Review Samyang-Rokinon 14mm f2-8
Review: Samyang/Rokinon 14mm f/2.8 ED AS IF UMC
Looking at the 14mm f/2.8 wide angle lens
Some our more striking commercial shots have come with the use of extreme wide angle views, and Keith regularly uses the Canon 14mm lens on a full frame 1Ds Mk3.
Unfortunately for those wanting to experiment with such views, the Canon lens comes in at around £1850
However, if you’re on a budget, the Samyang 14/2.8 lens is available for a bit over £300. What’s the differences and how well does the Samyang work?
Keith has been lent one for a few weeks to test out, and looks to see how well it works.
Update 2017: Keith has reviewed the all new Samyang XP 14mm f/2.4. If you are looking for a manual focus 14mm lens, do have a read of this review as well, since it’s still available at about 1/3 the price of the new XP
The lens appears elsewhere around the world under the Samyang, Bower, Rokinon, Walimex, Opteka, Falcon and other names, and is available from numerous suppliers, and in several different mounts (Nikon, Sony, 4/3, Pentax – only Canon EF tested here)
Back in the days I used 35mm film, I’d have considered 24mm as a fairly wide lens. With my 2004 Canon 1Ds (11MP full frame) I moved to the EF16-35 2.8L, and these days it’s the EF14mm and TS-E17mm (tilt/shift) that offer my widest rectilinear views.
I mention rectilinear (‘straight lines’) since I also have the Canon 8-15 zoom fisheye lens [review] which most definitely does not keep straight lines straight (unless they pass through the centre of the field).
As an architectural photographer, straight lines are important, and buildings, with their regular patterns, are great at showing lens distortions in this area.
I’d heard that the Samyang 14mm was a good lens but suffered from such serious distortions that it might not be good enough for ‘serious’ work. Was this so, could it be fixed, and how difficult was this ‘everything manual’ lens going to be to use?
I’m testing the lens on a ‘full frame’ 35mm/FX sensor. On a Canon APS-C body it gives an ‘equivalent’ of ~22mm and 21mm on a Nikon DX body (it is available for other mounts too)
This is not a ‘lens review’ full of charts and measurements, but more a feel for what the lens was like to use, and the kinds of results you can expect.
The 14mm f/2.8 ED AS IF UMC Lens
You may find this lens appearing under a number of different guises (Rokinon/Bower) around the world.
It’s made in Korea, and ours was kindly loaned by the UK (trade) distributor Intro 2020
The lens is supplied with a lens cap, and rear cap. The lens hood is fixed, and an integral part of the lens body.
There is a soft bag supplied for storing the lens.
I should note that I don’t do detailed measurement based reviews of lenses – I know there are people who love that sort of stuff, but MTF charts will have to be found elsewhere…
|Focal length:||14 mm|
|Angle of view:||115.7°|
|Minimum focusing distance:||0.28 m|
|Number of aperture blades:||6|
|Optical construction:||14 lens elements, 10 groups, 1 ASP, 1 Hybrid ASP, 2 ED, 3 HR|
|Available mounts:||Canon EF, Four Thirds, Nikon F, Pentax, Samsung NX, Sony A|
|Dimensions (for Canon EF):||96.1 x 87 mm|
|Weight (for Canon EF):||552g|
The field of view varies with mount (sensor size)
The Canon mount version has no electronic communication with the camera (the Nikon version does have electrical contacts and a chip that allows Focus Confirm and Auto Modes)
That means everything is manual. You need to manually focus and manually set the aperture.
There is no confirmation from the camera AF system, and you need to manually set your shutter speed and ISO – welcome to ‘real photography’ as some might like to call it ;-)
I have heard it said that you can glue a Canon AF confirm chip to the mount, but this wasn’t my lens to modify…
It also means that there is no EXIF data relating to the lens, so if such numbers matter to you, write down the settings.
The red alignment dot would be easier to use if it were also on the main lens body.
As you can see in the shot below, there is a distance scale on the lens, and combined with quite a long focus throw (how far the focus ring rotates) manual focus is smooth and easy to set.
Given the importance of that distance scale I’ll come back to issues of manual focus in more detail later.
Two aspects I’m aware that not everyone will agree with ;-)
- There is no hyperfocal distance scale – good!
