Luminous Landscape Forum
Equipment & Techniques => Medium Format / Film / Digital Backs – and Large Sensor Photography => Topic started by: marc aurel on September 08, 2014, 02:25:04 pm
-
I have been working with shift-lenses for architecture photography on Canon full frame cameras for a few years now. After buying the A7R und using it for a while I thought it was time to do an in-depth comparison of these lenses on a 36 MP chip and share the results. I know it’s not medium format but I decided to post it here because as far as I’ve seen most discussions about shift lenses take place in this forum.
Test candidates were the Contax-Zeiss PC-Distagon 35mm and the Canon TS-Es 17mm L, 24mm L II and 90mm. I added a combination of the TS-E 90mm with the Canon extender 1.4x III to the test (which results in a focal length of about 125mm).
As my „test target“ I chose a church with a brick facade. Supporting structures protrude a bit, but the main plane of the facade doesn’t offset much so there is a chance to test for field curvature. #0 shows the uncropped image, red frames mark the chosen areas for the 100% crops.
Technical parameters of the test were:
• Canon TS-Es adapted to the A7R with a flocked Metabones III adapter. Adapter mounted on tripod (not the camera).
• Contax-Zeiss PC-Distagon 35mm adapted with a Novoflex-Adapter. Adapter mounted on tripod with an Astat tripod collar.
• For longer lenses I chose a longer distance to the church to get identical framings for all images.
• Camera was on tripod (Gitzo 3541), backpack with some weight hang under the tripod. Arca D4 geared head. Shutter release with self timer.
• 10mm of shift were applied. The Canons can shift a bit further, but the PC-Distagon is limited to 10mm, so I chose this value for all lenses.
• I took sets of images with apertures from f5.6 to f16 in full stops.
• Exposure times were from 1/500s for f5.6 to 1/60s for f11.
• 3 different zones were analysed:
• Zone A is about 5mm from the optical centre of the unshifted lens.
• Zone B is about 15mm from the centre (about as far as the corner of an unshifted lens).
• Zone C is about 27mm from the centre (very close to the edge of the image circle).
• With each lens I first focused on Zone C and took a set of images with apertures from f5.6 to f11. Then I focused on Zone B and took another set of images. Then I focused on Zone A and took another set. With that I tried to test for the maximum sharpness that a lens is capable of in each zone – irrespective of field curvature effects.
• All images shot within less than 30 minutes. The light didn’t change much.
• RAWs converted with Lightroom 5. Identical white balance on all images.
• Sharpening settings: Amount: 50 / Radius: 0,7 / Detail: 30 / Masking 4.
• No correction of distortion or vignetting was applied.
• Crops shown at 100% (if you look at them make sure that your viewer shows them at 100%, the compiled crops are too wide for some screens)
The questions I tried to answer for myself were:
I: Which one is the best lens?
• Zone A (compilation #1): all lenses are quite good here (as expected), even at f5.6. The 90mm shows a bit more fine detail than the others. The 17mm and 24mm don’t loose much sharpness at f11, the others should be kept at f8 or f5.6 if possible. The 90mm+Extender is best at f8 too, but not as good as the others (as expected).
• Zone B (compilation #2): the 90mm is the best in this Zone, peaking at f8. It shows fine detail I see in none of the other lenses. The 24mm and 17mm are both very good, but need f11 for maximum sharpness. I was astonished that the PC-Distagon falls a bit behind in this zone. Still good, but it seems as if the designers had to make a small tradeoff here to achieve sharpness across the whole frame. The 90mm+Extender is OK at f8 and f11.
• Zone C (compilation #3): here the PC-Distagon shines - great sharpness, peaking at f8. Followed by the 90mm which is best at f8 too. The 90mm+Extender is OK at f11 and still usable at f8. For the wideangles this distance from the center is tough. The 17mm is slightly better than the 24mm, both need f11 that far from the centre. f16 would be better, but that would reduce sharpness in the other zones too much. (The different amount of vignetting makes it a bit difficult to judge contrast in this zone. With the wider lenses you look more from below so the shadow area appears larger. The flat part of the terracotta-relief reflects the bright sky in the back of the camera – again more with the steeper angles of the wider lenses. Impossible to rule this out, sorry).
