The GetDPI Photography Forum

Great to see you here. Join our insightful photographic forum today and start tapping into a huge wealth of photographic knowledge. Completing our simple registration process will allow you to gain access to exclusive content, add your own topics and posts, share your work and connect with other members through your own private inbox! And don’t forget to say hi!

ZA 24-70 test on SLRgear

douglasf13

New member
SLR gear updated its 24-70 test to include FF results, so I thought that I'd post a comparison with the Nikkor 24-70. Both seem like great lenses, with different strengths/weaknesses. The Nikon was shot on the D3, and the Sony on A900, and I'm not sure if the pixel count skews results. Enjoy. :)

ZA 24-70

http://www.slrgear.com/reviews/zproducts/nikon24-70f28/ff/tloader.htm

*edit. whoops, i accidentally put the Nikkor D200 link in. link fixed.
 
Last edited:

Quentin_Bargate

Well-known member
Hi Douglas,

The difference in pixel counts must render the Nikkor results highly suspect, even though I am sure its a great lens. Pretty close results.

Quentin
 

jonoslack

Active member
Thanks Douglas
I've used both on both (as it were). I think the results simply prove the point that test results aren't everything.

Quite clearly an Alfa 159 is an inferior car to a Ford Mondeo . . .but . . .but . . . but!
 

douglasf13

New member
Agreed, Jono. I think the best things about these tests are just the info about which focal length and aperture performs better than others. I now realize that, for landscape at 70mm, f11 is the way to go.
 

docmaas

Member
Having just bought the sal-2470 I'm a little disconcerted by these tests. They sort of confirm my earlier observations that the corners never do really get good. Aside from that I think a great deal of the areas where the zeiss seems poorer at higher apertures is probably due to diffraction which may impact the smaller pixels on the Sony more.

How does anyone whose shot both judge them on aspects other than resolution?

Mike
 

douglasf13

New member
I'd say that the corners look pretty good accept wide open or at 70mm until f11, which is what most have been saying all along. A zoom lens that is sharp in the center, yet soft in the corners at wide aperture, and pretty sharp all over at smaller apertures is pretty useful for my uses, although I am surprised the lens didn't test a little better wide open at 50mm and 70mm. Clearly, Zeiss/Sony made a decision for maximum center sharpness at the cost of corner sharpness (until smaller apertures,) which seems different from the Nikon design.
 

edwardkaraa

New member
In my opinion, the difference between 12 to 24 mp is huge. One lens that looks very sharp on 12 mp may very well lose the sharpness and become crappy on 24 mp. A good lens and a mediocre lens may perform equally well on a low count mp sensor. Moreover, these tests are all done on a flat target, so they tell you how a lens performs if all what you do is shoot painting reproductions. They are all taken at 1-2 meters distance, so there is no way to know how a lens performs at infinity or at closer distances. Some lenses are optimized for close distances, others for infinity, they don't tell you that in the lens litterature, nor in the tests. Field curvature has been very smartly used by Carl Zeiss since the 19th century to produce 3D effects that other manufacturers have been trying to eliminate thus giving you a flatter field corner to corner with no life nor character. Moreover, even 2 sensors of similar pixel count cannot be compared across systems, because of different AA filter strength, and demosaicing algorithms that may favour softness and film-like rendition (Sony jpg/IDC) or sharpness at the expense of stairstepping and artifacts (Canon jpg/DPP). All these tests are very intertaining, but in the end of the day, they are very useless.
 
Last edited:

Jorgen Udvang

Subscriber Member
In my opinion, the difference between 12 to 24 mp is huge. One lens that looks very sharp on 12 mp may very well lose the sharpness and become crappy on 24 mp. A good lens and a mediocre lens may perform equally well on a low count mp sensor. Moreover, these tests are all done on a flat target, so they tell you how a lens performs if all what you do is shoot painting reproductions. They are all taken at 1-2 meters distance, so there is no way to know how a lens performs at infinity or at closer distances. Some lenses are optimized for close distances, others for infinity, they don't tell you that in the lens litterature, nor in the tests. Field curvature has been very smartly used by Carl Zeiss since the 19th century to produce 3D effects that other manufacturers have been trying to eliminate thus giving you a flatter field corner to corner with no life nor character. Moreover, even 2 sensors of similar pixel count cannot be compared across systems, because of different AA filter strength, and demosaicing algorithms that may favour softness and film-like rendition (Sony jpg/IDC) or sharpness at the expense of stairstepping and artifacts (Canon jpg/DPP). All these tests are very intertaining, but in the end of the day, they are very useless.
Amen to that. Tests often seem to find flaws that have limited significance in real life, and fail to to evaluate features like field curvature, or even a simple thing like focus throw and precision. Most testers seem to hope for lenses that are universally good, but very few, if any, lenses are. It's understandable, since most people can only afford one lens within each focal length, but for enthusiasts and pro's, the 200 gram, 10-500mm f/1.2 Macro-portrait will remain a pipe dream.
 

docmaas

Member
All good points Edward. I don't really understand the field curvature impact/issue yet but I didn't realize the targets were so close.

