Perhaps because ultra-wide lenses are harder to manufacture, or simply by lack of popularity compared to longer lenses, there are very few options for ultra-wide angle lenses. Still, even with a handful of choices, someone has to choose.
This time, I wanted an ultra-wide rectilinear (non-fisheye) zoom for the recently-reviewed Pentax K10D. The options readily available are:
- Pentax DA 12-24 F4 – Digital Only – 77mm filter
- Sigma 10-20mm F4-5.6 EX DC – Digital Only – 77mm filter
- Sigma 12-24mm F4.5-5.6 EX DG – Full-Frame – Rear type gelatin filter
These three choices represent possible compromises fro ultra-wide zooms. Depending on your needs, you may consider other options like the Pentax 10-17mm f3.5-4.5 ED (IF) fisheye or a prime lens like the Pentax DA14 F2.8. On all Pentax DSLR cameras, which use a 1.5X focal-length multiplier, the Sigma 10-20 has the widest field-of-view. Although the difference is only 2mm at the wide-end, it represents an angle-of-view that is 20% wider. Compared to the Sigma 10-20 F4-5.6, the Pentax 12-24 F4 is narrower in terms of angle-of-view, but it is one full stop brighter at the telephoto end. Representing, the other end of the spectrum of compromises, the Sigma 12-24 F4.5-5.6 is possibly one of the most ambitious lenses to date. At 12mm, it represents the widest full-frame rectilinear field-of-view currently available. With respect to specifications, the side-effect is that this Sigma lens cannot support front-mounted filters and has a more narrow maximum aperture at the wide-end. Since Pentax has no full-frame DSLR cameras, the Sigma 12-24 is therefore the most limiting choice among these lenses: it neither has the largest field-of-view nor the largest maximum aperture.
Since specifications are not enough to obtain satisfactory results, research was required to choose between those lenses. Here Sigma lenses have an advantage because their availability for multiple lens-mounts means that they have a greater potential user-base and therefore more people write about them. First up, the formal testing at SLR gear is excellent and covers sharpness, vignetting, distortion and chromatic aberrations. They tested the Sigma 10-20. Although they apparently never test any Pentax lenses, they have a test for the Tokina 12-24 F4. Apparently, the optical elements of the Tokina 12-24 were co-designed with Pentax and should perform similarly. Although that is not certain, most user feedback on the Pentax 12-24 is quite similar to the Tokina 12-24.
Based on research, it seems that neither of these three lenses are poor performers. In terms of sharpness, the Pentax 12-24 F4 appears to be the sharpest, specially at the wide end. Second the the Pentax, the Sigma 10-20 is also quite sharp. Samples images posted on the internet from these two lenses show that both of them are acceptably sharp for more uses. A critical inspection of sharpness gives the Pentax 12-24 a clear advantage near the edges at wide apertures. The Sigma 12-24 is relatively sharp but it is less consistent across the frame than the wider Sigma 10-20. That can be seen in the MTF charts posted on Sigma’s website.
There are a number of optical characteristics which are not represented in MTF charts. Distortion, which is normally the biggest problem of wide-angle and ultra-wide-angle lenses, is controlled rather well again by the two digital-only lenses. Based on sharpness and distortion, the Sigma 12-24 somewhat falls behind the two other options. Plus, it does not bring anything in terms of specifications for Pentax DSLR cameras.
The major differences between the other two lenses fall under vignetting and chromatic aberrations. Choosing between them may be a matter of taste. Personally, I despise vignetting much more. It affects all images and gives an artificial tunnel-like look. The same effect is used in movies to show that the viewer is looking through a telescope or a pair of binoculars – a very artificial effect. Chromatic aberrations do not appear natural either but tend to only occur near areas of over-exposure. Obviously, those areas are very local and can be minimized by proper metering.
That being said, the lens which exhibits the most serious vignetting is the Sigma 10-20. According to SLR gear, it can vignette as much as 0.75 EV at its widest aperture. Closing the aperture, this Sigma lens only reaches 0.25 EV at F16. While at 0.25 EV vignetting is not very noticeable, F16 is rather dark and approaches the diffraction limit of the K10D. The Tokina 12-24, on the other hand, drops to 0.25 EV just at F5.6. This is consistent with samples from the Pentax 12-24 too. To the Sigma 10-20’s favor, it does suffer less from chromatic aberrations and flaring. However, those are intermittent problems based on lighting conditions which can be easily minimized, vignetting cannot be avoided.
The bottom line is that I chose the Pentax 12-24 F4 based on its excellent sharpness, low distortion and low vignetting. The consequence is that I use the supplied lens hood often to minimize vignetting. As for chromatic aberrations, the K10D’s metering system does a very good job at minimizing over-exposure which reduces the risk of chromatic aberrations.