Pentax DA 10-17 f3.5-4.5

The Pentax DA 10-17 is a bit of a novelty item: a fisheye zoom lens. With Canon releasing its first fisheye zoom (the very expensive 8-15mm), this type of glass might become be on the verge of popularity. Pentax pioneered the fisheye zoom with its F 17-28mm full-frame model, released in 1995. Pentax’s APS-C fisheye lens was released in 2006. Supposedly designed by legendary lens designer Jun Hirakawa, the architect of two of Pentax’s three FA limiteds (considered in some quarters as “ the very best AF SLR lenses made today”). Hirakawa was known for being a bit of a throw-back when it came to lens design. He thought lens making was more of an art than a science. He cared little for how a lens scored on tests. He wished to make lenses that could create striking images, that had “spice,” as he put it.

The DA 10-17 is not a lens that is going to do well on tests. In the first place, it’s not all that sharp. Consider the following image:

Patrick's Point SP11-30-2

This is pic was taken at the wide-end of the lens (10mm, 180º FOV). At web resolution side, it doesn’t look bad. So let’s look at a 100% crop:

Patrick's Point SP11-30-3

Now a sharp lens will provide fairly good detail even at 100% resolution, which is not what we find here. That is the bad news. The good news is that with a camera such as the Pentax K-5, which produces high resolution, low noise images (at least when shot at low ISOs), these images sharpen up very nicely in post, as follows:

Patrick's Point SP11-30-4

Here we have more than enough resolution for most purposes. So whatever handicap the lens might have in terms of sharpness (at least compared to other Pentax DA lenses) can easily be rectified with a camera like K-5. With some of the older cameras, difficulties may arise. On the K200D, I struggled to get any kind of sharpness out of this lens. Whether it was a consequence of the K200D’s mediocre AF, its lower resolution, its slightly noisier images, its CCD sensor, or some kind of front- or back-focusing issue, the DA 10-17 would not produce even modestly sharp results on that camera. But on the K-5, no issues arise at all: the images it produces sharpen up very nicely in post.

So while the DA 10-17 might not produce stellar results in resolution tests, it turns out to have more than enough sharpness for most practical needs. The lens is reasonably sharp from 10mm through 14mm. It loses a little bit of resolution at the long end of the lens, but not enough to really matter. Border to border sharpness is surprisingly good, considering the degree of distortion and light bending involved in this glass.

Another issue with the DA 10-17 is chromatic aberrations. According to

Lateral CAs may not be good but the really bad characteristic of this lens is purple fringing (a color blooming effect at extreme contrast transitions). At the time when I took the lens for a ride it was an overcast day which is among the worst situation when it comes to purple fringing due the extreme contrasts between your subject and the sky.

The problem is actually worse than the photozone reviewer suggests. It’s not just overcast days that lead to purple fringing. It will fringe in all kind of circumstances, even in the best of light. You can sometimes even see the fringing in the viewfinder! The lens could easily be described as a purple fringe monster, particularly towards the borders. So those who are uptight about such things should avoid this piece of glass. It will drive you insane. Here’s what the corner of the above image produces:

Patrick's Point SP11-30-5

Notice the jagged purple lines below some of the rocks. It’s subtle, but it’s there nonetheless. These are not particularly high-contrast lines. With branches against sky (blue or otherwise) — or any dark object against a patch of snow — the fringing is substantially worse.

The question, though, is whether this fringing can be removed in post. Here’s the image after it’s been subjected to defringing in Lightroom:

Patrick's Point SP11-30

All gone! I can’t promise that the purple fringing (or other types of CA) will always be removable in post. That simply isn’t true. But most of it can be removed; and if not removed, at least severely mitigated. In other words, I don’t regard such aberrations, however regrettable they may be, as an issue of immense importance. They are largely trivial and nothing, really, to get uptight about.

What’s of utmost importance is the quality of the images which the lens produces. And here the DA 10-17 really shines. While the lens will not score high on distortion tests (it distorts on purpose: it is, after all, a fisheye zoom lens!), or produce stunning numbers on resolution or aberrations tests, it will excel precisely at the most important test of all: namely, image quality. After a sufficient level of resolution has been achieved (which the DA 10-17, whatever its resolution limitations, does reach), the most important aspect of a lens is its color rendition. All things else being equal, a lens with superb color rendition will produce noticeably more stunning images than a lens with only mediocre color rendition. Add superb contrast and you have a lens that is, image wise, a real winner.

Patrick's Point AU10-15

Del Norte WI10-52

Southern Humboldt Bay-37

Columbia Gorge SM10-27-TIF

Nice enough images, to be sure; but what about the flare control of the lens? A lens with a 180º field of view will have trouble keeping glare from getting into the glass and creating flare issues. How well does the DA 10-17 handle the frontal assault of the sun? Judge for yourself:

Northcoast Misc WI10-520

Cascades SP11-233

Dry Lagoon WI10-50

Not bad. The only problematic flare shows up in the last image, where the sun is probably too bright (i.e., too strongly exposed) and too near the edge of the lens. Furthermore, the last image was taken at 13mm, while the previous two were taken at 11 and 10mm. Perhaps it’s more susceptible to flare at the middle or long end of the glass.

The most distinctive feature of the lens, apart from its ability to produce superb images, is the fact that it is a zoom lens. Is it really necessary for a fisheye lens to zoom? Don’t most people buy a fisheye lens for the distortion and the 180º field-of-view? Well, perhaps some people do. But it seems to me that it is precisely the fisheye lens, which is a kind of specialty lens of limited usefulness, where the versatility of a zoom is most welcome. Let me put it this way: the stereotypical fisheye focal length might be a nice place to visit on occasion, but it’s not really an FOV where I wish to stay for any length of time. It’s too limited and prohibitive. If I had a fisheye prime, I would find myself constantly taking it on and off the camera — so much so that I would end up not using it at all. The ability to alter one’s FOV (from 180º to 100º) is liberating. It means the lens stays on the camera much longer and, for this reason, can be rated as more than just a specialty lens: it’s a practical, useful ultra-wide zoom for those tight places when you can get away with either a little distortion (at the long end of the lens) or even more (at the wide end). And, to make it even more useful, it doubles as a traditional fisheye lens with the standard 180º field-of-view.

Although the Pentax DA 10-17 is designed specifically and solely for Pentax DSLRs, Pentax licensed the optical design to Tokina, which makes a version for Nikon and Canon mounts. The Tokina version cannot be regarded as identical to the Pentax version, since the Tokina version features different coatings, and coatings matter.

More images from the Pentax DA 10-17, the world’s first APS-C fisheye zoom lens. Note the rich, vibrant, intense colors produced by this fine piece of glass:

Cascades SM11-287
Cascades SM11-328

Cascades SM11-447