I set up a bellows attachment that I used to compare a Canon 5D to a Sigma SD10 for resolution and enlargement using a Schneider 150mm f2.8 Xenotar lens. The SD10 was the equal of the 5D at iso200 and below at 5D native resolution but above that the 5D left the SD10 behind. I just wonder why one of these testing houses can't set up a similar test to compare the systems independent of the lens.

thanks,

Mike

In my opinion, the difference between 12 to 24 mp is huge. One lens that looks very sharp on 12 mp may very well lose the sharpness and become crappy on 24 mp. A good lens and a mediocre lens may perform equally well on a low count mp sensor. Moreover, these tests are all done on a flat target, so they tell you how a lens performs if all what you do is shoot painting reproductions. They are all taken at 1-2 meters distance, so there is no way to know how a lens performs at infinity or at closer distances. Some lenses are optimized for close distances, others for infinity, they don't tell you that in the lens litterature, nor in the tests. Field curvature has been very smartly used by Carl Zeiss since the 19th century to produce 3D effects that other manufacturers have been trying to eliminate thus giving you a flatter field corner to corner with no life nor character. Moreover, even 2 sensors of similar pixel count cannot be compared across systems, because of different AA filter strength, and demosaicing algorithms that may favour softness and film-like rendition (Sony jpg/IDC) or sharpness at the expense of stairstepping and artifacts (Canon jpg/DPP). All these tests are very intertaining, but in the end of the day, they are very useless.
 

edwardkaraa

New member
All good points Edward. I don't really understand the field curvature impact/issue yet but I didn't realize the targets were so close.

Mike
Mike, field curvature is when the subject to sensor distance doesn't correspond to the focusing distance. The subject may be at 2 meter distance, but you have to set the lens to let's say 1 meter to make it sharp. It occurs usually at the frame periphery and mostly moves further away from the camera, making an arc shape (not always, some lenses such as the ZA zooms have a W shape curvature). The practical outcome is that when the center is focused at infinity, you may need to focus at a closer distance (the amount depends on the particular lens) to get sharp corners. Some extreme corners need less that 1 meter distance to be sharp on some lenses. If you know your lens well, you can use hyperfocal focusing and an appropriate f/stop to get every thing sharp in the photo. The common mistake is to use the center AF point to focus at infinity. Hope this makes sense :shocked::wtf:
 

edwardkaraa

New member
Btw, one of the advantages of field curvature if used properly is getting a very sharp center foreground and equally sharp peripheral background which adds to the 3D effect that Zeiss lenses are known for.
 

jonoslack

Active member
How does anyone whose shot both judge them on aspects other than resolution?

Mike
Hi Mike
I've shot them both - I was using the Nikkor 24-70 from last January until October, where it was certainly my most used lens. Since then I've been shooting the ZA 24-70 on the A900.

I'm sure Edward is right about the curvature of field. I really really really mind about sharp corners in my landscape work (which is often little slivers of land at the bottom of the frame with a big sky above). I immediately noticed that the Zeiss lens was not as sharp as the Nikon at close quarters with flat subjects. But for landscapes where the peripheries are often closer than the focal point (think foreground and corners) it seems to be at least as good. An extreme example of this is the Zeiss ZF 25mm f2.8, where the field curvature is extremely obvious (too obvious IMHO).

So, if you're thinking of using the ZA 24-70 for repro work, and for closeups of flat surfaces, then it probably isn't the best answer (but then it wouldn't be, even if it did have a flat field).

For me, this epitomises the difference between the A900 and the recent Nikons. As my wife once said about Ealing, "It has everything you need . . . but nothing you want". It seems to me that Nikon are concentrating on ticking all the testing boxes you need, whilst Sony (who would have believed it of them) and Zeiss are producing kit which takes the pictures you want.
 

edwardkaraa

New member
Two more extreme examples of how it is possible to get sharp corners. Of course with normal distances the DOF is so much greater that you can get fairly everything in focus with the right f/stop. Here the front lens was almost touching the bed and I was much closer than the lens MFD of 28cm in order to get corners in focus. Elementary stuff, but it seems not everyone is aware of it.

