I did not want to start a new thread or put a comparison between Nikon AF-S 14-24/2.8 G ED and Sigma 14/2.8 DG Art into a thread about the test results on Canon. As I already mentioned in this post, I'm wondering about a 4 ★ rating of the Nikkor and a pretty low one with 2 ½ ★ of the Sigma Art on Canon. Now, I don't like to just throw out my opinion but also see, if that opinion at least has some truth in it.
When I was trying to find a setup which can show differences of the 14 mm prime (close to the price of the zoom, btw.) and the zoom @ 14 mm, I didn't want to go for charts or brick walls, these are not my main subject with one of these lenses. In particular I wanted to find out how bad the borders and corners of the Sigma are wide open and at f/2.8.
Now, there was the first question: If I ever need to emphasize a border or a corner of a flat pane then where and how to focus? The difference of a point in the center of a pane and (if that point is exactly 1 m from the sensor away AND the sensor is aligned perfcetly parallel in to the pane) to a corner is roughly 66 cm farther from the sensor than the center of the picture. The horizontal border is 53 cm farther away from the sensors center - I didn't calculate, I used a laser distance meter.
That makes it clear, that I have to use LiveView to get the sharpness I would want. A big difference to focus on a chart, see what works best for center and hope the corners will catch up. They won't. Not even the most expensive DSLR lenses do so. So to me it makes a lot sense to put the sharpness where I need it to be. And that didn't end up bad for the Sigma - or well for the Nikkor.
Samples? I was hoping someone would ask. But here's the next thing: waterdrops on a window might not be the best subject to seek for sharpness? I agree, partly because it's difficult to focus manually at their low contrast. I don't do focus bracketing, so this might also be an issue. I didn't because the results to me are clear enough. The one with the yellow dot is always the Nikkor, the blue one the Sigma. See for yourself:
Both lenses wide open, lower border, more or less center of it.
Corners wide open - of both pictures there's a "f/2.8 vs f/2.8" version, with declasses the Nikkor very much - in this tiny crop and in this not very realistic situation. If someone wants to see, I post the f/2.8 versions as links, as the number of pictures per post are limited:
This sieve I took as something more structured than waterdrops. However, it does show another flaw of the Nikkor at 14mm and f/2.8. The fringing of the highlights is a bit more annoying than the Sigma, no?
Now, what did I learn?
- Don't trust ratings only (but that one is common knowledge at PZ/OpitcalLimits)
- A zoom still can have 4 ★ and being outstanding as a fast ultra-wide-angle zoom - at the extremes, no wide-angle zoom scratches the merits of great primes.
- But it's also true, it can't beat a great prime.
- In that case, the prime is even faster and stopping it down from f/1.8 to f/2.8 doesn't improve things for the zoom
- The Nikkor has more distortion than the Sigma (although the latter also shows a fair bit), but after switching the correction off in Capture One, the sharpness improved a little. And the distortion was much worse.
In my opinion, tests like here on PZ/OL should always only help if comparing more or less the same components.
- A chart stays the same - but lenses usually are corrected for ∞ . With a 50 × fl (50 × 14 mm = 700 mm) as a distance to the chart, one has to run into problems such as more vignetting and distortion at closer range.
- Also, the correction of panes must be difficult. In general as well but particularly at a close range.
- But all that is valid as well for the Nikon zoom, so why the huge gap? Difference in MP? 24 MP vs 50 MP?
I don't say, the Nikon zoom is worse than the Sigma prime - too much apples and pineapples... But I do say, my Sigma 14/1.8 prime is wide open in center, border and corner performance better than the Nikon zoomed out to 14 mm, has less distortion and CA. The big question remains: Will field curvature give a disadvantage the prime? To me, it's too much effort to test and not enough value in the result. I never use a wide open lens to reproduce a painting or drawing - and stopped down, things will become even.
But I saw the Sigma prime being better than I was hoping for - and that's not bad to find out after the purchase