Dan1 wrote:
An interesting observation - I think I see what you mean.
Yes I agree with this, if I may add that the resolving should be in the colours, and not only between light and dark. I found my 35L to "resolve very fine detail", but it was totally flat and uninteresting. The Contax 35/1.4 was a first round TKO. Slightly less in greyscale, but, oh, the colour separation!
A further consequence of this argument is that the characteristics of a lens canīt be separated from the resolving power of the sensor. It will be interesting to follow how Nikonlanders will find their lenses after taking the step from 12 to 36Mp, among peepers in particular.
Z250SA wrote:
A further consequence of this argument is that the characteristics of a lens canīt be separated from the resolving power of the sensor. It will be interesting to follow how Nikonlanders will find their lenses after taking the step from 12 to 36Mp, among peepers in particular.
Samuli, I've taken a look at your image, but I'm still not sure what you mean by
"having large DOF caused by mechanical vignetting resulting to smaller effective aperture in corners will ruin most of my photos". Maybe you could explain this a bit further.
johnahill wrote:
I've not been following this one much, what's the current opinion of this lens now that folks have had time to try it?
I've got the latest R28 and a Zf25/2.8 that both perform very well for landscape shots, but i'd really like to have something with a native EF mount.
How does the new 25 perform at near infinity stopped down to say f5.6 or f8?
Well, I and others like Roger C. have found that stopped down infinity distance landscape is not the strong suit of this lens.
It smudges the extreme corners which is ok for some and not for others.
It does better at close and medium distances. At infinity distance landscape stopped down I prefer my ZE 21/2.8.
Bobu wrote:
Samuli, I've taken a look at your image, but I'm still not sure what you mean by
"having large DOF caused by mechanical vignetting resulting to smaller effective aperture in corners will ruin most of my photos". Maybe you could explain this a bit further.
BorisSure, open these two images to two tabs: f/2 and f/11
Even at this miniature webthumbnail size both of the top corners are ruined in f/2 photo since the DOF in corners is at least f/5.6 equivalent due to aperture being effectively much smaller than f/2 in middle of the lens. Due to this "feature" I lost tens of good images and now I just try to avoid shooting 2/28 in the woods, where I would have sky against branches in top corners of image very often.
In order to fix I would need to:
- blur the corners
- lower the contrast in corners
Too much editing, I'm lazy, sorry...
Thanks Samuli, I see now what you mean. Are you sure that this is the result of "mechanical vignetting" and not simply based on field curvature?
By the way before you told me where to look in your image I've not found the effect disturbing at all. Maybe you should just stop looking for some strange DOF effects in corners ;-)
But the truth is that I fully understand you, when I know there is a certain weakness in a lens I can't stop looking for this effect in the images shot with this lens.
Bobu wrote:
Thanks Samuli, I see now what you mean. Are you sure that this is the result of "mechanical vignetting" and not simply based on field curvature?
It can be both but it's at least having larger DOF in corners, this is what makes it very annoying since it doesn't matter are the branches against sky 3 meters or 50 meter they are always in focus and have great contrast (even may have really poor sharpness). Also the DOF highlights show "clipped circle" shape in the corners.
Bobu wrote:
By the way before you told me where to look in your image I've not found the effect disturbing at all. Maybe you should just stop looking for some strange DOF effects in corners ;-)
But the truth is that I fully understand you, when I know there is a certain weakness in a lens I can't stop looking for this effect in the images shot with this lens.
Boris
>75% of my photos are based to my lens being able to draw the subject with high contrast and drawing the DOF with lower contrast, to focus the viewer to the subject (in this case the tree on foreground, into which I focused) - this is the reason why I shoot with Zeiss and Zeiss only. The f/2 version works really well except my eye is drawn to top of the image, otherwise it's just what I would want the lens to draw - if you compare the contrast and colors in DOF the f/11 has much stronger background and the picture no longer has same impact, in matter of fact I don't like at all the f/11 since it's just ordinary snapshot in the forest - a small landscape. And if the f/11 photo tries to be good it would have to be about the whole scene and not about the main subject (in which case I would have composed differently, more emphasizing the curve of the steam etc.)
I don't anymore print that much (I think my printer is jammed for not being used since last summer) and I mostly present my photos from 2560x1600 monitor or 55" LCD TV (1080P, naturally color calibrated ), and even at these sizes I and some viewers find this kind of corners disturbing, and if I would print this to A3 size then it would be really annoying.
Portrait (format, not subject) photos I would mostly prefer to compose them to 10:8 but I find it rather difficult without anything in viewfinder to help to compose my image well to this frame ratio - for me it's just not the same to crop afterwards. With landscapes I prefer sometimes 16:9 (and in live view movie mode I can frame to 16:9), and that might be enough to cut the corners off.
EDIT: Found the images from Apple Aperture; the f/2 image was actually f/3.2, so I had tried to mitigate the effect but even closing down 1 1/3 stop the corners are still like this!!! They are much worse wide open (if I remember correctly the vignetting makes the contrast even worse and these kind of "artefacts" show up more easily).
Samuli, I agree with most everything you wrote, especially I also vastly prefer the f/2 (f/3.2) version of your image. But I think you overemphasize the effect of the sharp upper right corner. My eye is still drawn on the tree in the foreground, at least at these web sizes. But I understand your point.
wayne seltzer wrote:
At infinity distance I could not get the extreme corners to be sharp at f8 on a planar church wall, even if adjusted focus using liveview on the corners. So I don't think it is just field curvature.
It is clear from the MTF that you need to choose between the best tangential focus or the best sagittal focus. Or something in between. You cannot get both in simultaneous focus.
At close distances it is fine.
Can you show some tack-sharp extreme corners at close distance?
Toothwalker, I am talking about the extreme corner smearing problem. I never saw it when taking pictures closer than infinity but when taking infinity shots with a level camera the extreme corners got smeared.
What aperture did you use above?
wayne seltzer wrote:
Toothwalker, I am talking about the extreme corner smearing problem. I never saw it when taking pictures closer than infinity but when taking infinity shots with a level camera the extreme corners got smeared.
What aperture did you use above?
I used the same aperture (f/2) and picture size that you used to demonstrate close-up performance.
Toothwalker wrote:
I used the same aperture (f/2) and picture size that you used to demonstrate close-up performance.
At the small sizes used it's hard to judge corner softness and other things. However it looks like the vignetting is larger at distance than it's close-up. Typical to Zeiss SLR lenses, specially 100MP has surprised me many times how big the difference is.
Ok, is the focal ring on finity mark? I also found that tilting the camera can help the corner smearing problem , so I would like it if you shot the building face square on with the top extrem corners still covering part of the building, and then produce 100% crops of the extreme corner.
wayne seltzer wrote:
Ok, is the focal ring on finity mark?
No. My lens focuses way past infinity.
I also found that tilting the camera can help the corner smearing problem, so I would like it if you shot the building face square on with the top extrem corners still covering part of the building, and then produce 100% crops of the extreme corner.
If tilting the camera improves performance, you have a ramshackle lens. Or how would you explain this?
Is there a reason why you need 100% crops to judge distance performance, whereas smaller images
are fine for close-ups?
Samuli Vahonen wrote:
At the small sizes used it's hard to judge corner softness and other things.
Yup.
However it looks like the vignetting is larger at distance than it's close-up.
Can I see a 100% crop of the bottom extreme corners of your shot, I think I see smearing in the murky, vignetted corner. Hard to tell at this size, 100% crop is easier to see the problem.
The midday landscape shot I posted earlier in the thread where I tilted the camera down so that the top corners would be on the background mountains ended up having slightly better extrem lower corners in the foreground than when not tilted. I don't remember if I posted the non-tilted shot or vice-versa.
Anyway, what are you trying to do here. Dispute my corner smearing problem with the lens which we have demonstrated by other people as well as me or what?
Or my statement about the lens being better at closer distances? The first thing has already been proven with different copies. The 2nd thing would require shooting targets at different distances outside.
I don't have the lens anymore, I was renting it, but maybe you could take a landscape shot in a forest with focus point being a close tree in the foreground and we will then get to see if it works for Samuli.
To test what Samuli was talking about requires shooting in a place where the trees are regularly spaced apart in increments so you can see the DOF affects.
wayne seltzer wrote:
Can I see a 100% crop of the bottom extreme corners of your shot, I think I see smearing in the murky, vignetted corner. Hard to tell at this size, 100% crop is easier to see the problem.
The midday landscape shot I posted earlier in the thread where I tilted the camera down so that the top corners would be on the background mountains ended up having slightly better extrem lower corners in the foreground than when not tilted. I don't remember if I posted the non-tilted shot or vice-versa.
Anyway, what are you trying to do here. Dispute my corner smearing problem with the lens which we have demonstrated by other people as well as me or what?
Or my statement about the lens being better at closer distances? The first thing has already been proven with different copies. The 2nd thing would require shooting targets at different distances outside. ...Show more →
Of course the corners of my shot are blurred. There is no dispute about the soft corners of this lens at large apertures. The MTF is clear and we have seen many 100% crops in this forum to illustrate this. I am responding to the practice of showing 100% crops to address the corner issue at infinity, and significantly downscaled images to show that the corners are fine at close distance.
There have been statements that the D25/2 is optimized for close range and that the corners are better at close distance than at infinity. The former is not true and the latter may or may not be true, but so far I have not seen any evidence in this forum in support of that statement.
I don't have the lens anymore, I was renting it, but maybe you could take a landscape shot in a forest with focus point being a close tree in the foreground and we will then get to see if it works for Samuli. To test what Samuli was talking about requires shooting in a place where the trees are regularly spaced apart in increments so you can see the DOF affects.
From his description I reckon he is talking about optical vignetting, which decreases the blur of tangential details in the corners. All normal photographic lenses do that, some more than other ones.
Ok, well the statement about being optimized or better at closer distances would have to be proved by shooting targets at different distances like I said before to prove or disprove.
The 2nd point about the heavy blurring in the extreme corners or corner smearing even at stopped down apertures like f8, that only happened to me on shots that were focused far away near infinity and not on closer shots like upper corners of my bar shot or the Chili's table tile shot I just posted. I am not talking about the gradual fall off sharpness to the corners but the case when all of sudden near the extreme corners all detail quickly gets much much blurrier. I don't see that in the Chili's shot which is very close focus.
Toothwalker wrote:
Of course the corners of my shot are blurred. There is no dispute about the soft corners of this lens at large apertures. The MTF is clear and we have seen many 100% crops in this forum to illustrate this. I am responding to the practice of showing 100% crops to address the corner issue at infinity, and significantly downscaled images to show that the corners are fine at close distance.
There have been statements that the D25/2 is optimized for close range and that the corners are better at close distance than at infinity. The former is not true and the latter may or may not be true, but so far I have not seen any evidence in this forum in support of that statement.
From his description I reckon he is talking about optical vignetting, which decreases the blur of tangential details in the corners. All normal photographic lenses do that, some more than other ones.
