Sigma ultrawide zooms – old and new 12-24s versus 8-16mm

For almost a decade the Sigma 12-24mm full frame ultra wide angle zoom has been unrivalled by any other makers – not Nikon, not Canon, not Tamron, not Tokina, not Sony. No maker has ventured where Sigma went, to the extremes of over 120° coverage combined with well corrected straight line geometry.

Today, the original 12-24mm is in its fourth incarnation, having progressed through EX to EX DG (digital) and then with added HSM hypersonic motor focusing, which never arrived for Alpha mount in the original design. The fourth version is an entirely new design, and does have HSM for Alpha. It is very similar to the new 8-16mm design, introduced three years ago for APS-C cameras, which also offers HSM focusing in Alpha mount.

Update five years later, 2017: there is now a constant aperture 12-24mm f/4 ART lens. This is a completely new design and has an almost-perfect performance, especially in terms of corner detail wide open. I’ve tested it in Cameracraft magzine. However, I did not have the various older models to make direct comparisons. The MkII remains available as the current non-ART, lower priced Sigma option.

Here you can see, from left to right, the EX DG 12-24mm f/4.5-56, the DC 8-16mm f/4.5-5.6, and the DG HSM II 12-24mm f/4.5-5.6. Don’t be fooled into thinking the original is wider in diameter, it actually shares exactly the same lens cap module as the new design; it’s smaller and around 100 grams lighter. Both the 8-16mm and new 12-24mm are surprisingly solid items.

First, we’ll look at the difference between old and new 12-24mms. I have used the old one in several versions on several different camera makes. and I’ve never had one which was truly sharp at all point across the frame wide open. The field is not perfectly flat, and autofocus modules are very bad at getting an exact focus at 12mm. Combined with lens mount tolerances, sensor flatness problems (mostly in Canon full-frame, which historically have not had very ‘plane’ sensors), sensor parallelism problems (all makes, Sony not excepted)… it was always a good idea to stop this lens down to f/11.

How bad is that? Perfectly normal for any lens covering over 85°. Even the best large format lenses, single focal lengths like Super Angulons, have always been used in the awareness that full aperture is for focusing and you stop down to between f/11 and f/32 for the actual shot. On 35mm format digital, using anything much below f/16 is counter productive for sharpness and my normal choices on the Alpha 900 have been f/10, f/11, or f/13.

The good news is that the latest version has a different kind of field flatness. The old one tended to have a zone, like a doughnut, of closer effective focus surrounding a sharp middle. At 24mm, where this older lens performed at its worst, this zone was pushed out to the far edges and could result in the corners looking softer than they do at 12mm. The new one has a simple barrel distortion in place of a wave-form distortion, and along with this goes a simple curvature of field.

The bad news is that the overall level of distortion is much higher than the old design. At 12mm, it’s close to needing the fingers of two hands. Adobe Camera Raw had a correction profile for this lens from Sigma almost the day it became available. That profile fixed the distortion perfectly but leaves you slightly less of a 12mm than you’ve paid for, because it reduces the angle of view.

Here are some comparative views. First of all, I’ve used only 10 megapixels of the Alpha 900 frame, cropping from the top of a vertical shot, to get this architecturally correct view. This is like using the 12-24mm as an extreme 12mm shift lens on an APS-C camera. As and when we get a 36 megapixel Alpha full framer, the crop to do so will be more  like 16 megapixels. This is the full frame:

Below you can see the crop used to 10 megapixels, and by rolling your cursor over the image, the change between a profiled conversion and a raw conversion with no lens corrections. On this crop it does not look extreme.

But this is a relatively kind way to use the new lens. Here is an example pushing straight lines into places where extreme wide angles don’t like ’em:

This is an uncorrected 12-24mm DG II HSM shot out of the Alpha 900 at 12mm. It’s not exactly what you want, and in fact, it’s not as ‘good’ as the old design despite being sharper. Hovering your cursor over the image shows the same raw file with the Adobe Camera Raw Sigma-generated Lens Profile (also works in Lightroom) applied. As you will see, straight lines have been restored along with even illumination. But – how much of that 12mm, 122° angle has been lost? Is it now really only a 12.5mm?

In practice, the new 12-24mm gives you a great range of creative choices when confronted with a building. Here is a revisit to the first subject, taken at different focal lengths, getting closer to the building with each shot:






When it comes to comparisons with the older design, the new one is much sharper at the edges. It does not need stopping down to f/16 to pull in the worst aberrations, though it does still display some around f/8 to f/11. Here’s an original 12-24mm EX DG design shot (12mm, f/9, vignetting corrected but distortion not corrected):

And here’s the new 12-24mm under the same conditions (small exposure difference due to changing light) processed similarly, without any geometric corrections:

On this shot, the corrugated barn sides have clean ribbing to the extreme ends, with some softening; on the old design, they begin to look a bit of a mess in the outer quarter of the frame.

Trying the 8-16mm extreme

But when doing these tests, I decided to throw a novelty into the mix. What if I put my 8-16m APS-C format Sigma DC HSM zoom on to the Alpha 900? Because it is not an Alpha lens, the 900 does not automatically crop the full frame. This is what I actually got with the lens set to 8mm:

And in Adobe Camera Raw, I just dialled up the Scale in Lens Correction to 146%, which blew up the central 12 or so megapixels of the frame to become a full 24 megapixel image:

And here, for comparison, is what the 12-24mm set to 12mm could produce:

This is a little tighter than the 8mm using the maximum I could get (including some extra image height), so the 8-16mm used this way can produce something closer to an 11.5mm full frame lens. However, I have not yet done the obvious – to get an engineer to remove the petal lens shade from the 8-16mm (it appears to be part of the front element assembly). This would enable even more angle without shading, and the possibility of square or 10 x 8 shape format crops.

What was particularly interesting about this experiment was the quality of the 24 megapixel file extracted from a smaller section of the Alpha 900 sensor by Adobe Camera Raw upscaling. Full size files are available to download for subscribers to Photoclubalpha – it’s well worth the $10 for a full year of access to any of the extras we provide. See the download links at the end of the article, which will become visible if you are a registered subscriber to the site.

The 8-16mm also achieves full frame coverage on the Alpha 900 when set to 16mm, though with fairly marked vignetting:

Here are some more samples from the 12-24mm DG II HSM: first, 17mm at f/8 – no geometry correction:

Next, at 12mm at f/13 which on the A900 seems to be the limit for good detail sharpness without extra effort in processing:

And 17mm at f/22 – beyond the diffraction limit, but processed carefully for detail:

And 12mm at f/9, an optimum setting for detail with plenty of depth of field for this subject:

So, what was my own decision? I own the 8-16mm and an Alpha 77. That’s what I use for travel and general work. I own an original EX DG 12-24mm. I decided not to buy the new 12-24mm because I concluded that the 8-16mm used on APS-C was effectively as good. The angle is not quite a match for the 12-24mm on full frame, as APS-C is not a true 16 x 24mm. For those occasions where a 12-24mm on full frame is needed, I’m nearly always able to work on a tripod at f/13 and focus manually (which overcomes most of the issues with the earlier lens). Since it needs less drastic geometric correction, it offers a very small angle of view advantage over the new lens in return for the risk of poor sharpness if not used well stopped down. I have no doubt the new lens is better, but it’s not £400 better which is what the ‘trade-up’ would cost – and the old lens is lighter and smaller, which I appreciate.

The test made me even happier with the 8-16mm, especially with the thought that some modification could make it a unique lens to use on the Alpha 900 or a future full frame EVF model (A99). After doing these tests, I decided it was not necessary to take the Alpha 900 and a 12-24mm despite the investment in two weeks’ shooting in the Sierras and Pacific Coast of California – the A77 and 8-16mm would do everything I needed.

But for those buying a 12-24mm, for full frame on any system, the new Sigma represents even better performance than the 8-16mm (better edge and corner sharpness at one stop down from wide open) and has none of the failings of the older lens even if it does need more post-process geometry correction.

Download full size images [private] 24 megapixels 12-24mm EX DG at 12mm f/9 Download Link
Download full size image 24 megapixels 12-24mm DGII HSM same as above Download Link
Download full size image 24 megapixels 8-16mm lens scaled to 24 megapixels from A900 ‘crop’ Download Link
Download full size image 24 megapixels 12-24mm DGII at 12mm compared to 8-16mm at 8mm cropped Download Link
Download full size image 24 megapixels 8-16mm at 16mm filling full A900 sensor Download Link
Download full size image 24 megapixels 17mm f/8 fence example shot boat Download Link
Download full size image 24 megapixels 12mm f/13 sequoia tree example shot Download Link
Download full size image24 megapixels 17mm f/22 boat Download Link
Download full size image24 megapixels 12mm f/9 riverside tree Download Link [/private]

To check the weight, specifications and other details of these three lenses we suggest you visit Sigma’s own site – for the new 12-24mm, here’s the UK site info. And here is the 8-16mm, which they oddly don’t class under wide zooms, but under DC lenses.

You can check worldwide shipped prices from B&H Photo in New York.

– David Kilpatrick

Chris Townsend on the NEX-7

Here at PCA we do not often redirect our loyal readership to other sites, unless it’s to buy stuff which earns us enough commission to enjoy a couple of pints and a burger every month!

But here is a review from one the UK’s most respected independent outdoor writers and photographers, Chris Townsend.

Leave only footprints, take only photographs – except that Chris is not that keen on footprints, he’d rather see you keep to paths which do not crush our wildflowers or erode fragile hillsides.

– DK

Sony Cyber-shot DSC-RX100 review

The first thing that is likely to strike you about Sony’s one-inch sensor Cyber-shot DSC-RX100 is size. It’s tiny, slightly smaller in body than the Nikon 1 series interchangeable lens mirrorless cameras using an identical size 2.7X factor, one-inch or 13.2 x 8.8mm sensor.

This just a fraction over half the area of a standard APS-C sensor, and where Nikon has chosen to have 10 megapixels of active imaging plus others unused or devoted to phase-detect focus on the silicon, Sony has opted for 20 megapixels.

At first this seems excessive, until the performance of other new smaller sensor cameras is considered. The Fuji X-10, for example, has a 12 megapixel sensor measuring 8.8 x 6.6mm and achieves a respectable balance of sharpness and noise-levels. The RX100 has a slightly lower pixel density. Compared to the Canon G12 it’s four times the sensor size and twice the sensel size.

The 1.0 type sensor also gives just that little bit more creative control over depth of field. With the usual third to two-thirds inch standard sensors in pocketable compacts, the lens must be used wide open at any given focal length to provide a degree of differential focus. To avoid sharpness loss, most such cameras can not be stopped down to settings like f/11 and sometimes have a choice between two apertures only, wide open and something moderate like f/8.

The Fuji X10 zoom only stops down to f/11 but offers a full continuous range of settings. So does the RX100, its 10.4-37.1mm lens ranging from f/1.8 to f/4.9 wide open but limited to f/11 minimum regardless of zoom setting. Since even f/11 can produce some diffraction-limit related softening, its performance around f/5.6 is critical. This would be the setting I would choose for routine Aperture-priority shooting.

At such a setting, the low ISO quality of the RX100 can be exploited. Unlike any of the Sony NEX models, the little RX100 has been given user control of maximum and minimum Auto ISO limits. The full auto range is from 125 to 6400. Manually set ISO can be extended downwards to either 80 or 100 (but these settings just overexpose the image and compensate in conversion). The camera seems to have been developed as well as manufactured in Japan, and the firmware and menu system resembles the mainstream Alpha DSLR/SLT camera line rather than the mirrorless NEX. Editor’s note: having sold my original RX100 I bought another, the second although made at the same date, is made in China. It seems either better or no worse.

The shutter is speeded to 1/2000th which is not a very fast high speed for a camera capable of 10 frames a second action bursts (or 2.5fps normal continuous shooting). The longest exposure possible is 30 seconds. By whatever means, aperture or shutter, Sony allow control to within 1/3rd EV step and compensation to ±3EV, but AE Bracketing is limited to three frames at either ±0.3 or ±0.6EV.

Control over settings is handled by a single top mode dial, a shutter release with power zoom lever to the front, a rear Control Wheel with four cardinal point click functions, four further surrounding buttons, and a Control Ring set round the lens bezel. This can be silent or make click sounds to mark setting changes but lacks physical resistance or detents. It doubles as a fine focus ring when the camera is set to manual focus, aided by focus peaking and on-screen magnification. Its action is very smooth indeed, and it can be operated easily by a single finger from either hand.

Real photo – the film and lens are entirely hidden behind the RX100 but imaged by its close focus ability at 10.4mm

Considering the 101 x 58mm footprint of the body, everything is designed efficiently to allow a 3 inch rear screen using 1.2k dots and an additional white-light augmenting RGB to improve sunny day use. It’s not a touch screen, nor is it articulated or hinged. But you will touch it, for sure… a cloth to wipe off your thumbprints is an essential accessory. It appears to be glass, but may just be a hard coated plastic layer, something with which Sony has a bad history.

Actual size next to a CF card, which this camera of course does not use – it takes SD or Memory Stick Pro Duo.

The new small battery type NP-BX1 allows 330 shots – better than many high pixel count consumer DSLRs and mirrorless models now – and can only be charged in the camera itself, via any USB 5v source and the supplied Micro USB connector cable. This is not a standard Mini USB, just as the Micro HDMI (cable not supplied) is not a commonly found fitting.

Against the disadvantage of in-camera charging you can set in-car charging, laptop or phone supply charging, and the camera’s ability to run without a battery installed when connected to its supplied AC charger. Both third party lith-ion cells and third party external chargers can be found on eBay. Anything which offers a standard, powered USB connection can charge the RX100.

The British charger is an old warhorse. The US charger is a neat monobloc transformer half this size with folding AC mains pins. This kludge is bigger and heavier than the camera…

A full charge takes 155 minutes using the charger with its high level USB-power output, but may take longer through a PC USB port or devices providing minimal USB power. You can leave it plugged in to USB all the time as the charge cycle is cut off when an orange charging light in the on/off switch extinguishes. When the camera is switched on to connect to the computer as Mass Storage (etc), this light turns green. It is possible to use the RX100 without the battery installed, connected to the charger.

In the box, you get no software, only an instruction manual which covers the bare bones. It seems to be assumed that what Sony call the best ‘professional’s compact’ ever will be bought by experienced digital camera users. Nearly all the functions on the RX100 from sweep panorama to HDR and noise reducing multishot modes are found on other cameras, and the location and nomenclature of all functions is at least familiar. Download links are given for a PDF identical to the bare manual, or a web-page based version with colour illustrations which is far better but can’t be downloaded.

Here is the link for the full colour, more detailed user manual:

A wrist strap is supplied, along with two neat cord and leatherette toggles to attach a regular camera neckstrap, as the body has two almost microscopic strap lugs. Nothing other than the very fine cords of Sony’s strap or strapholders would be likely to fit.

The body is solidly made and all access doors seal well, but it’s not resistant to anything wet, dusty or involving hard surfaces and heights. The lens’s rear glass is located very close to the sensor, and zooming appears to move only the middle and front groups. This should make it dust-free for life. Time will tell, and if any dust ever does get on the sensor, it will need a factory repair. But it looks to be designed so that will never happen.

For the professional user, the big appeal of this camera is its invisibility. Only 36mm thick with the lens collapsed, it’s just a fraction fatter and smaller than an iPhone, and with focusing down to under two inches there’s hardly anything it can’t capture. Users may criticise the 28-100mm equivalent focal length range, preferring 24mm if possible, but the focal length of the zoom is stated after allowing for some strong in-camera distortion corrections at the wide end.

To achieve a 28mm field of view (73°) for an in-camera JPEG, the corrections must deal with a very high level of barrel distortion. The raw file is uncorrected, and shows a diagonal field of view closer to that of a 24mm (85°). This may explain why Sony’s own information has claimed both 24mm and 28mm as the widest angle, when the stated focal length and sensor size clearly equate to 28mm. My measurements from the two image versions below indicate that if the correct equivalent is 28mm, the uncorrected diagonal angle is equivalent to a 24.8mm. Either way the RX100 should not be criticised if you could be happy with a new Canon EOS M – 1.6X sensor, 18-55mm lens, that’s a 28.8mm widest limit before applying Adobe Lens Profile corrections which will probably reduce the true angle to a 31mm.

And that of course applies to almost all wide angle lenses except the Sony NEX E 16mm f/2.8, which has pincushion not barrel distortion and therefore does not lose any of its diagonal angle (for that is how lens angles are measured) when corrected. There is an inbuilt profile for the latest ACR and Lightroom, but unlike other Adobe Lens Profiles, you can not adjust or turn off the disortion control. Apply the profile does not move any pixels, it simply corrects vignetting and CA. These programs are reading metadata in the raw file to apply the geometric correction automatically and you can’t disable it. To see the full field of view of the lens at its 10.4mm focal length, you must use a processor like Iridient Raw Developer (Mac only) which ignores the instructions.

Above: in-camera fully corrected JPEG at 10.4mm, and uncorrected raw conversion (by Iridient Raw Developer) showing full view angle of the lens before removing the high level of barrel distortion. Just move your cursor over the image to see the change. Adobe programs prevent the removal of the camera’s automatic correction – you can’t get to the ‘wide’ version.

When shooting video in 16:9 format with stabilisation set to Standard (optical) or Off, the lens range is trimmed to 29-105mm equivalent, and the image is cropped only slightly on pressing the Movie button. If you set Active stabilisation for video, optical stabilisation is replaced by pixel shift electronic stabilisation on the sensor. The crop is to 0.87X of the normal video field (measured here), meaning that the effective focal length range for Active video is 33-120mm. This 0.87X factor is exactly the same as the NEX-7 video crop factor.

The RX100 can be concealed in your hand and when used, with no eye-level viewfinder and composition on the rear screen instead, you look like any cameraphone user or tourist. In fact you are capturing what could be a highly detailed 20 megapixel image suitable for double page magazine or newspaper repro.

This is, of course, also a camera which won’t get you thrown out of sports stadiums or concert venues despite its ability to capture 50/60p HD1080 video with good quality stereo sound, and to capture full resolution JPEG still frames during video (17 or 24Mbps, not 28Mbps) without interruption. Writing the JPEG takes some time, parallel to video writing, and a faster SDHC or MS Pro Duo card is recommended. It can record AVCHD-2 format movies at up to 28 megabits per second, with AF during video and a good degree of setting control including manual exposure. It can not capture raw still files during filming. There is a faint click sound only during the video.


As for the quality of results, the lens may be letting the sensor down slightly; although very high in resolution even wide open, corners can lose detail because the focus plane is far from flat. Bright lights or overexposed details can produce a visible flare or glow, it’s possible to get purple fringes. Against this you must set pixel-crisp sharpness wide open, at any focal length, in many shots.

The exposure over-ride is excellent, and the screen really gives an exact view of what you are doing. Here, minus 2 stops was needed. This is at ISO 125, 1/25th at f/5.0 at 17mm (45mm equivalent) focal length. The original file has perfect detail corner to corner – every leaf sharp.

At the best – ISO 80 to 125, stopped down just one full step from full aperture – the RX100 can match or better the typical output of a 21 megapixel full frame DSLR with 24-105mm lens. At the worst it’s better than any smaller sensor compact, especially if the 10 megapixel JPEG shooting option is chosen or the file size is reduced to match a typical 12-16 megapixel 2/3rds inch sensor image.

One of my first tests, wide open at f/4.9 at 37.1mm and auto set to ISO 500, in camera JPEG. Just lovely colour and tone, perfect WB, perfect auto exposure. An early fallen leaf.

High ISO results are encouraging – using ISO 800 or 1600 should be no barrier to large clean reproductions, 3200 and 6400 remain clean in good light with detailed subjects but show coarse mottled grain in defocused areas with low light. Multishot modes are similar to NEX and can greatly improve results, but for my tests I stuck to raw files (though all the examples shown here are from in-camera JPEG) and single shot modes. Also, with f/1.8 apparently as sharp as most lenses well stopped down and having plenty of depth of field, I have tended to use low ISO settings in conditions where I’d set my Alpha 77 to ISO 800.

This shot was taken at 1/100th at f/1.8 at ISO 125, just because with this camera you CAN – no need for high ISO when you have f/1.8 at 10.4mm. But how about lens quality, how about depth of field? Take a look – all clips from the in-camera JPEG:

You can see the tendency to flare around light sources, and remember – this is an optically corrected image. Look at this in raw, and the purple fringes on those lights are the most colourful thing in the shot. This is from the middle of the frame.

Here’s the extreme right hand up to the very edge. Remember, it’s a 28mm f/1.8 equivalent.

Here’s the left hand, further away, a little bit in from the edge to catch the best detailed target.

And here is the bit you expect to be awful, more distant trees against the sky. Not bad for f/1.8?

Active video stabilisation is pixel-shift electronic, still stabilisation and standard video are in-lens optical. Both work well and the electronic variety is particularly good at dealing with small movements of your hand when holding such a small device for filming. Video quality is a match for any HD1080 DSLR, with a true 50p or 60p (USA) frame rate. The RX100 also has full user control over ISO, shutter speed, aperture and manual focus during video; the shutter-release zoom lever provides a smooth slow fixed speed zoom during filming. Beyond the 3.6X optical range, further digital zooming drops sharpness and can not be recommended. The point where digital takes over is well defined by a pause in zoom travel but you can not disable digital zoom to 14X maximum.

ISO 3200, 1/30th at f/1.8, 10.4mm, very low yellow pub light.

100% clip of in-camera 3200 JPEG – maybe a bit rough, but not bad at all…

White balance is generally well optimised, exposure is less predictable in difficult conditions. The multi-zone metering and focusing settings can produce unexpected results, spot and single point choices may not do any better as they will favour just the targeted tone. Access to +/- compensation is rapid. It can be assigned to the ring round the lens. This control ring is smooth in action and works well for adjusting exposure while viewing the rear screen.

Faults or flaws

The uncorrected image has fairly strong CA, which in defocused zones (especially that critical phase between sharp and truly out of focus) can create purple fringes on a large scale. The camera software turns these into white glow. Slightly defocused detail, especially if brightly lit beyond the clipping range of the sensor, can produce unpleasant bright fringes which are impossible to remove. Very bright areas even when well focused tend to flare into their surroundings.

I don’t really want to show what the fully lit bits of lichen at minimum focus look like – the highlights flare a fair amount.

Dynamic range is good, but not exceptional. Highlights clip readily, and recovery in either Raw Developer or IDC v4 did not pull in missing detail, it just darkened the value of a sharply clipped high bit. Though ISO 80 and 100 provide finer grain, they are less use than ISO 125 or 200 in contrasty light or with flash, as they clip more. Highlight colour recovery and use of DRO can produce some very odd effects. Editor’s note: since this original review, ACR/LR has been updated to process the raw files, and this is one reason I’ve bought an RX100 for the second time. I can now tames some of the lens and dynamic range issues very effectively.

