p.48 #1 · A thinner sensor stack may be possible after all!
Some more updates.
I thought I saw a hint of some sort of color cast, so I compared the unmodded and modded camera shots of the above.
The A7 used a WB of 5200K for proper WB while the modded camera used about 7200K.
That said, the modded camera largely takes care of this change with its AWB.
Here is the A7K shot with its AWB. A7K Pentax K24 f2.8 at f10 by jenkwang, on Flickr
Still a hint of magenta imo.
There might be several ways to get around this like playing with the WB and tint (though I've found it tricky with the AW picker tool in LR)
There are some adjustments in the camera profile that can be done to further tweak the colors (or create a custom profile)
p.48 #2 · A thinner sensor stack may be possible after all!
I wonder if I could ask a favour for anyone who has this mod with 35 FLE or 50 Lux Asph to take an image focusing at the corner points at wide apertures e. f/1.4, f/2.0, f/2.8
I am very tempted to send my A7S in but want to be sure that the 2 lenses I have are noticably better with the mod. I've seem some images posted on this thread which are stopped down but not wide open or close to wide open focusing towards the edges.
I have to stop my 35 FLE down to f/4.0 to get acceptable corners and 50 Lux to f/5.6. My other lens is the CV 15 III which I find acceptable stopped down from f/8 which is how I use the lens most of the time hence why I'm still a little on the fence.. Not helped that my A7S is only a month new..
p.48 #3 · A thinner sensor stack may be possible after all!
sootyvrs wrote:
I wonder if I could ask a favour for anyone who has this mod with 35 FLE or 50 Lux Asph to take an image focusing at the corner points at wide apertures e. f/1.4, f/2.0, f/2.8
I am very tempted to send my A7S in but want to be sure that the 2 lenses I have are noticably better with the mod. I've seem some images posted on this thread which are stopped down but not wide open or close to wide open focusing towards the edges.
I have to stop my 35 FLE down to f/4.0 to get acceptable corners and 50 Lux to f/5.6. My other lens is the CV 15 III which I find acceptable stopped down from f/8 which is how I use the lens most of the time hence why I'm still a little on the fence.. Not helped that my A7S is only a month new..
I don't know your acceptable level would be, but this lens has no induced field curvature after the mod; that is, you can't get sharper corners by changing the focus.
I don't know your acceptable level would be, but this lens has no induced field curvature after the mod; that is, you can't get sharper corners by changing the focus.
Great thanks for those images... I must have missed those as I have trawled through every page so I thought...
I was also hoping for some closer focusing subjects e.g using as a portrait lens with subject 1/3rd into the frame to the edge.. At the moment I'm reluctant to put a subject beyond the rule of thirds line when using f/1.4 between 2.8 due to smearing so it does limit its use..
I must admit if I'm using it for landscape I'm usualy beyond f/5.6 which I find acceptable but I'm disappointed I have to frame the subject between thirds to get acceptable shots.. I am finding the 35 FLE a little better than the 50 lux.
If the infinity shots you posted are anything linke what they will be like close up e.g. 2m and focusing close to the edge, then I think it's a mod worth doing.
Intersting what you mention about improving field curvature of the lens,
Here are some crops of a lamp using 50 lux wide open at approx 2-3m away to show the edge performance on my unmodified A7S compared to framing closer to the centre
This one taken focusing on the lamp at the edge of the frame
and this one which is focused on the 1/3rd line (I know not same composition)
p.48 #5 · A thinner sensor stack may be possible after all!
sootyvrs wrote:
Great thanks for those images... I must have missed those as I have trawled through every page so I thought...
I was also hoping for some closer focusing subjects e.g using as a portrait lens with subject 1/3rd into the frame to the edge.. At the moment I'm reluctant to put a subject beyond the rule of thirds line when using f/1.4 between 2.8 due to smearing so it does limit its use..
I must admit if I'm using it for landscape I'm usualy beyond f/5.6 which I find acceptable but I'm disappointed I have to frame the subject between thirds to get acceptable shots.. I am finding the 35 FLE a little better than the 50 lux.
