Thanks Juha for coming back to me, I missed your earlier posts while quickly browsing this thread. Good to know that it's probably not sample variation. And in Lightroom Classic I could eliminate the CA with the manual Defringe function.
I'll keep my sample, didn't want to go through the hassle of returning it anyway.
I bought the Voigtländer Apo-Lanthar 2/28 in E-mount version last week. The lens seems fine, sharp across the frame even at f/2 and no obvious centering issues.
However, I was surprised to see some purple fringing in the image corners at f/2, which only gradually disappeared on stopping down. Only at f/8 it disappeared almost completely. I fooled around with the built-in profile applied in Lightroom Classic as well as with a home-built profile, this didn't make much difference if at all. Below the upper left corner of an image taken at f/2 with the built-in profile and the home-built profile. The image center is clean. Did anyone else notice this? Could this be sample variation?
ustjwenew wrote:
I have the 28/2 on Nikon Z and have the idea so far that the fringing is much less than what I saw on Juha’s initial pictures taken with a Sony.
This is why I'm very interested to see the Septon on the Z...given that the E examples are fine but nothing special.
ustjwenew wrote:
I have the 28/2 on Nikon Z and have the idea so far that the fringing is much less than what I saw on Juha’s initial pictures taken with a Sony.
Well, the Nikon cover glas is slightly thinner than the Sony one.
For anyone with both the apo and the nokton for Sony e-mount, can you summarise the differences from your impressions so far...
I'm guessing the apo is similar sharpness in the center, significantly sharper at wider apertures towards the sides, smoother bokeh, but less bokeh - even with the nokton at f2(?), less CA....
Is that about right?
Is it worth keeping both for you, if so what's your situation for one over the other?
And if anyone has a rendering comparison between the two? I know there was one with the ultron f2 in Fred's review.
I've been leaning towards just keeping the 28/1.5 nokton for a bit more dual purpose flexibility but just would love a rendering comparison at f2 at maybe just under 1m and then another a little under 2m.
I got really annoyed with the nokton at first under harsh sunlight, but I'm hoping that was because of a filter that I suspect is causing other problems, will be checking it all when I get back. I know the bokeh isn't the smoothest in all situations but the flexibility is nice and if I really care about the shot, a little bit of pp isnt the end of the world.
Thinking maybe the APO would be nicer with stuff in the foreground appearing a little sharper at f8-f11 with slower fall off if it's like the 50 apo but reading some comments suggesting the opposite.
Yogifi wrote:
For anyone with both the apo and the nokton for Sony e-mount, can you summarise the differences from your impressions so far...
I'm guessing the apo is similar sharpness in the center, significantly sharper at wider apertures towards the sides, smoother bokeh, but less bokeh - even with the nokton at f2(?), less CA....
Is that about right?
Is it worth keeping both for you, if so what's your situation for one over the other?
And if anyone has a rendering comparison between the two? I know there was one with the ultron f2 in Fred's review.
I've been leaning towards just keeping the 28/1.5 nokton for a bit more dual purpose flexibility but just would love a rendering comparison at f2 at maybe just under 1m and then another a little under 2m.
I got really annoyed with the nokton at first under harsh sunlight, but I'm hoping that was because of a filter that I suspect is causing other problems, will be checking it all when I get back. I know the bokeh isn't the smoothest in all situations but the flexibility is nice and if I really care about the shot, a little bit of pp isnt the end of the world.
Thinking maybe the APO would be nicer with stuff in the foreground appearing a little sharper at f8-f11 with slower fall off if it's like the 50 apo but reading some comments suggesting the opposite....Show more →
I personally think that APO-Lanthar is better for landscapes and cityscapes where the target is to get great corner-to-corner sharpness as it's sharp all the way to corners at wider apertures. I personally find that I don't need to stop it down any further than f4 for those type of shooting scenarios whereas Nokton would need at least f5.6 for best corner-to-corner results. APO is also sharper centrally at close focusing distances thanks to the floating element design.
I've never tried to shoot those lenses at f8-f11 to maximise foreground sharpness so I can't really comment on how they compare in such scenario though.
Nokton seems to have a bit brighter T-stops at shared apertures, e.g. shutter speeds at Aperture priority mode are usually a bit higher with the Nokton at f2 vs. APO-Lanthar at f2 when shooting same scenarios with same Multi-metering settings. I think the main benefit of Nokton will be the brighter aperture for bokeh type shots.
