p.2 #1Kolari Ultra Thin conversion on Nikon Z6 / Z7
AcuteShadows wrote:
Looking at the cone of rays converging (roughly) at one point of the sensor, I'd say that for smaller apertures, you may get almost "pure" field curvature, but for larger apertures, the divergence of rays in the tangential plane induced by sensor glass thickness becomes relevant, and you get astigmatism (in addition to field curvature). Maybe someone should do a thorough calculation of these effects.
I think you are right. When stopping down astigmatism is usually reduced. Field curvature may still be present after stopping down, but that too will be reduced.
p.2 #2Kolari Ultra Thin conversion on Nikon Z6 / Z7
LarsHP wrote: AcuteShadows wroteLooking at the cone of rays converging (roughly) at one point of the sensor, I'd say that for smaller apertures, you may get almost "pure" field curvature, but for larger apertures, the divergence of rays in the tangential plane induced by sensor glass thickness becomes relevant, and you get astigmatism (in addition to field curvature). Maybe someone should do a thorough calculation of these effects.I think you are right. When stopping down astigmatism is usually reduced. Field curvature may still be present after stopping down, but that too will be reduced.
Is it reduced or just less visible due to a larger depth of field?
p.2 #4Kolari Ultra Thin conversion on Nikon Z6 / Z7
If you haven't seen this article by written by Roger Cicala from Lensrentals blog with lots of input from Brian Caldwell (the guy who designed the Coastal Optics 60 f/4 APO Macro lens), then you should read it. It explains the issues very well.
p.2 #5Kolari Ultra Thin conversion on Nikon Z6 / Z7
Thanks, Steve Spencer. (For some reason I can't quote from mobile. I get a message that I need at least 25 posts before I can reply here!)
Yes, I have read that article. Another article on LensRentals with test shots from Leica M wides on Sony camera made me think that the 28mm Summicron-M Asph II would work better on mirrorless than it actually does.
p.2 #6Kolari Ultra Thin conversion on Nikon Z6 / Z7
LarsHP wrote:
Thanks, Steve Spencer. (For some reason I can't quote from mobile. I get a message that I need at least 25 posts before I can reply here!)
Yes, I have read that article. Another article on LensRentals with test shots from Leica M wides on Sony camera made me think that the 28mm Summicron-M Asph II would work better on mirrorless than it actually does.
My 28mm Summicron-M Asph (11672) actually works fine on the Z7. Maybe it's because I try to focus at the hyperfocal distance rather than exact infinity.
p.2 #7Kolari Ultra Thin conversion on Nikon Z6 / Z7
Yes, I actually read your post from about a year ago where you said it works well with your Z7.
It could be interesting to hear what you see if you shoot a landscape wide open with focus at infinity in the extreme corner. In that scenario, I get reasonably sharp corners and sides, but totally blurred center because focus in the center is much, much closer.
p.2 #8Kolari Ultra Thin conversion on Nikon Z6 / Z7
LarsHP wrote:
Yes, I actually read your post from about a year ago where you said it works well with your Z7.
It could be interesting to hear what you see if you shoot a landscape wide open with focus at infinity in the extreme corner. In that scenario, I get reasonably sharp corners and sides, but totally blurred center because focus in the center is much, much closer.
Indeed, I do not use the Summicron 28mm wide-open at infinity. But I should test that when I get an opportunity.
p.2 #9Kolari Ultra Thin conversion on Nikon Z6 / Z7
Only two comments:
- How are people telling if similar poor corner performance for a given lens between Sony and Nikon is due to similar cover glass issues or just that the lens has poor corners? Wouldn't you need to shoot on a Leica body to really tell if this is because Sony and Nikon both impact the performance rather than it just being the characteristic of the lens on any camera?
- It seems like the UT performance improvement on the Z7 might still be limited by the fact that there is still 1.2 mm of epoxied on glass once Kolari swaps out the 1.1 mm Nikon cover with their 0.2 mm cover:
p.2 #10Kolari Ultra Thin conversion on Nikon Z6 / Z7
kwalsh wrote:
Only two comments:
- How are people telling if similar poor corner performance for a given lens between Sony and Nikon is due to similar cover glass issues or just that the lens has poor corners? Wouldn't you need to shoot on a Leica body to really tell if this is because Sony and Nikon both impact the performance rather than it just being the characteristic of the lens on any camera?
- It seems like the UT performance improvement on the Z7 might still be limited by the fact that there is still 1.2 mm of epoxied on glass once Kolari swaps out the 1.1 mm Nikon cover with their 0.2 mm cover:
So the UT conversion takes one from 2.3 mm to 1.4 mm rather than 1.1 mm to 0.2 mm. Still, presumably every bit helps?...Show more →
Lately Fred has been comparing a lot of lenses on both his Leica M10 and his Sony A7r IV. Those comparisons have been very helpful in knowing how a lot of lenses perform on Leica vs. Sony. Then Sony to Nikon comparison can be more meaningful.
p.2 #11Kolari Ultra Thin conversion on Nikon Z6 / Z7
mapgraphs wrote:
I've been using LTM, M and legacy F lenses, ultra wides, wides , normal and short tele on my Z and M-240 since November 2018. There is no difference in the IQ between the two. I haven't seen any Z or R user reports of the types of problems encountered with E bodies since 2018. The mod may be a solution looking for a problem...