• So which one is the best lens? This is a difficult question, it depends on what you need. I love the PC-Distagon because there is no really weak point in the whole image circle. The 90mm is the sharpest lens nearly everywhere, but in Zone C it’s a bit behind the Distagon. The 17mm and 24mm are great in most of the image circle, but they have a hard time in Zone C. They cover very extreme angles, so you’ve got to put that in perpective. These are 100% crops from a 36MP sensor.
(continues in part II)
-
II: Is field curvature relevant for these lenses?
• Field curvature is not much of a problem with these lenses when they are not shifted. But since field curvature usually grows faster when you get further away from the centre - I was wondering if this was relevant for my shooting.
• Compilation #4 shows crops of Zone A at f8, in the upper line I focused on Zone A, in the lower line on Zone C. I see just a very small sharpness difference with the 17mm and 24mm, mostly because of their greater depth of field I think. With the PC-Distagon I see clear differences which indicates some field curvature. I see no difference with the 90mm so it seems to be well corrected for field curvature. The 90mm+Extender is not sharp enough for such subtle differences.
• Compilation #5 shows the same comparison at f11. The differences are nearly gone, the PC-Distagon still shows a small amount of difference.
• No matter which lens – I usually focus on a spot that is in a medium distance from the optical centre (here: Zone B) to keep everything within depth of field.
III: What is the optimal aperture for each lens in unshifted and in shifted position?
• Unshifted position: for a flat subject I would use all lenses at f8. This will change of course if you have a 3-dimensional subject that needs more depth of field.
• Shifted position (+10mm): for a flat subject the PC-Distagon and the 90mm can be used at f8 even shifted. The 90mm+Extender is better at f8 too. The 17mm and 24mm need to be stopped down to f11 for better corner sharpness. From about 3 or 4mm of shift f11 is better with the 17mm and 24mm.
• Compilation #6 shows the crops at my recommended apertures. In contrast to the compilations #1 to #3 the three crops for each lens are from the same image which was focused at Zone B.
IV: Can I improve the results further by selective sharpening?
• My standard sharpening settings (Amount: 50 / Radius: 0,7 / Detail: 30 / Masking 4) are not enough in Zone C for most lenses when they are shifted. To get optimal results I often use gradient filters or adjustment brushes in lightoom to raise sharpening levels in selected areas.
• Compilation #7 shows my effort to get optimal results from each lens with this technique. Vignetting is corrected to a certain degree. By using gradient filters the amount of sharpening is raised individually. Sharpening levels for Zone A are now at 60 in Zone A, 70 to 80 in Zone B, 75 to 80 in Zone C (values depending on lens). For the 17mm and the 24mm I added clarity in Zone C (+10 for the 17mm and +20 for the 24mm).
• If you want to invest very much in an image from the 17mm and 24mm you could combine an f16 image for Zone C and f11 for A and B. But that's a lot of work.
• I am curious: what sharpening settings do you use for these lenses? Do you have propositions for better image quality in the weaker areas?
I would be glad if you would comment and add your own findings on these lenses.
Marc
-
Thanks for the elaborate tests! Very interesting.
The PC-Distagon 35mm looks fantastic. Unfortunate that it cannot be tilted.
Haven't seen it before but I realize that the TS-E 17 and also the 24 to some extent push the A7r sensor into crosstalk despite being strongly retrofocus, the duller colors in the extreme corners of those lenses is most likely due to crosstalk.
-
Thanks for the elaborate tests! Very interesting.
The PC-Distagon 35mm looks fantastic. Unfortunate that it cannot be tilted.
Haven't seen it before but I realize that the TS-E 17 and also the 24 to some extent push the A7r sensor into crosstalk despite being strongly retrofocus, the duller colors in the extreme corners of those lenses is most likely due to crosstalk.
Hallo Torger,
thanks.
I was confused about the duller colours in Zone C with the 17mm and 24mm too. I'm not sure that this is crosstalk. Could that be with such a strong retrofocus design? According to LensRentals.com the exit pupil distance is 90.5mm for the TS-E 17mm and 86mm for the TS-E 24mm, so even fully shifted the angle in which the light hits the sensor is not very steep. When I look inside the PC-Distagon I would guess exit pupil distance is less for this lens.
