MJKoski wrote:
Yea well, if you stop down let's say classic Z-series DSLR lenses, they are all critically sharp in the center and mid regions while the edges do suffer when paired to modern high MP cameras. Wide open is not that harsh. I tend to stop down the lenses to get acceptably sharp face instead of a fraction of eyelash or just the iris. What's the next step? To soften the image in post... Gosh! Instead of studio conditions, my opinion is that they are nice lenses for highly detailed content in the nature. Of course the opinion is wrong from some other PoV.
People tend to be more or less hairy (not totally plastic) without some care before a photoshoot. 50MP+ tends to show up some texture compared to 6MP or 12MP sensor cameras....Show more →
Just to help you understand more about Zeiss lenses....three of the most famous portraits lenses are Zeiss....the 50mm Biotar, the 75mm Biotar, and 50mm Planar. Oh...for Hasselblad there's also the Zeiss 150 Sonnar...and oh ...the absolutely classic Zeiss 80mm/2 Planar is also out there making people look hairy for decades in a way so in demand that the lens sells for thousands of dollars now. I'll leave out the 135 and 100mm Zeiss lenses that are also famous for hair magnification follicle accentualization portraiture.
I don't always use Zeiss lenses...but when I do bald people have a full head of hair all over again.
highdesertmesa wrote:
Note the Leica lens hits infinity when adapted much sooner than the CV, but on my M10-R/M10M they both hit infinity at the hard stop:
Interesting. My guess is that the sensor stack is so much thicker on the GFX that it affects where infinity ends up, depending on the distance to the exit pupil and the focal length. Which in that case means that optimal adapter thickness is not the same for all lenses even when these lenses are perfect on the mount/stack they were designed for.
Makten wrote:
Interesting. My guess is that the sensor stack is so much thicker on the GFX that it affects where infinity ends up, depending on the distance to the exit pupil and the focal length. Which in that case means that optimal adapter thickness is not the same for all lenses even when these lenses are perfect on the mount/stack they were designed for.
Not just thicker, but the cover glass is also spaced further away from the sensor than any other digital camera of which I'm aware at 9mm. The GFX cover glass acts almost like a last rear lens element, and IQ results (and apparently in some cases infinity focus) vary depending on the focal length and design of the adapted lens.
Makten wrote:
Interesting. My guess is that the sensor stack is so much thicker on the GFX that it affects where infinity ends up, depending on the distance to the exit pupil and the focal length. Which in that case means that optimal adapter thickness is not the same for all lenses even when these lenses are perfect on the mount/stack they were designed for.
This is going to just be due to two different focal lengths. A 0.5mm lens shift on a wide angle has a different effect on focus position (larger) compared to the same lens shift on a telephoto. Check any two Leica mount, unit focused lenses (that have same MFD) and the longer focal length lens will physically extend further than the shorter focal length lens when moved from MFD to infinity. Edit: talking about the in/out movement of the lens itself, not the pitch (degrees of turn) of the helicoid. The same distance shift (due to a wrong sized adapter) will have a larger focus shift on a wide angle compared to a telephoto.
sfogg wrote:
This is going to just be due to two different focal lengths. A 0.5mm lens shift on a wide angle has a different effect on focus position (larger) compared to the same lens shift on a telephoto. Check any two Leica mount, unit focused lenses (that have same MFD) and the longer focal length lens will physically extend further than the shorter focal length lens when moved from MFD to infinity. Edit: talking about the in/out movement of the lens itself, not the pitch (degrees of turn) of the helicoid. The same distance shift (due to a wrong sized adapter) will have a larger focus shift on a wide angle compared to a telephoto.
Not in this case, since both lenses were accurate at the infinity mark on a Leica camera. Then there can't be anything else than the sensor stack that makes them behave differently on an other camera with the same adapter for both lenses.
Edit: Oh, I see what you mean now. Yes, that might be the case actually.
Yup, I did that on my Fotodiox tilt/shift and no-name Pentax 645 adapters. Except on those I did it on the GFX side to avoid potentially screwing up the spring/release portion of the mount. The difference on a Pentax 2 ring zoom is really obvious. With the mount short if you turn the zoom ring you go out of focus. The closer the adapter is to the proper length the less focus shift occurs when you zoom.
On the Novoflex I will have to try and shim it on the Leica side.
Makten wrote:
Edit: Oh, I see what you mean now. Yes, that might be the case actually.