From my experience they are usually pretty meaningless decoration, which fool the unwary into believing they offer some deep insight into how lenses focus ;-)
[I’ve written more about why, generally, I avoid ‘hyperfocal’ techniques]
- Filters – pah, who needs them (no, not even the option for ‘gel’ filters at the back to reduce image quality)
The aperture blades are not round, so do give a slight shape to OOF highlights at anything other than f/2.8
Move your mouse over the image to see the six sided aperture at f/5.6
If you were wondering why the first view wasn’t a full circle, it’s because we’re looking in at a very oblique angle, and there is internal vignetting – this is perfectly normal for any wide angle lens. It’s known as optical vignetting, and as you can see from the f/5.6 image, usually goes if you stop down a bit.
There is an example later, showing the considerable difference that going from f/2.8 to f/8 makes in an actual shot.
The lens is of plastic construction, but it’s a very solid plastic, and matched to a metal lens mount. If you drop it, don’t expect to find a convenient repair service.
Wide open at f/2.8 the lens focuses moderately easily using the viewfinder of my Canon 1Ds Mk3, but that’s a very good viewfinder. I’ve also quite a bit of experience using manual focus lenses (tilt/shift for example) so it’s not an issue.
If this is your first manual focus lens, then expect to put in quite a bit of practice. A combination of prefocusing and working at f/8 means that many of the example shots I’ll show, needed little real thought about focus when I wanted to press the shutter.
You get a lot of depth of field at f/8 on a 14mm lens – learn to use it.
Note that I’m testing the lens here with manual setting of exposure – you can use aperture priority or some automatic setting of ISO or combination, but do experiment to see how your particular camera works with a lens like this. I use manual exposure settings a lot, so this aspect didn’t trouble me (YMMV)
The distance settings…
For a quick check of the accuracy of the distance settings, I set up the camera in our kitchen, photographing a pack of thin batteries on a door (they were at hand, had no depth and fine writing on them).
The distance I set, was to the film plane (that little ø mark just below the hot shoe).
At only half a metre, using 10x liveview at f/2.8, there is plenty of detail and the focus is easy to adjust with live view.
The view from the camera shows a rather bowed view of my kitchen door.
At one metre you get a feel for just how wide the coverage is for a 14mm lens. However, the distortion is now quite noticeable, as is the fact that it’s not a nice simple barrel distortion, which is easy to ‘fix’.
At 2 metres, the fine detail isn’t showing clearly, even at 10x magnification on the rear view. It’s still fairly easy to focus at f/2.8 though. At f/8 it does take quite a bit of rocking the focus back and forth to decide the sharpest point.
The view at two metres – all the walls and doors really are straight and true, even if the house dates from the 1880’s.
I’d originally thought of providing a basic table, showing the difference between indicated focus (on the lens) and the actual distance.
However, it appears that the effective focus point also varies slightly with aperture.
So, I’d suggest you try this one out for yourself, if you are going to rely on the distance scale. In particular, use something like the moon to find out where the infinity focus point is for your lens at several different apertures. At f/8 on my setup, infinity focus was just slightly in (shorter) from where marked.
Now, if this sounds bad, it’s probably because those of you in the past who used distance scales, never actually checked their accuracy ;-) Accuracy in this area comes at a price (think cine lenses with their extended focus throw)
It so happens that my preferred way of using such wide angles lenses often relies on setting the lens to infinity, rather than the rather vague ‘hyperfocal’ technique (I’ll show an example in a bit).
As well as some test shots, I decided to just go for a walk round near where I live, along the canal in Leicester, and try and give a feel for the sort of view that 14mm gives (shots at f/8, infinity focus, unless noted).
I like using the EF14mm for ‘street’ photography every so often. The powerful lines and perspective, along with being able to get in really close are fun (and not part of my day to day photo work).
Now you’ve seen the really obvious distortion that the lens shows in my kitchen, look at the subsequent examples and see how often it shows up in more general shots.
The churchyard at St. Mary de Castro late one winter afternoon (more pics) (colour not boosted by nearly as much as you might think in this one).
The large horse chestnut tree in the churchyard.
Blocks of flats along the canal.
Afternoon colours on a day like this are intense. A slight chill in the air adds to the calmness.
There are always plenty of swans about, who assume that if you are holding anything, it’s possibly some bread…
This shot manually set to 2 metres focus on the lens, with the camera just held out and pointed in the direction of the swan.
At f/8 you’ve a lot of depth of field, so even if the distance scale isn’t perfect, a rough guess will work.
Let’s get back to that serious warping of straight lines…
The shot below, of an absolutely superb local hardware store, shows the wavy ‘moustache’ distortion of lines across the frame.