I suspected that with the 17mm and 24mm the upper part of the facade reflected a different part of the sky (with the bright sun still high in the sky behind me). The wider the angle of the lens, the closer I had to get to the subject which changes the angle between camera and Zone C high up. In the crops that are corrected for vignetting (compilation 7) the flat part of the terracotta-relief is brighter with the wides than with the longer lenses, maybe because of these reflections. I think to a lesser degree that could have happened to the bricks too. But that's just a theory.
Another question to users of the IQ 250 (or another highres 44x33-chip):
Has anyone used these lenses with a MF digital back on an Alpa FPS or an HCam? I would really like to know how they behave on these sensors compared to the A7R. I had the idea of investing in such a system and one of the things I wanted to find out with this test was if image quality in the corners is good enough. You can shift less of course. But even within the reduced shift range you would use a larger portion of the weaker part of the image circle than with the A7R. I'm confident that the PC-Distagon and the TS-E 90mm would have no problems. But with the 17mm and the 24mm I'm afraid that this would more than outweigh the advantages that the larger image area has. Has anyone tried this for architecture or landscape?
Marc
-
I believe it's Crosstalk. If you put a tech lens on the A7r (say 40mm Rodenstock), then you will really see this problem from 10mm of shift out. The TS-E's are giving you a bit of relief due to the metabones adapter which is moving the lens further away from the sensor. The 40mm Rodenstock at infinity is about 1/4" from the the A7r sensor, which really pulls in crosstalk issues. At around 13mm of shift about 50 percent of the color is gong and by 18mm the shifted image starts to appear monochromatic. Same problem with the A7r chip's big brother 50MP Sony in MFD.
What I seen in your images is IMO totally correctable in post. Did you shoot an LCC on the shifts you took, as it will help somewhat on the light falloff correction. Capture One with an LCC does an excellent job on the A7r files.
Paul
-
I will try. Just have to borrow a lens cast calibration plate somewhere. Never used one before. I will report on the results (that will take a few days).
Marc
-
I was confused about the duller colours in Zone C with the 17mm and 24mm too. I'm not sure that this is crosstalk. Could that be with such a strong retrofocus design?
Yes I think so. Even if strong retrofocus if you shift to the edge the angle is increased and with current sensor design it does not need to be that steep. Sensors have a metric called "critical crosstalk angle" which is the maximum angle they can handle before crosstalk occurs. This angle is most often really small. In the MF world Kodak KAF-39000 is one of the best concerning this particular metric and it has 39 degrees. Unfortunately these numbers are rarely published, but I manage to find a cross-section photo of a D800 sensor, it should be about exactly the same as A7r:
(http://www.chipworks.com/en/technical-competitive-analysis/resources/blog/technologyblog/files/2012/10/Figure_1b-300x223.jpg?file=2012/10/Figure_1b-300x223.jpg)
and from top:
(http://www.chipworks.com/en/technical-competitive-analysis/resources/blog/technologyblog/files/2012/10/Figure_1a-300x223.jpg?file=2012/10/Figure_1a-300x223.jpg)
What you can see here is that the pixel is considerably higher than wide (common especially with small pixels) and that there is no light shield mask to isolate adjacent pixels (common), and that while wiring stop the light somewhat in the horizontal direction there's no stopping in the vertical, ie probably you get more crosstalk in vertical orientation of the sensor (or the other way around, from these images we cannot know which side is the short side).
If you want to surely see at which angle the lens delivers light at the image circle lens you can mount the lens on a tripod exactly perpendicular to a wall, then use a laser pointer and point it from the extreme side at the front of the lens (ie at the image circle edge), and measure the angle the laser comes out at the other end. As one cannot see the exit pupil with this method the best way to see the angle is probably to put a table with a white paper on behind the lens and try to cut the laser so you see how it shines over the paper and measure the angle directly on that (then you don't need to know where the exit pupil is you get the angle directly if you have the paper properly lined up).
Anyway, assuming it's a triangle with to the exit pupil the TS-E 17 with its 67mm image circle that would be "only" 20 degrees, not that much but still probably a little bit too much for this sensor.
-
Another question to users of the IQ 250 (or another highres 44x33-chip):
Has anyone used these lenses with a MF digital back on an Alpa FPS or an HCam? I would really like to know how they behave on these sensors compared to the A7R. I had the idea of investing in such a system and one of the things I wanted to find out with this test was if image quality in the corners is good enough. You can shift less of course. But even within the reduced shift range you would use a larger portion of the weaker part of the image circle than with the A7R. I'm confident that the PC-Distagon and the TS-E 90mm would have no problems. But with the 17mm and the 24mm I'm afraid that this would more than outweigh the advantages that the larger image area has. Has anyone tried this for architecture or landscape?