If they are both unit focused lenses and one wanted to get extra geeky you could focus them at optical infinity on the wrong length adapter and then measure the length of each lens. Then focus the lenses at the physical infinity stop position and remeasure their length. This will be tough to measure since the difference is going to be small.
Given that both lenses hit infinity at the stops on an M camera the two lenses will likely show basically the same difference in length from the optical infinity position - physical infinity stop position. That should correspond to how off the adapter is.
sfogg wrote:
No, I won't. I don't have a single Leica mount lens that can't hit infinity on a M body. Do you?
Making the adapter short is just an excuse to give the manufacturer wiggle room for poor tolerances in the build of their adapter.
Shawn
I do – one such is the CV 35 f/1.2 III. I happened to be pretty lucky in that my copy is so close as to be a non-issue, but I may be the only one. Others, including Fred Miranda, have reported their copy will not hit infinity wide open (which is the test for infinity alignment, wide open). But since most people don't shoot infinity at f/1.2, it's not a big deal since stopping down a little can compensate.
Another case for making the adapter shorter is for the tendency for some M lenses to lose infinity calibration over time. Leica M lenses go back to what 1954? Novoflex has to account for everything. Just because your adapter lets all your current lenses hit infinity doesn't mean it might not be an issue at some point.
Also it's quite common to have an M lens that hits infinity before the mark. The CV 21 f/3.5 is one.
Hard stop infinity is nice. It's convenient. But it's not universal and cannot be relied upon 100%.
Novoflex is pretty good with manufacturing – I would assume they're trying to account for variance from spec for the lens mount of the camera you're adapting it to. Maybe the GFX mount is pretty dead-on, but some mounts may not be, so Novoflex probably does this across the board – not because they're lazy but because they can't possibly know which mounts are more prone to vary from spec than others.
sfogg wrote:
Yup, I did that on my Fotodiox tilt/shift and no-name Pentax 645 adapters. Except on those I did it on the GFX side to avoid potentially screwing up the spring/release portion of the mount. The difference on a Pentax 2 ring zoom is really obvious. With the mount short if you turn the zoom ring you go out of focus. The closer the adapter is to the proper length the less focus shift occurs when you zoom.
On the Novoflex I will have to try and shim it on the Leica side.
Shawn
Don't you run the risk of introducing skew when shimming the adapter?
Anyway as I mentioned earlier, maybe shimming would help greatly when adapting to a sensor with thin cover glass like the Z-mount or a thin-sensor modified A7 variant. But on the GFX with its thick cover glass, shimming is a wasted effort if your goal is IQ at infinity.
But since it sounds like you may shim your Novoflex for M to G-mount, please test before and after for infinity edge and corner sharpness and close distance for corner smearing (or whatever IQ issue you think the short mount is inducing at close distance). I guess test it with a floating element lens like the 50 Lux or CV 35/50 APOs.
highdesertmesa wrote:
Another case for making the adapter shorter is for the tendency for some M lenses to lose infinity calibration over time. Leica M lenses go back to what 1954? Novoflex has to account for everything. Just because your adapter lets all your current lenses hit infinity doesn't mean it might not be an issue at some point.
Also it's quite common to have an M lens that hits infinity before the mark. The CV 21 f/3.5 is one.
Hard stop infinity is nice. It's convenient. But it's not universal and cannot be relied upon 100%.
Novoflex is pretty good with manufacturing – I would assume they're trying to account for variance from spec for the lens mount of the camera you're adapting it to. Maybe the GFX mount is pretty dead-on, but some mounts may not be, so Novoflex probably does this across the board – not because they're lazy but because they can't possibly know which mounts are more prone to vary from spec than others....Show more →
What lens has the tendency to optically lost infinity calibration? To do that the optical group would have to shift within the lens barrel. The optics are screwed into place and spacing is set with shims. Take a lens apart and then put it back together and as long as you put the shims back where they were the lens will still focus the same. I have Leica glass going back to 1932, they hit infinity at infinity. What does happen more is the RF system in camera gets out of alignment over time where it no longer lines up at infinity. But that is just more reason why Leica lenses have a hard stop at infinity. Stop at infinity isn't just nice, on a RF lens it is needed for the RF focusing to be accurate and to calibrate the RF system in camera. Otherwise you are fighting between the optical system and the RF cam.
No way camera lens mounts have to be accounted for by Novoflex. If cameras had that much drift lenses would be way off too and SLR focusing would be wrong as well.
