, the corrections could be maintained evenly from 0.5m to infinity. I don't find that 1.5m at f/5.6 is a disaster, but the degree of correction does vary with focus distance. I guess that floating elements would have doubled the price of the lens.

David

[tom power 53]

Thanks for your insight on this lens David. This NEX system seems to become more and more interesting to me based on IQ and portability. I love my A700 and lenses but am tired of carrying them around and so rarely do now days. That is not a good thing.

[agorabasta]

When I say harsh things about Nex and its lenses, it doesn't mean I hate the system. It simply means that loads of good stuff that system has are just right, OK, so they don't require my criticism
I surely love the 16mm for convenience and portability. And I can live with its problems, more so because carrying a Nex around as a 16mm P&S I hardly ever need to push it into unfavourable combination of settings/conditions.

There are some bad things I can say of the 18-55 due to its OSS that causes a considerable additional CA when the stabiliser active element gets far off-centre. But it's the nature of in-lens IS that shows in many expensive Canon/Nikon lenses as well.

Many good things have some ugly features/properties. But it can't invalidate all the good that's surely there.

[supernovak]

I hope David you could advice me here - I need to make decision for my trip to Barcelona between getting either Voigtlander Heliar 15mm (new one) or Sony's 16mm.
For my urban shooting, the wider is better and here Heliar fits better - not sure about overall image quality and sharpness - yes I need sharp!
in canada sony is 299, and i can get heliar for 499, so the difference is not huge.
thank you in advance!
novak

[David Kilpatrick]

Hi Novak. I would say the Sony any day. It's cheaper, it does not require an additional $50 adaptor. and it gives perfect colours across the frame - the 15mm Heliar must be processed from raw, and the colour shifts to the corners must be corrected using special software. The Heliar is no sharper centrally, and may not necessarily be sharper at the edges unless you work around f/11.

The NEX 16mm can be used in all modes, and allows you to use the panorama functions with perfect results as the camera can map the distortion - you can't guarantee equally clean sweep panoramas with the Heliar, as the camera has no idea which focal length is being used. The 16mm is f/2.8 and although the corners are blurred, the centre is dead sharp even wide open. The Heliar is only f/4.5 wide open, and needs stopping down to f/5.6 to tidy up the image. By f/5.6 the NEX lens is already pretty good, with better illumination/colour.

The E-lens will autofocus, and its 49mm filter thread accepts any good wide-angle polarizer or slimline pol, only 49mm thread so not expensive, and fitting a filter does not cut off the corners. The Heliar does not accept filters and has a built-in petal hood which is designed for full frame (and therefore almost useless on APS-C). In contrast, Sony designed the 16mm with exactly the same bayonet as the 18-55mm zoom, and you can just fit the hood from the 18-55mm on the 16mm when the light is tricky. It does not cut off, it gives perfect protection.

Finally, the 16mm will accept the 12mm converter lens when this is available, and you can also get an optical finder for city/street photography, and you can make full use of the auto-ISO program/aperture/shutter priority combinations on the NEX for rapid shooting but always getting the best ISO.

David