If the infinity shots you posted are anything linke what they will be like close up e.g. 2m and focusing close to the edge, then I think it's a mod worth doing.
Intersting what you mention about improving field curvature of the lens,
Here are some crops of a lamp using 50 lux wide open at approx 2-3m away to show the edge performance on my unmodified A7S compared to framing closer to the centre
I no longer have the lens so I can't test it out for you. However, I don't think you would have any problem framing the subject in the outer third. Here is a couple of pictures with subject placement not at the center, WO:
p.48 #7 · A thinner sensor stack may be possible after all!
FYI, and not sure if this has been posted elsewhere, but Jim Kasson blogged some images from a prototype Kolari modded a7II. This is the first I've seen of any recent news to indicate progress on modding the Mark II camera with a thinner cover glass.
p.48 #8 · A thinner sensor stack may be possible after all!
Add to this thread again.
(as usual, full size picts on the flickr links)
Managed to test out the CV15/5.6 Heliar v1 (ltm version).
All done at f11 to be sure that the edges balance up in sharpness as best as is possible on these cameras.
Unfortunately, I did not realize until shooting the series of lenses that my friend's A7 was set to Std Jpg (all settings at 0) instead of RAW.
So I also set the A7k to shoot at the same settings.
I did shoot in RAW+JPG though, and to be fair, I'd say that if I processed the RAW, I'd get a little better results with the more powerful on computer JPG engine, but its really slight and would not really invalidate the test shots done in JPG.
The lighting conditions varied through the test shots.
From bright Sun to Shaded between A7 and A7K shots. (just for additional info)
For color cast on corners, those familiar with it will know that exposure and type of scene can influence it.
So the change in lighting condition between the test can be a factor.
I'd say that it generally does not improve between modded and unmodded cameras.
Sharpness in general, the A7K does give a little bit better edge definition.
But this is really just a little bit to me.
On the sides, also a little bit better edge definition, but imo, not enough to make me go buy a CV15v1 after I have a modded camera.
That said, I am not familiar with the CV15, as I don't own this lens.
Perhaps this really is the max that the lens can do on whatever format (ie. film; M9 )??
Or is the copy I tried not as good?
I'm also curious how this lens performs on a camera that is supposed to be catered for RF lenses (ie. the M9)
Do give me some inputs if possible. TIA
p.48 #10 · A thinner sensor stack may be possible after all!
I once owned a Heliar 15/4.5 first version. I used it on a NEX camera (can't recall which model) and a Ricoh GXR-M. Therefore, my experience is limited to APS-C sensors. I expect that a full frame camera will show even more differences between center and corners..
My five cents :
1- Excellent relationship between size/price/optical performances but not a match for far more expensive glass .like a ZM 15 or a Leica WATE -The Heliar 15/4,5 isn't a sharpness champion at any f/stop. ( best aperture seems to be between f/5.6 and f/8. F/11 isn't the best choice for assessing sharpness).
2-The lens has a noticeable field curvature -even on film cameras. If you focus for better sharpness at the centre, corners/edges won't be pin sharp. Likewise, focusing at the corners degrades somehow performances at the centre. You may choose to focus in-between or stop down to f/11-f/16, where the effect almost disappears due to depth of field, but the general performance past f/8 worsens, as stated above.
3-Field curvature with the Ricoh GXR-M was visibly lower than with the NEX and it was easier to get a fair balance between centre and corners This was surely because of the thinner filter stack of the Ricoh
4-Copy variance might play a significant role as well. Cosina's QC seem's to be less strict with the Voigtlander line than with the lenses they make for Zeiss. (ZM, ZF, etc..) This is logical, considering the price difference.
p.48 #11 · A thinner sensor stack may be possible after all!
rscheffler wrote:
FYI, and not sure if this has been posted elsewhere, but Jim Kasson blogged some images from a prototype Kolari modded a7II. This is the first I've seen of any recent news to indicate progress on modding the Mark II camera with a thinner cover glass.