At minimum focus, the aberrations get worse and overall sharpness is reduced, especially around the wide angle and two or three inches working distance with the lens wide open. It is easy enough to get know the lens, and its substantial sweet spot (almost anything not close-up, not contrasty or with patches of extreme overexposure). Having said that, you can also obtain stunning close ups at 10.4mm:


Click on this, and you can download the full size (probably crunched a bit from Level 10 JPEG by WordPress) image file. You will see a world of detail to amaze you and some fascinating aberrations and artefacts as well – perfect in a way, imperfect no doubt, but a wonderful thing to be able to do with a camera so small you can get it down into the a subject like grass. f/11, 10.4mm, hand-held, 1/40th at ISO 125, ACR processing.

High ISO JPEGs look clean in good light with hard detail. They look very mottled and mushy in darker softly focus areas of smooth tone. You may want to avoid using 3200 or 6400, but remember – the lens is f/1.8 to f/4.9, covering a range which is typically represented by an f/3.5 to f/5.6. At the wider to middle end, there is a two-stop advantage fading to a third-stop at the tele setting. If you stick to the wide angle end, you can use ISO 800 with as much success as 3200 would achieve in a DSLR, and pretty well the same depth of field too.

Design – the most annoying single thing is the pop-up flash which sits exactly where you are likely to hold the camera body at the left hand end. You will just have to learn not to hold it that way! That’s a penalty paid for such a small body. The tripod bush is also off centre to the lens. This only matters for specialised multi shot assembly or macro stepping.

Does it work?

Yes! The RX100 is actually a great little companion camera, and after getting it, I stopped using my NEX-5n kit for casual everyday snaps. The RX100 lives in my wallet beltpack or a carefully emptied and cleaned-out pocket, wrapped in a microfibre cloth. I may shoot a few pix or a video clip, and every day, I just connect the camera to my iMac and use iMovie to Archive the entire media contents. This copies all movies and also all stills. I then format the card before the next use, and the camera is always fully charged when I pick it up off the desk.

My best pictures are every bit as a good as a typical NEX-5n with 18-55mm shot, my worst results are better than most consumer pocket cameras and no worse than the worst NEX shots. You can take bad pictures with any camera! My videos are as good as any of the NEX or Alpha models so far, and streets ahead of Canon, even including Canon DSLRs used by professionals. I would give the Nikon D800 videos the edge over RX100, and NEX-7 or Alpha 77 videos equal status. All are far more detailed and crisp than Canon’s HD1080, yet that is now a bit of an industry standard. I predict that the RX100 will gain a bit of a cult following for video making. Its movie setting on the mode dial allows user-set aperture and shutter speed, full control once you add manual focus.

Suggestions that it may supplant NEX are groundless. You can fit wonderful glass on NEX, and get 24 megapixels to the highest standard. You can’t fit wonderful glass on the RX100 and the zoom it comes with, Carl Zeiss or not, has clearly visible distortion and aberration issues that depend on firmware or software for correction.

Is it worth the money?

Maybe. I think the RX100 has been overpriced by around £100 in the UK but I see that many retailers are already dropping the price by that amount. Around £400-450 seems a fair price, the official £550-580 is high. Update: we sold our (Japan, June 2012) RX100 for £400 in August 2012. A replacement (China, July 2012) was found as new for £385, eBay used, in February 2013. New prices are now more of less where we suggested they should be, mid £400s UK, cheaper USA.

– David Kilpatrick

Check out the price of the RX100 from B&H


NEX-5n – sweet sixteen and expandable at a cost

Much of my NEX-7 critique was written while also using a NEX-5n outfit. I was lucky enough to find an opened NEX-5n 18-55mm kit missing its mini flash at a very low price, as new. By the time I had finished completing my NEX-5n system with bells and whistles, the total would almost have paid for a NEX-7.

The final kit consisted of the 5n, the FDA-EV1S electronic OLED viewfinder, the ECM-SST1 stereo microphone, and the HVL-F20S flash. The microphone was inherited from my NEX-5 so maybe doesn’t count. In the NEX-7 review, I start by suggesting that the 7 is really more part of the Alpha A-mount system and not the NEX system. The 7 either doesn’t need, or doesn’t accept, any of the accessories shown.

The NEX-7 finally became available after the 5n kit, and for a while both were used together. Just as the 5n can not use a plug-in stereo microphone or an Alpha system flash, the 7 can not use the NEX microphone or the HVL-F20S flash. Although I had one ‘NEX system’ with two bodies, the NEX-7 needed my HVL-F20AM flash originally bought for the Alpha 900 or one of the larger guns, and my Rode Video Pro mic, bought for the Alpha 77.

That’s why I count the NEX-7 as a hybrid, partly ‘big Alpha’ in heritage. It does not integrate with other NEX accessories, and vice-versa. Sony shows no sign of dropping the NEX Smart Accessory Terminal from 3 and 5 series bodies so this parallel range situation continues. If you’ve bought a mic or a GN 20 flash for your 5, it will not be usable with the 7 you plan to buy tomorrow.

The OLED Tru-Finder

Harking back to the wonderful Minolta Dimage 7i and later the Konica Minolta Dimage A2, the FDA-EV1S slightly resembles the fragile hinged EVF of those cameras. Like them, it can be flipped into a 90° upright position or used at angles between, so the eye looks down rather like using a waist-level finder with magnifier, or a Hasselblad 45° prism.

The finder is supposedly identical to the Alpha 77 or NEX-7. It has a different dioptre adjustment, a small slider which has a huge effect for very little travel. It’s not easy to adjust but stays put when set. If you wear and remove spectacles at random, and use the camera with both the naked eye and glasses, it’s one of the least ergonomic adjustments. The eyepiece surround is a semi-hard plastic and not as comfortable or efficient as a larger soft eyecup (which Sony does not make, but would be so easy to add to the system).

The add-on finder appeared to be slightly less clear and smaller to the eye than the A77. The same goes for the NEX-7 finder. The difference seems to be in the optical train, how the lenses are arranged in the ocular itself. It may even be nothing more than an eye-surround and eyepoint issue. The OLED screen is identical but I do not seem to get precisely the same viewing experience. Maybe it’s also a little dimmer by default to conserve battery power.

In use, the way the finder sticks out behind the NEX-5n rear screen is a bonus. I’d love to get an Alpha 580, my wife Shirley uses one and she is blessed with a small nose. I’m not! The 580 eyepiece is set forward of the screen surface by a good distance. It makes a very uncomfortable viewing position for me, and add-on magnifier eyepieces don’t help all that much. The FDA-EV1S in contrast is almost perfect. The NEX-7 is better because my nose can end up beside the camera not touching it.

In practice, I ended up hardly ever using the vertical viewing position. The finder sits forward when flipped this way, and somehow my hand position wasn’t all that comfortable holding the camera and looking down at a normal scene. Instead, I found the flipped-out rear screen and a waist-level camera position more useful. Then, of course, the EVF gets in the way. The sticking out eyepiece which is so comfortably in use can obscure your view of the screen slightly.

Finally, I ended up removing the finder most of the time. It seemed a bit vulnerable, it reduced battery life greatly, it prevented pocketability with my favourite 16mm lens (or at least, felt even more vulnerable in a pocket) and much of the time I realised I was composing on the rear screen anyway. As a result the buyer of my kit got a very little-used EVF.

It is the best EVF made, or at least on a level with the best other EVFs using new technology. I can work with an EVF. Some just can’t and almost need an optical finder. But I’m not sure I would ever fork out well over £200 (or around $300 before tax in the USA) on this accessory again. I’m looking at getting a 5n back into the fold, especially after going back to my raw files. I don’t think this small accessory should cost more than a lens, and the 16mm optical finder is equally overpriced. Sony’s accessory prices generally are a negative customer experience and do not create evangelists for the system.

The vertically-angled finder was tried, but not used, for this shot with the camera near ground level. It was far more comfortable just to use the rear screen.

Enhanced vision

Somewhere out in webland, it’s been pointed out I’m old and that my opinions on EVFs (etc) may be irrelevant. That’s a bit of an own goal, as EVFs have maximum appeal to those with ageing bad eyesight. Old eyes tend to be longsighted, and can’t accommodate to close focusing, needing reading glasses as well as distance glasses in many cases. Older people find composing pictures on rear LCD screens difficult, they may have to hold the camera right at arm’s length (you see it all the time!) and even then, they may not be reading the screen menus clearly or seeing the picture at pixel-sharp quality.

It’s young eyes which work best with phones and compact cameras lacking a viewfinder. They can focus on a screen held inches away from the eye. So can older eyes with serious short sight – just remove your specs, and you are away!

The EVF, especially in the NEX-7 and as an add-on to the 5n, is a boon to these with presbyopia. Suddenly, menus can be seen sharply and pictures composed and reviewed in better detail. The dioptre slider allows correction for the most common range of near and far sight, though it can’t correct for other conditions like astigmatism. For those who must always wear specs, just removing the eyepiece ‘cup’ can help.

As with the NEX-7, one key step is to disable image review when using the finder. It is disconcerting to have the image you have just shot block your view for even 2 seconds, especially when it prevents photography. The effect is different to having the same happen on the rear screen, because while the camera is to your eye, it becomes your window on the world.


Here’s something not often mentioned, and once again, I end up knocking the NEX-7. The NEX-5n and previous models have simple slot-type strap connectors mounted so that the camera always hangs with the lens down, LCD screen up. Even with a 16mm only, this lens-to-the-ground position protects your lens. You can even walk around in light rain and be confident it won’t get on the glass.

The NEX-7 with its magnesium body shell uses the higher end traditional post and triangular D-ring found on the Alpha 77, Dynax 7D, Alpha 700 and 900. And it does not hang lens down like the 3 and 5 models. The strap also gets twisted more easily. No need to say which I prefer. The cheap connector may be cheap, but it has a better function.

The rear screen mounting

The mechanism of the screen on the NEX-5n is slightly better than the earlier 3 and 5, perhaps because the EVF demands it must be able to move outwards in a slightly different way to be seen clearly for waist-level shots. It is my own view that Sony missed a trick, as the EVF on this camera would certainly have allowed a reversible, fully articulated Alpha 55 type screen and its extra thickness, without impeding EVF use.

Because the rear screens of all the NEX models do not twist to allow vertical composition combined with waist or overhead viewing, it makes less difference to me whether they are hinged at all. This probably reflects the emphasis on video shooting, where vertical composition is not needed. For the still photographer, cameras with articulated screens that can orient for verticals and also aim forwards for composing self-timer groups are most desirable. The screen is there. It’s already detached from the body. All that’s lacking is the correct mechanism, even when Sony has shown they have the necessary rights or patents, and can make them.


The NEX-5n in addition to a 16 megapixel sensor offers lens corrections (for JPEG), AF correction for Alpha lenses attached via the LA-EA2, automatic sensing of DMF (manual focus taking over from confirmed AF) with magnification and focus peaking, electronic front curtain shutter, true 50/60p HD1080 video, extended sensitivity from ISO 100 (instead of 200 minimum) to 3200 (instead of 1600 maximum), and high-speed 10fps sequence shooting (this not really matched by the focusing abilities of the camera, any more than the 1/50th second shutter lag is).

To the earlier 3 and 5 models, firmware updates have retrofitted focus peaking with or without magnification, and AF correction for lenses with the adaptor. They can’t add lens corrections, DMF, new video modes, better low light and HDR multishot modes, or change the louder double-action shutter with its 1/20th second delay.

The NEX-5n also has a touch screen. I disabled this function from the start, along with smile shutter. Face detection I leave on as this does help with focus and exposure for many shots. Since I’d parted company with my NEX-5n before using the touch screen, I can’t comment on its value. It’s just something I don’t like using.

The only advance I would argue against is the extension of Auto ISO to 3200 with no ability to control the range. Though 3200 on the NEX-5n is not unlike 1600 on the NEX-7, both these settings are too high to allow a perfectly clean image from raw after the best processing. In-camera JPEGs confirm that. I found the NEX-5n working at 3200 in many conditions where I would have been confident of a good image at 800 or even 400. This obliged me to use manually set ISO, or put up with the 3200 quality.

NEX-5n, 16mm, 1/30th at f/11, closest focus, ISO 3200, ACR conversion from raw

100% clip showing how the presence of sharp detail (wallpaper) reduces the appearance of grain compared to defocused tones (right hand side).

How good is that quality? Compared to the 24 megapixel sensor, I’d still say it’s better than a one stop advantage. One of my magazine reviews (f2 Freelance Photographer) was accompanied by a near double page spread from the NEX-5n shot above taken at ISO 3200. It’s certainly good enough for that. However, ISO 1600 is much better. ISO 1600 is so good that in decent light, you could easily be using ISO 200 off a camera of the Nikon D200 or Sony Alpha 100 era, even at 100% pixel comparison. 3200 is amazing, as you can see. But it’s definitely a grainy look where 1600 can almost be noiseless.

If a firmware update could ever achieve it, I’d like to see Sony put a maximum and minimum auto ISO selection into the NEX models, as they have done in the Alpha 77. Failing that just a maximum limit would be useful.

There is no doubt that the 16 megapixel Sony CMOS is one of the best sensors yet made, and a great balance between pixel count and image quality. See below…

16 versus 24 with ACR

Adobe Camera Raw has the ability to open files with a set of fixed size interpolations from the raw data. In this respect, it is better than Lightroom, which can export files to different sizes on demand but shows (at least from my observation) a slightly lower quality. ACR’s image sizes are slightly arbitrary and clearly are not related to the pixel dimensions of the raw image. You can open a 17.5 megapixel or 25 megapixel image from a 12, 14, 16, 18 or 21 megapixel raw.

When you select an enlarged or reduced conversion, the large image preview and editing window reflects this. Your 100% view changes to be a 100% view of the size you are producing. In the case of the NEX-7, the 24 megapixel image size is the largest option on the list. You can not open to 25 megapixels as you can with the NEX-5n. You can reduce to 17.5, 11.2, 6.3, 2.8 or 1.6 megapixels.

The largest size is always 6144 pixels wide (longest dimension), or the native size of the raw file. So a Nikon D800 image which is 7360 pixels wide also shows up with only smaller options, and rather oddly skips the 6144 24 megapixel choice. With ACR, you can not open or preview a D800 raw at 24 megapixels, only at full size, 17.5 megapixels and the smaller choice.

If the raw file has unusual dimensions – 4:3 or square for example – you may get interesting options. The 21.3 megapixel Dalsa medium format backs show a 6144-wide 28 megapixel maximum size output option. The 16 megapixel Kodak MF backs allow 5120 square or 6144 square output, the largest size being 37.7 megapixels. As these backs have no AA filter and are teamed up with unrivalled lenses, a 16 megapixel Hasselblad 80mm Planar shot scaled up to 28 megapixels is hard to tell from a native Nikon D800+zoom lens image. D800E with top grade prime beats either.

Using the NEX-5n (or the earlier 14 megapixel 3 and 5 models) ACR offers 17.5 and 25 megapixel conversion, viewing and export or opening as well as the native size and the smaller ones mentioned above. It is largely my experience using the resizing functions along with NR and sharpening that makes me prefer the 16 megapixel sensor to the 24.

This view was taken with the NEX-5n and 18-55mm at 55mm, ISO 100, happen to end at f/11 though is was intended to be at f/10 (the non-lockable controls of the NEX-5n did this to me far too often).

This view was taken a few seconds later using the NEX-7 with a different (black) 18-55mm at f/10, all other parameters being similar, and both as raw files.

Here are 100% clip sections of both images. The 5n is top, the 7 is bottom. Both files are 6000 x 4000, the 5n image was exported by Adobe Camera Raw to 6144 pixels wide. Both images use Sharpen 25, Radius 0.5, Detail 50 and zero for both luminance and colour NR. In the very subtle low contrast texture of the lamp-post and the definition of the hex nut, the 7 clearly wins but it’s a surprisingly fine margin.

Yes, the difference is obvious. The 7 wins. Think again – for this clip, I’ve put the NEX-7 native size image TOP and the NEX-5n interpolated 24 megapixel output BOTTOM. What is the conclusion? That the lens you use – even the individual sample of the lens used as no two 18-55mms will perform identically – has far more effect on usable image detail than 24 megapixels versus 16.

Scaling images down in size

When I have been editing NEX-7 images at their native 24 megapixels the NEX-5n upscaled image has sometimes looked better overall, and the native size NEX-5n image nearly always wins. The NEX-7 image misses the mark for me maybe 30% of the time. For all ISO settings above 400, I tend to set the ACR output to 11.2 megapixels. Setting it to 17.5 doesn’t lose the granular feel. ACR’s 2012 process (CS6) has noise reduction and sharpening controls which work faster and better than any of the plugins or other raw conversion programs I’ve used.

Checking as I write: for 66 images just processed from the NEX-7, Alpha 77 and Alpha 580 (16 megapixels) I counted that 33 of the 24 megapixel images had been downsized to 3600 x 2400 pixels or thereabouts to resolve issues with noise or sharpness. Not one of the Alpha 580 images had needed downsizing. The 6000 x 4000 shots could, perhaps, have been downsized to the 4912 x 3264 of the 16 megapixel sensor or the 4076 x 2731 offered by ACR; the 3600 x 2400 size is the minimum for image library Alamy. If you send them anything except the sharpest and most noise-free images, you risk having all your work rejected, your submissions placed in a slow queue, or your entire account deleted for repeatedly less than perfect technical standards.

For stock library use, this image was reduced from 24 megapixels to a slightly cropped version just over 9 megapixels. The 18-200mm Tamron lens at 66mm, with the NEX-7, was used at f/8 and the focus was on the gold pan. A 24 megapixels, the degree of softening on the young girl’s face is beyond the acceptable limit; scaled down, it looks natural when viewed at 100%. Shooting with 16 megapixels instead of 24 will give an impression of greater sharpness, or greater depth of field, when checked at 100% though this is an illusion and two prints made at the same size will show no difference.

Working with the 16 megapixel sensor, across the usual range of conditions from daylight to night scenes, hardly any images need reducing in size to hit Alamy’s QC mark. Working with the 24 megapixel sensor, every image has to go through ‘is this really OK at full size?’.

While I definitely want my 24 megapixel Alphas for studio, tripod, architectural, landscape, artwork copying and similar tasks, having sold both the NEX-5n outfit and the NEX-7 if I was to purchase again it would be a NEX-5n with reservations.


After using the NEX-7 interface, especially with the settings lock function, it is very difficult to go back to using the 5n or earlier interface. On all NEX cameras the menus are very easy to navigate, consistent, and reasonably fast. But on all NEX cameras except the 7, the rear controller is much too prone to accidental operation. The vertical straphanging reduces this a bit, as it avoids the control touching your clothing. The action is so light that just brushing against a coat or the fabric inside a camera pouch can be enough to change the aperture in A-Mode, the shutter in S-Mode, or the EV exposure correction (requires a light pressure to the bottom of the control followed by rotation).

It would be good if Sony could make a firmware modification which locked the primary function of rear controller rotation – something like a two-second press on a specified button – while leaving the compass-point click/rotate functions (located North, East, South, West or 12, 3, 6 and 9 o-clock on the controller) available.

The final reservation is where I started, about the Smart Accessory Terminal and how it differentiates the 3 and 5 models from the 7. The 7 fits well into any existing Alpha setup, the lesser models are only a partial match. The terminal has appeal. It’s long been thought that Sony could use it to add other functions, such as a GPS module or a wireless flash commander. No such accessories have appeared yet. Does that mean it’s a dead end, to be replaced by 7-style interfacing – or is the 7 a level on its own?

Sony has provided some kind of road map for lenses and would perhaps be giving too much away if they issued a roadmap for NEX development. It would make planning today’s purchases less of a gamble for existing system owners and new adopters if they could.

– David Kilpatrick

To see NEX-5n specs and prices at B&H follow this link. Clicking on our Amazon or B&H search boxes can benefit this site (many thanks to the person who keeps ordering dozens of academic books from Amazon!).

Sony NEX-7: the high-end hybrid

The Sony NEX-7 is not a NEX at heart. It’s part of the rest of the Alpha system in every respect except its lens mount, and even that can be converted with a choice of two adaptors. A NEX-7 with an LA-EA2 phase detection autofocus adaptor is little different from an Alpha 65.

The NEX-7 does not accept any of the smart accessory terminal add-ons common to the NEX-3 and NEX-5 models, because it lacks the accessory slot. That means it does not use the same flash models, or the same microphones. Instead the NEX-7 has the iISO Minolta flash shoe, accepts all the flash accessories from the Alpha range and is capable of wireless remote flash operation.

Its external microphone, if required, is also from the Alpha system or any suitable 3.5mm stereo jack connected model mounted on a bracket or an accessory shoe adaptor.

Packaging sequence

Click this for an enlarged view of the opened box, which was full of white bits. Though sealed and brand new, the NEX-7 purchased needed a good dusting and blowing before removing its front cap and fitting the lens.

You can’t see here, but when the cap was removed from the body there were specks of the same white packaging filler on the matt black baffle between the lens mount and the sensor. But, once blown clean, the NEX-7 proved remarkably free from dust-on-sensor problems. It’s one of the least dust-prone cameras I have ever used, despite the exposed position of the sensor which is not even covered when lenses are changed or when the camera is turned off.

The dust on the camera may also be seen here. The eyepiece surround is supplied separately packed and sealed. It is not soft rubber, and makes pretty sharp contract with my brow, almost demanding to be shoved into my eye like an eyecup, with specs removed. I found the finder good to use but started removing my glasses permanently when working with the 7 outdoors.

Battery and stamina

Given the relatively short battery life of any camera using an EVF as sophisticated as the Sony OLED device, I feel it’s a pity they did not go the whole distance and have a larger right-hand grip housing the 500-series lith-ion used by the 77, 900, 700, 5xx and other larger more ‘professional’ camera bodies. The little 1050mAh 50-series cell shared with the lesser Alpha 55, 33 and all NEX models is stretched to the limit by a 24 megapixel EVF camera. Third party 1300mAh versions don’t actually seem to last any longer.

What’s interesting is that losing the flapping mirror and mechanical focus and aperture operation, found in the earlier Alphas, has not doubled battery life. You would imagine all this heavy mechanical stuff would drain power fast, but in fact the electronic alternative of live view, with the sensor active all the time feeding a digital viewfinder, proves far less efficient.

Because the NEX-7 uses an external power option with a dummy battery (a Canon approach I have never much liked, always preferring dedicated DC input) it would be possible to design an add-on base with a better power source. But as with other NEX models, the SD card slot lives in the battery compartment and such a ‘power grip’ would either need to be removed to change cards, or incorporate more adaptation to provide a relayed memory slot.