Novoflex simply didn't make their adapter the right length, which is fairly common unfortunately. But like I said, the 7artisans adapter is correct, as is my Nikon F to GFX adapter.
highdesertmesa wrote:
Anyway as I mentioned earlier, maybe shimming would help greatly when adapting to a sensor with thin cover glass like the Z-mount or a thin-sensor modified A7 variant. But on the GFX with its thick cover glass, shimming is a wasted effort if your goal is IQ at infinity.
.
Precise mounting makes a huge difference on my GFX 50R setup. Only one of my lenses has floating elements, so it really wants to be at the correct flange distance. The lenses with simpler designs are less fussy about that. All are very sensitive at wide apertures to any lack of parallelism. Because my main adapter is a Toyo VX23D, this is of course a major challenge... Recently I spent an evening adjusting the VX23D. What a huge difference that made in wide-open performance.
Now when I read someone saying their adapted lens has "tilted elements" or what not, I'm likely to think it's an adapter problem rather than a lens problem.
highdesertmesa wrote:
Don't you run the risk of introducing skew when shimming the adapter?
Anyway as I mentioned earlier, maybe shimming would help greatly when adapting to a sensor with thin cover glass like the Z-mount or a thin-sensor modified A7 variant. But on the GFX with its thick cover glass, shimming is a wasted effort if your goal is IQ at infinity.
But since it sounds like you may shim your Novoflex for M to G-mount, please test before and after for infinity edge and corner sharpness and close distance for corner smearing. I guess test it with a floating element lens like the 50 Lux or CV 35/50 APOs....Show more →
Sure, but with a micrometer you are measuring to hundreds of a mm so pretty easy to see skew.
Why exactly is shimming a wasted effort if my goal is IQ at infinity? Optically there is ZERO difference on a unit focused lens between an adapter at the proper thickness (shimmed or just built properly) with a lens at infinity and a thinner adapter where you have to focus out to get infinity focused. The optics are EXACTLY the same distance away from the sensor in both cases, they have to be or you aren't actually focused at infinity. The difference of course is if the adapter is proper I can focus to infinity and KNOW I am at infinity if the adapter is correct.
Where you will see differences is in floating element lenses or zoom lenses. If the lens is mounted the wrong distance away from the sensor the lens is now being used outside of how it was designed (the floating element won't be at the proper distance from the sensor) and it will impact the performance of the lens. On a zoom lens you will see this as focus shift when zooming.
sfogg wrote:
Sure, but with a micrometer you are measuring to hundreds of a mm so pretty easy to see skew.
Why exactly is shimming a wasted effort if my goal is IQ at infinity? Optically there is ZERO difference on a unit focused lens between an adapter at the proper thickness (shimmed or just built properly) with a lens at infinity and a thinner adapter where you have to focus out to get infinity focused. The optics are EXACTLY the same distance away from the sensor in both cases, they have to be or you aren't actually focused at infinity. The difference of course is if the adapter is proper I can focus to infinity and KNOW I am at infinity if the adapter is correct.
Where you will see differences is in floating element lenses or zoom lenses. If the lens is mounted the wrong distance away from the sensor the lens is now being used outside of how it was designed (the floating element won't be at the proper distance from the sensor) and it will impact the performance of the lens. On a zoom lens you will see this as focus shift when zooming.
It's a wasted effort at infinity (for ~50mm and wider) because on anything but the native M mount (or SL, Z, or thin-mod A7 are ok, but still not as good as native M), thick cover glass ruins edge/corner performance at infinity. Quite a few current M lenses don't do well even on the SL2 for corner performance at infinity (21 SEM f/3.4, etc.). And it's not just the thick cover glass, the M sensors have a unique microlens design that is made specifically for M lenses.
sfogg wrote:
Sure, but with a micrometer you are measuring to hundreds of a mm so pretty easy to see skew.
At the risk of being pedantic, .1 mm is approx. .004" -- roughly the thickness of a typical piece of copy paper or a single human hair -- so I doubt anyone is measuring accurately "hundreds of a mm" using a micrometer. Heck, accurately measuring to .1 mm using a micrometer can be tricky enough!
Where you will see differences is in floating element lenses or zoom lenses. If the lens is mounted the wrong distance away from the sensor the lens is now being used outside of how it was designed (the floating element won't be at the proper distance from the sensor) and it will impact the performance of the lens. On a zoom lens you will see this as focus shift when zooming.
With my FrankenKameras, I use only adapted lenses and most of them are zooms (so I can take advantage of the oversize image circles they project and use them to provide in-camera, rear rise / fall movements.)