Fortunately, software exists that, with appropriate profiles’ can fix this to some extent. I wish that DxO optics Pro supported this lens with my 1Ds mk3, since its corrections are really good. It doesn’t, so what about ACR in Photoshop/Lightroom?
It turns out that you can make profiles to fix distortions (incl. vignetting and chromatic aberration).
It’s much easier though, to find an existing profile and modify it – I took one for the Canon 5D Mk2 and edited it to work with my 1Ds3.
Note that the corrected version shows how I was not quite square on to the bridge structure (all the outdoor shots are hand held). Some care is needed here, since it’s the distorted version of the image that you see in the viewfinder.
Another bridge shot… is this corrected?
Another view of the red girders. Is this corrected?
And finally, a look along the canal. How about this one? Any geometry correction?
Actually, none of the last three had any corrections applied.
The distortions are quite strong, but just not very noticeable in many images, but when they are, it’s painfully obvious ;-)
The correction settings file I used with Photoshop CS5 came from Thomas Berndt at www.photo-worX.de
Here it is in the lens correction filter, in Photoshop, at the bottom of the list.
My own modified version is here if anyone wants to experiment [Samsung 14mm Lens correction profile for Canon 1Ds Mk3 – zip file].
There are some versions available for newer versions of Photoshop/Lightroom from Adobe – unfortunately, when I last checked, not for many cameras.
The shot below at f/2.8 was focused on the text by moving backwards/forwards after setting the lens to minimum focus – it’s hand held. At f/2.8 the round aperture gives a nice out of focus blur.
The cropped shot below was focused on one of the window highlights on the salad.
The swan is very close – manually set at 1 metre and f/8 – camera held at arms length, no viewfinder used. Did I mention not putting your hand too near hungry swans?
In the shot below, I’ve pushed up the colour, to show the flare, shooting directly into the sun.
Still at f/8, the six blade aperture shows in the lights and the flare from the just out of shot street lamp.
It’s worth noting that at 14mm it only takes a slight pointing upwards to get very noticeable convergence of verticals (see my Why I use shift lenses article for more).
One more example with the sun in the corner of the frame shows the effect, but also shows the minimal chromatic aberration you get with this lens.
Detail at 100%.
Move your mouse over the image to see the difference.
It’s reduced further at f/11, but you get significant vignetting with any wide lens.
The shot below, is at f/8, with the lens set at infinity (actually slightly less on this lens at f/8). You can see a bit of vignetting, but much less noticeable, even compared to f/5.6
The bottom of the frame is obviously not sharply focused. However, at f/8, I can work out the resolution (when focused at infinity) as 14/8mm or 1.75mm – more than enough to show blades of grass. This simple calculation lies at the heart of my preferred choice of infinity focus for most of my landscape work – See my Why I don’t bother with hyperfocal distance charts article for more).
The top left hand corner is shown below at 100%
There have been no corrections of any sort applied. Chromatic aberration is virtually absent.
The centre of the frame (100%) shows good detail, on what was a pretty low contrast day.
Do go back to the full frame shot above and note just how small a part of the image this is.
Compared to my much more expensive Canon lens, this is good. Sharpness and contrast is slightly down with the Samyang at f/8, but in the corners there is distinct (albeit minor) chromatic aberration with the EF14 2.8L II.
This isn’t going to be a side by side shoot out, but it’s worth briefly looking at some other comparisons.
A brief comparison with the EF14 2.8L II
I’ve set up the camera on a tripod, looking down the street where I live.
I actually took rather a lot of pictures, but reduced size shots on a web page show very little IMHO, and my boredom threshold in such testing is crossed very rapidly ;-)
The Samyang lens is lighter and plastic bodied, whilst the Canon just feels very solid and hefty. With just a quick look into the front of the lenses, there are more reflections visible in the Samyang, but not a lot, and experience with using it doesn’t show too much flare or reduced contrast.
Move your mouse over the image of the street below (Canon) to see the Samyang version of the shot.
Note too the warmer look of the Samyang lens and its slightly wider coverage (both shots processed at the same colour temperture).
Image contrast is better and residual slight barrel correction is very easily corrected with the Canon. At f/2.8 or f/4 the differences are pronounced, but by f/11 it’s close (if you’re not shooting architecture ;-)
The Samyang beats the EF14 2.8L II for chromatic aberration at all settings
So, the EF14 is much better in many ways (autofocus for example) – but so I should hope, for £300 vs £1800.
My suspicion is that for many users the Samyang represents exceedingly good value, particularly if auto focus isn’t vital and you can perfect your software lens correction workflow.