Despite its odd position between the Alpha and NEX ‘systems’ and the clear drawbacks of some aspects of its design, the NEX-7 is a compelling camera. It’s got some of the handling qualities of a classic screw-thread Leica, from the left-hand eye position to its overall dimensions and a reassuringly solid feel. It does not surprise me that so many owners enjoy fitting Leica lenses of all eras; they look correctly proportioned, and in the case of late designs for the CL/CLE they were designed to fit a body which may even have inspired the 7.

 Missing the point

The hammering taken by the little battery probably accounts for why no GPS was built in to the 7. This must have caused many buyers and owners much frustration, especially if their previous camera or other camera happens to be an Alpha 55, 65 or 77. The NEX-7 can make these redundant for travel and landscape work, but those are exactly the times you want GPS. The NEX-7 is not so much use in the studio, or for action sports, or domestic shots… the times when you do not need GPS!

This is the one mismatch in the specifications which has caused me problems in making decisions about what to use when, and when gear to keep for the future. It has left me unable to part with other cameras despite the fact that I don’t really need them; and it leaves me obliged to take my A55 or A77, instead of the NEX-7, on expeditions where the NEX-7 might be more convenient. In the end it is what has persuaded me to part with the NEX-7; I have ended up using the Alpha 77 all the time instead.

The other omission in the NEX-7 is in-body stabilisation, not just sensor-movement but the pixel-tracking electronic variant used only for video in the A65/77. This prevents the 7 from being a true alternative to the larger DSLR/SLT models even when the LA-EA2 adaptor is fitted, unless you are very lucky and can find a compatible Sigma OS lens from the period before Sigma decided to drop optical stabilisation from Alpha mount products. It also limits its use with many manual classic lenses.

What you do get is a bigger sensor area used for HD video than on the 77 or 65. It’s interesting that this was possible, as video readout can demand a limited choice of source pixel dimensions for best quality. With a bigger source frame, perhaps the NEX produces better video, but it’s not a difference I can detect.

It would be wonderful if it turned out that the processor and firmware of the NEX-7 allowed Sony to issue an upgrade to add pixel-tracking electronic stabilisation (even with the inevitable crop of video to 1.87X factor). Is there any reason why pixel-tracking can not be used instead of sensor shift for stills? Others are using it. It certainly can work for viewfinder stabilisation with an EVF, as it does with the A77.

Great features of the NEX-7 – that ‘real’ hot shoe, the pop-up flash, the twin control drums, and the EVF – are let down by lack of GPS and video stabilisation for manual lenses, both offered by Alpha SLT models.

Cynics will believe that Sony could have implemented all this, but preferred to limit the NEX-7 and avoid cannibalising Alpha SLT sales. I think they simply didn’t think it through, or realise how the NEX-7 would change the profile of the whole NEX system to the point where some photographers could consider using nothing else. And they may have needed to use the processor power in other ways.

The killer button

While the first reaction to NEX-7 design tends to be delight at the provision of two unmarked, identical control wheels on the top right rear edge the new owner quickly discovers there is one tiny button which changes the reliability and usefulness of the camera beyond all else.

Seen here with the pop-up flash raised, showing also the infra-red remote front facing sensor and the on-off switch surrounding the shutter release, is that very important small unlabelled flat button.

That is the flat black button to the right of the shutter release, unlabelled. You could assume that its ability to invoke a series of the most useful screen menus, to change key settings in a manner close to Sony’s Alpha 700 QuickNavi, would be most praised.

But no. It is the ability of this button to LOCK the camera setting controls – to disable operation of the TriNavi control wheels (rear multi-way optional, and two top as a minimum). Just hold the button down for two seconds, and a message appears saying the settings are locked. That means that if you have set Aperture Priority, f/9, ISO 200, no exposure over-ride then the camera will stay that way until you unlock it and change things. That just needs another two-second pressure, something you are not likely to do by mistake.

The handy tips are not optional (more detailed ones are). The rear screen is bright and very sharp. I fitted a GGS-type glass screen surface shortly after the product shots were taken. That movie record button is a bit of an issue, see later comments.

You do get continued access to the exposure over-ride and ISO change even when the controls are locked, but only with enough deliberate action to prevent your clothing or your camera bag from doing what it does so well on the NEX-5n – setting several stops of underexposure. It rarely goes the other way for me. I’ve had this happen between two shots without apparently touching the camera! The rear controller which handles this function is so light in its action, and indeed the dials of the NEX-7 are equally free.

This single point about the NEX-7 puts it ahead of the NEX-5n and all earlier NEX models for me, despite the fact that NEX-5n images are often better in low light, and all previous NEX sensors seem to produce rather smoother sky noise at minimum ISO.

What it does not lock is the other killer button – the movie shoot red button, placed to catch your thumb or the camera strap or anything else passing. It is ridiculously easy to start shooting video accidentally; it happens often enough on all NEX models, but the 7 takes it to a new extreme. It needs to be included in the LOCK function with the next firmware update, or its operation changed to a two-press action; first press changes to movie mode and crops the finder view, second press starts filming; half pressure on shutter release when in pre-shoot mode returns immediately to stills mode.

It’s surprisingly difficult to hit the movie the button with the camera at your eye when you want to. For a button which is so easy to hit by mistake, it scores top marks for being hard to find when you need it.

The AVCHD-2 file structure makes the situation worse, by putting the camera into video directory mode if you accidentally record a second or two of video. If you don’t immediately hit the delete button and remove this, but instead start shooting stills, it is a bit tedious to get back into the video playback mode and delete the unwanted clip. The fastest way to get there is to shoot another brief instant of video, then playback and delete this and your previous accidental clip; menu diving to change between still and movie playback takes much longer.

Hitting the movie button to shoot short clips, like these taking using the 18-55mm OSS lens and high quality 1080p, is not always easy as you must move your secure firm thumb-position on the grip to press it.

This is a failing of the dual directory structure, which maintains an entirely separate ‘database’ for AVCHD movies, preventing you ever playing back mixed stills and video or accessing both at the same time for file management (delete!) purposes.

While the Lock/Unlock function does improve the handling of the camera, it has mysterious lapses. I’m still trying to work out exactly when and how my ISO setting is changed on a locked camera, nearly always to something unreasonably high. It doesn’t happen often but when it does, it frequently manages to be the next frame after one shot I have taken normally, and done nothing more than drop the camera from my eye and lift it again.

The sensor dilemma

ISO 1600, natural light, handheld – no noise reduction at all when processed from raw using Adobe Camera Raw. A full size version can be downloaded by photoclubalpha subscribers.

The same with 25/50 noise reduction for both luminance and colour settings in ACR. A full size version of this is also available to members.

This brings me to the question of the 24 megapixel sensor. As with the Alpha 77, it only provides true advantages when used at or close to its lowest ISO setting. On the NEX-7 this is 100 not 50, and the overall performance of the NEX is ahead of the SLT design through the range of ISO up to 16,000 maximum. It’s not a doubling of speed for the same noise level, more a matter of getting slightly better images at the same setting. However, using the Alpha 77 ISO 50 setting puts it ahead of the NEX in practice.

The strength of any AA filter is determined by two factors, its diffracting or diffusing power, and the gap between the filter and the sensor. If you increase the gap, the strength of the filter must be reduced; put it very close to the sensor surface, and a strong filter is needed.

Today’s designers prefer a weaker filter and a larger gap, as this reduces the effect of dust on the filter, a major cause of user dissatisfaction. Anyone who has used a camera such as the Canon 5D MkII which has a very weak AA filter very close to the silicon will know the problem, but that camera takes the effect so far you can often see moiré patterns too. The AA filter is too weak, and video makers often get the camera customised expensively, replacing the front glasses over the sensor with a stronger low-pass.

A second effect of a filter closer to the sensor is that at the corners, the stronger diffraction structure may be further strengthened by the angle of the rays passing through. The distance from filter to sensel is greater with rays at an acute angle than those passing through on axis.

The strength and distance of the filter are also linked to the density of the sensels on the sensor (pixel pitch). The 16 megapixel of the NEX-5n and 24 megapixel sensors of the NEX-7 have subtly different AA filter assemblies. This leads to some lenses performing better on the 5n, some on the 7. Whatever the complex mix of underlying reasons, there are many who would love to see the robust and versatile 16 megapixel sensor find its way into a NEX-7 body because they want to use third-party manual lenses such as the Voigtlander 12mm or 15mm Leica mount designs.

From my point of view, I like the 7. It seems to have a weaker filter and maybe a greater gap between glass and silicon, if the dust-on-sensor results are anything to judge by. I have not noticed any serious colour shift with, for example, the 16mm pancake lens but my experience with the lens is so different from many others. I rate it as one of the better f/2.8 85° angle lenses around, not the “it sucks” offering often implied.

Even after downsizing and crunching for the web, the moiré on the flyscreen of this diner on the run up to the Mojave desert can be seen. If you are a photoworld member you can access a full size, level 12, AdobeRGB version without sharpening or NR and see just how well the 16mm has performed at f/10, a sensible working aperture.

NEX or Alpha?

Apart from issues of stabilisation and GPS, the choice between NEX-7 and the similarly priced Alpha 77 involves a few other considerations.

First of all, there’s weatherproofing or ruggedness. The 77 is a very tough, splash or rainproof camera with a ‘skinned’ body, externally finished to be fairly resistant to minor scuffs. The NEX is a bare metal body without any special attention to dust or moisture proofing.

Then, there’s the duty cycle. The NEX has the familiar basic 1/4000th shutter, admittedly with the electronic front curtain option that doubles its expected life if you use it all the time. The 77 has exactly the same option but based on the 1/8000th shutter only found in top-end Alphas 700, 850 and 900. That gives it probably the longest expectation of shutter life yet in any Alpha, time will tell.

There are a not many functions on the 77 not found on the NEX, but there’s one big physical difference – the rear screen.

Inarticulate viewing

The NEX-7, a major redesign compared to the 3 and 5 series, sticks to exactly the same rear screen frame and hinge setup as those with a minor adjustment to angles and a slightly tougher construction. This is not one of the best design decisions made by Sony, and has drawn some users to other makes.

The articulated screen of the Alpha 77 is not just good for viewing portrait compositions at waist-level or aiming ahead to see yourself when doing a self-timer group or making a video ‘to camera’. It is good for not using at all! Nearly all the time, when not wanted for a specific purpose, the screen of my Alpha 77 stays reversed to the camera. It feels more comfortable, it never lights up when working indoors or draws attention because of its glow, it does not get marked by my hands or face. Same goes for my older 55.

The NEX-7 screen, in contrast, is permanently exposed and also limited in its movement. It is not good for vertical compositions, and it can’t be used for viewing from the front. It also can not be protected by facing towards the camera.

How different the 7 would be, had the screen been designed like the Alpha 77! It would have felt like a pure rangefinder camera with the screen reversed and I’ve considered getting or making some kind of cover just to hide it away. The surface of the NEX-7 screen is very easily marked, and I picked up a single visible scratch line on it within a week. A month later I finally obtained a GGS MkII type glass screen protector (the model labelled NEX-5C is correct) and felt able to use the NEX freely in the real world, instead of treating it as a fragile object.

I use the angled screen occasionally on the NEX-7 or 5n, but don’t use the angled viewfinder of the NEX-5n with accessory EVF. Most times I need an angled screen, it’s because I want to hold the camera overhead, at ground level or at waist level. Not up to my eye.

There is one button press which doesn’t exist on the NEX series, but exists on all the SLT models – switch between EVF and rear screen. With auto switching set, the EVF takes over when you raise the camera to your eye, but the rear screen continues to operate after the camera is returned to strap-hung position. The only way to prevent it from continuing to operate is to use the power save delay setting, and reduce this to the minimum ten seconds.

You can go into the menus, and switch the camera to use either the EVF only, or the LCD only. But there’s no over-ride if you do so. Set it EVF only, and the EVF is the only way to see the menus needed to get back to using the LCD. Set it to use LCD only and there is no quick way to use the EVF, you’ve got to menu dive.

There is not even an option which enables the LCD to cycle, through its display modes, to OFF. The closest I have got is to set the LCD for information display only and turn the brightness down to minimum manual. What I’d like to see is an EVF/LCD button, just as on the A77, because the NEX-7 is basically the same kind of camera. I don’t have to use that button on the A77 as I just reverse the screen, and flip it round when it’s needed.

Though there are many custom functions you can assign to buttons, EVF/LCD switch is not one of them. This may seem like nitpicking, but it is an omission that wastes battery power. Using the ten seconds power save timing still leaves the sensor and the display/s operating far too long. There is no state, unless you use the EVF only option, where the camera shuts off as soon as you take it away from your eye and take your finger off the shutter button.

I  find it significant that without taking a single picture, while writing this section of the article only, checking the operation and changing menu settings the battery in my 7 has dropped from 33% to 17% power. The EVF uses more power than the LCD.

Shooting speed

Though the NEX-7 has a minimal shutter lag (20 milliseconds, or 1/50th) between pressing the shutter and achieving image capture, this figure is deceptive. The camera may not be ready to have the shutter pressed, indeed you may not even be able to see and compose your quickly-observed action shot before it is too late.

This may be why Sony has made the LCD/EVF aspect so restricting, and why the default settings use Auto switching and leave power on for 20 seconds before sleep. If you use these settings, the response of the EVF is much faster than it is when the EVF only is selected. The camera is already operating, feeding an image to the rear screen, and switches this image rapidly.

If you are already framing and viewing, shutter timing can be very precise. Indeed, with many subjects I’m so used to SLR-type delays I missed the moment by firing too soon – with the electronic first curtain, the camera sound happens AFTER the shot is taken, which is deceptive. But if you lift a sleeping 7 to the eye and expect to grab a street shot, you’ll be frustrated. It can take two, three seconds or more to get it alive, viewing, exposure set and focus happening. Bike and board action show at Knott’s Berry Farm, Tamron 18-200mm.

The slowest setup is to use the EVF only and set the image review to any time (it does not matter whether it’s for the minimum 2 seconds or longer). If you enable image review, you’ll be locked out for a second or so from taking another image in fast succession, and you’ll also see the image in the finder. This is a case where the functions of the EVF and the LCD need to be separated. The image review needs to be able appear on the LCD and never block the viewfinder or the shooting pipeline.

For fast shooting, disable image review entirely and leave the EVF/LCD set to auto with the power save mode set to longer than the gap between any two shots. I don’t think you should ever need longer than the 5 minute setting and for me one minute is enough. You can set up to an hour but I don’t see much point in this, unless you were waiting and watching unpredictable wildlife with the camera on a tripod.

Though the evidence is nothing more than observation, setting focus peaking may also very slightly delay shooting response, and using a single central focus spot actually seems slower than selecting the 25-point multi area AF (but that so often picks a foreground zone and misses your target).

When you get the NEX-7 set up correctly for maximum fast response and minimum possible interruption or delay, it’s a fast enough camera to use. Much also depends on the lens; the 16mm f/2.8 is almost instant in response, the 18-55mm OSS is hesitant, and the 18-200mm Tamron likes to wake up, stretch, yawn and then focus. Apparently the Sony 18-200mm behaves much the same way.

And then, what counts

OK, I’ve spent a lot of time looking at why a camera which could be nearly perfect falls just short, partly through small details of interface programming and default settings. Most reviews don’t even go into this stuff, apart from describing what can be found in the manual (which you don’t get in a physical form, only as a PDF on the supplied CD).

In practice, the NEX-7 produces stunningly good pictures at ISO settings under 400 and is definitely at its best with either 100 or 200 set. It is after all a high resolution camera, and in old-fashioned terms it is the Kodachrome 25 or the Pan F of the digital world.

A studio shot to show the hazards of electrical fires, with smoke detector. These are not contrived subjects, both ‘energy saving’ bulbs fizzled out this way (one hanging down vertically, the other upright in a large lamp) and the NiMH cell got hot enough to burn its covering and split it, in a charger. The NEX-7 at ISO 100, with a 50mm f/2 Russian tilt-lens, reveals a level of detail well beyond most full-frame DSLRs.

As we have found out repeatedly with DSLRs, there is little point in having expensive lenses and high resolution if the image is not correctly focused. The great strength of contrast-detect AF, and magnified assisted manual focus, is that both offer near-perfect focus regardless of aperture related shifts. You can check directly through the viewfinder by setting AF to DMF, which will automatically magnify the image if you touch the manual focus control of the lens after AF is confirmed.

The 24 megapixel sensor does not reward apertures smaller than f/11; I normally set f/8 or f/9 with the 16mm f/2.8 or 18-55mm, and use f/11 with the Tamron 18-200mm because the reduction in unsharpness towards the edges outweighs any loss of ultimate resolution.

The EVF serves perfectly well if you stick to the kind of subjects the NEX-7 is ideal for. While it has some great functions like Face Recognition including registering up to eight ‘known’ people, that sort of function is mainly for people who buy the 7 because it’s the best and looks the part. It is not really a great people camera despite the Smile Shutter and the convenient built-in flash. The NEX-5N is better because people tend to hang around in low light, indoors as well as out, in the evening as well as the day. The 5n has a near two-stop advantage in real terms for effortless high ISO quality.

Mono Lake, windy summer day; the NEX-7 is great for subjects like this at ISO 100. But so was the Box Brownie… and so is almost every camera made today.

Nor is any NEX model the ideal choice for pets, kids, school sports or the usual domestic stuff. It’s actually a better camera for creative still life, macro, architecture, landscape, fine art found studies, formal portraiture and of course top grade 1080/50-60p video. Unlike DSLRs (even SLTs) the 7 can focus very smoothly during video if the subject changes distance. No camera can be quieter in focusing or stabilisation.

If you do decide to stretch the NEX-7 to the limit, I’ve found that as an example the 70-400mm SSM G when fitted using the LA-EA1 contrast detect focus adaptor works well. Focus is achieved surprisingly quickly and far more accurately than on DSLR bodies without Micro AF adjustment. Still, the lens is almost useless. Beyond 200mm, an unstabilised hand-held tele looks worse through an EVF than I can ever remember with optical finders. You really notice the jerky image and if you use the AF-A DMF function with focus peaking for a magnified manual fine tuning at 400mm, it’s like trying to hand hold a 5000mm lens. If you want to play with this, you need a tripod, and a good steady one at that.

There is no point in having a the best focus or lens quality if the image quality lets you down. I like to use Auto ISO, and find that the NEX-5n for example holds its quality acceptably right up to the high 3200 setting which you can’t limit; it would be so much better if you could stop it going over 1600, too. The NEX-7 also has an Auto range which you can’t customise. It will run from 100 to 1600, and for whatever reason I find it tends to be at the extremes of the range. Far too many images are at 1600, and 1600 simply loses too much detail. With a little care, raw conversion and reduction in image size to under 10 megapixels equivalent can yield an impressive result. But I don’t buy a 24 megapixel camera in order to get 10 megapixels.

I find I use the NEX cameras a lot in the evening or at night, when I don’t want a full sized SLR style camera on me. That makes low light, high ISO performance important. Here’s a shot in San Francisco at dusk, just enough light to hand-hold the 16mm at f/4 and ISO1600. By reducing the  image size to 10 megapixels after processing from raw, I get a very acceptable noise-grain structure and excellent sharpness (below).

The greatest contrast is with a camera like the Canon 5D MkIII. I was using this alongside the NEX-7 for a while, and maybe that colours my view. At 1600, 3200 or 6400 the 5D MkIII is still useful for quality images. The NEX-5n does well up to 1600 and is better than the 7’s 1600, at 3200. I was also using the Nikon D4, and various other cameras. Even the Nikon D800 for a while with its 36 megapixels. The NEX-7 needed to be locked down to ISO 800 or under to make use of the full 24 megapixel native size. I would have liked to have limited its Auto ISO range to 100-400, or 200-800.

Why 200? The camera does not have in-body stabilisation. This has been another issue for me, with all NEX bodies. The rest of the world goes mad for legacy rangefinder lenses, legacy manual vintage SLR lenses and similar stuff. I don’t. I bought a couple to try at the longer end and realise from the magnified focusing view that if you fit a 200mm, you will need to use 1/1000th shutter speed or even shorter to get any kind of sharpness.

The Alpha 77 has control over auto ISO range, and has SS, and has in-sensor video stabilisation which by using pixel tracking can cope with any lens you fit even if not identified to the processor. Pixel tracking would have been invaluable in the NEX.

This shot may look superficially OK, but it was hand-held at 0.6s (2/3rds) with the 16mm. I could almost guarantee that with two frames taken, I would get a sharp one doing the same with the Alpha 77 and 8-16mm Sigma. With the NEX and no anti-shake at all, this was the better of two frames taken with careful support though hand-held. Below, 100% clip from the centre. I very rarely ever have to reject images for this reason. The most serious NEX competition, Olympus’s OM-D system, has sensor stabilisation which works with any lens – just the advantage Alpha users have always claimed. Knotts Berry Farm, museum.

Ultimately, the NEX-7 demands both stabilisation in the lens used (which the 24mm f/1.8 CZ and Sigma 19mm and 30mm lack) and constant attention to manual setting of ISO to secure the optimum settings for image quality. You can not even set a slowest shutter speed for Auto ISO.

When you get it right, it’s hard to beat. Turn the viewfinder ‘show effect’ off, so you get Auto Gain in the EVF, set the camera to ISO 100, set a manual exposure for the brightest shot you are likely to encounter and just shoot raw. Push process anything underexposed.

The Tamron 18-200mm lens

The NEX-7 with black Tamron 18-200mm, and Alpha shoe fitting HVL-F20AM flashgun mounted.

I tried two different Tamron 18-200mm lens on Tamron’s stand at Focus on Imaging, one black and one chrome, on my NEX-7. There seemed to be a marginal difference in which side of the image was not quite as sharp, between the lenses, but VC stabilisation could account for this and it was only visible wide open at 200mm.

I bought one shortly afterwards at the show – apparently the last one there – and the performance was much as expected – as good at 18mm as the 18-55mm kit lens, as good at 200mm as the Sigma 18-250mm OS we use on the Alpha 77. It is as good at 200mm as the 55-210mm SEL lens I tried during the NEX launch event in 2011. When occasional shots show unsharpness towards one side for no obvious reason, I’m pretty sure that this is a result of stabilisation decentering a group. Nearly all stabilised camera and lens combos give me occasionally ‘soft on the right’ results, maybe it’s down to my personal camera shake tendencies.

The lens was so useful that I stopped bothering to use the 18-55mm at all, and ultimately sold that. If you are going to have an unpocketable camera, it might as well be a little bigger and have the range of lens you need. I did have to carry my HVL-F20AM flashgun to use on the 7 with this lens, as the pop-up flash is barely able to avoid shadows from the 18-55mm (lens hood removal obligatory). The Tamron 18-200mm – and Sony 18-200mm to an even greater degree – will cast a huge shadow from the pop-up.

This is not just a minor shadow. Lens shade on at 18mm, you have a mound of shadow occupying a third of the frame to the left. You get a shadow in shot right up to just before 200mm with the hood on, and up to around 40mm with it removed. In practice, you can’t use the pop-up safely at settings below 50mm and you can’t use the lens hood if you do.