One thing I've noticed with zoom lenses is that it's nearly impossible to achieve infinity focus at the same point across their focal length ranges, including those that don't have any floating or aspherical elements. I've always assumed this is because very few zooms are truly parfocal, but your comment has piqued my curiosity and I'd appreciate it if you could expand / elaborate on it and explain exactly what aspect of their design causes this effect?
Guys, there is one problem with infinity calibration that I think you are forgetting: how recently the helicoid was lubricated.
I have several old lenses that have ran out of grease so that they have a wobble and/or play in the helicoid, probably also making the infinity stop not being where it is supposed to. If the grease forms a film that can take up the play, it must be thick enough to affect the infinity stop.
And yes, they should be relubed, but there is no one around here that provides that service, and many lenses are a PITA to disassemble (and more so to assemble).
It's a wasted effort at infinity (for ~50mm and wider) because on anything but the native M mount (or SL, Z, or thin-mod A7 are ok, but still not as good as native M), thick cover glass ruins edge/corner performance at infinity. Quite a few current M lenses don't do well even on the SL2 for corner performance at infinity (21 SEM f/3.4, etc.). And it's not just the thick cover glass, the M sensors have a unique microlens design that is made specifically for M lenses.
Again, there is no difference from you under focusing to achieve optical infinity focus on a thinner adapter vs the same lens that is optically focused at infinity (at the stop) on the proper length adapter.
The optics are the exact same distance away from the sensor. They have to be or the lens isn't focused at the same point.
The difference is knowing you hit infinity (good for quick shots or shooting from a moving platform), having the focus scale be more accurate, better performance closer in (if the lens has a floating element or a zoom), and not altering the MFD.
The difference you are seeing in infinity focus position is not from sensor glass. It is due to the same shift being applied to two different focal length lenses. It is no different than using an extension tube with two different focal length lenses. The extension tube has a greater shift on focus of the wider angle lens than it does on a telephoto lens. Just on the adapter it is a negative extension tube.
Again, there is no difference from you under focusing to achieve optical infinity focus on a thinner adapter vs the same lens that is optically focused at infinity (at the stop) on the proper length adapter.
The optics are the exact same distance away from the sensor. They have to be or the lens isn't focused at the same point.
The difference is knowing you hit infinity (good for quick shots or shooting from a moving platform), having the focus scale be more accurate, better performance closer in (if the lens has a floating element or a zoom), and not altering the MFD.
The difference you are seeing in infinity focus position is not from sensor glass. It is due to the same shift being applied to two different focal length lenses. It is no different than using an extension tube with two different focal length lenses. The extension tube has a greater shift on focus of the wider angle lens than it does on a telephoto lens. Just on the adapter it is a negative extension tube.
You two are talking about different things. One is that the adapter thickness doesn't matter as long as the lens is unit focusing and you don't care about a calibrated infinity stop. The other is that adapter thickness doesn't matter because you won't get good results anyway on a sensor with much thicker stack than what the lens was designed for.
Floating lens elements will of course screw the performance at infinity too, if the infinity stop isn't correct. So the adapter thickness matters in that case regardless of focusing distance. It might matter less at infinity than at MFD though.
You two are talking about different things. One is that the adapter thickness doesn't matter as long as the lens is unit focusing and you don't care about a calibrated infinity stop. The other is that adapter thickness doesn't matter because you won't get good results anyway on a sensor with much thicker stack than what the lens was designed for.
Floating lens elements will of course screw the performance at infinity too, if the infinity stop isn't correct. So the adapter thickness matters in that case regardless of focusing distance. It might matter less at infinity than at MFD though....Show more →
Thanks.
I’ve ordered the TTArtisan M adapters for both RF and G-mounts. If they get infinity focus closer to the hard stop, I’ll compare results with my Novoflex (G) and Kipon (RF). Since they’re only around $35 each, it was worth a try.
I’ve ordered the TTArtisan M adapters for both RF and G-mounts. If they get infinity focus closer to the hard stop, I’ll compare results with my Novoflex (G) and Kipon (RF). Since they’re only around $35 each, it was worth a try.
I might order one of those too, because I want to try the 7artisans 28/1.4 FE+ on the 50R. Should work in 1:1 mode at least, and perhaps the FE+ doesn't give too much backwards curvature of field since it's tweaked for the Sony sensor stack. I know the GFX stack is even thicker, but somehow it doesn't seem to give worse results than on Sony if you use the same portion of the image circle. Maybe because of different refractive index, or that it's three separate plates with quite a bit of distance from the sensor.