So far Jim's posts have been on corner performance, which is as-expected from the mod. He is also using the same WB settings so you can see the color difference, which is also expected.
I'm looking forward to him testing the image stabilization, which he has a good methodology for based on his earlier tests on the A7II. Since the thinner filter changes the weight of the sensor assembly, it has to be verified that the performance of the IBIS is not impacted. I've seen some test results from Kolari that look positive, but Jim has a more quantitative method for checking it.
p.48 #12 · A thinner sensor stack may be possible after all!
artur5 wrote:
I once owned a Heliar 15/4.5 first version. I used it on a NEX camera (can't recall which model) and a Ricoch GXR-M. Therefore, my experience is limited to APS-C sensors. I expect that a full frame camera will show even more differences between center and corners..
My five cents :
1- Excellent relationship between size/price/optical performances but not a match for far more expensive glass .like a ZM 15 or a Leica WATE -The Heliar 15/4,5 isn't a sharpness champion at any f/stop. ( best aperture seems to be between f/5.6 and f/8. F/11 isn't the best choice for assessing sharpness).
2-The lens has a noticeable field curvature -even on film cameras. If you focus for better sharpness at the centre, corners/edges won't be pin sharp. Likewise, focusing at the corners degrades somehow performances at the centre. You may choose to focus in-between or stop down to f/11-f/16, where the effect almost disappears due to depth of field, but the general performance past f/8 worsens, as stated above.
3-Field curvature with the Ricoh GXR-M was visibly lower than with the NEX and it was easier to get a fair balance between centre and corners This was surely because of the thinner filter stack of the Ricoh
4-Copy variance might play a significant role as well. Cosina's QC seem's to be less strict with the Voigtlander line than with the lenses they make for Zeiss. (ZM, ZF, etc..) This is logical, considering the price difference....Show more →
Thanks for your inputs.
They are very valuable to me as I try to make sense of what I am observing so far, trying out the lenses.
The lens I tested is an old one.
Its marked as Cosina and not Voigtlander.
I too have certainly found field curvature to be the reason for many shots being even weaker off the center on many wides (when just the center is used for focus).
Thats why I shot focusing using the off center and stopped down to f11 instead of f8 to try getting the DOF to be able to cover the whole image.
I was more interested to see how off center parts have improved, rather than absolute sharpness, in which the latter, probably f8 being the better setting.
What you see on the display looks very promising for legacy lens collectors! It gives strong indication that the sensor of the new Sony A7RII has no problem with magenta color shift at the borders anymore. It also looked like the borders were less blurred compared to the sample from the A7S.
Though color cast might be a PITA, it's fixable. What troublesome is the smearing. I don't know why they couldn't just switch to JPEG so that they can actually zoom in to check whether the corners are any good. The Hologon smeared like no others, and "less blurred" doesn't mean much.
Thanks. So, it will be interesting to find out how much Leica M lenses in the 18 to 50 mm focal range will do better on the A7RII as compared to the A7R. The WATE 16-18-21/4 should be a stellar performer on the A7RII as it was on the A7R. :-)
p.48 #16 · A thinner sensor stack may be possible after all!
For me, it's how well will the Voigtlander 12mm and 15mm-II do with the A7RII - will the improvement be enough to use the -II of the 15mm instead of the excellent larger/heavier (relatively speaking) -III version.
Looks like the color cast will be (mostly) gone, but how much better will be the smearing at large apertures?
p.48 #17 · A thinner sensor stack may be possible after all!
pinholecam wrote:
That said, I am not familiar with the CV15, as I don't own this lens.
Perhaps this really is the max that the lens can do on whatever format (ie. film; M9 )??
Or is the copy I tried not as good?
I'm also curious how this lens performs on a camera that is supposed to be catered for RF lenses (ie. the M9)
Do give me some inputs if possible. TIA
I primarily shoot Leica digital M - M240 and M9 and own both M mount versions of the 15mm. I have shot some comparisons on the Sony a7 series cameras with a range of RF lenses, including the first 15mm M mount version (optically identical to the LTM version). You can see how it does on the a7R and a7S. Both of those links include comparison images with the lens on the respective digital Leica I owned at the time.