The HVL-F20AM gives no shadow at all even at 18mm with the Tamron with the hood fitted, and having separate batteries it does not further exhaust the hard-worked NEX-7. Folded down to the off position, it sits neatly above the Tamron lens body. So the recommendation has be Tamron or Sony plus this accessory flash, if you want to go the 18-200mm route.

The Tamron may have an acceptable close-up ability when set to 200mm, but at 18mm it lags behind the 18-55mm and way behind the 16mm pancake. If you are used to the 16mm’s ability to focus on subjects barely a hand away from the front rim, the 18-200mm’s inability to be used for this kind of wide-angle close-up will frustrate you. I would never consider leaving the 16mm behind – or its ultrawide and fisheye converters.

If you set the camera to use auto correction for vignetting, distortion and CA it appears to recognise the Tamron and to apply appropriate adjustments. I don’t know if this is because the Tamron is an authorised SEL lens, and Sony have data embedded in the camera firmware, or because Tamron is used the same identity as the (very different) Sony lens and it’s pure chance that corrections are similar.

I saw several children running towards this scene, but these were the last two. I’ve retouched a badly positioned child and a notice board out of the shot. The NEX-7 with 18-200mm Tamron was only just fast enough in operation to enable this shot. Anything much more spontaneous was rarely caught on time.

Whatever the case, if you want fast viewing and focusing and only shoot raw like me, disable lens corrections in firmware. It is difficult to judge or measure exactly what happens, but I find the 18-200mm tends to have a seek and find action when focused to start with which makes it slow, and that a further ‘wobble’ is created by the corrections as they are applied to the live image. It’s almost like a small auto zooming effect, depending on focal length, and if you zoom the processor may apply a new correction.

The slowest overall combination involves using ‘Effect On’ in the live image, AF, Face Regnition, Object Tracking, Auto ISO, and auto exposure setting like A or P, stabilisation on, and lens correction. This is not specific to the Tamron, but it tends to show the slowing-down effect most.

The NEX-7 will work fastest if you turn all this off and work manually, with manual ISO, manual focus. But of course you can’t really do that in practice. And, if you have used any EVF Sony camera, you’ll know well enough that startup times in tests are irrelevant. You are just as likely to have a second wasted as the viewfinder switches from burn-out blank to normal exposure as you are to have an AF lens do its yawn and stretch routine before finally ambling down the stairs to get breakfast.

Having recently used the Canon G12, G1X and Fujifilm X10, I can confirm that there are substantial benefits to optical finder operation as long as the camera is genuinely able to autofocus during a hasty shutter press. Sony has yet to achieve this, and it does not surprise me that the useful 16mm optical finder designed to pair up with the pancake lens does not fit the NEX-7, which lacks the smart accessory connector. So you don’t have that option with it as you do with the 3 and 5 series. The 16mm optical finder could also be used, reasonably well, with the 18-200mm locked at 18mm (the Tamron lens has a lock).

The comparison in size between the 70-400mm Sony G lens and the 18-200mm Tamron, the Alpha being shown fitted to the 7 via an LA-EA1 adaptor (which does enable it to focus, very accurately and not too slowly).

Comparing A77 use to NEX-7 use, a superzoom on the A77 focuses in the time it takes to press the shutter, with phase detection AF, but the A77 finder is every bit as slow as the 7 thanks to the 24 megapixel sensor and the way it provides your live view. I’d have to rate the NEX-7 plus 18-200mm as the slowest combination I have ever used, and if this camera had been provided to me for review in the guise of a compact with a built-in zoom of this range, I would have dismissed it as unusable from the start.

Such is the appeal of the camera with 18-200mm that I never felt that way. There are a few other ‘likes’ to help, like the Tamron’s filter thread of 62mm matching my Alpha 77’s CZ16-80mm so it can share one polariser.

Finally, like the Sony 18-200mm the Tamron has its OSS and focusing both optimised for video. The slow and occasionally odd behaviour of AF for stills may be due to the smooth, damped AF during video which hardly ever hunts off-target and always transitions between planes without jumping or overshoot. The OSS during video is amazingly stable, and both functions are so close to being silent they make other systems seem crude. This did not prevent focus from eventually drifting for no apparent reason during some long clips.

This video was shot at a photo trade show using the NEX-7 and 18-200mm, hand-held and walking while shooting. The refocusing and stabilistion can both be judged from it.

There is none of the noisy IS you can find in a Canon lens, and when comparing the sound picked up that system’s new 5D MkIII with a 70-200mm IS L lens against other options, the NEX-7 with Tamron 18-200mm emerged as the quietest of all possibilities. It was almost matched by the A77 with an SSM lens, but the sudden and fast focus responses given by the PD-AF systen during video both caused loss of focus with sudden changes, and more audible operation.

NEX-7 with Tamron 18-200mm and Rode Pro Videomic with ‘dead cat’ wind baffle.

The NEX-7 with this lens is uniquely good as an HD camcorder. Even if you bought the lens and never used it for still work, it would be the lens which completes the camera as a decent spec video rig. The same applies to Sony’s own 18-200mm. This isn’t to say you will not encounter some unusual effects if you choose to shoot freehand, to pan with subjects or zoom with AF and OSS active. Occasional distortions or apparent jumps in position of the subject can happen. It’s unfair to criticise any system for this, as the solution is to disable stabilisation and AF, mount the camera on a fluid head tripod, lock in manual exposure settings and shoot like a pro.

Tamron updated my lens under warranty to solve a definite problem with shooting panoramas (the original batch of lenses simply didn’t work – areas of blur and bad stitching). The service is handled in Germany not the UK. It eliminated the problem.

Parting with the NEX-7

Despite everything, after my final two-week trip using the NEX-7 and Alpha 77 side by side (or for different situations) in California I made the decision to sell the 7 and Tamron. I had not been able to afford the camera and the zoom to start with, especially so soon after investing in the Alpha 77. Based only on the three months spent with the NEX, it didn’t make many exposures; a mere 1500 or so. That’s because of the way I use NEX cameras generally. They are my pocket notebook, my out-shopping, business travel or evening out camera.

With the Tamron, or indeed the 18-55mm which I’d sold almost immediately in order to finance the Tamron, the 7 is not a pocket camera or even an under-jacket camera. Whenever I was in an unknown location – a stop on the road, a wildnerness view, a beach with no sign – I used the Alpha 77 and waited for a GPS lock before shooting. I wear an old Lowepro Sideline Shooter belt-pack bag which leaves hands free and places no strain on neck or shoulders, and having first packed it with NEX-7, filter, batteries, cards, 18-200mm, flash, two converter lenses and 16mm I found it was just as happy with A77 and 16-80mm, 8-16mm and 70-300m Sigmas and accessories. It weighs more but proved just as convenient.

Both the NEX and the A77 went through twice as many batteries as Shirley’s Alpha 580 despite not taking as many shots, per camera. EVF cameras are power-hungry. Both definitely lost me shots I wanted to grab and would have secured with a conventional DSLR or a faster compact. Having a shutter response time of 0.02 (1/50th) second means little if you have an aperture-focus-exposure cycle taking seconds when the camera is raised to the eye and first pressure is taken on the shutter.

Both A77 and NEX-7 also went through my stock of 16GB memory cards, shooting raw only, rather too quickly. For the A77 I’ve now bought a 32GB. You do not need to do much shooting to fill up cards with a 24 megapixel camera.

Having packed up the 7 and sent it to its new owner, I very nearly bought another. It’s a camera like that. When you consider that extremely fast response – so fast that unless you retrain yourself, you will anticipate action too early – and the amazing low light AF ability, high resolution, controls and handling, video quality, high grade shutter, near-silent operation… there’s nothing else like it. Then you remember that the A77 and A65 match nearly all aspects that matter except the ‘pure’ mirrorless design and compatibility with all kinds of optics.

I had already owned a NEX-5n with accessory EVF and other kit, and despite liking the quality delivered, decided that this was not the pocketable solution either. For a while I had both together. I’m reverting to a first generation NEX-3 taken in part payment for the NEX-7, which will be happy with my 16mm; I’ve checked some of my 2010 14 megapixel files, and find the Adobe Camera Raw 2012 process and new lens profile handling improve them significantly. I guess my argument is that I had considered a Canon G-1X, but remembered that it’s only 14 megapixels, its noise performance is no better than a first generation NEX, it doesn’t do close focusing the same way, doesn’t have wide and fisheye converters, and if anything it’s bigger.

Reverting to a ‘traditional’ old NEX without the flash shoe and auto-gain live preview means I can’t use it with studio flash. But I never wanted to use the NEX-7 with studio flash (AC main strobe) except for the purposes of testing that function. I have an Alpha 900 and an Alpha 77!

So what advice?

Here’s my view. If you do not own an Alpha SLT 24 megapixel camera, or a compact of high quality, the NEX-7 is a star buy. You just have to be aware that this is absolutely NOT the camera for birds in flight, dogs running, kiddies scooting round the living room, sports, candids or street shooting. All are possible and owners have good examples.

It’s a really great buy if you have some vintage short focal length Leica, Contax or similar glass and want to make good use it without spending twice as much on a Fujifilm X-Pro1 or a Leica.

It’s the only camera of its type that can do smooth refocusing during video, silently, maybe 75% of the time. Unfortunately, the 25% when it wanders off focus completely for no apparent reason makes life with AF video not much better than it is with every other unsatisfactory solution. If you shoot video, you’ll get some of the best ever quality 1080p from the 7, but you’ll end up using a tripod with manual focus and exposure for anything beyond casual clips.

The NEX is a cat, so don’t expect it to do dog tricks. Or obey!

Reasons why I am wrong include the 50mm f/1.8 OSS lens which I don’t have. If I had the cash to hand I might have bought one, and who knows? It might have been a compelling reason to stick with the NEX-7. But I’ve got a great 50mm f/1.4 Sony for the Alpha 77. The adaptor LA-EA2, which puts an Alpha 65-style AF module and SLT mirror on to the body to work with Alpha mount lenses, could also have tipped me in favour. But once it’s all assembled, it’s approaching Alpha 77 size without the ergonomics or bigger battery.

You buy a NEX to do the things a compact system cameras does well, like being small and portable, unobtrusive and precise in feel. It happens, uniquely, to beat most DSLRs in two or three aspects – image resolution and quality including dynamic range, versatile lens compatibility, and focusing accuracy whether AF or manual (based on the near-perfect precision of contrast detection or magnified visual with peaking indication).

If you DO own an Alpha 77 or 65, and thus have access to almost the entire feature set of the NEX-7 already, I suggest that your money – the better part of $2000 or £1400 with a zoom like the Tamron – is better spent on reinforcing your Alpha system if necessary, and acquiring a large sensor compact of the new 1X generation. For my outlay I could, admittedly six months after my NEX-7 purchase, have a spare Alpha 65 body and an RX100.

Here are some parting thoughts:

1) Battery Compatibility – the NEX models share a battery size with the Alpha 55, 37 and similar bodies. If you own a 55 without too much investment in lenses, moving to NEX-7 could be neat. I enjoyed travelling with my A55 and NEX-5 plus a pocketful of small batteries charged up!

2) Electronics Break Down! If I really wanted to go with NEX, I’d be better off buying two NEX bodies, an LA-EA2, having one battery type but also a coherent backup. If I really want to major on Alpha a-mount SLTs, my first backup investment should be a second Alpha body. I have that (A55 and A77) but guess what, I risked two weeks travel and shooting with A77 and NEX and never thought to sling the A55 in the case.

3) EVFs suck. They are wonderful, but if all you have is EVF, you miss half your potential. We are keeping our Alpha 580 and even our 700 and 900 for the moment. See also – Electronics Break Down!

4) 24 megapixels is a dozen too many. Most of the time you only need 6 megapixels, sometimes you need 10 or more. Very rarely does anyone need 24, except if they are shooting sports and wildlife, or something where a crop helps. Guess what the NEX-7 is not so good at. When you do need 24 (or even more) you don’t often need a pocketable camera. Note to Sony: RX100 – 20 is probably too many, too.

– David Kilpatrick

See NEX-7 at B&H (if you use this link or our B&H sidebar adverts, or Amazon ads, we do make something – not a great deal, and it does not cost you anything, but it can add up to help pay site costs).

Creep no more! The LensBand.

Though their website shows a very different way of using this theme park wristband repurposed – the LensBand anti zoom creep device – it just happens to fit many Sony zooms perfectly without stretching it into position to lock barrel to collar.

By fitting the LensBand exactly over the joint between the zoom ring and static lens barrel, you can create a much smoother damped zoom action and prevent the zoom from tromboning when carried, or creeping during long exposures when the camera is angled up or down. This issue is not tackled by stabilisation and very small amounts of zoom creep can affect exposures as short as 1/30th without your realising. You don’t get why the zoom appears to be not so sharp – that’s because you do not realise you, or gravity, have caused a tiny shift in focal length during the exposure.

These silicone rubber bands are not expensive, and they can also be used to provide extra grip on zoom or focusing rings. We chose two colours as the best for Sony Alpha – orange and black! At the end of this article you’ll find a link to our associates at The Photostore (Adrian, for a decade the manager of the Minolta Club) who have both these wonderful rubbery colours in stock.

Here’s Mr Orange fitted to a very first batch 16-80mm CZ zoom which has become slack enough to extend itself when strapdangled. By positioning the band centrally over the seam, leaving the focal lengths just visible and the focusing window clear, the lens is transformed. The zoom action becomes super-smooth and the lens does not shift at all. The elasticity of the LensBand is 100% perfect for this zoom. You can yank it over the back of the zoom ring closer to the camera, at one side, to increase the lock effect but in practice this position works full time. It locks, and prevents creep, and also radically improves the feel of the lens zoom.

As you can see, it also looks pretty funky and matches the orange lens not-a-seal-just-a-decoration bit on the Alpha 55 mount.

But for a Sigma 70-300mm Apo Macro – one of our favourite lenses despite past histories of stripped gears, since no other lens does the same thing in terms of focus range and pure sharpness – the black band looked better. This time, the fit is a little tighter but the band forms a cone shape gripping between zoomring and barrel. It prevents the rather slack zoom of a brand new 70-300mm Siggy from causing havoc. Between 150 and 250mm, this zoom will tend to self-extend if aimed downward (or collapse back when tracking those same old BIFs and airshow targets beloved of dPreview boringphoto posters!).

You can see that this new band collected dust and skinbits instantly – it really doesn’t matter if you have just spent 15 minutes in a bath, the human body sheds a couple of million sensor dust specks every hour and your fingers are a magnet for silicone rubber. Black always looks worse – the orange band stands up to our macro lens photography a bit better.

The band bridges a fatter zoom ring to a narrow diameter lens barrel. This still works. By moving the band just 1mm either way, the tension or damping effect on the zoom can be controlled precisely. This particular lens was improved 200% in feel by fitting the band. It was also made usable on a tripod when aimed down or up.

Well, there it is. Could not really be a simpler concept (we’ve done it with fat rubber bands – one Photoclubalpha member reports using a band found holding together a bunch of broccoli). But these robust, one-size-fits-all-except-tiny-microfourthirds LensBand alternatives are clearly a better choice.

No, they don’t fit the 70-400mm SSM G. For that you need the inner tube of a hobbit bicycle tyre, or perhaps a couple of sand eels knotted together. Or one big one in the fashion of the Worm Orobouros. One day, we’ll tackle that; at the moment, our 70-400mm SSM G is solid enough not to slip. Another really rocksolid lens seems to the 16-105mm Sony SAL zoom, lovely firm action; and the Tamron 18-200mm E for NEX feels like it won’t get loose for years.

You find the LensBand for sale all over the place, but in UK, for orange and black, visit:

Adrian Paul also has loads of small accessories and he can get almost ANY Sony Alpha part or product to order – faster than most retailers. He is an authorised Sony Alpha supplier and has 25 years plus in-depth knowledge of the system.

For worldwide orders, B&H in the USA stocks all the colours, even the ugly ones which look nice on Canon lenses!


Sony DT 16-50mm f/2.8 SSM

I’m about to offend myself. I own this lens, and I know how upset owners of brand new lenses get when someone says it’s not perfect. Well, the 16-50mm SSM is far from perfect and if you know how to check out lenses, you’ll agree should you be lucky enough to own one. It’s a compromise. But I love it.

Here’s the problem; this lens has such soft corners and complex distortion at 16mm and f/2.8 that it makes the NEX’s legendarily reviled* 16mm pancake look like a Super Angulon in disguise. It’s got a curved field at 50mm and stopping down does not always bring distant scenes into perfect focus across the frame. It suffers from rampant chromatic aberration which just becomes a dead-sharp fringe on stopping down. *Not by me!

This shot was taken on a preproduction A77 and 16-50mm. I was not supposed ever to show it. But I know there is no fault with the shot, the pre-release gear was just fine. And I really like the minimum focus, at 50mm, at f/2.8!

Yet it also has exceptionally high central sharpness, great colour and contrast, and a lovely quality to its differential focus. That’s the old traditional English-language term for the context in which people over-use the term ‘bokeh’, and deserves to be revived. With f/2.8 to play with across the entire zoom range, you can use differential focus creatively. At medium settings, 24-35mm, the distortion disappears and the sharpness extends corner to corner wide open. You have to set it to 50mm to lose the edge.

More than this, the 16-50mm SSM is a video-tuned lens. Its natural host camera, the Alpha 77, crops the frame considerably when shooting HD video. The soft corners and even most of the distortion don’t get a look in, they are outside the video area. The standard and 3D pan modes of the A77 also crop out the problems. The focus action and silent supersonic motor of the 16-50mm are ideal for A77 video shooting with active AF (if you want it) during takes. The f/2.8 aperture allows the lens to be stopped down to the optimum f/3.5 used for movies and also for high speed (12fps) mode, and have no issues with aperture shifts if the focal length is changed.

The Carl Zeiss 16-80mm, left, is smaller than the Sony 16-50mm SSM.

After testing the lens, I decided to keep my 16-80mm CZ which is now five years old. It’s not just the different quality of image produced by the CZ coatings and design, or the very slighter better close-up ability (you can’t get quite as close but at 80mm the subject scale is a touch bigger on the CZ – the 16-50mm wins at 16mm, where getting two and half inches closer to the subject makes a real difference). The CZ is lighter, takes 62mm filters rather than 72mm, and is considerably smaller with lens hood size adding to the difference. Working in the field, it is a lens which can easily be held in the hand with fingers free to operate the lens-mount release button, hold a rear cap, or even another lens – the usual juggling of two lenses which photographers get used to.

With lens hoods fitted, the overall relative sizes become more obvious. The SSM lens has an attractive metal front ring, a new trademark of higher-end Sony lenses, shared with the 70-400mm G.

The 16-50mm is at the limit of diameter, shape, balance and weight to be safely gripped with another lens in the same hand, even briefly during the process of swapping over. That’s not to say it is cumbersome, just that the 16-80mm is faster and more secure to work with because it’s that little bit smaller and lighter.

Once on the camera, I have to say I like the overall balance created by the 16-50mm. It tends to help the A77 hang lens-down, a position I prefer with the camera under my left arm and the strap over my shoulder. The zoom action is super-smooth and well damped, and also has a lock which operates at 16mm to prevent gravity-fed creep, and keep the action firm in future.

No creepy zooming – thanks to Royal Mail, and their neat Sony-coloured rubber bands which are a perfect fit to go on the CZ 16-80 and make the zoom action super-smooth and stay put!

My CZ is now well used and over-free in action. A rubber band to go over the front end of the zoom ring is the cure! You can get proper broad Alpha-ish orange silicone rubber ones from Lens Band as well as the free orangey-red ones used in the UK to hold our postal deliveries together. My way of using a rubber band is not quite the same as Lens Band’s method, it goes over the flush seam between zoom ring and lens barrel on the 16-80mm and it doesn’t just hold the zoom, it smooths the zoom action.

The zoom lock on the 16-50mm was missed from the 16-80mm… missed by all owners, that is. The 16-50mm has a type of raised  moulded marking. Durable? Maybe not. The similar raised ‘P’ on my A77 mode dial is now a ‘D’ having lost its stalk.

The best shots I’ve got from the 16-50mm are as good as the best from the 16-80mm, but I can trust the CZ more in the 35-80mm range. From 35-50mm the SSM becomes increasingly soft and sharpness towards the edges of the frame can be poor. At first I thought this was only at full aperture, but shots at apertures like f/5 and f/7.1 were affected. I compared my own lens with two pre-production Sony samples I had used months earlier; we were told not to release images taken with these. The degree and type of sharpness loss was identical, enough for me to conclude this is a characteristic of the lens and not a coincidental case of rogue lenses.

Major plus points for the 16-50mm include focus accuracy, which is much better than the 16-80mm on most Alpha bodies. The f/2.8 aperture activates higher accuracy sensors, such as the Alpha 700’s central point and the extended range of the 11 cross sensors of the Alpha 77. When used on the Alpha 580 for live view pre-shot AF, or on the NEX models with the original LA-EA1 contrast-detect AF adaptor, both focus speed and accuracy are optimum.

The SSM lens has an AF/MF switch but no on-lens button control. Direct Manual Focus is supported and unlike SAM (conventional in-lens motor) lenses, the supersonic drive is not damaged by moving the focus ring without engaging MF.

Despite the large area of glass, the 16-50mm is no more prone to flare than the 16-80mm. The new Sony coatings used for this lens (water and oil/dirt resistant, very hard, similar to Nikon’s NanoCrystal) do a great job. And, of course, they are part of the final reason I am keeping this objective. It’s weatherproofed to some degree, as is the Alpha 77 body. Reports vary between dowsing with a bucket of water without harm, to reluctant use in slight drizzle. I think I’ll get myself a Sigma EX DG filter for my lens, as these have the same coating now and they are about the best slim-mount UV filters made for optical quality without paying Hoya Pro1 Digital prices.

Also, with the 72mm filter thread, there seems to be less need for a super-slim filter. The CZ lens suffers from very strong mechanical vignetting at both ends of the scale, producing dark corners at 16mm or 80mm alike. At 16mm, depending on the position of the SSS/AS sensor-based stabilisation, a dark corner can be well enough defined to need cloning out or the image cropping. The 16-50mm SSM has no such issues. Not only is optical vignetting well-controlled, the mount does not create any dark corners.

These dark corners are created by adding vignetting and grads in raw processing. The 16-50mm, at 24mm, turns in great shadow to highlight detail without a hint of flare; 1/50th at f/9, ISO 100, hand-held with SSS – mid-January in the Scottish Borders. When I pulled up to shoot this, a car with two camouflage-kitted big Nikon and Canon multi lens ‘serious enthusiast’ shooters pulled in alongside. They were still struggling with tripods, a kissing-gate, a stone wall, lenses and car by the time I’d got the sunray shot (which disappeared in seconds) and left. I just carry my Alpha 77 – but then, I’m not a ‘serious enthusiast’ and my ideal camera would be invisible and with me all the time. I’m a panda – sees shoots and leaves.