For use on a crop sensor body, the distortions are even easier to fix, but IMHO, a crop body defeats the whole idea of getting a 14mm lens…
If you’re interested in the Samyang 14mm (or whatever it’s called where you are) then consider what it is you want to use it for?
For some people the lack of AF and having to manually stop down the lens to the working aperture will be a show stopper.
However I found it perfectly easy to use out and about, by making use of good depth of field at f/8 and an ability to roughly estimate distances.
Exposure was set manually in all the tests here, and relies on the fact that if the light changes enough to need to alter the shutter speed, ISO or aperture, then you should easily be able to notice it. It’s hardly difficult with a bit of practice, and you’ll develop a much better intuitive feel for scene lighting, which is no bad thing. You could use aperture priority (or auto ISO) if you must, but I’d suggest taking the time to learn a bit more about light, since as well as manual focus, it isn’t going to harm your photography ;-)
Close up working needs care, or a tripod and liveview. In many respects, difficulties I’ve heard about using a lens like this reflect lack of experience and skills – the cost of digital images (i.e. effectively nothing) mean that if you want to learn the skills, it’s easy to just go out and experiment (it’s where the majority of my skills come from…)
Getting the best from ultra wide lenses like this takes a lot of thought – expect quite a steep learning curve when it come to appreciating just what works and what doesn’t. The strong perspective and emphasis of what’s close, need some serious thought about composition, particularly if your previous idea of ‘wide’ was 24mm ;-)
As an architectural photographer I really noticed the complex geometric distortions of this lens, enough to render it completely useless for some of my work. But hold on, add in software correction, and I could find much more use for it.
Update: Samyang now offer a number of .lcp lens correction profiles for their lenses for Photoshop/LR [Samyang download]
There is also the fact that in many photos I took, even though I knew how bad the distortion is, I couldn’t see it…
Take a bit of care in reading any reviews with tables of figures about lens performance, such wide lenses are exceedingly difficult to test accurately and repeatably without a very expensive test bench setup.
Taking several shots hand-held, it was easily able to stitch them into much wider views, such as this one of St. Mary de Castro.
Stitched with Autopano Giga, which handles the lens distortions with ease.
Despite being a plastic bodied lens, it’s robust and solidly built, but to my mind, could be improved in a few areas. The red dot for aligning the lens with my camera mount is on the metal part of the lens hidden by the lens mount – I could do with it at the side of the lens.
The lens cap fits firmly, and showed no inclination to fall off.
The precise accuracy of the focus scale has been widely questioned – test it for yourself if you get the lens. Of particular importance would be the infinity point (which will also vary with larger changes of temperture)
The range of lens mounts that this lens is available for makes it an interesting option to explore, but as I’ve said, I feel it needs a bigger sensor to get the real benefit of its ‘width’.
An absolute bargain, if its features (fully manual action, no AF confirm) and foibles (geometrical distortion) are not a problem for you.
Buying the Samyang/Rokinon/Bower 14mm f/2.8 lens
We make a specific point of not selling hardware, but if you found the review of help please consider buying the lens, or any other items at all, via our links. It won’t cost more, but we do get a small contribution towards running the site (thanks!).
In the UK
The Samyang (Rokinon) 14mm f/2.8 ED AS IF UMC Lens provides an extremely wide angle view. With good optical quality, but noticeable geometric distortions that may or may not be an issue.
The fully manual lens is available from numerous suppliers, and in several different mounts
(Nikon, Sony, 4/3, Pentax – Canon EF tested here)
Article first published December 2013 – Comment below, or discuss with Keith on Google+, where there are also many much higher resolution images taken with this lens.
If you’re new to the site, it has many hundreds of such reviews and articles – Articles index page.
Never miss a new article or review - Sign up for our Newsletter (2-4 a month max.)
Enjoyed this article?
All articles and reviews are listed on our main Articles and Reviews page, or use the search box at the top of any page. Experimental items, hacks and how-to articles are all listed in the Photo-hacks category.
Some articles that may be of interest:
- Using a tilt/shift lens - what it is they actually do
- Focus with tilted lenses - lots more information about what's going on when you tilt a lens. See also: Focusing the view camera - External link to [very] detailed coverage of camera movements
- Keith's tilt table spreadsheet (zipped file)
- Using old lenses on your DSLR - fun with adapters
- Keith's lens reviews and lens related articles
More of Keith's articles/reviews (Google's picks to match this page)