Bottom line is I think the CV15 vII performs very similarly on unmodded a7 series cameras compared to digital Leica, with respect to edge smearing.
I recently acquired the CV15 vIII and did a side by side test of it on the Leica M240 against the previous version (haven't gotten around to publishing it). I assume M240 results will be very similar to a modded a7 series camera. The biggest difference was complete elimination of the edge color shift without any special processing by the camera or in software afterwards. Distortion correction was also better in the new version. Otherwise, in respect to color, contrast, sharpness, etc., I barely saw any differences. Peak central sharpness was already by f/8, but if focused centrally, edges needed f/11 for improvement. I haven't looked into this much, but am guessing the vIII also has some field curvature and that focusing somewhere between the center and the edge will improve 'average' focus/sharpness across the frame at f/8.
Is the vIII worth it? Yes, if you hate correcting color shift in post (which I believe results in additional artifacts and some quality degradation). The convenience factor here is huge. Otherwise, I don't see a noticeable image quality improvement other than better distortion correction. I dislike how much larger the lens has become. If you can live with the color shift and get a decent copy, the LTM (both 12mm and 15mm) is so much more compact, where it's almost no thought about bringing one or both along. But with the M versions, particularly vIII, the lens is to the point where it is fairly chunky (for an RF lens), particularly thanks to the built-in hood... With the 3D-Kraft a7RII preview, I suppose it might be possible the color shift with the older versions of the lens will be less problematic, which will be very interesting to see once the camera is out in the wild.
As for Cosina quality control of the Voigtlander lenses.... my impression is it has improved with the latest lenses I've bought, such as the 35/1.2 II, 21/1.8, vs. the older lenses I also own or have tried. When I bought the 12/5.6 LTM, I tested a couple and one was horrendous... not sure it's an indication of general QC problems with Cosina back then, but if you're buying an older Voigtlander lens, you definitely want to test before committing.
p.48 #18 · A thinner sensor stack may be possible after all!
I read this hopeful news this morning and was immediately curious about the Contax G 21 Distagon, a lens I have been resisting buying for sometime, strictly because of the performance on unmmoded sensors.
I am waiting to see if a) the smearing improves significantly without the modification of the A7rII b) if Kolari can mod a sensor with IBIS and c) if the the sensor does show dramatic improvement with smearing/color, if there will be a synergistic effect with the sensor is moded from the A7IIr.
p.48 #19 · A thinner sensor stack may be possible after all!
nehemiahphoto wrote:
I read this hopeful news this morning and was immediately curious about the Contax G 21 Distagon, a lens I have been resisting buying for sometime, strictly because of the performance on unmmoded sensors.
I am waiting to see if a) the smearing improves significantly without the modification of the A7rII b) if Kolari can mod a sensor with IBIS and c) if the the sensor does show dramatic improvement with smearing/color, if there will be a synergistic effect with the sensor is moded from the A7IIr.
Thoughts anyone?
Nehemiah,
a) I'll let you know when I get mine from Amazon
b) Ilija can mod the A7II now, so I think it would be the same for the A7rII
c) Modded 1st gen still can't cope with the likes of G 21 and 28 at WO to f/5.6. So if this cam can do what the modded 1st gen. do, I'm already happy enough. Though early Hologon samples indicates it might be even better. If it is so, then it means Sony did change the stack thickness. At that point, Kolari mod might not make sense anymore since the gain would not be dramatic enough to pay. Until I see a direct comparison between the modded cam vs. the M9/M240, I would say any improvement on the native platform would not be that much.
p.48 #20 · A thinner sensor stack may be possible after all!
Just a reminder that I'm planning to score and document RF lens performance on the A7r2 when it starts shipping if people are prepared to take test photos. https://www.fredmiranda.com/forum/topic/1374669/0
I'll revive the thread, or people can post photos directly into it.
I placed my order 15 minutes "late" so I don't know if I'll make the 1st cut (B&H) next week.