Though Sony owners may be reluctant to admit it, the SSS mechanism can decentre the sensor and if the lens coverage is so tight it barely covers the corners of the frame (16-80mm and 16-105mm both guilty) you can get the occasional asymmetric dark corner. I’ve never seen this yet from the 16-50mm. But when I check the 16-80mm against the 16-50mm using the Alpha 900 full frame finder to examine the image circles, if anything the 16-80mm has more apparent clearance round the extremes of APS-C, with a softer gradation. The 16-50mm has a tight exact circle.

I have also checked the way the 16-80mm and 16-50mm focus as you zoom. Though the CZ is not perfectly parfocal. That term describes a zoom which retains exactly the same focus point, as you zoom. Video and TV camera lenses are parfocal, otherwise, the focus might ‘go off’ during a zoom. The CZ is nearly parfocal, just a touch varifocal. That’s the opposite term, and means a zoom which changes the focus as you change focal length. At one time, varifocal lenses were not actually called zooms; they date back to the 1920s, and J H Dallmeyer’s adjustable telephoto lenses. Konica made a famous 35-100mm f/2.8 Varifocal in the 1970s.

Silent focus, silent A77 camera (almost), 16mm and ISO 800 at f/3.2 – with ACR profile correction. Café society, Hawick.

The 16-50mm is either a perfect parfocal zoom, or so close you will never know. It is possible to focus at 16mm, and zoom in to the subject. This can only happen with very accurate focus, and a parfocal zoom. Try it with the 16-80mm CZ and you will see the image go out of focus, not to mention clicks and jumps in brightness changing as the aperture adjusts (that’s because the CZ is a variable maximum aperture lens, f/3.5-4.5). The 16-50mm can zoom during video, in or out, without losing the original focus point and without any brightness change or aperture adjustment.

Now you may understand why I want to keep this lens even though – unlike some enthusiastic new owners – I find that its sharpness across the field is not actually as consistent as the CZ. It is a far better overall match to the Alpha 77 especially for video work. But in January, I chose the CZ in preference for a week abroad, and I would most likely do so again.

The Alpha 77 (and 65) include built-in correction profiles for this lens. They are so effective that when I first saw JPEGs from it, I thought the geometry was perfect. If you intend to use the lens for JPEG and movie shooting, any criticisms can be moderated. The correction profile can not improve sharpness, and it does change the effective focal length slightly so than you don’t get a true 16mm.

This is a straight-on shot of the Adobe chart used (not this way, shot nine times per full frame) for profile creation and it shows how very bendy the 16mm f/2.8 setting is at this range, the target is A2 size. Click image to see full size.

This is the same, but JPEG with the in-camera correction enabled.

This is the same, with the Adobe Lens Profile I have created and sent to Adobe, applied in raw conversion of the first example. Please note that the Adobe profile applies to shots taken at three times this distance or more – these profiles, like the in-camera profile, are never much good at rigorous correction of geometric targets shot a couple of feet from the lens.

For Adobe Camera Raw, I have made a profile for the lens which covers three apertures (f/2.8, f/5.6, f/11) and three focal length settings (16mm, 24mm, 50mm) between all of which ACR will interpolate correction values. Because the extreme corners of the image go so much out of focus when shooting the target (refocusing ruins the profiling process) I don’t think this profile handles chromatic aberration as well as it could. The profiling program needs a sharp image of the RGB colour channels to work out their relative scale, which is how CA is corrected. Applying 150% CA correction, instead of the default 100%, seems to improve the conversion.

Here is an uncorrected real-life shot on the 16-50 and 16mm, 1/125th at f/9, ISO 200 (click image for full size 24 megapixel view, and note the chromatic aberration at the left end of the shot especially).

This is the same raw file processed using the Adobe Lens Profile I have produced for the lens.

You can dowload from here the 16-50mmA77rawAdobeLensProfile, hopefully it will also become available from Adobe’s user-created download area. Unzip the file to extract the .lcp file, and place this in your Application Support/Adobe/CameraRaw/LensProfiles/1.0/Sony directory. You require Photoshop CS5 to use the profile.

So what is my conclusion? I do not agree with some of the over-the-top reviews including one to be found on the Sony store USA site claiming it’s the best zoom of this range and aperture for any system. It is not, you get more than you pay for (much less than a lens of this specification might cost from others) but not an optical miracle. You get a very well designed optical compromise housed in a particularly good mechanical design. I would compare it favourably with Olympus’s waterproof ‘Top Pro’ range fast lenses for 4/3rds. I think it can claim to match Canon’s 17-55mm f/2.8 and Nikon’s similar lens, I’ve used both and the Sony is rather neater. It’s probably a little better than the Pentax/Tokina 16-50mm f/2.8, which it most resembles but definitely is not related to.

It’s different from the CZ 16-80mm, not better or worse; it has a different mix of good qualities and failings. The obvious competitors are Sigma’s 17-50mm f/2.8 OS and the Tamron 17-50mm f/2.8. The Sigma offers Optical Stabilisation. The Tamron is now an older design, replaced by a new VC stabilised version for other mounts, but still issued without stabilisation at about 60% of the price of their VC versions, for Alpha. It is the lowest-cost option in this range.

The Sony Alpha SSM 16-50mm f/2.8 DT lens is supplied with rear cap, 72mm lens cap, and bayonet petal hood. It does not come with a case or pouch. My lens was purchased ‘white boxed’ – that is, split off from an Alpha 77+16-50mm kit by a dealer and priced accordingly. The lens is only available with the A77, or as a separate item; it is not currently offered as a kit option with the Alpha 65 or other models.

– David Kilpatrick

Check the current price from B&H Photo Video – remember, B&H ship worldwide and for the UK buyers, offer a UK service.

Technical Data (Sony information) amended to remove nonsense

  • Lens Type : Standard Zoom
  • Focal Length 16-50mm (35mm equivalent 24-75mm)
  • Lens Mount Type : Sony A-mount, SSM in-lens supersonic motor focusing, electronic coupling
  • Aperture (Max.) : f/2.8
  • Aperture (Min.) : f/22
  • Filter Diameter : 72mm
  • Lens Groups-Elements : 13 groups, 16 elements
  • Minimum Focus Distance : 12″ (30cm)
  • Distance Encoder : Yes
  • Distance Scale: Yes
  • Angle of View: 83°-32°
  • Non-rotating Filter Thread : Yes
  • Aperture : 7 blades (Circular aperture)
  • Lens Weight : 20.4 oz (577g)
  • Maximum Magnification : 0.2x
  • Dimensions (Approx.) : 3-1/4 x 3-1/2” (81 x 88mm)

Compare the 16-80mm Carl Zeiss technical data:

  • Lens Type : Standard Zoom
  • Focal Length 16-80mm (35mm equivalent 24 – 120mm)
  • Lens Mount Type: Sony A-mount, in-body motor focusing via mechanical drive coupling
  • Aperture (Max.) : f/3.5 – 4.5
  • Aperture (Min.) : f/22 – 29
  • Filter Diameter : 62mm
  • Lens Groups-Elements : 10 groups, 14 elements
  • Minimum Focus Distance : 14.4” (36cm)
  • Aspheric Elements : 2 aspheric
  • Distance Encoder : Yes
  • Distance Scale : Yes
  • Angle of View: 83°-20°
  • Non-rotating Filter Thread : Yes
  • Aperture : 7 blades (Circular aperture)
  • Lens Weight : 15.7 oz (445g)
  • Magnification : x 0.24
  • Dimensions (Approx.) : 2 7/8 x 3 3/8” (72 x 83mm)

Sony’s Zeiss 24mm f/2 Distagon ZA SSM T* reviewed

The Sony Zeiss 24mm f/2 SSM Distagon ZA T* is probably the best, or equal to the best, in its class. It may perhaps be the best ever 84° angle fast lens ever made for the general SLR system market, and I would happy to pitch it against any of the current equivalent offerings for medium format digital.

The initial journey with the 24mm f/2 was not one of intensive companionship – I am long past the stage of getting hold of a wonderful lens and then shoehorning all my photographs into that lens’s view just because I love the glass. I’ve been through that phase. I remember when I was 18 and my then fiancée (Shirley – still here!) bought me a brand new 35mm f/3.5 SMC Takumar, my first ever multicoated lens as well as my first new boxed product. I shot almost everything with that lens for a month…

A full-frame Alpha 900 study at full f/2 aperture. Check the sharpness in the central – very limited – sharp focus zone by clicking the image for a full size version.

My review of the 24mm appears in the British Journal of Photography for January 2012 but was written in November, and at the end I comment that I do not think I would buy one. Well, between writing that and publication – after returning the test lens loaned to me by Paul Genge of Sony UK – I placed my order. I sold a set of lenses including a 28mm f/2 Minolta RS and a 17-35mm Konica Minolta D to pay for it.

Check current availability and price at B&H Photo Video (opens in a new window will not lose this page).


It was partly medium format which persuaded me. I’ve been experimenting with MF digital, first using a Hasselblad with a Phase One P20 and then shifting to a Mamiya 645 AFII with a 22 megapixel ZD 37 x 49mm back. Once you put the Zeiss on the Alpha 900, the image quality jumps to match the level of a similar MF pixel count. And without spending into the tens of thousands you can’t match the angle of view at a higher pixel count.

These two cameras both shoot 22 megapixels over a 16 x 12″ print shape (the Alpha 900 being cropped) and both were current in 2008 – though the Mamiya ZD model was shortly to disappear. And the two lenses have similar coverage.

I looked at the corners of my MF shots on a 35mm lens (nearly identical angle of view) – to be clean, they demanded f/11. And then I looked at the corners on the Zeiss, which are even cleaner by f/4. Finally, I considered what Sony may have in store – 36 megapixels on full frame. Everything I’ve seen from the 24mm – including its performance on the A77 and A55 – indicates it will not run out of resolution even if full frame goes well over 50 megapixels.

Then I had the job of looking back over the Alpha 900, Alpha 55 and Alpha 77 pictures taken with the 24mm, and preparing some comparison shots. This was when I realised that my normal line-up of zooms, no matter how good, never got the same from any camera – APS-C or full frame – as this CZ prime. It may be bulky, take large filters, and cost nearly £1,000 but no other solution on any format from NEX through A77 to MF offered the same as the 24mm on Alpha 900. You will, however, be surprised later on to see just how well the tiny NEX 16mm f/2.8 does in comparison when both lenses are stopped down to f/8.

The 35mm 2:3 format shape offers a bit of vertical composition ‘rise or fall’ potential compared to to 3:4 shape of my Mamiya with 35mm wide–angle. Beyond this, the 24mm offers both CD and PD focus with different adaptors on the NEX system, and smooth near-silent AF during video on the Alpha 65/77 and future models. It’s both future-proof and a future classic.

Photojournalism or architecture

Because the 24mm has a fast f/2 maximum aperture, it’s seen as a choice for news, documentary, reportage, sports, and close quarters party or family shooting. Though a little vulnerable because of its size, it does this job well. Unlike tele lenses, any mark on the front glass of a wide-angle like this will show in pictures when the aperture is stopped down. Special care should always be taken of retrofocus and fisheye lenses with vulnerable front elements, my own lens will get a Sigma EX DG 72mm UV filter. Why Sigma? I ran a series of ad hoc tests on filters and these turned out to be just as good as Hoya Pro 1 Digital at half the price, and with better multicoating.

At f/2, struggling with light for a hand-held shot with 1/40th at ISO 1600 on the Alpha 55, the 24mm showed surprisingly clean imaging from the boat to the lights on the cliff top.

Here’s a shot taken at f/2.5, 2/3rds of a stop down from wide open – a sensible aperture to give that hint of extra depth of field and improved optical performance. Click the image to view a full size A55 image on pBase.

When fitted to my A55 or A77, the 35mm-equivalent field of view is also a good general lens for photojournalism (what you get is more or less a Fuji X100 equivalent, but hardly pocketable). The performance over the APS-C field of view is so good that working at full aperture carries little penalty at all except restricted depth of field. The geometry and field flatness over the restricted field mean  you could use the lens for artwork copying and get a better result than the 50mm f/1.4 of 30mm f/2.8 SAM macro will produce.

Over full frame, this technical excellence makes the lens attractive to the commercial, industrial and architectural photographer. Whenever you need to apply a strong software correction, focal length figures are thrown out of the window. For example, once the on-board lens correction in the A77 is applied to the 16-50mm f/2.8 SSM lens at 16mm the true minimum focal length equivalent becomes close to 17mm not 16mm.

Hasselblad’s 28mm superwide for its HD series cameras has strong barrel distortion, relying on in-camera and Phocus raw software converter functions to remove it. So while the lens claims to be a 17mm equivalent, that is only true over absolute full-frame 645. On their digital sensors, it’s only equal to a 21mm and the correction means the true crop is more like a 23mm.

A second effect of applying any in-camera or post-process distortion correction is loss of true image pixels. Either you crop the frame after sampling down, or the image is interpolated upwards to fill the frame. Both solutions are far from satisfactory because unlike a fixed interpolation, the value ranges from 0 to whatever maximum is involved (typically between 3% and 7%) and all of this is never a clean ratio.

Above: a sea horizon (the top of the crop is the top of the frame, and it is full width). Top, CZ 16-80mm at 16mm 0n Alpha 77, uncorrected, showing complex wave-form distortion as well as vignetting despite stopping down to f/11. Centre: CZ 24mm on Alpha 900, uncorrected, at f/13. Bottom: 24mm after applying a 2% barrel distortion correction. Click image to view a larger version.

Here the 24mm CZ shines. It really uses all the 24 megapixels of the A900 or indeed the A77, because geometric correction rarely needs to be applied. It has a true 24mm focal length which does not need to be quietly changed to 25mm or 26mm by applying a lens profile. If a 35mm retrofocus AF lens was made for MF digital to this standard, even without the f/2 aperture, it would be hailed as a world-beater. The most that’s needed is a correction of 2% (+, removing barrel distortion) in Adobe Camera Raw and this restores something like a sea horizon near the top of a landscape format frame to a perfect straight line.

No correction is applied here to this full frame 24mm Alpha 900 image – a central horizon, and straight lines which are not parallel to the frame edge, make the 2% distortion (similar to many standard 50mm lenses) no issue at all.

For many subjects, depending on the distance and a ‘rigour’ of the shot (the sea horizon is the most demanding example) no correction at all will be needed. This applies to most interiors, and always to scenes like mountain views or forest landscapes where there is no perfectly flat horizon.

The Alpha 900 is so close to MF digital quality I should really forget the attractions of MF systems. Nearly everything I see from them which impresses me is down to using prime lenses of first quality like the Zeiss and Mamiya 80mm f/2.8 standards and working in a methodical way often using a tripod, minimum ISO, mirror-up operation. Applying the same parameters to Alpha full frame lifts the end result to match – and the CZ 24mm f/2 is a key to unlock that quality.

At f/14, the 24mm is not losing detail sharpness on the Alpha 900 as long as the correct raw processing parameters are applied. To secure this depth of field, f/14 was needed – a medium format camera would require f/27. Holding the camera, viewing and composing this shot were all aided by the ergonomics, weight and viewfinder quality of the Alpha 900. Click image for a full size version on pBase.

This is a dual-purpose or multi-purpose lens. Where the 16mm focal length of the NEX SEL 16mm f/2.8, the Alpha SAL 16-50mm f/2.8, the CZ 16-80mm or SAL 16-105mm all cover the same nominal angle not one of these has the same neutral geometry, even illumination and good corner to corner sharpness at wider apertures. Corrected by software, they don’t have the same true angle and the outer field can become noisy because of extra sensor-mapping gain applied to reduce vignetting.

The size and SEL comparison!

But I would like to show you something surprising. I am a great fan of the 16mm NEX f/2.8 pancake, which is one of the few such lenses made to have a positive (pincushion) simple distortion pattern and a cup not cap shaped field of focus. It is a revolutionary inverted telephoto design of great simplicity, with only 5 elements, enabling the lens to be 16mm focal length yet have a rear node position over 20mm from the sensor – thus avoiding all kinds of vignetting and colour shift problems.

People who don’t understand how to use a focus plane where the corners are focused FURTHER than the centre – the exact opposite of the CZ 24mm f/2 where the corners are focused CLOSER than the centre – do tests like landscapes wide open and wonder why the grass either side of their feet dissolves into blur. Actually all the little 16mm needs is modest stopping down, as would be applied by any professional using a Super Angulon for that matter, to f/8.

First of all, have a look at some lens sizes. I like this shot, as it shows just how big CZ had to make the 24mm to get what they did. It dwarfs the SEL 16mm for NEX and the classic Minolta 28mm f/2 RS:

I’d like you to see the exact comparison between Alpha 900 with 24mm CZ and NEX-5 with SEL 16mm.

This is the A900 and 24mm, entirely uncorrected and uncropped – the building on the right actually does not have a straight wall, don’t be fooled into thinking there’s a sudden burst of barrel distortion! Aperture f/8.

This the NEX with 16mm, corrected in ACR; I’ve tried to keep the camera positions very close but this was real-time shooting and with viewfinder versus screen composition, not so easy. You can see that the 16mm has slightly less true angle of view when corrected but don’t judge from the foreground flower tub, just check the horizontal angle. This is also at f/8.

You can click each image and view a full size JPEG. I have made both of them 24 megapixels, exporting from the NEX to the same size file as the Alpha 900. That may be unfair but you can judge. My opinion is that both the NEX 14 megapixel sensor and the SEL 16mm are underestimated by far too many owners; as far as ISO noise handling goes, the 16mm f/2.8 on NEX is actually as ‘fast’ as the 24mm f/2 on Alpha 900 but that comparison may change with future full frame bodies. As for depth of field, the f/8 shot on APS-C would need to be at f/13 on full frame to match, but in practice both are well covered.

Using the NEX 16mm in different conditions would produce a different result – wide open in a room interior, the corners would be likely to look very blurred. My scene above conforms to the cup-shape focus plane of the NEX lens, and works against the cap-shape focus plane of the CZ 24mm.

Remember as a general rule: barrel distortion = corners focused close than centre. Pincushion distortion = corners focused further away than centre. Moustache or wave form = a doughnut normally of closer focus between centre and corners, but when a full frame lens with this type of distortion (like the 16-35mm CZf/2.8 – or a more extreme example, Canon’s 24-105mm f/4 L) is used on APS-C, you get this doughnut at the corners and more or less have straight barrel distortion not waveform. No distortion at a given distance usually means a flat focus field, the quality which Carl Zeiss highlighted when naming the Planar lens.

Alternatives to the 24mm

The best way to get the 84° coverage with similar near-perfect rendering is to go for the mid-range of a high end zoom. As it happens, Sigma’s 8-16mm is better at 16mm than any of the above-mentioned APS-C options and you can also get a pretty good 16mm from their 10-20mm options and Tamron’s 10-24mm. Tokina’s 11-16mm f/2.8 is weakest at 16mm, best at 11mm. The older Sony 11-18mm is not wonderful at the longer end.

On full format, 24mm at the bottom end of the 24-70mm CZ is no match, it has more distortion and softer corners; 24mm in the middle of the 16-35mm CZ f/2.8’s range is better but with strong complex distortion, more even than the Konica Minolta 17-35mm f/2.8-4 D lens (which manages f/3.2 wide open at 24mm). You might think Sigma’s 12-24mm full frame zoom could be good at 24mm, and perhaps version II HSM when it finally become available for Alpha will prove to be. The original, which I still use mainly for its superb 12mm results, places its worst extreme of field flatness deviation at the image edge when set to 24mm.

I have used Canon’s 24mm f/1.4 USMII and this is faster, larger and more expensive than the Sony CZ lens in almost perfect proportion. Like the CZ f/2 it is a nearly perfect lens, with a hint more barrel distortion and slightly soft extreme corners on full frame wide open. The same goes for the Nikon 24mm f/1.4. I’ve also used Canon’s 24mm TSE tilt-shift and this lens betters the CZ for technical and architectural uses, as it should – so does their 17mm f/4 TSE, which has no match in any format. But such lenses can’t also be used for everyday autofocus image grabbing whether professional or family.

Last question, then. If such a perfect lens can be made at f/2, surely all the affordable 24mm f/2.8 designs could be just as good? We wish! Wouldn’t it be great if the classic Minolta 24mm f/2.8 AF which Sony never transferred to the new Alpha range proved to have the same optical excellence as the CZ? It does not. Nor do the Canon 24mm f/2.8, or the Nikon, or anything made by Pentax or Olympus, or even Leica.

The 24mm f/2 used at f/2.8 on the Alpha 55. Try this with a classic Minolta 24mm f/2.8 and even on APS-C you won’t get the same corner to corner even illumination. Here the focus is on the distance, not the tourists – they are also showing a surprising amount of movement at 1/40th. Click the image for a full size view.

This 24mm is the most recent AF 24mm prime lens to have been designed for full format. Zeiss have designed a slightly more complex manual focus 25mm f/2 Distagon for Cosina partnered manufacture, available for Canon and Nikon, since Sony showed the 24mm at photokina 2010. But Sony’s full-frame DSLR rivals, Canon and Nikon, have not gone for this sub-£1,000 RRP ‘moderately fast’ 24mm niche.

If there’s one competitor, it is Sigma’s excellent 24mm f/1.8 EX DG, which uses a larger 77mm front diameter glass unit to reduce vignetting to the absolute minimum. Distortion is higher, and the lens at present has no HSM version. This makes it less future-proof for Alpha system owners, and also less compatible with NEX and with video shooting in general.

Features of the 24mm

Because it’s a fixed focal length, the 24mm is a very plain lens – it has only two controls and one moving ring. There is an AF/MF switch, though unlike SAM lenses this lens can always be controlled from the body. With SAM type lenses (built in non-supersonic focus motor) it is essential to use only the lens switch, and never to use the body switch instead while leaving the lens set to AF. This is because any attempt to focus manually may damage the gears and motor unless the switch on the lens is specifically disengaged.

Manual focus or held focus can be set or toggled using the single on-lens button. New Alpha models like the 77 allow a wider range of functions to be assigned to the lens button, which is described in the menus as a Focus Hold button. Direct Manual Focus is also supported on bodies which offer DMF, meaning that once focus is confirmed and locked by your pressure on the shutter button, you can fine-tune focus by eye before firing.

The manual focus action is very smooth and well balanced, not too light and not too short in throw (which can be an issue with shorter focal lengths. The focus scale is minimal, behind a traditional Minolta-style clear window, with a depth of field indicator to the minimum f/22 aperture. Really, such markings mean little today as we expect so much from higher resolution sensors. It is time that Sony, and others, built parameter-governed DoF calculation into firmware.

