Sigma 14-24mm f/2.8 DG HSM Art Review
I will attempt to contain myself during this lens review, but it won’t be easy. This lens is something special. Sigma had a really, really big challenge to try and beat the legendary Nikkor 14-24mm f/2.8 lens. They won, pure and simple.
My praise is primarily centered on the resolution measurements I conducted using the MTFMapper program. Simply amazing results.
Sigma 14-24 f/2.8 on Nikon D610
Everybody has come to expect Sigma to be simply first-rate with build quality, and they don’t disappoint. I have read that Sigma made this “Art” lens as weather-resistant as their “Sports” series, although I didn’t personally do any tests to back up that claim. The lens comes with a nice zippered semi-soft case and strap, plus a slip-on lens cap.
If you look at some other reviews of this lens, such as Lenstip here and LensRentals here, you will find that they both agree with me about this lens. Except mine generally tested better than their copies did! Astonishing.
The Sigma 14-24mm f/2.8 DG HSM lens initials mean that it's full-frame (DG) and it uses Sigma's hypersonic motor (HSM) for auto-focus. It omits the "OS" designation, which means the lens doesn't support image stabillization. Since this lens is a member of the Art, Sports, and Contemporary series, it means that the lens can converted by Sigma between Nikon, Canon, Sigma, Sony, and Pentax mounts.
I have to admit that this is a tough lens to test. It has such an extreme wide field of view that you have to be super close to the test chart. My resolution results are at a subject distance of 0.7 meters to 1.2 meters, even using my large A0 (4x5 foot) resolution chart. The chart also had to be aligned parallel to the image sensor with incredible precision, or the else measurements were affected. This means that lens resolution at infinity focus can only be inferred.
About the only things missing from this lens are optical stabilization (Nikon doesn't, either) and the ability to use front-mounted filters. The Canon version of the lens can be fitted with rear-mounted gels, if you buy their little add-on filter kit. You can also buy third-party adapters for breathtakingly huge front-mounted filters, if you really want them.
Sigma also has this massive advantage over Nikon and Canon: their USB dock to program new firmware and focus calibration for both different distances and multiple focal lengths. Customize the calibration at 14, 18, 20, and 24mm with 4 distances each, or 16 total calibrations. You can also get Sigma to convert the lens to a different camera mount, for a fee.
To achieve their optical and mechanical quality, Sigma was forced to make this lens large and fairly heavy. I’m used to big lenses, so it’s a non-issue for me. I always use battery grips; the lens balance is just slightly front-heavy on my D610.
Zoom rotation: 80 degrees through the 1.7X zoom range.
Focus rotation: 110 degrees
Non-removable petal-shaped lens hood. (Sigma can convert it to a shorter round hood for a fee).
Fluorine front lens surface coating to repel dirt and water.
Angle of view: 114.2 through 84.1 degrees.
Close-focus 0.26 meters (6cm working distance)
9 diaphragm blades
17 elements in 11 groups
3 FLD elements, 3 low-dispersion SLD elements, 3 aspherics.
Minimum f-stop: 22.
Electromagnetic aperture, like the Nikkor “E” lenses.
1150 grams (different weights for different camera mounts)
96 x 132 mm dimensions.
Auto-focus is with Sigma’s hypersonic motor (HSM).
Manual focus override by just turning the focus ring.
The only lens switch is for auto or manual focus (I just leave it parked at “AF”).
There’s a distance scale (meters, feet) but no depth-of-field scale.
I never noticed the delay while focusing, so I’d have to say it’s “fast”. Focus speed becomes a non-issue in an ultra-wide lens, anyway. It’s also quiet, although I’m not a picky cinematographer. I never had any missed focus shots, either.
The focus ring felt silky smooth and perfectly damped; you could easily rotate the focus ring with a single finger. You’ll notice that the image magnifies a little bit as you focus closer, but this “focus-breathing” isn’t objectionable.
The zoom ring is quite a bit more damped than the focus ring, and the zoom ring is nearest the camera. It’s still a totally smooth rotation, but is a bit challenging to rotate with one finger if you’re in a hurry. The rubber zoom ring has a wider spacing on its ribs compared to the focus ring, for additional tactile feedback about which ring is which. I think it’s just the right distance from the camera body for zooming with my thumb and middle finger while the camera body/grip rests on my palm.
I didn’t notice objectionable vignetting in my photos, although it certainly exists; you’ll only tend to notice it at all with a clear blue sky at wide apertures. As long as it doesn’t draw attention to itself, I tend to ignore vignetting.
