Big Nikon 17-35 v Contax 17-35 comparison going on today (the converted 17-35N arrived this morning). Focus accuracy is rather hit and miss with the 5D – not enough to spoil everyday shots, but a pain in testing.
I've never seen a pair of lenses show such a close relationship to each other: they're more like twins than cousins. Given that the Nikon predated the Zeiss by two years, one wonders whether Zeiss used the Nikon as a benchmark. Both are ten group designs with three aspherical lenses (the Nikon has 13 elements; the Zeiss has 10), but they have exactly the same peaks and troughs in performance. I've seen more variation between samples of the same design than I have between the Nikon and the Zeiss at 21mm thus far . . . more details coming.
The Nikon has 13 elements in 10 groups. Two of the elements are ED glass, 2 crown-molded aspherical, 1 hybrid aspherical.
The Nikon is famous for the radically shaped, large-diameter aspherical front element. Look at the lens construction diagram at the end of this web page:
Zeiss has 15 elements in 10 groups, which "incorporated 3 aspheric surfaces along with lens elements from special types of optical glass". There is no mention whether this is 2 aspheric elements one of which with both surfaces aspherical, or how many lens elements from "special types of optical glass" there are.
Your comment made we wonder if Zeiss copied Nikon outright to take revenge for what Nikon did to them after WWII (although you did not suggest that), but comparing the two lens construction diagrams, they seem to be very different optical designs. The Zeiss design is original.
My mistake . . . misread the CZ 17-35 description chart: 15 elements, not 10.
I'm not done even with the first stage test yet, but strong similarities have emerged: the lenses have similar optical 'signatures', and both appear to have what I can only describe as 'non-linear', waveform focus fields at 17mm, which curiously are mirror images of each other. I've only previouisly seen Nikon lenses behave that way.
They have very similar distortion characteristics; in many ways both peak at around 21mm and deteriorate at longer focal lengths - in other words, both were optimised for wide performance (unlike the 16-35L). They are both 17-35/2.8. They're similar in size and weight . . . at this point it's really tough to choose a favourite!
Much testing remains to do, but part one should be online tomorrow.
Does anyone have a schematic of the optical design of the Nikon lens? I don't suspect that there are as many similarities as you are suggesting (or even that one might be a copy of the other). Most ultra-wide angle zoom lens with have similar numbers of elements in similat numbers of groups. (Canon's 16-35 f2.8 zoom is 14 elements in 10 groups, and the canon 17-35 is 15 in 10 groups).
Zeiss, quite differently than Nikon employs a much bigger front element (Zeiss flter size is 95mm and Nikon is only 77mm). That's going to make a significant difference in design, and I would suspect (if we can ever find a Nikon lens schematic) that other differences are pretty significant as well. Chances are slim to none that Zeiss actually "copied" the Nikon lens. Was it a benchmark? I think is probably unlikely as well. If the Nikon only appeared 2 years before the Zeiss, Zeiss was probably already working on their lens before the Nikon appeared.
I've no way of knowing what goes on at briefing meetings for the creation of a new lens design, but I'd be surprised if they didn't have on the table the best currently available. I'm frequently struck by the similarity between Nikon and Zeiss lenses – in practice – and similarly, it seems to me that Leica's 'signature' has more in common with Canon's than anyone else. Whether this is simple convergence, or coincidence, I couldn't say.
Shooting a Canon 16-35L and the Contax 17-35N back to back gives discernibly different results, but the Nikon 17-35 already has a certain 'Zeiss-ness' about it. The schematics are quite different, but by accident or design their aesthetic intentions seem to have been similar. http://www.16-9.net/schematics.gif
Nikon's ED elements are marked yellow; aspherical elements blue.
Here's weird for you . . . in testing several Nikon lenses, I've noticed irregular waveform variations in resolution across the frame: ripples in the focal plane. The wider the lens, and the wider the aperture (the thinner the focus), the more obvious the problem becomes. At smaller apertures, peak resolution decreases and the frame edges improve, stretching the MTFs flat, and minimising this variation.
Comparing the Contax N with the Nikon AFS, which both suffer from the problem to some extent at 17mm, is particularly difficult because their focal planes are diametrically opposite shapes. Rather than try to describe the phenomenon, I've had to resort to a diagram. This is how both lenses compare at 17mm, compared with a typical flat field in which the image circle uniformly decreases in resolution as it moves away from the centre frame.
http://www.16-9.net/fieldfocus.gif
In regions A and C, the Contax at 17mm is sharper, at all apertures. In region B, the Nikon is visibly better, even at f11. It's interesting to compare this fag-packet sketch with the official MTF charts at the top of this thread, which suddenly make sense!
Like the Nikon, my Canon 24mm L similarly has a hole in the middle: an area of low resolution centre frame that takes until f5.6 to 'heal over'.
I read somewhere that distortion and resolution are coupled. Regarding Zeiss Distagon lenses this would make sense, they are known rather for sharpness than missing distortion.
Here's a quick shot from last night, shot at the long end (34mm) indoors without flash, f/2.8 and 1/30 sec. I like the sharpness, color, bokeh, and the vase on the right edge is very sharp. It is only slightly cropped, I took off a bit from the top and right sides only. http://www.pbase.com/sparkytom/image/73178571.jpg
you can see the original here if interested.
I found that by 35mm, mine is much improved relative to the Nikon and Canon equivalents: the performance seems to dip a bit at 24mm, but comes on strong again at the long end.
hubsand wrote:
I found that by 35mm, mine is much improved relative to the Nikon and Canon equivalents: the performance seems to dip a bit at 24mm, but comes on strong again at the long end.
It seems quite interesting, hubsand. Waiting for the final test report.
Thanks again for your effort!
Emanuele, I think hubsand was nuts to sell it. I know I will keep mine. It's a thrill to have a true EOS autoeverything lens by Zeiss, especially on the wide end where Canon is so terrible.
tom in mpls wrote:
Emanuele, I think hubsand was nuts to sell it. I know I will keep mine. It's a thrill to have a true EOS autoeverything lens by Zeiss, especially on the wide end where Canon is so terrible.
Well, reading his auction, he describes this lens as a stunning performer (something like the best wide zoom on earth) so I was just curious to know why he decided to sold it. I suppose that hubsand's first goal was to test this alternative solution: test ended so no more interest in having the lens. This only my humble opinion.
I am not sure if hubsand is selling because of the region A, B and C thing.
He has established that, at 17mm, Contax wins in regions A and C, but Nikon wins in region B. In terms of total area, region B is much larger than regions A and C combined. That means, over most of the image frame, Nikon is the sharper lens at 17mm, "so I am gonna sell the Contax and keep the Nikon".
BUT...
Although the Contax did not defeat Nikon in region B, it is totally usable across regions A, B, and C. Nikon's region C showed a significant drop. In comparison, the Contax is a true 1.0x crop lens, but the Nikon begins to look more like a 1.3x crop lens.