Personally for me it's definitely worth it keeping both, as I've been longing for native 28mm options for Sony E-mount for a long time and I tried a lot of adapted alternatives while we didn't have too many native options (especially manual focus lenses).
I wish Cosina will release one more 28mm option for E-mount, i.e. a pancake lens to go along with Septon 40/2. Nothing's been announced on that front but they have corresponding pair of pancakes for X-mount and have had a similar pair of 40mm and 28mm pancakes for some SLR mounts. Their current F-mount 40/2 and 28/2.8 are slightly too big to be called pancakes, but some of the earlier versions were definitely qualified.
Juha Kannisto wrote:
APO is also sharper centrally at close focusing distances thanks to the floating element design.
The lack of sharpness at wider apertures outside the central areas — specifically along the lines and at the intersections formed by the rule of thirds — at medium-to-short distances (where it’s more likely to get a blurred background with such a short focal length), combined with LoCA, was the main reason I returned the 28/1.5 Nokton...
Juha Kannisto wrote:
Nokton seems to have a bit brighter T-stops at shared apertures, e.g. shutter speeds at Aperture priority mode are usually a bit higher with the Nokton at f2 vs. APO-Lanthar at f2 when shooting same scenarios with same Multi-metering settings. I think the main benefit of Nokton will be the brighter aperture for bokeh type shots.
The difference at f/2 could be down to T-stops but it could also be due to higher vignetting on the APO as it’s wide-open whereas the Nokton is stopped down. The camera meter could get a bit confused by the darker corners of the APO in that case. Could also be a mix of both effects.
I personally think that APO-Lanthar is better for landscapes and cityscapes where the target is to get great corner-to-corner sharpness as it's sharp all the way to corners at wider apertures. I personally find that I don't need to stop it down any further than f4 for those type of shooting scenarios whereas Nokton would need at least f5.6 for best corner-to-corner results. APO is also sharper centrally at close focusing distances thanks to the floating element design.
I've never tried to shoot those lenses at f8-f11 to maximise foreground sharpness so I can't really comment on how they compare in such scenario though.
Nokton seems to have a bit brighter T-stops at shared apertures, e.g. shutter speeds at Aperture priority mode are usually a bit higher with the Nokton at f2 vs. APO-Lanthar at f2 when shooting same scenarios with same Multi-metering settings. I think the main benefit of Nokton will be the brighter aperture for bokeh type shots.
Personally for me it's definitely worth it keeping both, as I've been longing for native 28mm options for Sony E-mount for a long time and I tried a lot of adapted alternatives while we didn't have too many native options (especially manual focus lenses).
I wish Cosina will release one more 28mm option for E-mount, i.e. a pancake lens to go along with Septon 40/2. Nothing's been announced on that front but they have corresponding pair of pancakes for X-mount and have had a similar pair of 40mm and 28mm pancakes for some SLR mounts. Their current F-mount 40/2 and 28/2.8 are slightly too big to be called pancakes, but some of the earlier versions were definitely qualified....Show more →
Thank you for your thoughts, and the comparison albums again - I had seen this a while and forgotten about them. Perfect.
This image (croped below) stood out for me, foreground appears way sharper even at f2:
Often saw the nokton lose some colour in the sky whilst the apo kept more of it across the sky.
Justin Stone wrote:
I mean it looks pretty good to me. Camera labs liked it. Short of perfect in the corners, but that’s a decent result.
The test by cameralabs was run on a Nikon body.
Given that the optical design of the Apo Lanthar 28mm for Nikon and Sony is exactly the same, I wonder if, because of the different thickness of the filter stack Sony vs. Nikon, the image performance might be different.
Maximilian wrote:
The test by cameralabs was run on a Nikon body.
Given that the optical design of the Apo Lanthar 28mm for Nikon and Sony is exactly the same, I wonder if, because of the different thickness of the filter stack Sony vs. Nikon, the image performance might be different.
The optical design looks exactly same based on the optical design images provided by Cosina for VM, E and Z but the lenses are optimized for each filter stack. MTF curves look identical for E and Z and somewhat different for VM. Performance of VM version and E-mount version (I had VM version before E) when both are used on Sony cameras is quite different. Lots of sensor stack induced field curvature when using VM on Sony (needs to be stopped down to f8 or more for best corner performance) while E-mount version is optimal by f4. Cosina always say that they optimize the lenses for each filter stack and surely they also do it between E and Z versions.