I also didn’t notice any lateral or longitudinal chromatic aberration, although my editing software would rid it automatically, anyway.
The pronounced bulbous front lens element is prone to catching the sun in shots; a lens with this many elements is inevitably going to show multiple reflections, which mostly look green. I try to use my hand to cast a shadow and help out the lens shade when possible. You’re bound to find yourself using the “healing brush” after a day of shooting, however, to clean up lines of green reflections.
Bokeh isn’t really a useful discussion with a lens this wide. What there is of it looks decent, but not excellent. Consider it to be a non-issue.
The lens has low barrel distortion at 14mm, and virtually none beyond 15mm. It’s a simple type of distortion that can be removed easily in your photo editor.
Worst-case 14mm barrel distortion hardly detectable
Distortion 14mm corrected in editor
24mm no distortion seen
Like everybody else, I read a few reviews on this lens before I decided to buy it. Those other reviews agreed that the resolution was a bit better than the Nikkor 14-24 f/2.8 lens. My own tests make me think that I got a better copy of the Sigma than most of the other reviewers did; the resolution was striking for a super-wide lens.
I use the MTFMapper program to perform resolution and focus tests, which you can get here.
I have an article about the MTFMapper use here.
All of my resolution tests are done using unsharpened, raw-format from my Nikon D610 (24 MP). I use live view and contrast-detect focus, to eliminate any concerns about focus calibration. I’m showing the best results from about 10 shots at each focal length and aperture tested.
I halted each resolution test after stopping down to f/16, because the diffraction effects ruin the resolution beyond this aperture. The lens stops down to f/22, if you really need it.
I haven’t seen resolution characteristics quite like these before. The results at shorter focal lengths show a sort of sine wave pattern of sharpness, versus the expected gradual tapering-off as you move away from the lens center. You might want to avoid using a focus point at 6mm and 12mm from the screen center, unless you use live view or stop down the lens to about f/5.6 or beyond.
The resolution numbers in the center of this lens are quite astounding. The edges are better than expected, although the meridional versus sagittal results at a few focal lengths show a fairly large separation (astigmatism).
I’m guessing that the use of the 3 aspherical lens elements causes the unusual-looking resolution progression as you move from the lens center toward the edge.
Seeing the 14mm f/2.8 resolution results, I can’t wait to travel to the mountains and try out some starscapes away from the city lights that I unfortunately have to contend with.
14mm f/2.8 MTF50 lp/mm resolution
These results are just remarkable. I show the results separated out into both sagittal and meridional directions across the whole image sensor (Nikon D610). This is a very unusual pattern that I haven’t seen before. Note the subject distance is only 0.71 meters to cover the test chart dimensions of about 4 by 5 feet. I can’t answer how the resolution numbers might change when focused on infinity.
The usual MTF contrast plot, but made from real measurements
The contrast plot above (14mm f/2.8) shows how lens resolution is usually depicted, except MTF contrast plots from most manufacturers are “theoretical performance” without considering the camera sensor. This plot is from the actual measurements on the camera sensor.
A close-up from the middle of the resolution target
The shot above shows a portion of the resolution target photo center, although it is overlaid with the edge resolution measurements in “cycles per pixel”. You have to know the sensor pixel size (5.95 microns here) to convert into other resolution units.
As noted in the picture, I saw a peak resolution of 65.3 lp/mm. This is the same as 3133 lines per picture height, in the sagittal direction. The meridional direction peak resolution was 63.6 lp/mm or 0.38 cycles per pixel.
Test chart corner, 14mm f/2.8
The extreme corner of the resolution chart is shown above. On many lenses, you only see numbers like these in the chart center. Wow. It’s crisp right out to the corners.
14mm f/4.0 resolution and MTF contrast plot
I use a little laser behind a tiny hole in aluminum foil to substitute as a star. The lens is focused on infinity, and the target is about 15 feet away. The images shown are at the pixel level (100% magnification). All shots are in the extreme corners of the sensor.
I included some comparison shots using my Tokina 11-16 f/2.8 lens. It’s considered very good, but you’ll see that the Sigma clearly beats it, even though the Tokina results are on a DX sensor and the Sigma is on an FX sensor.
I also included a shot at f/4.0 to compare to f/2.8 results. There’s almost no difference when stopping down, because the Sigma is so good even wide open.
The size of the coma looks the same in all corners of the sensor. Only the orientation of the coma changes.
Tokina 11-16 coma at 11mm and 16mm reference images
Sigma 14mm wide open versus f/4.0 coma