Here, f/5.6 was judged to be fine for the degree of differential focus wanted – at ISO 400, by tungsten kitchen spotlights and window light mixed, on the Alpha 77 hand-held with SteadyShot and manual ‘peaking’ focus.

The CZ design is clearly corrected for medium distance work but retains its performance for close-ups. Unlike Sigma’s design which achieves 1:2.7 image scale, or the new manual Zeiss 25mm which focuses down to 18cm and 1:4, the Alpha lens focuses to 19cm (actually, I make it 18cm as the scale goes beyond the 19cm marking) and manages a 1:3.4 image. Don’t be fooled by distances! The front element of the CZ is already 12.2cm from the sensor plane, and the lens hood takes another 3cm or so. The actual clearance when shooting at close range is minimal. For comparison, the SEL 16mm f/2.8 for NEX will only focus down to 24cm, and the front of this lens is only 40mm from the sensor, leaving a clear 20cm between camera and subject. The Nikon and Canon f/1.4 designs are limited to 25cm and are, quite simply, nothing like as useful for close-ups as the CZ.

You might think that the 16-50mm f/2.8 or the 16-80mm CZ could match the combination of wide angle and close focus found on the 24mm – but not so. To get similar close-ups even at a 24mm setting is not possible – an extra 6 or 7cm in minimum focus distance, when you are talking an 83-84° angle of view, makes a big difference.

Moving in to minimum focus, the bottom wing of the lens hood was only 1cm away from the subject – under 19cm from bread roll to sensor, but only 6.8cm from bread roll to front element. At f/3.2, a hand-held 1/40th was needed (the closer you get, the less you can rely on SS to handle speeds like 1/15th). Focus peaking again enabled the manual focus point to be precisely judged. Great bokeh too.

With a non-rotating front thread, 72mm is one of the classic Minolta sizes. It is necessary to use slimline filters, as with the 20mm f/2.8. It’s interesting to compare the revived older lens with the newer one. The 20mm has only five mount contacts, being non-D specification where the 24mm has eight and reports much more accurate focus data. The 20mm has no lens button, uses screw drive focus, and has a close limit of 25cm at which it has a 1:7.7 image scale. There is also a considerable difference in the build and feel of the CZ; I have no doubt it contains some plastic, but it feels like a good solid piece of engineering and is stated by Sony to have a metal lens barrel. Not metal-skinned plastic, like NEX lenses.

As for coatings, Minolta’s legacy was a use of multiple layer (super achromatic) coatings to rebalance both the contrast and the colour transmission of the entire AF lens range (except designs made by third parties, like the 100-400mm APO). This advantage over other makes was never capitalised on, and made some Minolta designs seem lower in contrast than competitor’s equivalents. No-one ever complained about the colour though! Zeiss’s path from 1975 onwards was to use multicoatings a different way, maximising contrast and light transmission but permitting each lens design to have its own colour transmission quality and variation in contrast. Contax RTS lenses were always praised for their resistance to flare and their extreme macrocontrast.

Since the advent of digital, both overall contrast and colour transmission have become less critical – no need for packs of filters to balance lenses for repro purposes, no need to test Kodachrome with a clip-test to set this up. Just post process or shoot a WB card to taste. Also, Sony Alpha lenses are made in many places – the old Minolta unit, the new CZ-Sony collaboration, co-developed with Tamron and apparently also with Sigma, built by Shanghai Optical or some other owned and partnership facilities in China, made in Thailand but not apparently any more in Malaysia…

While distortion associated with viewpoint and perspective perception is always a companion to shorter focal lengths, over the field of the Alpha 77 (equal to a 35mm lens view or so, in full-frame terms) shapes and solids look natural. At f/4, and ISO 1250, I’ve chosen to downsize this 77 file to 3600 x 2400 pixels (click the image to open). This still allows you to see how clean the light sources in-shot are, with absence of colour fringes. Depending on conditions 1 pixel CA cancelling may be needed with the 24mm.

So, we have here a lens with a Zeiss design and a T* coating which is entirely unlike any Minolta legacy design and will surprise those used to the way ex-Minolta lenses perform. It is fairly immune to flare, not entirely so when confronted with bright sources just outside the image margin, but without the strings of coloured patches associated with 24mms and light sources in the shot. It focuses silently and at a speed which means you may not notice it.

The lens itself weighs 555g, and at 76mm length and 78mm diameter it’s smaller than the 16-50mm f/2.8 SSM which weighs 22g more. I’m not a big fan of lenses you can not clasp in one hand while also operating the lens release mount of a camera; optics this size and weight are about the safe limit. You can not compared the lens-juggling friendliness of the 28mm f/2, for example, with either the 24mm or 16-50mm and even the 16-80mm zoom is much easier to handle in the field. It’s best to remove or fit the hood before changing the lens, don’t leave it in storage position.

The hood reverses over the lens neatly. The whole item, when in this configuration, is a bit large to handle for safe and secure lens changing.

The finish is lustrous, with rubber rib grips that collect dust and dander readily. The supplied lens hood is surprisingly flexible plastic, with a slight spatter finish to the exterior and a kind of semi-flock paint on the inside. It is efficient, but a poor fit with a not very firm bayonet locking action. It’s easy to get the alignment wrong and it’s not as firm or solid as most other Sony hoods. The rear lens cap is still the frustrating one-orientation only design inherited from Minolta, which leaves even those with a quarter of a century of lenscap-fitting experience fumbling for the correct position.

There is of course a Zeiss front lens cap and you get a free blue badge on the lens itself!

Format, pixel count and cropping

For many years when using film I found wide-angle zooms were not essential, standard zooms were useful, and tele zooms were vital. Generally, with any wide-angle you can zoom with your feet or by doing little more than leaning forward or back a bit. Either that or you simply need the widest lens you can get. Whenever I fit my Sigma 8-16mm or 12-24mm on their respective formats it’s the 8mm or 12mm end which is needed. I only end up zooming in if for some reason I decide to leave the lens on, and move to a different situation without time to switch lenses.

With film, you could crop and enlarge. Small pixel count DSLRs made that difficult or impossible – when you are trying to make 6 megapixels do a full page magazine image, cropping is not an option. Zooming in to fill the frame every time became vital from 2000 to 2008 when the first full frame 24 megapixel models arrived.

I think that 24 megapixels has finally made cropping an alternative to zooming. You may need 9 or maybe 12 megapixels, or if you are shooting entirely for the web you may need no more than 2 megapixels. Fixed focal lengths of exceptional quality, sharp all over the frame in the plane of focus, start to be useful. It has never been a good option to crop wide-angle zoom shots asymmetrically, using just one corner. With a lens like the 24mm you can crop any composition out of the high resolution frame and it will not look so different from an on-axis shot with a narrow angle lens.

Lens resolution really does count, as I have found. For three years I used the Alpha 900 with a range of lenses, including the 24-85mm Minolta RS I keep for convenience. When working with medium format lenses on adaptors, I could see that zooms while ‘sharp enough’ usually came nowhere near realising the potential of the 900. Then, using the 24mm, I saw the same pixel-level sharpness pop out. After a month using the 24mm (kindly loaned by Paul Genge) my ordered Alpha 77 finally arrived. I had already seen how the 24mm got the maximum from 16 megapixel APS-C, and this was followed by discovering its power to do the same at 24 megapixel APS-C.

A standard Sony leather-look lens posing pouch is supplied.

How far can this go? If Sony’s 24 megapixel APS-C sensor formed the basis for a full-framer, it would be a 60 megapixel monster and match all but the most expensive medium format image sizes. I believe the 24mm CZ could go there if Sony chose to.

And that, in the end, is why I changed my mind about owning one. The hour or two of useful daylight and howling gales outside have not allowed me to make much use of it yet – but this is a lens for the long term. And for tomorrow’s Alphas as well as today’s.

– David Kilpatrick

Footnote: added February 2016 – I’m now selling this lens, as I don’t think Sony is likely to produce an A99 model II with functions that will restore what I want to have (notably, GPS – they are most likely to drop this). I’m looking at a move to native FE-mount lenses and probably the 25mm f/2 CZ Batis, even though it’s weaker for close-ups, vignetting and distortion.

Here is a recent example of a full aperture shot on the A7RII with LA-EA3 adaptor –

Sony Alpha 77 review – tomorrow today

It must be two years ago, at least, that an Australian sports photographer confided he had seen a Sony prototype which would blow away everything – an Alpha which could shoot at incredible frame rates (he mentioned 15fps) and follow focus. It may have been something unlike the Alpha 77, which follows focus 12fps with locked preset exposure, or in any appropriate exposure mode at 8fps. Or it may have been an early experiment. Whatever it was, the 77 is close to this rumoured prototype .

The Alpha 77 is a successor to the Alpha 700. Here is beside my old and well-worn 700.

The viewfinder

The Alpha 77 is a camera that points the way for future development, whether of DSLT (Sony ‘Translucent’ mirror technology) or entirely mirrorless SLR-mount bodies. It’s the OLED viewfinder with 2.4 megapixels of self-illuminating RGB which changes the game. It makes the transition from optical to electronic viewfinders likely for everything except a limited choice of professional optical viewfinder DSLRs. EVFs will not be unique to Sony and it will continue to develop in resolution, colour fidelity and refresh rate.

But this type of viewfinder has a specific limitation. Unlike earlier EVFs, the Sony OLED has a threshold below which it simply turns the pixel illumination off entirely. To save battery power, this is set to a relatively high black level and clips the three-quarter tones straight to d-Max. All EVFs are slightly unsatisfactory to the human eye because there’s no real shadow detail to see if you switch your glance away from well-illuminated parts of the view. The new OLED has dead black shadows and it doesn’t matter how much your brighten it, the cutoff is based on exposure level in the image.

If you own an A55/33/35 then the new finder is so far removed it might as well be an entirely different era, not just a generation. The area surrounding the huge visual image is dead black, not milky and luminescent like the A55. The shooting information is set neatly outside the image, not overlaid at the edge overlapping picture detail. The colours are bright and the information text, though smaller to the eye, is crisper and far more legible.

One comment (added after this was first published) – the A55/33/35 may be considered better in very low light. In good light, the OLED/24 megapixel combination is excellent. In low light, it shows noise until focus and exposure are confirmed by shutter pressure, at which point the view clarifies and the noise reduces. In near darkness, its shows very strong noise, mainly red, which largely obscures any visible detail. In conditions where exposure was 15 seconds at f/2.8, ISO 100, using the A55 and A77 side by side the A55 was better. It looked grey and flat, but surprisingly detailed and bright into shadows which were too dark to penetrate by eye. The A77 gave a contrasty screen or finder image with little useful detail, nothing in the shadows, and very strong noise. Clearly the CMOS sensors are responsible for whatever level of detail is visible, and the A55 sensor seems to me have a couple of stops more lift to tones on the threshold of its sensitivity. The EVF types differ in the A55 showing more shadow detail with lower contrast, and lower colour saturation so the noise does not look bad. Bottom line, the A55 is half way to night vision; the A77 is like turning on an old UHF analogue colour TV, no signal, just noise.

There will be users who complain that having vital information just above and below the image area means they don’t see it because of their specs. But the A77 has the best eyepoint and visibility for this info of any model to date. All I can say is that if you ensure you are using the EVF correctly, you will always be in touch with what the camera is doing.

This shows a ‘busy’ finder display – in fact, everything overlaid on the photo here can be turned off leaving just the active focus point (a single one, if you use centre spot focus) and the neat information bars above and below the image area. The rest of the field is dead black too, not milky grey like the Alpha 55, and the contrast is high. This image is dropped in and has no connection to the information displayed!

The result is a very graphic view of your composition. Despite the impressive size of the apparent viewfinder image, it is not as large visually as the Alpha 900. The A700 for example is 95% field at 0.90X of a 1.5X sensor coverage, a viewfinder ‘size index’ of 0.57X (0.95X0.9/1.5). The A900 is 100% of full frame at 0.74X, index 0.74X. The A77 is 100% of 1.5X format at 1.09X, index 0.726X. It’s therefore closer to the A900 than the A700.

The A77 eyepiece glass is much closer to the eye than the A55 or previous SLR models, and the upper positioned eye-sensor turns the finder on and off with precision saving power but causing no delays.

But has Sony got the figures right? The A55 claims to be 1.1X and 100% view. Each camera has a different eyepoint – 27mm for the A77 compared to a mere 19mm for the A55 – but this should not be allowed to influence the stated apparent magnification. Placing one camera to each eye, the A77 has an almost 20% larger apparent viewfinder field (linear) than the A55 and about 30% larger than the A700. It is just a fraction smaller than the A900 as the calculations indicate.

It looks to me as if the A55/35 manual misrepresents the EVF size in that camera, it very clearly is not larger than the A77. Perhaps they include the dead space not used for the image but for data. In the A77, the data display is tight and efficient and it can show everything you ever want including the ISO in use when you select Auto ISO.

All the other figures agree. It also makes the A77 finder view an almost perfect match for Nikon’s full frame DSLRs, which have a slightly smaller apparent screen size than the A900.

The less-shadow-detailed aspect of EVF works for composition much the same way early Leica viewfinders did. The scene is simplified, and this tends to concentrate the eye on impact and simple compositions. Using the A900 alongside the A77, I was struck by the way the A900 puts you in touch with texture, colour, subtle light, and fine details. The A77 reduces the world to simpler tones and connects you to shapes, composition and impact.

As for flicker, shearing when panning, clarity of focused detail – well, these are all limitations, but much reduced in this new finder. What is not so limiting as the A55 or the earlier NEX models is the speed of startup from sleep. The A77 finder really does go to sleep (the A55 is very good at failing to switch off) and wakes up so rapidly, as your eye approaches the finder, you don’t miss the shot. Combined with the mere 1/20th second shutter lag, this SLT gets back to the responsiveness of classic non-AF SLRs.

You can cycle through these displays or choose to skip one or more when pressing the Display button (there’s a menu item to configure exactly what information each press shows you, with separate settings for the EVF and the rear screen). These diagrams from the manual do not represent the finder very well. Our colour version, using a finder overlay file supplied by Sony, gives an accurate view of what looking through an A77 finder is like.

Here I am, I’ve written stacks about the viewfinder and not even begun to describe it fully. Tell you what – go and find one, try it. I can’t photograph it adequately (I have tried) and it would be exhaustive to go through the deep menu settings of the A77 which let you decide for example how many different information-display setups you scroll through when you press the Display button. Don’t want EVER to see the horizon level guage? Then set the camera to skip it. Hate the graphic display of f-number and shutter speed scales against each other? Deselect that too.


From the Alpha 77 Karma Sutra – left, position for portrait photography; right, position for those who like to video themselves and still look as if they are in eye contact with the viewer.

And then there’s the rear screen. It has a double hinge and rotate action, of which more later. It is a top grade screen, 3 inches and nearly a megapixel. If you plug an HDMI monitor in to the camera, that will take over providing a better solution for studio or video previewing. Even a 20 inch or larger HDTV set looks sharp when fed from the Alpha 77.

The menus of the A77 require the attention to website presentation detail best reserved for the galley-slaves* of dPreview. Trust me, if there’s anything which can be set on any other Alpha, it’s probably settable on the 77 or not there at all. I’m sad to see that I can only control my auto ISO range minimum between 100 and 12800, and my maximum between… 100 and 12800! Of course this is great. I can have auto ISO over any possible sensible range but not ISO 50.

*Queried on dPreview forums – I’m an ex-newspaperman. Galleys are proofs of type (or the metal itself) and if you’ve ever worked on the TV listings or the Sports pages, you’ll know what it means to be a galley-slave in editorial terms – form before function, and function before fun! Well, I can have fun. I do not have to reproduce every single screen and menu because there’s no big boss saying that’s how it is done! It’s great work that dPreview undertake. But as someone else has already done it, I prefer to spend my energy doing the stuff they haven’t done. Then you can read both.

ISO 50? What use is that, with less dynamic range than 100?* It’s a great deal of use. In my studio with powerful flash – which this camera can cope with perfectly, having a setting to over-ride exposure simulation in the EVF/screen and do auto gain for modelling lights – it gives me an extra stop instead of being forced to work at a setting like f/13 or get overexposure. Outdoors in bright light it combines with the 1/8,000th top shutter speed to remove the need for a 4X ND filter to get superwide apertures, but due to reduced dynamic range, it’ll still clip whatever highlight would have been clipped with an overexposed ISO 100 shot. In the studio I can control my contrast to use it well.

*This is Sony’s own statement in the manual, page 139: “The recordable range of the brightness of a subject (dynamic range) is slightly narrower for areas where ISO is less than 100.”. DxO Labs tests actually say that colour depth, tone depth, dynamic range and signal to noise ratio are all at their best if you set ISO 50, but they also show that ISO 50 is really ISO 63 overexposed a bit and ISO 100 is really ISO 80 underexposed a bit. Where that leaves the intermediate 64 and 80 settings on the camera we can only guess.

As well as all the info in the finder and on the back LCD, you get this top plate LCD which means you can close the back one. Even when the camera is asleep (power save mode) the sensor must still be receiving and handling the image, as the exposure display will change when you move the camera around. I’d guess this is a low power mode which also enables the system to continue to monitor exposure between frames during sequence shooting.

At first I did not fall in love with the Alpha 77 – when my purchased firmware 1.02 version body locked up on me in front of one Canon and one Nikon user I was just a little angry with it. It could have done this somewhere quiet, on its own, without spectators. Even now with firmware 1.03 I am not sure it won’t repeat the cataleptic fit, but it shows no signs of doing so. See my section on the Movie mode later on, though…

The SLT mirror and ghosting

I’ve tested the Alpha 77 in the most demanding lighting conditions. Sony says the SLT mirror (a very thin film of plastic stretched on a frame) has new mirror and antireflection coatings. They have also paid extra attention to the design of the AF module and the interior of the darkchamber, to avoid reflections.

Though some colour fringes on the bright water definitely hit this 16-80mm CZ shot (along with a colour bokeh issue making those in the foreground magenta and those to the rear greenish) there was no hint of any ghosts or flare in dozens of shots taken in conditions like this.

This light is extreme, and the patches of sun and reflection are placed exactly where the Alpha 55 tended to produce flare. The Alpha 77 shows no sign of it, and has not in any of our photographs so far.

As far as loss of sharpness goes, I do not believe there is any more significant sharpness loss from the SLT pellicle than there is from, for example, the rear filter permanently fitted into a 300mm f/2.8 Apo G tele. Both are between the lens and the sensor and both are plane clear optical elements. If anything the glass thickness and distinct double air to glass surfaces of a rear mounted filter make it far more likely to degrade an image than the SLT. Almost any filter you fit in front of the lens is going to have a greater effect (unless you spend a stack of money, a really bad effect on 24 megapixels – we’ve replaced our older Minolta, pre-digital Hoya and other filters with the latest Sigma EX DG after testing them).

This is just a routine test I ran at all ISO settings to check colour and tonal response. Not noise. I already knew before doing this that the noise thing was a non-issue for the simple reason that this sensor beats anything else out there; I’ll start pixel-peeping for noise when another maker comes along and shows they can do this pixel density better. The colour is also very consistent indeed across the ISO range and the feathers in the mask retain the expected detail up to ISO 3200.

This picture was taken using a setting I quickly discovered is just perfect for news, PR, presentations and images needed quickly from events – the Small JPEG in camera, with DRO enabled, at ISO 3200. Sampling down to one quarter of the file size (still large enough for an A4 print) creates a dead sharp, low noise image. Photographers do some good, occasionally – this is a cheque for £3000 being presented to DJ Dave Lee Travis for the PACE Centre charity, by the Master Photographers Association. Their annual dinner happened just a week after I got my Alpha 77. I was confident enough to risk taking all my press and PR shots on the new camera.

And this is a 100% pixel level view of that shot without any post processing.

Input and feedback

Then again, having to set up the camera and realising the full extent of the customisation possible through the Menu and Fn buttons, I felt depressed. This was almost like handling a Canon 7D – one of those cameras where, if someone passes it to you, you can never be sure if it has been configured only for photographing flocks of ibis flying behind bare poplar trees. Would its 19 AF points and 11 cross type sensors do the 7D trick of locking on to a sweet wrapper someone dropped on the lawn instead of the wedding group a yard further away?

Well, yes. The A77 can do that sort of wrong stuff if left on wide area focus – but it doesn’t light up the wrong AF points in the finder to fool you, and it does not require programming of AF preferences to avoid proximity or response speed errors (both Canon and Nikon pro models can disappoint if used ‘out of the box’, and need their defaults changing depending on your typical photo situations). Instead, if provides very accurate feedback about which sensors are being used. And it has Face Detection which really works, because this is a live view camera 100%. I have never liked Face Detection much until the SLT EVF generation arrived. Even then, not much.

Face detection kept the focus on photographer Paul Cooper (right) accepting the president’s ribbon of the MPA from Henk van Kooten (left) despite Henk’s focus-target jacket moving into the foreground of the AF zones. Taken by stage lights at ISO 3200

I had to take a few shots where a person receiving a presentation was facing the camera, and the presenter often stepped into shot with back of head to the lens. With Face Detection on, the A77 never once switched focus to the nearer person, and always stayed locked on to the subject facing me. This is a situation where the above-mentioned makes, if allowed to use wide area or multi point AF, tend to shift focus to the foreground intruder because as conventional DSLRs they don’t have Face Detection in optical viewfinder mode.

The auto exposure of the A77 seems to be more closely linked to active AF points than any previous model. It may have 1,200 metering zones on the CMOS sensor but it will bias strongly towards correct exposure at the point or points of focus, especially if the central point coincides with a very bright are. I do not mean it is literally spot metering. I mean that, for example, in my office with medium lighting and a very bright computer screen if the camera is aimed at the screen the exposure in matrix mode, with centre spot focus, becomes correct for the screen and the rest dark. When the screen is moved away from the centre zone, exposure increases by two stops even though the overall image contains about the same brightness.

What I’m seeing may not be the same sort of meter-linking-to-AF that is found in the Canon EOS 400D, as an example. This will give you over or under exposure if the focus point hits a dark or light area. The A77 biases towards avoiding overexposure. A dark subject at the central focus point does not seem to brighten the image the same way a very light subject, like a screen, darkens it.

This is not like separate metering cell TTL, the classic Minolta honeycomb. It isn’t even like a camera with centre weighted or spot optically fed meter cells. Every point on the sensor is a spot meter even though you can only ‘spot meter’ from the centre. Every point is equally sensitive down to EV–2 (ISO 100 with f/1.4 lens) and up to EV17. It is four times more sensitive as a matrix/centre-weighted meter than the Alpha 900 and a staggering 16 times more sensitive than A900 spot metering mode. It also has +5 to -5 EV exposure compensation compared to the A700’s+/-3EV (A900 – 3EV, expanded to 5EV by later firmware, but the 700 was never improved – see comments, originally I referred to A900 as 3EV either way, as that’s in the manual). And whatever things the SLT mirror does, it seems to feed the new AF module plenty of light – it’s able to focus in conditions half as bright as the A900. Added comment: the A77 metering is four times as sensitive as the new Canon D1 X, so although that camera has amazing sensitivity up to ISO 204,800 the A77 will actually meter exposure in lower light.