I compared my M240 to my Z7 w my Leica 50mm Asph, 35mm cron Asph, 28mm Elmarit Asph, 18mm Asph and every one was better on the Leica.
p.2 #13Kolari Ultra Thin conversion on Nikon Z6 / Z7
@ kwalsh
The exact epoxied-on-sensor glass thickness is unclear, so Diglloyd is just guessing. Even if he's right about that measure, we don't know the exact total thickness of the Leica sensors either, but I doubt Leicas epoxied layer is much different from the Sony sensors. Since Kolari Vision has measured the removable Sony glass to 1.75mm and the total thickness is estimated to about 2.3 or 2.5mm by LensRentals, the epoxied layer might be around 0.7. If that's true, then the total thickness of an UT converted sensor will be about 0.9mm.
In short, we only know how thick the removable layers are and have to rely on side by side tests. In my case, I don't have a digital Leica M for comparison, so I use my full spectrum converted Sony a7 (while having a hot mirror filter on the lens), and the 28mm Summicron-M Asph II simply looks stellar on that camera. Really impressive. So I feel confident that a Ultra Thin conversion will improve the sides and corners on my Z6 quite dramatically.
When focusing the 28mm Asph II in the extreme corner at 1.5-2m (6 feet), the sharpness @ f/2 on the a7FS is jawdroppingly good, but depressingly mediocre on the Z6. How it will be on a Leica, I don't know, but stellar on a UT converted Z6 is good enough for me
The reason why I want to convert my Z6 instead of buying a Leica M camera is primarily the 5 stops of built in image stabilization. If the M 240 had almost any kind of IBIS, I wouldn't have spent a second on considering doing the conversion.
p.2 #14Kolari Ultra Thin conversion on Nikon Z6 / Z7
@ AcuteShadows
I don't shoot landscape wide open either, but the described test will show the amount of field curvature and if it is reverse or not.
Reverse field curvature will diminish or in extreme situations even cancel out background blur. When shooting a portrait for instance, it will reduce the background blur outside the image center relative to the used f-stop. In other words, the extra you put in having a f2 or f1.4 lens will not be visible (or at least less visible) in the image compared to a f2.8 or f4 lens.
p.2 #17Kolari Ultra Thin conversion on Nikon Z6 / Z7
@Steve and @LarsHP - Thanks a bunch for the explanations/information. That all makes sense now.
@LarsHP - Since you mention wide angle, edge/corner performance and five stop IBIS just be sure to remember that IBIS really does not work very well at all on the edges/corners for wide angle lenses. So if you'll be getting much less than five stops in those use cases.
Here is the Z7 using the 14-30/4S shot at 30mm and 1 second handheld with a pinhole in front of a flashlight as the target. Below are crops from the full images showing the amount of motion corrected. Top row is center with IBIS off, middle row is center with IBIS on and the bottom row is the upper left corner with IBIS on. Note how in the corner IBIS is barely effective while in the center it is quite effective.
Below is the same setup, also one second, now at 14mm. Both rows are with IBIS on, again the upper row is the center the lower row the upper left corner. The center is doing a beautiful job here at essentially four stops of IBIS performance, the corners meanwhile are perfectly useless.
Anyway, just something to be aware of. Naturally you won't always being using IBIS, wide angle and care about the corners at the same time. But don't imagine you'll be able to do UWA landscape or architecture with critical corners and have IBIS help you much at all.
p.2 #18Kolari Ultra Thin conversion on Nikon Z6 / Z7
The 28mm Elmarit-M Asph is plainly awful on a Nikon Z camera. Sharp in the center, but increasingly blurred the more away from the center you look. I got some test shots from a Z7 user at DPreview and it was very demonstrative of how bad a Leica M wide angle can be on a mirrorless camera. It's a definite no-go with a Nikon Z7 and allegedly even worse on a Sony a7 series camera.
p.2 #19Kolari Ultra Thin conversion on Nikon Z6 / Z7
@ kwalsh
Thanks for posting the tests.
I wasn't aware of that limitation from the IBIS, but have seen some odd things going on in the sides and towards the corners in some images that might be explained by this.
Are the center and corner crops from the same image (w/ IBIS on) in each vertical row?
p.2 #20Kolari Ultra Thin conversion on Nikon Z6 / Z7
@LarsHP - No each crop is a separate exposure. That's why I did 9 or 10 of each because of course with the human body things are a bit different each time so wanted to see the variation and distribution of results to make any conclusion. I'd need to dig out another pinhole and then place it carefully to do a test showing performance in the center and edges simultaneously. That said I've done about 300 exposures at different shutter speeds and focal lengths and they all show the same effect. Also see below for a link to another thread where someone else shows the impact in a single exposure.
An important thing to note is that you can see in the corners the uncompensated blur is primarily in the radial direction. This is because the camera is correcting for angular motion by shifting in the image plane and when we have a rectilinear projection the same angular motion translates to different shifts in the image plane depending on the distance from the optical axis. For normal to telephoto cases where the angle of view is not large this difference is small because tan(t)~=t but for wide angle that's not true and the error gets quite noticeable.
Worth noting that for an equidistant fisheye projection this wouldn't be an issue because in that case the projection results in equal displacement across the frame for an angular displacement.
This topic just came up recently in a decent thread over at DPR:
The OP there has laid out the math nicely and another contributor distilled the results down into a fairly nice graph showing how the impact varies with focal length. The OP over there also posted a nice single exposure shot using a string of Christmas lights clearly showing the effect across the frame.