View attachment 15878

View attachment 15879
 
Last edited:

douglasf13

New member
Excellent points, Edward, and they outline the reason that I posted the graphs in the first place, although, admittedly, adding the Nikon graphs was a mistake, because I accidentally made this into a contest, for which I apologize. These tests were relevant to me, because, although I've been aware of the field curvature of the 16-35, I wasn't aware that there were similar design elements to the 24-70. This is mostly my fault, because I've had the lens over a year, and the curvature isn't noticeable on APS-C. Since receiving the A900 at the end of last year, I've only shot people with the lens, so I haven't done much testing that would show the curvature. I appreciate your info on this, because my upcoming trip to Hawaii with that lens would have probably taken me a bit by surprise. :bugeyes: Cheers.
 

DavidL

New member
I may have a solution, well it would be for me.
I don't want 24mp, honest, so treating the Sony as a 1.3 crop camera and doing the necessary additional zooming by walking, I may well have sharpness edge to edge. I don't always make sense but think I do on this. Actually If there was a fixed 21mm f2.8 Zeiss then I would get a A900 with the 135 f1.8, then I could be a happy man. Well contented for a while with an old Minolta 28-105 to keep me happy in the middle while snapping around.
David
 

edwardkaraa

New member
Thank you very much Douglas. I believe that understanding how our lenses work helps us in making better photos with them. Don't think that the 24-70 or the 16-35 are not able to resolve sharp corners. The photo below was taken at 22mm and f/2.8 on last weekend. The couple are friends of mine so I hope they don't mind me posting their photo here. At this f/stop, the corners are not supposed to be sharp, however, the top right corner is extremely sharp by any standards (at 100% magnification). Curiously the focus is exactly on the man's face which is also very sharp. The wall behind is blurry of course, and gets gradually sharper until you reach the top right corner where the focus is just right on. Field curvature is indeed a very fascinating characteristic, notably with Zeiss designs.

View attachment 15889

Excellent points, Edward, and they outline the reason that I posted the graphs in the first place, although, admittedly, adding the Nikon graphs was a mistake, because I accidentally made this into a contest, for which I apologize. These tests were relevant to me, because, although I've been aware of the field curvature of the 16-35, I wasn't aware that there were similar design elements to the 24-70. This is mostly my fault, because I've had the lens over a year, and the curvature isn't noticeable on APS-C. Since receiving the A900 at the end of last year, I've only shot people with the lens, so I haven't done much testing that would show the curvature. I appreciate your info on this, because my upcoming trip to Hawaii with that lens would have probably taken me a bit by surprise. :bugeyes: Cheers.
 
Last edited:

douglasf13

New member
Great example, thanks. I've been using primes a bunch this year, but it looks time to go give my 24-70 some proper testing on the A900. I think your theory about Zeiss' field curvature relating to "3D-ness" is an interesting one.
 

jonoslack

Active member
I may have a solution, well it would be for me.
I don't want 24mp, honest, so treating the Sony as a 1.3 crop camera and doing the necessary additional zooming by walking, I may well have sharpness edge to edge. I don't always make sense but think I do on this. Actually If there was a fixed 21mm f2.8 Zeiss then I would get a A900 with the 135 f1.8, then I could be a happy man. Well contented for a while with an old Minolta 28-105 to keep me happy in the middle while snapping around.
David
HI David
You know me - I'm obsessive about sharp corners, and in most circumstances I'm getting them with the 24-70 . . . except on flat field at close distances.
 

edwardkaraa

New member
I think your theory about Zeiss' field curvature relating to "3D-ness" is an interesting one.
I think it is a combination of 2 factors, both considered as flaws in this world of computer designed very well corrected modern lenses: field curvature and LoCA. The effect is mostly visible with the 85 and 135 that produce in many photos a dark outline around the subject, almost like pencil drawn, that helps separate the subject from the OOF background. This has been discussed in detail on FM sometime ago in regards to Contax lenses. It's like an accentuated PS edge sharpening, only optical. But of course this and field curvature produce poor results in controlled tests :D

This must be the reason why so many people wonder why Zeiss still insists on introducing "flawed" lenses that perform so poorly on testing benches and compare unfavourably to the competitors, but simply these flaws are what makes Zeiss glass stand out in real life photos and gives it its distinguished look.
 
Top