So, after a couple of weeks, I began to realise that the A77 was giving an even lower failure-rate than the A55. I had learned which settings to prefer – three zone focus for example is far better than old-style wide area and almost makes single centre spot focus redundant. I was finding that exposure is generous but never highlight-clipped, because it’s read from the actual imaging sensor; you can trust the simulation given by the EVF, too, and adust the +/- over-ride with confidence.

The camera stopped being complex and started to suit my declining mental powers. Life is a curve. You start just learning to set shutter, aperture and focus. In your prime, you want to set twenty different things for every shot and switch from P to A to S to M with C or S or A and -2 sharpness and ten stop HDR then portrait look for the next one. Ultimately as the brain cells sneak off for a nap you find good old shutter speed, aperture and focus do you just fine.

If you need reminding in big print, let the rear LCD see the light of day, and you get this big clear information pane on demand.

Now some cameras have fooled me, there are no dials and they just hid this stuff from me so I’d end up with bad things like the optically soft set of landscape pix I shot at 1/2000th and f/5.6 (wide open) on the NEX-5 last month. But the Alpha 77 with its top plate LCD info display, its ‘come to life’ burst of finder shooting information when you take first shutter pressure and confirm focus – well, it is constantly reminding me what I am doing. I know other cameras and other Alphas have finder displays, even the NEX was probably telling my longsighted eyes what it was up to, but the Alpha 77 presents working information better than any camera I’ve used. It is simply a very clear and well designed display both in-camera and on the rear screen.

If you enable image review, the SLT cameras slow down. I fold the rear screen to face the camera back. I have turned off image review. I shoot with confidence just as I would once have done on film and sometimes I do not check a single image until I’ve copied the card contents to my computer.

Button pushing

The Alpha 77 has loads of buttons despite Sony’s one-time insistence that they planned to have fewer mechanical components in future. There are nine push buttons, one rocker button and one control wheel on the back of the camera alone; five push buttons, one collar switch, one shutter and control wheel on the RH top. Then there’s the stray stop-down button, the lens change release and the AF-mode switch living round the mount.

Rear screen folded away and protected – that’s how I use the camera all the time. Plenty of buttons to push – and you can have fun swapping their functions round to confuse your friends!

Several of these buttons can have their functions modified so they no longer do what it says on the silkscreened white or blue print (white for shooting mode, blue for playback). If you are particularly odd you can even swap round functions and confuse people who borrow your camera (shades of Canon!). If you are relatively normal you can leave this well alone. You may customise the stop-down preview button to show the final picture effect instead (stop-down plus picture style and shutter speed result) and through the menus you can change the behaviour of lens-resident Focus Hold. The instruction manual omits to mention these are on the lens, and not on all lenses – some owners have spent ages looking for the Focus Hold button which does not exist on the camera.

What’s most odd about the A77 is that three of the main dedicated-function buttons are completely interchangeable. ISO (next to the shutter) AEL and AF/MF (under your thumb) can all be changed to do anything from the following long list of functions:

AEL Hold*
AEL Toggle*
Spot Meter with AEL Hold*
Spot Meter with AEL Toggle*
AF/MF Control Hold*
AF/MF Control Toggle*
Object Tracking
AF Lock*
Aperture Preview (stop down)*
Shot Result Preview (final picture simulation)*
Smart Teleconverter*
Focus Magnifier*
Exposure Compensation*
Drive Mode*
Flash Mode
AF Area
Face Detection
Smile Shutter
Metering Mode
Flash Compensation
White Balance*
Creative Style
Picture Effect
Image Size

*The entries I’ve marked with an asterisk already have their own dedicated buttons for which these are normal function choices (AEL button, for example, covers all the first four but can only do one function, preset in menus).

There is one button you may want to modify if you own lenses with a Focus Hold button. The AF/MF button, by default, performs this function with most lenses (it switches to MF when pressed, the same effect as holding focus). So it’s almost a spare button, given that there’s also an AF/MF switch on SSM/SAM lenses and a body AF/MF/S/A/C switch too. Since it sits right next to the AEL button and closest to the rear control wheel, it a natural choice for any function you might want to use in a hurry.

You can not change the function of the Finder/LCD manual switch button, the Drive Mode, the White Balance or the Exposure Override. You can switch the Preview and Smart Teleconverter buttons between two functions each only. The Fn button accesses all the parameters you can’t reach directly through any custom button (like setting the Auto ISO range) and most that you can (like Face Detection). Its full function list is:

Memory Recall (only present when mode dial set to MR)
Scene Selection (only present when mode dial set to SCN)
Movie (only present when mode dial set to Movie)
Drive Mode
Flash Mode
AF Area
Object Tracking
Face Detection
Smile Shutter
ISO and ISO Auto setup
Metering Mode
Flash Compensation
White Balance
Creative Style
Picture Effect (only active for JPEG-only shooting)

The Display button can not be customised and only serves to cycle through Display setups – but you can customise those, and thus what the button does for you (above, Menu to set which finder display states you wish to cycle through, ticked). The Help (?) button can also not be customised, which if you don’t want potted hints and tips makes it redundant in shooting mode, though it serves as the Trash button when reviewing images. The Playback also can’t be customised, nor can the Menu button (which can return to Last Used or Top by setting a preference), nor the top LCD illuminator button. This one is interesting because it toggles – the panel light remains on until the camera goes to sleep, or it’s pressed again to turn off.

In use, I found there was one button missing which would make a huge difference to this camera. The 12fps ultra high speed shooting mode is only accessible through the mode dial. Because of the way the camera handles bursts of frames and buffering, it would be useful to be able to shoot normally in any mode (single frame, or other continuous speed) and switch to 12fps by holding down or toggling a button (preferably holding down, say, the AF/MF lock reassigned for this purpose).

Here’s a neat touch – as long as you have a lens with no MF/AF switch on it, you can set M focus on the body selector, and pressing the AF/MF button will do autofocus for you, letting it go will lock the focus back to manual. That’s another reason to like my 16-80mm CZ on the A77 – ideal for studio products or architecture, or indeed for landscape. It does not work with the 18-55mm SAM. There is in fact a bit of an overall mess with SAM, SSM, and standard lens focus types including the assignable DMF (direct manual focus) to the AF-A mode. There are some lenses where you are warned never to use body MF setting always only to use the lens switch. What you will find is that some menu items are greyed out, and some buttons don’t work, if a mismatched combination is set.

Lens compatibility

The new 16-50mm f/2.8 is enabled for in-camera lens corrections

This leads to the general question of lenses and the A77 generation. It seems there’s some additional information chipped into some but not all lenses which enables the Lens Correction function (Vignetting, Chromatic Aberration and Distortion) for in-camera JPEGs. Raw files are unaffected, and I don’t know if this information is used to enable better panoramas but that would be a practical fringe benefit. But since the 18-55mm SAM, 55-200mm mk2 SAM, and 18-250mm (an old design relatively) are in the release firmware along with the 16-50mm SSM maybe there’s no info in the lenses themselves, and future firmware will add more.

The oddest incompatibility is the manual’s statement that Front Curtain Shutter should not be set on ON for ‘Konica Minolta’ lenses (added note – see Comments at the end of this article, some discussion of this). First of all, the description of Front Curtain Shutter is misleading. When this is set to ON, it means NO front curtain shutter – electronic gating instead. When it is off, you are using the physical Front Curtain, the blades uncover the sensor to start the exposure. The manual does make it clear that by Front Curtain Shutter Sony means No Front Curtain Shutter.

But what is a Konica Minolta lens? There are very few, nearly all were made by Tamron. Konica Minolta never even got round to rebranding the Apo G lenses, they just changed the box and the lens cap and left Minolta as the name on the lens. I have a 28-75mm Konica Minolta, a 17-35mm Konica Minolta and at one point I had an 18-200mm, and of course, the kit 18-70mm and various horrible full frame plastic lenses like the 28-100mm. I think there are 75-300mms in KM guise.

Everything else prior to Sony was Minolta because KM simply never made any, or if they did, it amounted to no more than a box label change. To confuse things, some of these Konica Minolta like the 18-70mm, 18-200mm and 75-300mm became Sony lenses without a single substantial change. And many Minolta lenses became Sony lenses while retaining a heritage right back to pre-D days or the origins of the AF system itself. The 28mm f/2.8 is the most obvious example, the 50mm f/1.4 another.

Why would an electronic front curtain produce overexposure or inconsistent exposure only with Konica Minolta lenses? Given the very fast response time (1/20th of a second) and high speed of the system (1/250th flash sync, 1/8,000th shutter) the only thing I can think of would be the speed of aperture closing action. It is also something which would be invisible in revised lenses; maybe the Sony 18-70mm kit lens actually has aperture blades which close 50 milliseconds sooner than the KM equivalent.

This would also mean earlier Minolta lenses, not just KM, might produce overexposure (aperture still not fully closed when exposure commences) or uneven exposure (aperture continues to close down during part of the electronic progressive gate-opening). The same would apply to many third party lenses.

I’m pretty sure this is why the warning is made, and that singling out Konica Minolta lenses is an error. Any one individual lens may have sluggish diaphragm, indeed a common cause of overexposure in all A-mount lenses is incomplete stop-down. So the advice should be don’t use ‘Front Curtain Shutter’ set to ON with anything except Sony A-mount lenses – or test your independent lens before use.

A warning about not setting Micro AF adjustment with third party lenses is given, as usual. This is because the makers borrow lens identity codes. The Sigma 12-24mm f/4.5-5.6 Mk1 has the same code as the Minolta/KM 28-80mm kit lens. This issue will also affect the behaviour of the Auto Lens Correction register, which in our body does not have any effect on the CZ 16-80mm for example but does correct the new 16-50mm. I would guess ‘correctable’ lenses have extra information, older and third party lens won’t. So don’t bank on this function fixing JPEGs from your ‘heritage’ of early Sony glass!

Sigma has issued a list of lenses which are known to have AF problems with the Alpha 77/65, and will upgrade them free of charge. You can read the list here: It includes the 18-250mm HSM OS, which we have. Ours does not seem to have any issues at all on our A77 with firmware 1.03. I’ve also tested the 8-16mm and 100-300mm OS, 70mm macro and 70-200mm f/2.8 HSM Macro MkII. These are not in the warning list and all seem to work well, even though they do not work reliably with the Alpha 55.

The A77 has Fast/Slow AF options and I’ve used Fast. The accuracy of the AF is much better than any previous Alpha model.

Autofocus and exposure

The 19-sensor, 11-cross AF module is not most densely populated – the A900 has 10 extra hidden ‘tracking assist’ line sensors bringing it up the same total (they are there, they just don’t have screen markings) and its central double cross sensor is technically the best type around. But having eleven cross sensors does more good in practice.

AF has a new mode, Zone. This creates three groups of sensors left centre and right which act like mini wide zones. Wide Area focus can, of course, locate widely spaced details which are concurrently in focus. Zone identifies the zone with the most focused points, then works within this area. It can use information about the change in confirmed focus within one zone to help track the subject into the next zone. It also gets the exposure right more of the time.

Because the sensor is feeding image data to an analysing computer, Face Recognition and Object Tracking can be linked to the AF. There must be some theoretical speed penalty – something must be slowed down by microseconds if you enable these functions. I can’t detect it. The only slowdown is the time you take to press the central controller button to register an subject for tracking. Smile Shutter is also possible. My subjects normally scowl so I don’t use this.

The focus point – here, centre focus spot aimed first at the family and then held using the shutter release to recompose the scene – helps determine exposure. The 1200-zone metering has correctly placed the baby’s white clothes in the value range 250 to 254 RGB.

The most reliable and accurate focusing method remains central single spot focus, or local selectable single spot. Now that all of these are cross-type, there is no compulsion to stick with the centre and recompose, but it’s a habit hard to lose. I have already observed that the metering in spot mode is 16 times more sensitive than spot metering in the 900, and I would guess that when the system biases matrix metering values to the centre focus point, there’s a related gain if not that much. There is a proper near-IR AF illuminator in the camera body – Minolta tradition lives on! Without flash, it really is possble to focus in the dark and the EVF does a very grainy but usable job as a night vision viewer too. The illuminator also enables focusing on plain surfaces as it projects a pattern.

The SLT mirror of the 77 diverts 30% of light to the AF module, but this is actually more than the old semisilvered patch and double mirror system used to let through in SLRs. The AF sensor itself may not be any more sensitive, it’s simply getting a much better image feed. This 30%, by the way, means the light reaching the sensor is reduced not by the 1/3rd to 1/2 a stop sometimes mentioned, but by .6 of a stop or nearly 2/3rds of a stop. 50% would be one stop. (See comments at the end of the article for some more precise figures on the mirror split provided by Dr Daniel Oi).

My experience so far with the camera indicates that exposure is very reliable over the entire EV range. I simply leave it on the matrix multi-zone setting. The EVF warns me if it’s going to bias too much to the focus point.

High speed shooting

This brings me to the aspect of high speed sequences. No cameras in this class has ever achieved 12fps, let alone at 24 megapixels. In theory you get 13 raw or 11 RAW+JPEG frames before the camera slows down, and it looks as if the buffer must hold about 320MB. The camera does not have a dual processor like the Alpha 900, and it does not benefit from the robust performance of fast CF cards.

Autofocus is provided in Hi drive mode, along with AE (8fps, accessed via the Drive function button, in any shooting mode). Using this shooting speed you have full control.

In Speed Priority AE mode (the 12fps setting on the mode dial) the focus is locked before frame 1 if you have the camera set to Single (S) AF, but in exchange for this, you can set both the ISO and the aperture. You can also set these if you use Manual focus.

If you set the focus mode to C (Continuous) then both AE and AF continue during shooting. You can set the ISO, but not the aperture, so the ISO is your only way control the shutter speed. Added note: the Canon 1D X has now bettered this record high speed shooting by providing 14fps. This mode in the Canon locks both AE and Focus with the first frame, locks the mirror up, and you must use Live View on the rear screen to compose the shot. For focus tracking with viewfinder, the 1D X is limited to 12fps.

12fps is very impressive. It makes a huge difference in action work. I often test sequence shooting on the local races, and I quickly found that even tracking a horse (necessary to keep it in the frame at all for more than one shot at 3fps) certain frame rates just produced two stages in its stride, repeated. The horse was galloping at 2.5 clops per second and I was shooting at 5fps. With 7fps it gets better, 8fps or 10fps still better and with 12 fps you reach the point where four different positions of the legs are recorded.

Also, it becomes possible to aim the camera at a fixed spot like a hurdle, and fire, capturing several positions of the horse before it leaves the field of view. It is not as necessary to pan with the subject every time to get more than one shot.

Animation of three hand-held frames, cropped from a Sigma 70-200mm shot taken at 70mm, showing how 12fps captures very fast action in relatively small steps. At 5fps, the second frame would have the horse leaving the right-hand edge.

In practice, you certainly get your 13 raw or 11 R+J shots at 12fps or the slower AF-capable Hi 8fps setting – or indeed at the slower 3fps rate. But you don’t get anything like the same continuous shooting capability as past models even if you knock the JPEG size right down. You’ll get around 18 Normal Small JPEGs (6 megapixels, lowest quality) at 10 to 12fps before the rate slows down to an erratic 3fps with occasional half to one second pauses. For raw files, after your 13-ish burst is up, you may get between 0.5 and 1 frame per second with occasional one to two second pauses. With an average SD card (20MB/s write) you will wait 15-20 seconds after the last shot before being able to shoot fast bursts again.

Here’s another sequence, this time as stills without the annoying animation you can’t turn off 🙂

And here, below, is a 100% crop from the original ISO 800 raw file processed using Adobe Camera Raw 6.5 (Sharpness 50, Radius 0.5, Detail 0, Masking 0; Luminance NR 25, Luminance Detail 50, same from Chroma NR)

Sharpness? The 12fps C-AF setting forced the Sigma 70-200mm to be at f/3.5, two thirds of a stop down from full aperture, but also gave a shutter speed of 1/6400th. I could have perhaps picked another detail with slightly more punch, and looking at all the shots, my prefocused point was actually about 1 metre behind the horse (if the AF refocused during this sequence, I can’t see any evidence in the images). The 8fps or Single-Shot AF 12fps settings allow control of aperture, and I know that f/5.6 would have cleaned up. Just remember you are looking at a section of an image something between 6ft and 8ft wide, it’s very easy to view a tiny clip like this as if it was just another digital image.

I tried one technique, shoot 2, 3, 4, 6 frames with brief pauses – as if catching different moments of an event, in bursts. Even though I spread the 10fps bursts over a ten second interval, by the time I had totalled 20 raw frames I was down to the single shot per second or worse situation.

To follow up, I shot a burst then allowed the buffer to write for about 8 seconds before firing again. I got seven frames at 12fps, which fits in with the card in use taking about 15 seconds to finish writing from a 13-frame burst.

For one of the horse racing tests, I shot one burst of frames and as the camera slowed down, two horses fell and two jockeys were injured, one requiring a stretcher. I was unable to get ANY pictures of the incident as it happened, and by the time the Alpha 77 was able to shoot again, the ambulance crew was on the track. Each race gave me just two chances to shoot a burst so I’m afraid that testing every single setting combination on the camera was not possible.

Panorama speed mystery

There’s something I don’t understand about the raw, file and buffer handling of the Alpha 77. Shooting panoramas – which have to start with exactly the same frame by frame 24 megapixel data readout – I counted 42 frames apparently firing at something close to the 12fps maximum, then creating a panoramic JPEG, and the finished 6.7MB JPEG file was written to card and the buffer cleared before I had time to see if the light was still on.

What exactly is happening here? How could the processor and the buffer somehow handle the throughput of the wide panorama with at least double the number of continuous burst frames I could get with even the smallest JPEG – and then do all the computing to assemble the panorama and write it to card, leaving me ready to shoot immediately?

I ask this because when shooting panoramas with the Alpha 55, my 15MB/s SanDisk Ultra II SD card failed – it was not fast enough, could not handle the data and became corrupted. Clearly panorama shooting is data intensive one way or another. But in the Alpha 77 it appears to be allocated buffer and processing power which is denied to more useful motordrive sequence shooting.

Movie shooting does not enjoy the same fast buffer clearing. Shoot any higher quality movie beyond a mere blip on the button, and you can’t fire a still frame for some seconds. With some HD-movie systems, you actually shot a still frame during the movie and lose nothing except a couple of movie frames; with others, you can end the movie by pressing the shutter and capturing a still. In yet other makes, you can shoot a still but lose two seconds during the movie. All these solutions are valuable when still shots could be important. The A77 movie function does not permit any such choice and may block all shooting by occupying the buffer to card writing process for many seconds.

And, in reverse, you can not initiate a movie while the card write light is on. I tried this with one of my horse race test subjects. I decided I would shoot the front runners going over a hurdle, get my 12fps burst, then do a few seconds of movie of the stragglers who reach the spot a few seconds later. Although I could have shot further (faltering) still frames, movie shooting was blocked out with a warning message telling me ‘Writing to Memory Card – Unable to Operate’. When raw shooting was set, this lasted many seconds, but interestingly with the small JPEG option only two or three seconds were blocked out.

After my tests of the high speed shooting functions, I conclude that to cover some sports events well you would need a pair of Alpha 77s, or the 77 and some other camera – and you would need to keep a close eye on the card writing light. As a result of the performance with my SD cards giving write speeds around 20-30MB/s I ordered a SanDisk 45MB/s Extreme Pro, and plan to get a 95MB/s card when they are available.

Added after receiving the SanDisk Extreme Pro 16GB 45MB/s card: burst shooting is not extended, using raw it varies from capturing 10 to 12 raw frames at maximum rate, sometimes with a single frame jumped (two groups of 5-6 frames) which never happens with my ordinary Transcend 20MB/s card. Buffer to card writing is reduced to around 10 seconds from around 20, but at least with this extra card speed, there is no way to extend the approximate 1 second unbroken raw shooting burst.

The Alpha 65 and 77 are the only DSLR/T cameras so far made which use the USH-1 SDXC specifiction to allow writing data at this kind of rate. They are ahead of the card game. But they need to have this capability. Anyone expecting to make full use of the high speed drive functions and best video quality of the 77 with cards like the faithful wallet full of Transcend 16GB Class 10s we have been using will be disappointed.

The dedicated Movie mode

On the mode dial of the A77 there is a movie position. This does not mean it has changed, like some Canon models, to be incapable of shooting unplanned movie clips – you can do this at the press of a button, like other Sony models. What the Movie icon means is that you gain access to manual control of ISO, shutter and aperture. As a penalty, you lose AF.

Why? That is easy to answer. The AF sensors in the 77 are rated for f/5.6 aperture use. Default movie apertures range between f/3.5 (typically set on fast lenses) and f/6.3 (well, with an f/6.3 zoom lens there is little choice). AF works best in this range. If you really want to shoot AF movies at f/2.8 on a 200mm lens, try by all means. I have tried at f/3.5 and for every clip which has a smooth focus transition, there will be another where the SSM, SAM, HSM or plain old screw drive makes a sudden shift.

So if you want to work at f/1.4, leaving focus to AF would be a disaster. Apart from making constant shifts, there’s a big chance of hunting. I worked with the 24mm f/2 SSM Carl Zeiss for a while, and this lens does not find focus easily in low light with any camera. On the A77 for video it was auto-set to between f/2.8 and f/4 in low light. Video AF happens at the actual working aperture, not wide open like still AF. Being stopped down a little improved the 24mm’s accuracy.

And if you want to work at a smaller aperture than f/6.3 – say f/16 for a deep focus effect with a superwide lens – then AF simply would not work at all. The Manual Focus restriction placed by using the Manual Movie mode is necessary despite the howls of protest it’s produced from those who don’t understand the technology involved. A side benefit of setting Movie mode on the dial is that your view through the finder is cropped correctly to the HD area before you start filming, which makes composition easier – see below.

In the Movie position on the Mode dial you might believe you can use the self-timer. The manual ticks that box. We had to check it, but although you can set the Drive state, including self-timer, doing so has no effect on Movie shooting. Nor can you shoot stills with the Mode dial in this position. Various other manual details, such as indicating flash can be usedin this mode, are also incorrect or there’s a small firmware glitch with the camera. I would expect the shutter release to be operational with the mode dial set to Movie, so that stills could be captured. It is disabled and you can’t take still shots at all if the dial is in this position. Be warned!

Not only that, but after I had pressed the Self Timer 10 second setting when in Movie mode, the LCD top display showed a single frame symbol plus the 10 from the self-timer mode, and the shutter release was disabled even after returning the mode dial to Program or other settings. The Drive mode had to be reset to get it out of this tangle.

