Rokinon 10mm f2.8

The Rokinon 10mm f/2.8 ED AS NCS CS is an ultra-wide lens designed for APS-C DSLRs. It is produced by the South Korean company Samyang. Its 14 elements 9 groups design features one extra-low dispersion element and two aspherical elements. It weighs 580 grams (1.28 lbs) and features a “Nano Coating System” to improve light transmission and suppress lens flare. Like all the Samyang designed lens, the Rokinon is exclusively manual focus. Just like the manual focus lenses of old, it features a generous focus throw and an impressively dampened focus ring. An adjustable hyper-focal scale allows for greater fine tuning, as it rarely comes out of the box lined up correctly and the lens does focus past infinity.

Samyang has become known for making large, fast primes and selling them at (relatively) bargain prices. Some people rave about the quality of their lenses, comparing them to Zeiss primes. Others complain about quality control and accuse proponents of Samyang glass of hyping the product.

In light of this, how does the Samyang derived Rokinon 10mm f2.8 perform on a 16 MP DSLR (i.e., the Pentax K-5iis)? Let’s look at some pictures.

Sayang lenses are (generally) known for being sharp. The Rokinon 10mm is no exception. Obviously, it would be unrealistic to expect an ultra-wide lens to be as sharp as a 50mm. But granting the extreme width of this lens, it’s pretty darn sharp. Oddly enough, because of the extreme wide angle field of view offered by the lens, coupled with the fact that it is a manual focus lens, the Rokinon 10mm can be both very easy and very difficult to focus. It’s easy in the sense that the wide angle offers a very wide depth of field. Stop the lens down to f8 or f11, and nearly everything’s in focus. But if you absolutely want to nail the focus, thereby getting the very best resolution out of the lens that is possible, that can be more challenging. Wide angle lenses push the detail far away, making it difficult to determine, within an optical viewfinder, whether one has attained optimal focus. So the following images should be seen, not as scrupulous tests to determine precisely what can be achieved with the lens using the very best technique (say, by making use of magnification in live view or in an electronic viewfinder), but rather what can be achieved through ordinary use, attaining the best focus possible by focusing through the optical viewfinder.

The first image demonstrates the level of detail that can be captured by the lens, when shooting at f8. First, the entire image:

Now the center crop:

And the corner crop:

While the corner crop is not as sharp as the center crop, it’s really quite good. I don’t see anything to complain about here. The lens is commendably sharp. If you need more sharpness out of an ultra-wide angle, you’re probably better off moving to a full-frame camera and shooting with either Nikon’s or Canon’s 14mm prime lens. Or, better yet, the Zeiss 15mm. Expect to pay huge sums not only for the FF camera, but for the ultra-wide full frame lens.

Another example at f8, starting with the full image:

In this shot, the focus is on the foreground, especially the tire. So the center crop will also be in the foreground, to match the focus:

Reasonably sharpness — again, nothing to complain about. How does the lens fare away from the center but not all the way to the edge? Let’s take a look:

Seems to be doing well enough. What about the far corners?

Here the resolution doesn’t seem quite so good. Indeed, it seems less sharp than the previous example. But that’s what we sometimes run into with ultra-wide angles. You not only have to deal with reduced resolution toward the edges and corners, but the fact that the camera may be much further from corner than from the center, particularly when photographing something close, leading to variable performance in resolution.

One more example at f8:

Center crop:

Corner crop:

How does the lens perform at f5.6. Let’s take a look, starting with the full image:

Center crop:

And the corner crop:

Here the corners are much closer to the camera than the center. Yet it’s not bad at all. Resolution does seem to be weak in the very far corner. Otherwise quite respectable performance.

What about at f2.8? Don’t you buy an f2.8 lens to use it a f2.8?

I’m personally not a fan of shooting landscapes at f2.8. I see the f2.8 on this lens as being useful for astrophotography, but otherwise not all that essential. Nonetheless, here’s a couple of images shot at f2.8, just to satisfy curiosity. The first one is an image of a pier:

The center crop:

Edge crop:

As to be expected, edges are not as good as they would be if the lens were stopped down. That’s fairly typical performance for an f2.8 or faster prime lens. Now let’s look at a second image. This one is taken inside, up close to the object being photographed, so one can appreciate what level of bokeh can be drawn from the lens:

The cup is the point of focus. Here’s a crop:

And the far corner crop:

Of course, one of the problems of attempting to judge the wide open performance of an ultra-wide angle lens is that it’s very unlikely that your corners are going to fall within the plane of focus. This is why the mania for corner sharpness wide open is a bit foolish. What are the odds that the interior shot that you have to photograph handheld is going to match the field curvature of your lens, thereby allowing you to get sharp edges and corners (assuming that the edges are sharp wide open)? It’s not very likely. So it is best to tolerate “soft” corners when shooting lenses wide open. The demand for sharp lenses wide open is merely driving the prices of lenses into ridiculous heights, as lens manufacturers attempt to satisfy what is essentially a false demand.

According to numerical tests, the Rokinon 10mm suffers from 4.65% barrel distortion. Although that’s a fair amount, it’s to be expected in an ultra-wide angle lens, particularly one priced under $500. The barrel distortion will mostly be a problem when photographing straight horizons, such as views looking out to sea. Lightroom does not feature profiles with this lens, so you’ll have to fix the distortion yourself.

There’s another type of distortion that afflicts all ultra-wide angle lenses, regardless of cost. It’s not correctible via lens design. Let’s call it perspective distortion. Here’s an example:

You can try to correct this tolerably well in post, but you lose some of your ultra-wide-angle field-of-view:

The Rokinon 10mm does suffer from a certain degree of chromatic aberrations. That is to be expected. They clean up nicely in post, so they’re not worth complaining about. However, there’s another issue that sometimes arises that is not so easily dispensed with. This involves transitions between hard, dark edges and blue sky. It’s like chromatic aberrations, but more subtle. It can only be removed through cloning in Photoshop. Here’s the type of image where it occurs:

Along the edge, you get a whitish blue line:

It’s small, it’s subtle, but it’s there nonetheless. This is a legitimate weakness in the lens, precisely because it cannot be so easily removed, unlike most CA’s.

The Rokinon 10mm features what Samyang calls it’s “Nano Coating System,” which is supposed to improve light transmission and control for flare. Nano coatings are generally considered the best of the best. How do these coatings work on the Rokinon 10mm?

In short, not very well. I’m not sure what sort of nano coatings Samyang, the manufacturer of the Rokinon 10mm, is using. They’re clearly not as good as the nano coatings used by Nikon, Pentax, and other traditional Japanese camera companies. Consider a couple of examples, shooting into strong sunlight:

Strong, direct sunlight simply doesn’t work with Rokinon 10mm and should be avoided.

Otherwise, the Rokinon 10mm is a fairly contrasty lens with good color rendition. It does, to be sure, seem to have a slight bias toward the aqua side of the color spectrum. Nor does it quite match, in terms of overall tonality, any of my Pentax lenses featuring the ghostless coating technology, first introduced in 1997. This in no way is to suggest that the Rokinon lens has poor or mediocre contrast and color. The contrast and color of these Samayang-derived lenses is actually quite good — just not quite as good as what Pentax routinely offers. The design goals of Samyang lenses are primarily to provide a generous amount of resolution at a lower cost. Undoubtedly, compromises must be made to achieve these goals. Thus we get sharp, inexpensive lenses with less than stellar contrast and color and a few minor but troublesome issues with transitions between dark objects and blue skies.

Here’s a set of sample images, all post-processed in Lightroom, taken with the Rokinon 10mm f2.8, all shot at f8: