mapgraphs wrote:
M-240: 5952 x 3968 Pixels (23.62 MPixels) (3:2)
A7ii: 6000 x 4000 Pixels (24.00 MPixels) (3:2)
Edit:
My bad. To explain more concisely, if one takes the same lens, for same shot, with both bodies, the M-240 shot is cropped (sides, top and bottom) compared to the 6000x4000 image...
You are comparing pixel density not the size of the sensor. Just like we do not say a Sony A7 III (6000 X 4000 pixels) is a cropped version of a Sony A7r III (7952 X 5304 pixels) we should not say that the M240 is a cropped version of the A7 II. Neither the super small differences in the number of pixels (between the M240 vs, the Sony A7 II), nor the huge difference (between the A7 III and A7r III) tells us the image is cropped. You need to look at the actual size of the sensor and a 35.9 X 23.9 sense size is very common in FF cameras. As Lars has pointed out, it turns out the Sony is actually the cropped version of the Leica, but the crop is so small it is inconsequential. It is about a 1.008 crop and is so small that with the numbers the manufacturers provide it can't even be calculated precisely because the numbers are rounded too much.
Another excursion with the Ultron II. Something I did with the 75mm Nokton. This is looking at the rendering and edge of the frame with different points of focus, both near and distant (not infinity hard stop).
The Maremma pack wasn't out and barking so something of a tranquil lens swap. The fence/tree line runs diagonally to the distance, it is not parallel with the road so you can get an idea of the depth of field with a distant point-of-focus.
Both sets are Iso 200, auto AWB and wide open; unfortunately shutter speeds aren't matched so one set is plus one stop and the other set is minus one stop in post (it's not a one-to one comparison), note the crop with set two...
Fred Miranda wrote:
Against the light:
Warm Luau in Kona, Hawaii.
Love your family photos, it's bringing me back. If you get a chance, grab breakfast at Hawaiian Style Cafe in Waimea. Get the Stew Luau and a stack of pancakes. Most memorable location we visited was the Papakolea green sand beach, long drive but very worth the trip.
Arvaker wrote:
So, I posted a few examples from the Leica M10 until now. This time it is the Ultron on the Nikon Z6, @f2.4, i-ON-ii
Looks great as well at infinity, from f2.8 and smaller in extreme corners (daytime foliage; have not tested night sky).
It shows quite a bit of reverse field curvature on the stock Z6 - something I usually look for when checking M mount wides used on FF mirrorless cameras. To me this is expected, and it strongly suggests that Voigtländer will not make this lens in E mount (without an optical redesign at least).
Looks great as well at infinity, from f2.8 and smaller in extreme corners (daytime foliage; have not tested night sky).
On this tractor scene, did you notice how the left corner foliage sharpens up? That's the effect of induced field curvature due to using the CV 28/2 on a thicker sensor stack. There is also a difference in micro-lenses design between the sensors.
This will be apparent when shooting at wide apertures when the subject is a mid distance and not so much as close distance since blur will mask the issue.
Fred Miranda wrote:
On this tractor scene, did you notice how the left corner foliage sharpens up? That's the effect of induced field curvature due to using the CV 28/2 on a thicker sensor stack. There is also a difference in micro-lenses design between the sensors.
This will be apparent when shooting at wide apertures when the subject is a mid distance and not so much as close distance since blur will mask the issue.
Yes I am aware, it is not perfect on non-M-cameras, but man oh man, it is like night and day compared with the Zeiss Biogon ZM 28/2.8 adapted on mirrorless. It was a disaster until f8, and the extreme corners at f11.
And I think the Ultron only shows minor curvature, close to the edge, and it shapes up already at 3.5, and as good as gone by f5.6. Per example, SOOC jpg below is the Ultron 28/2, on the Z6, @f4:
Arvaker wrote:
Yes I am aware, it is not perfect on non-M-cameras, but man oh man, it is like night and day compared with the Zeiss Biogon ZM 28/2.8 adapted on mirrorless. It was a disaster until f8, and the extreme corners at f11.
And I think the Ultron only shows minor curvature, close to the edge, and it shapes up already at 3.5, and as good as gone by f5.6. Per example, SOOC jpg below is the Ultron 28/2, on the Z6, @f4:
Fred Miranda wrote:
On this tractor scene, did you notice how the left corner foliage sharpens up? That's the effect of induced field curvature due to using the CV 28/2 on a thicker sensor stack. There is also a difference in micro-lenses design between the sensors.
This will be apparent when shooting at wide apertures when the subject is a mid distance and not so much as close distance since blur will mask the issue.
Interesting. I went back and tried some field curvature exercises with the 28mm Ultron II on my Z7, tests for the thicker stack, and as with my 28mm Ultron f/1.9 LTM, there's no induced curvature - no field curvature penalty as shown on my A7. Completely different sensor structures between the Z6 & Z7 apparently...
(Edit: I could "dial-in and out" the induced field curvature with the 28 Ultron LTM and other wides on the A7)
Try a landscape scene, focusing in the corner at infinity. The central part of the image should still be sharp if there is no field curvature.
When comparing my (at the time) stock Z6 with Sony a7II, there was a slightly better performance with M mount wide angle lenses on the Z6, but no big difference.
Another and useful way to test for field curvature, since it will tell you if it is regular or reverse, is this:
1) Determine where exactly infinity is on the distance scale by using EVF magnification in the center point. Do it several times and check each time where (on the scale) you found it. You will then get a good idea where the best infinity focus is.
2) Do the same as in 1), but now by finding focus in at least two of the corners (on opposite sides).
If infinity focus in the corners is further away compared to where you found it in the center, you have regular field curvature.
If infinity in the corners is found at a shorter distance on the scale, you have reverse field curvature.
If infinity focus is found at the same place in both center and corners, you of course have a lens / camera combo that yields no field curvature (which is actually quite rare in my experience).
With my Sony I'd test by having a subject at 2-3m with structures 60-100/200m distant.
Depending on the lens, focusing on the near subject will bring the background into focus on the edges where it should be out-of-focus. It's very simple. Changing the subject distance alters how much is in focus on the edges, one can dial-in the effect. Variations in background distances will change the effect as well.
With my Sony I'd test by having a subject at 2-3m with structures 60-100/200m distant.
Depending on the lens, focusing on the near subject will bring the background into focus on the edges where it should be out-of-focus. It's very simple. Changing the subject distance alters how much is in focus on the edges, one can dial-in the effect. Variations in background distances will change the effect as well.
I am unsure what you are saying here when you start off saying "still sharp" and then continue confirming field curvature. Are you saying that you see no field curvature using the Nikon Z7, but you do see it with Sony?
I have seen full size images from Nikon Z7 with Leica 28mm Elmarit-M Asph with terribly awful reverse field curvature, so the Z7 certainly can't compare to a Leica M camera regarding sensor glass stack.
Check out post #4, comparison between M10/Z6/Z7. Naturally Z6 will give the worst rendering due to AA filter on top of the sensor incompatibility, the Z7 fair much better. Judging from the 21sem, the Z7 did pretty darn good judging from the corners of the building! It also seem to need a bit more sharpening and they'll be about equal in the center portion.
On a side note, the 21sem on M10 seem to have deeper dof, which I speculate it's due to the RF calibration - Leica don't calibrate their lenses to always let the focus point fall right in the middle of the focus plane on all apertures, some lenses simply impossible to do as the focus plane will shift backwards as one stops down. That's where RF become superior method to EVF focusing - lots of optimum focus plane thoughts put into the RF calibration.
So the Z7 does indeed induce some field curvature judging from the smeary edges. Though I'd be happy to have Z7 as an alternative body to use with M glasses with some penalty.
With my Sony I'd test by having a subject at 2-3m with structures 60-100/200m distant.
Depending on the lens, focusing on the near subject will bring the background into focus on the edges where it should be out-of-focus. It's very simple. Changing the subject distance alters how much is in focus on the edges, one can dial-in the effect. Variations in background distances will change the effect as well.
Regarding field curvature on non-Leica cameras, it should be noted that some M mount wide angle lenses perform awfully and others admirably, depending on the exit pupil distance and speed of the lens in question. Retro focus designs perform much better than symmetrical ones for this very reason.
It appears to me that the 28mm focal length is particularly troublesome. I speculate that it's because many of these lenses are not retro focus designs while having a naturally given short exit pupil. 21mm and wider lenses are typically retro focus, while 35mm lenses and longer don't need it.
Check out post #4, comparison between M10/Z6/Z7. ...
Interesting thread. After spending some time with Leica and Zeiss lenses, I no longer have high expectations for either, when used on an M. Pick the right lens and the right version and it's good to great (depending on the intent). I'm happy with my VMs.
For my own use, I want the out-of-focus parts of the image to actually be out of focus. Especially if I'm using the lens for reportage purposes - as it was designed to be used. So, I guess I'm talking figs to an apple and orange discussion... I can get edge to edge sharpness on the Z7 wide open, if I want it. More importantly, the background is out of focus, when I want it to be... I'm good. ; - )
mapgraphs wrote:
(...) I can get edge to edge sharpness on the Z7 wide open, if I want it. More importantly, the background is out of focus, when I want it to be... I'm good. ; - )
Yes. For landscapes, you can just stop down the lens and the problem is (largely) solved.
However, the reason why I am so concerned with field curvature is exactly what happens to background blur. Reverse field curvature reduces - or even cancels out - the background blur, and this happens when we use cameras with thicker sensor glass than the (wide angle or normal) lens was designed for.
The 7Artisans 50mm f/1.1 had so much reverse field curvature that the background blur was less than the Nokton f/1.5 I had (which also induces reverse field curvature on MLC, but not as much).