As for the instant Movie button, it’s not in the best place – a long thumb reach for on and off actions to start and end clips and the placing tends to make me tilt or move the camera needlessly. When adjusting the viewfinder dioptre to switch from working with spectacles to working without, it’s so close to the dioptre control I push it occasionally by mistake. I would like the Movie mode to switch operation to the main shutter release OR the movie button.

It’s worth noting that picture effects (see later pages) can be applied to movies, as can all other image adjustments, and will be seen in the finder as you shoot.

But what you won’t see until the moment you press the movie button is the working HD movie area. It is a surprisingly fierce crop from the full sensor, not just a top and bottom slice to HD 16:9 format. This is required for the Steady Shot digital, pixel-shift based movie stabilization which does not move the sensor like regular SS. It reduces the 1.5X area to something like a 1.8X area, not as much ‘zoom in’ as using the 1.4X smart converter but enough to cut heads and feet off subjects you have pre-composed using the full finder. There are indicator marks on the display, that’s true, but you will find them hard to see in many conditions even if you remember to use them.

Shot with the 24mm f/2 lens – no way to move back or zoom out for the movie, but plenty of space you would think for the 16:9 HD crop.

This is what the HD movie viewfinder field switches to when the Movie button is pressed (slightly re-composed horizontally but not cropped at all). The movie stabilization need the surrounding pixels. It highly effective, both through the live viewfinder and when viewing the results, and totally silent unlike sensor-shift SS. It also does not warm the sensor assembly up as much.

A caveat to movie shooters. Select the highest quality AVCHD 2.0 formats (the camera will warn you) and you may not be able to play them, burn them to disc other than Blu-Ray, or import them into HD movie editing with programs like Apple iMovie. No doubt this will change with updates, but right now apart from Sony’s PC-only Picture Motion Browser (they have now had half a decade to port it to Mac) there are few programs which can recognise the 1080/50 or 60p 28Mbps format and even fewer which edit it. You will need to buy software to do so. On my system, Toast Titanium 11 includes Roxio HD viewer and because it is a Blu-Ray compatible disc authoring package, this was able to handle the .MTS files in all formats, and convert them to formats editable using iMovie. Roxio offers similar PC utilities.

Sony PMB is also very useful for GPS data handling, map viewing and correcting GPS data. While Adobe Bridge with the GPS Panel (download from Adobe) installed allows viewing and editing of co-ordinates, it still has no link to Google Maps, Microsoft Virtual Earth, or any other useful location display.


There is no provision for audio notes or audio only recording, which is a bit of a waste of built-in functions as this would be possible and can be useful. Audio notations attached to images are popular with travellers and journalists.

The stereo sound is not much different to the NEX or Alpha 55 despite being capable of 48kHz (better than CD) sampling. The microphone under its unusual top grille seems less prone to wind noise than most DSLRs; the wind noise reduction feature, a bass cut filter, is additionally effective. Body handling noise is present, so is lens focus noise.

The external mic socket provides 5v phantom power for popular condensor mics, and is stereo too. No provision for fixed gain is made – auto level or gain is always on. The camera needs an option to disable auto gain so that a preamp or mic with dB cut choices can be used to control a fixed sound level; better would be three levels, like Nikon; even better still a proper sound level monitor display and full control in fine steps like Canon. Without at least one of these options the Alpha 77 can not considered for semi-pro or professional live sound video work.

GPS Data

While I think that the GPS on the A77 has so far proved faster in locking and more accurate in positioning than the A55, it’s hard to quantify as I have not used both together for long enough in difficult locations. To improve your GPS, download this file:

Create a folder on your SD card inside the PRIVATE/SONY directory and call it GPS. Copy this file into it, put the card in your camera, switch on. Go to GPS in the main menus and make sure it’s turned on, then look at the GPS ASSIST entry and you will find the assist data starts on the day you download and lasts a month. Visit the assist.dat download once a month or more often to keep your GPS working with the fastest and most accurate lockon and co-ordinates. Do not worry if you format or change cards, once in the camera the assist.dat file is copied to internal GPS memory. Your original card does not have to be present.

Sony state 15 seconds or more to acquire or change position, and variable accuracy due to GPS being a US military provision which can be subject to deliberate degrading. In this model, when no GPS signal can be detected on power up, the camera simply turns off GPS embedding (on the A55, it uses the last co-ordinate). But if you are out and about, the last position may be shown on some pictures. I don’t think I move that far in 15 seconds


An example of GPS map location from Media Pro and Google Maps – click on the image for a full size screen shot.

Adobe gripe – it’s long overdue for Bridge to have a GPS map function when Lightroom does. As it happens I use Media Pro for all my digital asset management. When this was Microsoft Expression Media 2 it has its own Virtual Earth window, but now it’s been taken over by Phase One, that has been replaced by auto-opening a web browser Google Earth window. I don’t really need to see maps at raw file stage, but it could help with filenames. I like my filenames to be a ‘catchline’ format – an alphanumeric string which contains a key word about the subject. That could be simple like venice2011-15.jpg or a bit more precise like guideccasangiorgiovenice2011-15.jpg.

Mouthful? Not ISO compliant? Not ancient PC friendly? Sure. But very useful indeed many years later when searching for stuff. And access to GPS map location helps me decide filenames, then later on input metadata for caption, keywords, description, and much more. With 15,000 finished images stored on my system everything which helps me identify them is valuable.

Image formats and styling

The Alpha 77 has a stack of great functions and features I will never use. That’s because they are not available if you shoot either RAW or RAW+JPEG. They include multi-shot modes (combining tonal range for HDR, or reducing noise for low light and high ISO) and in-camera post processing effects.

There are some of these JPEG-only modes I feel comfortable with. Panoramas, as an example, don’t provide a raw file and you can’t bracket exposures. You have to trust the camera despite the huge range of tones and light a wide panorama can cover. If you choose your start position well (including the brightest highlight area of importance) exposure is very reliable, and at low ISO with Fine quality, the JPEG is of a professional standard.

Standard JPEG (click images for 1000 pixel wide version)

Three-shot HDR using 3EV spread

Three-shot HDR also works well, especially at low ISO settings of 50 or 100 and in Extra Fine JPEG. The 24 megapixel file gives plenty of scope for reducing to a smaller final result. There is also a special HDR Painting mode, which processes the file with a masking effect to create what is currently a popular ‘look’. Unlike the standard range of HDR settings, this is far from being a straight image and the lack of a raw file or normal JPEG to back it up means it’s only for fun.

Over the top with HDR Painting style, High strength – it actually works best on dull, wet days with grungy subjects!

More ‘only for fun’ stuff includes soft focus, selective colour against mono (called ‘Partial Color’), toy camera, miniature effect, two monochrome looks including one which uses three exposures, ‘pop color’, posterization, retro photo, soft high key. All of these are irreversible real time post processing. You have no normal backup when shooting.

The post-processing method offered by Nikon and others, where you can apply similar effects to raw or JPEG files already stored on your memory card and create a new version, is preferable. For standard HDR shots on Sony’s current models, a normal JPEG is saved along with the three-shot HDR and that’s good. For any of the Picture Effect post-processes, even multishot, no standard result is saved – all you get is the processed file, after a wait of 10 seconds or considerably less.

Selective yellow on monochrome – but that’s the only shot I have. No raw, no standard JPEG…

You do get a pretty accurate preview of the result in the EVF. If the stop-down preview button is set to ‘Shot Preview’ mode instead of ‘Aperture Preview’, you also get a simulation of the effect of your shutter speed – so flowing water brief time exposures (up to 30 seconds) can be previewed to see exactly which shutter speed suits the water movement best. Along with exposure simulation, there should be no reason why your shot ever goes wrong.

The Auto ISO Multi-Shot mode, Multi-Frame Noise Reduction, captures six frames and creates one JPEG. The pixel alignment seems very accurate and shots at settings like ISO 3200 show an improvement in detail which would be hard to obtain even by good raw processing. It’s not so much the noise that is reduced, it’s the overall quality of the image which improves. Using the high 25,600 ISO setting which can only be accessed in this mode shows that it’s slightly inferior to a straight 16,000 ISO shot despite the six-frame synthesis.

For all these multi-shot modes, the 12fps function of the Alpha 77, SSS, and the quiet, mirrorless shutter action combine well. They are all usable without much effort or worry, hand-held. The one ‘tonal range’ adjustment which does operate in RAW+JPEG mode, though only the JPEG is changed, is the DRO or DRO+ setting which uses a single shot.

High speed shooting also benefits exposure, DRO and white balance bracketing. Early information and the use manual state that you can define the number of exposures for bracketing, and the range covered, with the Alpha 65. In fact it is limited to three shots. The Alpha 77 gives you a choice of 3 shots at +/-3EV, the same at 2EV, then 3 or 5 shots at 0.7, 0.5 or 0.3 EV intervals. It is missing the obvious 1 EV step choices and that will baffle many, especially HDR raw users who would like 5 shots at 1 EV intervals.


The Alpha 77 has a proper, threaded, high grade studio flash sync terminal as well as the usual Minolta i-type hot shoe. Like the Alpha 700 and 900 (and unlike the consumer level cameras, including the Alpha 580) it can have the HVL-F58AM or 43AM wireless control capable flash mounted on the camera to control group/channel wireless strobes with power ratio. It can also use HSS (burst flash with shutter speeds up to 1/8,000th and corresponding power attenuation). It can not officially use the HVL-F20AM as a wireless controller, but owners have found it works – with a slightly longer than normal delay in flash firing, according to Gary Friedman, who has compared it with the pop-up flash wireless control.

The flash sync Prontor-Compur coaxial connector (PC flash socket) is sealed behind a cover shared with the Remote Release socket. This cover was so tightly sealed it threateed to break a fingernail opening it the first time, a small screwdriver was needed.

The most important change for professional and enthusiast owners is the long-overdue addition of a menu item which prevents the EVF or LCD live view from showing actual exposure when Manual aperture and shutter are set.  This item is under Live View Display, and is called ‘Setting Effect’ – off or on. While this nomenclature is not exactly transparent, it describes the function well as all picture styles and creative effects normally shown in the finder are also bypassed. The important thing for studio flash users is that you can set 1/125 at f/16 with modelling lights, and see a normal finder view not a black hole. You still must remember to set white balance to Flash or Daylight, otherwise the camera will set it from the modelling light K.

It is best to use one of the three Memory registers (accessible through Menu screen after turning the Mode dial to MR, Memory Recall or Register) to store a manual exposure, fixed WB, low ISO, Setting Effect OFF preset for studio work. Then you can return to any other setting and get your accurate exposure and ‘look’ preview back again.

The internal pop-up flash (GN12) has the usual range of first, second curtain, fill-in, off, auto options; TTL Pre-Flash, ADI, and also manual power control down to 1/16th which can be useful for triggering slave flashes if you don’t have a cable or a wireless trigger (and Minolta shoe adaptor). Because the body is weatherproofed, the flash shoe cover is a softer plastic type which seals tightly. Don’t lose this shoe cover or swap it for one of your others.

And the rest

By the time I’ve written this single review article, it will be one-third the length of the complete camera guide books we used to do for Hove twenty years ago. The Alpha 77 has so much more to discuss.

You will be concerned about high ISO quality, diffraction, resolution, having good enough lenses. I would question whether the new 16-50mm, used wide open, is a ‘good enough’ lens – let alone the 18-55mm SAM also being offered as a kit lens. Just don’t worry. Whatever your existing lenses are capable of doing, the 24 megapixel sensor will give you more of it. Let’s say your favourite lens is really only good up to 12 megapixels. It will be just as good if you use the 12 megapixel Medium size JPEG option on the Alpha 77, and if you do that, the 1.4X Smart Teleconverter function will also deliver a 12 megapixel drawn from the centre of the field only – so most likely just as good.

Rather too distant heron, shot using the 2X Smart Converter for JPEGs (this is actually a clip from a raw file processed in aCR to match). ISO 3200, 70-300mm SSM G lens. Click image for 1000 pixel version

Click image for 1000 pixel 100% size clip from ACR processed version (my density choice)

Click image for Capture One Pro 100% clip from raw (ditto)

Click image for in-camera processed (JPEG Fine, Low level of NR) 100% clip (camera’s density)

I am now shooting with auto ISO set to go from 100 to 3200 instead of 1600, I have started using Medium and Small JPEGs with DRO+ to ensure exposure correction for events type shots, I’ve tried all the lenses I have and the only thought is that I need to stick around f/8 to f/11 for safety. Balancing extra depth of field with a hint of diffraction loss. I’m using the manual focus ‘peaking’ function to check the accuracy of my AF (this shows a coloured line on correctly focused details, when the AF/MF button is pressed in). I am not so worried about low light, high ISO as I first thought. It’s actually as good as the 16 megapixel sensor when needed, and when it’s not, the extra resolution repays careful low ISO technique.

The new tilt, hinge, flip, swivel and cartwheel rear screen is just great for the few times I need to use it. The EVF may consume more power (470 images versus 530 per battery official rating) than the big rear screen but I no longer need to switch between the two for menu and function operations. Because of the new design, all positions found on other cameras from hanging-under to almost flat on top (R-1 style) are possible except facing forwards and positioned to the side. There are firmware or orientation sensor errors, as the imager can appear upside down in more than twisted position. The hinge design makes a vertical grip possible and also allows a wider range of tripods or quickmount plates.

The A77 has all the focus and AE hold and lock, slow sync, focus point shift, exposure over-ride and other key functions I need. It claims to be weatherproof, and having nearly broken that thumnbnail off opening the flash sync cover  I do believe the seal is tight. The card slot door is not so reassuring and I see no trace of any proper sealing, not even a labyrinth design.

I am baffled by Sony’s indecision about ON/OFF switch design, the camera labelling is the reverse of the Alpha 55/33/35 or 580/560 etc, though the action is the same. The direction is the reverse of the NEX-5. But there is one consistency, to turn any camera on the movement is always from left to right – whether Alpha 100, 700, 900 and whether the switch is rotary or a slider. Maybe this is the rule they stick to.

You can not configure the directionality of the two control wheels, as you can with Nikon, and for some reason I have always tried to open or close the aperture by taking the wrong directione. That is because the wheel directions go against the old Minolta protocol that turning the aperture ring to the right opens up, turn to the left stops down.

Like the Alpha 700, the Alpha 77 has magnesium alloy body shell combined with other metal and plastic components. It has the proper strap-lug fixed into the mag alloy casting, like the 700 and 900. This lug and triangle-ring design, as opposed to the slot-type strap fixing of the lesser camera bodies, is always a clue that the structure is based on a good solid metal skeleton. The overall design and balance of the Alpha 77 are as good as any Alpha I’ve used. There are hints of the 700 and also some memories of the Dynax 7xi present in the sculpturing of the body. To those who say it looks a bit like a Canon, yes, it’s true that Canon design has caught up with 1990s Minolta style in the last couple of years…


From the initial press meeting with Sony, where cameras were prototypes and the images were not allowed to be shown, I decided that if I could work for a year with the Alpha 55 and have no problems then the Alpha 77 was a safe investment. The viewfinder is a pleasure to use, though EVFs differ from optical screens in one important respect, that the eye can not compensate for small errors in the dioptre setting. With an OVF like the Alpha 900, I can set the dioptre midway between what’s needed for my sight with and without glasses, and get along fine with either. That can not be done with the EVF and it demands a precise dioptre setting for each. I have found it more comfortable to use without specs, so they spend too much time perched on my head, hanging from my collar or stuffed into a pocket.

It will be another year before I know just how wise the decision to go with EVF SLT models has been. And maybe another ten thousand words.

– David KiIpatrick

Please read the comments for some notes on corrections, which I will continue to make.






Tamron 18-270mm – a hero, but no VC…

After using Sigma’s 18-250mm optically stabilised zoom on Alpha bodies for a year and more, the first thing which strikes about the Tamron 18-270mm for Sony mount is the lack of the VC (Vibration Control) stabiliser found on the same lens made for Canon or Nikon.

Tamron’s lenses come without a case, but with a custom fit petal lens hood, front and rear caps. Design is clean with a Nikon-like sleeve grip and Canon-ish gold ring. The PiezoDrive focusing is similar to Nikon AF-S/Silent Wave or Canon USM, or Sony SSM, but not identical and on Sony models it can contrast-detect autofocus reliably. Sigma’s HSM hunts.

With Sigma facing patent claims by Nikon – that parts of their OS technology infringe on Nikon VR – Tamron VC is a mature system not so far challenged in the same way. It is also a very solid kind of stabilisation, free from swimming effects, and in this respect closely matches Sigma’s approach. Both are generally more comfortable than Canon’s IS which often seems to attach the image by a bungee cord to the viewfinder screen.

For video work, in-lens stabilisation is generally better than in-body as long as there is a good stable view which does not tend to float free when you pan slowly. For long lens work in general – over 200mm – in-lens stabilisation provides a view which is easier to aim and compose. We had already checked the lens out on Canon, with its smaller sensor area missing off the extreme corners (and therefore doing the lens favoured compared to other brands) but to compare with Sigma’s lens, needed to look at it in Sony mount.

The lack of VC in the Sony version of this lens is regrettable. There is no corresponding reduction in retail price.

Against this the Tamron has a longer zoom range, and it’s much smaller and lighter than the Sigma, taking regular 62mm filters not the unusual and large 72mm size. It also offers Piezo Drive focusing, which almost as quiet as SSM yet as fast as SAM. Small adjustments make a sort of faint clicking sound and focus travel is unusually fast, but a range of freehand refocusing tests using the Tamron showed that it is just as reliable in locking on to difficult targets as any other lens. Usually fast focusing means lots of overshooting or hunting, but not on the Alpha 580 used for this test.

Although the size and weight difference between this and the Sigma doesn’t look all that extreme when photographed in the studio, the heft in your hand (volume) is much less for the Tamron. It does not really seem any bigger than the Tamron/Sony 18-250mm design or the earlier 18-200mm.

The design of the lens follows these, with the LOCK switch for holding the lens at 18mm when walking round positioned for the right hand to operate, a long way from the AF/M switch (which should be used instead of the body switch for changing to manual focus).

This is a better design than the Sigma which clusters the AF/M, OS on/off and Lock controls together on the left hand side. Even after a year of use, both Shirley and I regularly turn the lens OS off, or turn AF off, instead of operating the Lock. All three controls move in the same way and are intended for the same fingers. Tamron’s location of Lock on the right hand side is ergonomically better.

However, both lenses fail to do the one simple thing which would improve such zooms – make the Lock control operate at ALL focal lengths not just 18mm. The Tamron is firm as we test it, so was the Sigma when new, but our Sigma can not now be used to pan with a plane or bird flying overhead unless one hand is used to keep the zoom from collapsing to 18mm immediately the lens is aimed upwards. To do the studio shot, the Sigma had to be taped to keep the zoom extended. Otherwise, it can’t even sit on a table set to 250mm.

You can’t see the sticky tape stopping the year-old Sigma zoom from deflating itself to 18mm every time when placed in the studio for this shot. The new Tamron is still young and firm. But we need locks which work at ALL settings.

It can not be difficult to devise a zoom lock which works at intermediate settings and it would transform the functionality of lenses like this.

Apart from ergonomics, there is no significant difference in build quality. Sigma feels more solid but heavier in action, Sigma’s exterior finish is difficult to clean and collects marks and dust easily. Tamron feels more plastic in build but has a high quality metal bayonet just like its rival.


Just studying the lens coatings shows why the Tamron can be more contrasty and less prone to flare in some light – especially if you fit a cheap filter to the Sigma and get contrast-eroding reflection for that front element.

The Tamron lens has visibly higher detail contrast than the Sigma, and in the centre of its field produces a very sharp image. The edge of the image lets it down, however, rather badly. The detail is soft at longer focal lengths unless stopped well down (ƒ/11 or so) and red-green chromatic fringes are serious enough to spoil JPEGs. They are not even very well corrected by using Adobe Lens Profile to process from raw (there is no Sony profile but Nikon, using similar sensors, can be selected).

This is a Sony Alpha 55 ISO 400shot, deliberately off centred in composition, with the Tamron set to f/9 (a good compromise between diffraction and stop-down sharpness) and 270mm.

The focus point is away from the centre of the image, and the lens displays good contrast and sharpness, but even here there is a slightly dirty look to the detail and chromatic fringes hit the white edge. This is NOT by the way anything to do with the Alpha 55 translucent mirror!

Here’s the edge of the shot at 270mm and f/9. I feel it would be almost unfair to Tamron to publish some of the worst results I got wide open. This is a defocused distance, of course, but this is also real-life imaging. This is why we did not switch from the bulky, heavy Sigma to the neater, lighter travel-friendly Tamron.

At full aperture and 270mm the performance is markedly inferior to the Sigma at 250mm wide open. The lens has better multicoating but poor field flatness, which creates the softening to the edges and corners.

The Tamron at 18mm has pretty strong barrel distortion which, when corrected using a lens profile in Adobe Camera Raw or Lightroom, lost some of the wide-angle coverage.

At wide to medium focal lengths, the difference is less marked and the Tamron is more equal the Sigma or other ‘best’ superzooms. But this is a lens bought for its extra reach at 270mm; given the performance, it’s not all that much use unless your subject is centred and surrounded by out of focus background.

Tamron at 270mm.

Sigma 250mm view – at near-infinity, the Tamron is longer the Sigma but not quite as much as 270mm would indicate.

Another issue is that of focal length, above and below examples. If the Sigma is a true 250mm (which it is not, all such zooms are shorter than their stated figures) then the Tamron is actually 265mm not 270mm at infinity.

This is unscientific, but the baby owl did not move and both lenses were placed in turn against the wire of its enclosure ensuring the same shooting distance to within a centimetre or so (with lens hoods removed). Tamron at 270mm.

By this distance, the Sigma at 250mm really is no different in focal length than the Tamron at 270mm, due to internal focusing differences. And it focuses closer than the Tamron for a larger maximum subject scale.

Although the close focus is good, at 49cm and 1:3.8 scale it’s not as good as the Sigma with 45cm and 1:3.4 scale – the true focal length at closer distances also seems to be shorter than the Sigma, though this is hard to evaluate.

As for bokeh, that’s not why you buy these lenses:

How many stumps? Wiry would be a fair bokeh description at medium apertures and longer focal lengths (270mm again, above, at f/9).

The Tamron PZD focus does work on the LA-EA1 Alpha adaptor for NEX; it’s not fast, but can lock autofocus perfectly even in difficult light. The Sigma can not do this at all and is not AF-compatible with the NEX adaptor. But… manually focused, the Sigma has OS. Vital!

Most telling is the weight difference when mounted on a light body like the A55. The Tamron is a far better match even if not as ‘good’ a lens – 970g for A55+Tamron, 1400g for A580+Sigma. Check prices, and work out your priorities.

– David & Shirley Kilpatrick

1 2 3 4 5 8