Today I tested my new Zeiss 21mm ZE against my Nikon 14-24mm. This was just a test of the sharpness of the lenses. Other aspects like CA, colors, flare etc. were not tested. The aim of this test was to check, whether my new 21mm ZE lens has any faults.
The test-object was the side-wall of my house. The test was done with a Canon 5D II on a Gitzo series 5 tripod with a Burzynski head, mirror lock-up and manual focus with 10x live view on the wall at the center of the image. All pictures were shot in raw at ISO 100, converted with LR with the sharping parameters (50, 0.7, 100), contrast set to +75 and then exported to jpg with 100% quality. The Nikon zoom was adapted to the 5D II with a 16-9-adapter V2.
Summary:
The Zeiss 21mm has a much stronger vignetting until f/5.6 compared to the 14-24mm lens. The Zeiss has avery good sharpness between f/2.8 and f/8.0 (peak performance between f/4.0 and f/8.0). From f/11 the performance is reduced due diffraction (f/22 is very bad and in my opinion not useable at all). Between f/5.6 and f/16 the Nikon Zoom has nearly the same sharpness. At f/2.8 and f/4.0 the Zeiss is slightly better in the corners. At f/22 the Nikon is much better than the Zeiss (allthough the sharpness is also reduced at these apertures due to diffraction).
Both are very good lenses and I will keep both.
Two world class lenses - elbow to elbow. I believe the 14-24 is optimized for the wide end - just goes to show how nice that lens really is - keeping up with one of the best primes. The Zeiss contrast is a little stronger - could change that around in post for the zoom. The only downside I see to the Zeiss is the center distortion (not shown in this test) - don't want to shoot any straight lines with that lens.
Wow, the Zeiss also renders a much cooler image. That vigneting is pretty poor on the Zeiss. The Nikon is so good for a zoom it's hard to ignore. Canon Canon's possible offering (see the patent thread) match the Nikon? I bet they'll be using it as their benchmark.
Bobu wrote:
Today I tested my new Zeiss 21mm ZE against my Nikon 14-24mm. This was just a test of the sharpness of the lenses. Other aspects like CA, colors, flare etc. were not tested.
Understood. However the old 21 mm f/2.8 Distagon had much less lateral CA than the Nikkor 14-24 mm. I suspect the new one isn't quite as good as the original in that regard, though awfully close. Certainly better than the Nikkor zoom. It would be interesting to see a CA comparison.
Bobu wrote:
Summary:
The Zeiss 21mm has a much stronger vignetting until f/5.6 compared to the 14-24mm lens. The Zeiss has avery good sharpness between f/2.8 and f/8.0 (peak performance between f/4.0 and f/8.0). From f/11 the performance is reduced due diffraction (f/22 is very bad and in my opinion not useable at all). Between f/5.6 and f/16 the Nikon Zoom has nearly the same sharpness. At f/2.8 and f/4.0 the Zeiss is slightly better in the corners. At f/22 the Nikon is much better than the Zeiss (allthough the sharpness is also reduced at these apertures due to diffraction).
The effects of diffraction can't be avoided, so if one lens looks much better than another it's likely using a different f-stop. In this case your Nikon shot had a shutter speed of 1/13th while the Zeiss had 1/8th, yet the images are similarly bright. That suggests the Zeiss correctly stopped down to f/22, while the Nikkor did not, for whatever reason.
Thanks for the test, by the way. From what I've seen from these lenses, it's clear to me that the Zeiss is sharper in the centre and much shaper at the corners at f/2.8, but suffers from higher light fall-off at the edges (aggravated by today's digital sensors - see below). At f/5.6 the Zeiss is probably a fair bit sharper too, but the Nikkor is so good that we don't see a big difference on a 21-megapixel full-frame sensor. I would guess though, that the Distagon will leave the others trailing in its wake when we hit 50-megapixel sensors and beyond.
Pixel Perfect wrote:
Wow, the Zeiss also renders a much cooler image. That vigneting is pretty poor on the Zeiss. The Nikon is so good for a zoom it's hard to ignore. Canon Canon's possible offering (see the patent thread) match the Nikon? I bet they'll be using it as their benchmark.
The Zeiss isn't much cooler: the centre crops actually show the Zeiss being very slightly redder, with the Nikon greener. But the differences are tiny. The colour cast in the corner of the Zeiss image is down to the Canon 5D Mark II's sensor not coping well with non-perpendicular light rays (film doesn't have this problem, of course). The lens didn't project those colours.
I don't know enough about sensor design to know if we're likely to get rid of this problem in the future (maybe at the expense of sensitivity by redesigning or eliminating the micro-lenses). For now though, the Nikkor zoom obviously requires less post-processing (i.e. none) to get even colour rendition across the frame.
mMontag wrote:
The only downside I see to the Zeiss is the center distortion (not shown in this test) - don't want to shoot any straight lines with that lens.
I don't have the 21mm - most of the web posted images with that lens are of landscape subject - but of the Architectural or common straight featured images I've seen does show what is referred to as the "mustache distortion".
mMontag, I don't own it either - so we're on the same terrain there. My understanding was the strength of this lens was landscape (or at least that was the reason I was interested in it) - but any super-wide angle has problems reproducing the super straight lines of architecture, no?
Thanks, I remember the thread & actually, I thought that kind of distortion in wides was pretty common - regardless, I think if architecture was your intended target, a TS-E would be a better bet. I'm interested in this lens for landscapes (on a 5D II) - for which I think it seems ideal.
globalkiwi wrote:
Thanks, I remember the thread & actually, I thought that kind of distortion in wides was pretty common - regardless, I think if architecture was your intended target, a TS-E would be a better bet. I'm interested in this lens for landscapes (on a 5D II) - for which I think it seems ideal.
What center distortion are you refering to?!