Juha Kannisto wrote:
The optical design looks exactly same based on the optical design images provided by Cosina for VM, E and Z but the lenses are optimized for each filter stack. MTF curves look identical for E and Z and somewhat different for VM. Performance of VM version and E-mount version (I had VM version before E) when both are used on Sony cameras is quite different. Lots of sensor stack induced field curvature when using VM on Sony (needs to be stopped down to f8 or more for best corner performance) while E-mount version is optimal by f4. Cosina always say that they optimize the lenses for each filter stack and surely they also do it between E and Z versions.
That's awesome they they optimize differently between the E and Z mount versions. I think Z mount filter stack is thinner than E, but thicker than M. I often wonder now that some 3rd party manufacturers are supporting both E and Z, whether they truly optimize fully for each. Some E mount lenses don't perform as well on Z as they do on E, so it does seem like there is room to optimize specifically for E and Z independently.
Yes, but I have a couple of comments.
First, the MTF plots of E- and Z-mount versions seem identical; it's statistically impossible that two lenses optimized for different sensor stack thickness perform identically at the optical test bench.
Then, the VM version seems to suffers less from astigmatism and shows slightly better corner performance.
Is Cosina telling us the whole truth?
Ripolini wrote:
Yes, but I have a couple of comments.
First, the MTF plots of E- and Z-mount versions seem identical; it's statistically impossible that two lenses optimized for different sensor stack thickness perform identically at the optical test bench.
Then, the VM version seems to suffers less from astigmatism and shows slightly better corner performance.
Is Cosina telling us the whole truth?
I think the VM vs. E/Z MTFs were already analyzed early in this thread (p.3 #6). Search results indicate that Cosina explicitly labels their MTF charts as "Calculated Values" 計算値 in their detailed product specifications and technical guides, so they are probably not measured results from a test bench.
It seems that e.g. Sigma also generally presents identical MTF charts for the E-mount and L-mount versions of the same lens model. I believe they also optimize (their modern mirrorless lenses) for each sensor stack as L-mount is for their own cameras but E-mount is where they sell more. I've had some of their lenses in both E and L (Sigma fp) and never noticed any visible IQ performance differences between the mounts.
Pure speculation: It is conceivable that some specific lens designs may be relatively "immune" to sensor stack issues, while others may suffer more noticeably. Sort of like the idea that lenses can be designed to be less sensitive to physical assembly errors. The Thypoch Simera Z 28/1.4, which *everyone says* is the exact same optic as the M, E, Z, and RF versions, seems not to have the lowered overall sharpness and/or the irretrievably blurred peripheral zones (when used on my Z cameras) that are commonly attributed to stack thickness issues when lenses are switched between camera brands.
Keith B. wrote:
Pure speculation: It is conceivable that some specific lens designs may be relatively "immune" to sensor stack issues, while others may suffer more noticeably. Sort of like the idea that lenses can be designed to be less sensitive to physical assembly errors. The Thypoch Simera Z 28/1.4, which *everyone says* is the exact same optic as the M, E, Z, and RF versions, seems not to have the lowered overall sharpness and/or the irretrievably blurred peripheral zones (when used on my Z cameras) that are commonly attributed to stack thickness issues when lenses are switched between camera brands.
The peripheral zones of my Simera E mount version don't looked optimized for E mount, but given that I don't have an M mount version to compare it too, I can't be sure. But I think someone has tested the M mount version on both an M mount camera and E mount camera, and the difference was miniscule. The Z mount filter stack is a lot thinner than Sony though (which means an M mount lens can perform a bit better than when mounted on an E mount camera, at least that is my general impression of participating here over the years.)
From an M mount lens use on E mount perspective, the wider the focal length, the more chance there is an impact. I'm sure there are always exceptions, but that's the generally trend I've noticed in comparisons over the years.
And I have seen comparisons where E mount lenses haven't performed as well on Z mount cameras, although there was a question whether it was a flange distance issue or filter stack. Suffice it to say there still could be minor differences in performance of an E stack optimized lens on a Z mount camera.
But the design of the lens definitely has an impact, as you speculate (and I do agree that different lenses designs can be more sensitive to filter stack differences), so you can't make a complete generalization on performance.
