2. Perceived Sharpness
This is the #1 claimed reason for the new generation of 50s like the Zeiss and the Sigma: high perceived sharpness even wide open. Whoever thinks that sharpness is for pixel peepers and people who hang out in forums discussing sharpness rather than actually shooting pictures: You’re right. But it’s also for people who actually shoot pictures. There’s a great quote by Mark Crislip aka PitchBlack who sums it up perfectly.
Still Life Shot: Detail Comparison
Ok, so here’s a still life motive that I was shooting and that I modified a bit for this test. Delicious hot pasta from ½” audio tape tossed with crunchy stereo jacks:
This was shot wide open with all lenses. For the 58 mm and 55 mm lenses, the camera was slightly pulled back to compensate for the focal length increase. The focus was set on logo of the Horch studio microphone on the right (see last paragraph of this section for details). The whole scene is lit by three lights plus the candles: A large soft box on a boom from above, a 15°-gridded highlight from the back right on the glass and a medium striplight behind a diffusor panel (for grading) on camera right. Yes, there’s lots of room for improvement, but this isn’t about the scene itself.
Let’s just look at a crop so we can compare perceived sharpness. This is really unbelievable. Check it out, these are 100% crops (as opposed to all the other images on this site, they’re not in high-dpi, so people with “retina” displays will find it a bit blurry).
- As you can see, there’s just an extremely slight advantage of the Zeiss over the Sigma, but it’s hardly noticable. (Note that the focus plane is something like a mm further away with the Sigma image.) It’s absolutely breathtaking what level of detail these lenses both produce – wide open! Keep in mind that the images are on a D800, it must be even better on a D800E or D810.
- Just to give you an idea what league we’re talking about here, click next for a comparison of the Sigma and the horrible Nikkor 58/1.4G.
- Just because comparing the worst with the best is a little unfair, the next one shows the Nikkor 58 vs. an average lens, the 50/1.4G. I guess you get the idea now.
- Then, you can see a comparison of the two Nikkor 50s. They are not really that much different from each other, except for the color fringing (quite visible with the 1.4, not so with the 1.8).
- The next comparison is the Sigma with each of the 50 mm NIkkors.
- Finally, you get a comparison of the Zeiss with each of the Nikkor 50s.
Do these differences matter? Sometimes. Check out this separate article if you want to see how they do.
Addendum: Because some people asked how the focusing was done, I’ll tell you. As said, I used the center of the logo on the mic as a focusing point. With each lens, I took four shots. They were focused 1. with AF in Live View (this uses contrast detection), 2. manually after focusing to infinity, 3. & 4. repeat one and two. With the manual-focus Zeiss, I took four shots focused manually in Live View, focusing to infinity in between. I then compared the results in 100% view in post, and took the “best” shot, which was the first one in all cases – the Live View focusing did a good job. In the above crops you can see that the focus plane nevertheless isn’t precisely the same, but varies by a millimeter. However, as the selected crop intentionally contains a curved surface with a layered structure, you’ll be able to easily locate the in-focus structure in every shot.
You may interpret the failure to focus 100% accurately in two ways: a) This should not happen in a test! Correct. Well guess what: As you can imagine, I tried my very best, because, as you correctly noticed – this is a test. So if you consider the mm of focus point variation an issue, then the other interpretation is for you: b) Even if you try your best, use a heavy studio tripod, Live View, a still life scene and good light, this millimeter is what you can achieve in practice. At least in case you’re a photographer and not a test lab. So you may take this as an additional result.
Focus Test Chart Comparison
Yeah well, test charts. One quote I like from Ken Rockwell is: “…if you are shooting test charts, a flatbed scanner is still far better than a D800.” And I’d add that if you’re shooting brick walls, the thing you need to worry about most is your taste, not image quality. Anyways:
I always use the Focus Test Chart by Tim Jackson to fine-tune the AF for my lenses before I first use them. The thing is, you can see quite a lot of your lens qualities from that simple chart, e.g. sharpness and contrast, longitudinal CA, focus shift. With this chart, the camera is pointed at a 45° angle to the chart, producing an image like the one on the right. The frame marks the crop area from which you see the images from the respective lens.
Keep in mind that the “results” that you can see concern specifically the marked area of the image frame, i.e. off-center. Since the focus point of a good image is rarely in the exact middle, I don’t care much about this exact center, off-center like the marked area is actually much more relevant in my photography. (To get an idea of how sharpness varies across the image frame, you can check out the field maps on DxOmark: Choose a lens, and go to Measurements -> Sharpness -> Field map.)
Click through this gallery (100% crops @ high-dpi), a summary of which is provided below:
- The Zeiss is tack-sharp wide open, has very moderate spectrochromatism, and absolutely no focus shift.
- The Sigma is pretty much the same in this respect, only that there is noticable focus shift (which is normal for many lenses, unfortunately). Too bad, for a lens of this sharpness class, you might think. Well, the thing is, in practice, you won’t notice too much of this. Maybe this is something for a separate article.
- The 58mm Nikkor… as I said in the beginning: It’s a joke. It’s worse than the 50/1.4G. Useable results start at f/2.8. In the above examples you can see the heaps of CA, too. The only good thing about this lens is that there is no focus shift at all…
- …which again the Nikkor 50/1.4G has quite a lot of, the highest amount of all lenses. In the comparison slider you can see how the focus shifts backwards as the lens is stopped down. If you get a good copy (there seems to be considerable sample variation with these), it is as sharp as the 1.8G, with two third stops advantage. At f/1.4 it’s as sharp as the 1.8 at f/1.8, and so on. This goes for the off-center. In the center, they’re the same at the same aperture. Again, keep in mind that this varies with the quality of the 1.4G copy. What doesn’t vary:
- (Longitudinal) Chromatic aberration is substantially worse with the 1.4 than the 1.8. The 1.4G is infamous for this, too. It’s the main reason why I and many people I know got the 1.8G instead.
- Also, the Nikkor 50/1.8G has a lot less focus shift.
Before I go ahead and compare all the thousands of combinations, you can download the whole set (crops from all five lenses at f/1.4 through f/5.6) and compare yourself.
Go back to page 1: Introduction
Go to next page 3: Bokeh quality
Proceed to page 4: Sensitivity to bright light inside/outside the frame
Proceed to page 5: Vignetting, Coma, Sunstars, Color Rendition
Proceed to page 6: Handling & Quality
Proceed to page 7: Summary