Marc
Hi, Marcus,
here is a colleague from Berlin. You asked für a 44x33-chip (test-) picture? Here you go:
Horizontal picture, Ts-e 17mm, f11, full Shift (12mm), Aptus II 8, HCAM.
focusing scale: 50m (good sharpness, except corners):
Link (http://www.adrianschulz.de/lula/L_005087.jpg)
focusing scale: ~3m (medium sharpness in the center, bad sharpness in the distance, corners better):
Link (http://www.adrianschulz.de/lula/L_005089.jpg)
Vielleicht sollten wir uns irgendwann mal treffen und testen.
-
Another question to users of the IQ 250 (or another highres 44x33-chip):
Has anyone used these lenses with a MF digital back on an Alpa FPS or an HCam? I would really like to know how they behave on these sensors compared to the A7R. I had the idea of investing in such a system and one of the things I wanted to find out with this test was if image quality in the corners is good enough. You can shift less of course. But even within the reduced shift range you would use a larger portion of the weaker part of the image circle than with the A7R. I'm confident that the PC-Distagon and the TS-E 90mm would have no problems. But with the 17mm and the 24mm I'm afraid that this would more than outweigh the advantages that the larger image area has. Has anyone tried this for architecture or landscape?
Marc
Marc:
On the main Phase One site, they have featured some of the work by Gerald and his night shots. He is using the HCam and 17mm TS-E. He has had some very good results on the IQ250. Phase has featured some of his downtown Dubai night panorama work.
Alpa on their website has featured several blog posts, on the usage of the IQ250/FPS and the 17mm TS-E. They did have some shots that could be downloaded.
There has been a sort of debate on this forum and others as to can the Canon TS-E lenses/IQ250 stand up a tech camera/IQ250 or IQ280.
The 50MP chip in the IQ250 (which I believe is the same chip in the just announced Credo 50 and the Pentax 645Z) has crosstalk issues with tech wide lenses, which limits the amount of movement. Gerald has had some very good success with both his 250 and 280 and the TS-E lenses with movements, so I feel from looking at his work, the TS-E lenses (17 and 24) are a better solution for the 50MP CMOS chip.
You can find more links to Gerald's work over on the www.getdpi.com forum.
Paul
-
Adrian - thank you for your test shots.
The second one looks sharp to me everywhere - of course there is some falloff in sharpness towards the corners like on full frame. A direct comparison with the A7R would be a good thing. Pixel width of your Aptus is practically the same as with the A7R. I sent you a personal message so we could meet in Berlin and try to find out how they compare in the different zones.
I think the vignetting in the top corners of your image shows limitations of the image circle and the vignetting in the lower corners is hard vignetting of the mount when shifted too far? How much you can shift up without vignetting in the lower corners?
Torger - thank you for your detailed response.
I'm really curious what will happen when I use an LCC - but you said that P1 will correct it well, so I'm hopefull. When the rumors of an even higher resolution Sony 35mm chip come true - maybe full frame (at least with shift lenses) will run into the same problems as medium format needing an LCC for every shot.
Paul - I didn't know most of these works. Thank you for the links. I will take a close look because this is the most probable upgrade path for me at this moment.
No matter what system you choose - there are always tradeoffs when you want too shoot architecture digitally and need wide lenses with shift:
1. Pancake cameras with a 54x40mm CCD chip and Rodenstock wides: probably the sharpest images as far as I've seen. But no liveview, precise focusing needs more time. LCC needed. HR32 is the widest that can be used with more than just a bit of shift.
2. Hasseblad SLR with 54x40mm CCD chip: retrofocus design of lenses makes LCC unnecessary. But shift only available with the 1.5x tilt-shift-converter which limits wideangle to 24mm equivalent on 35mm. Not as sharp as the rodenstocks as far as I know.
3. Pancake cameras with a 44x33mm CCD chip and Rodenstock wides: for wide angles you need the HR23 and 28 with their more limited image circle and with stronger lens cast. LCC needed. No liveview.
4. Tilt-shift lenses for 35mm on HCam or Alpa FPS with a 44x33mm CCD chip: Widest usable angles. I'm not sure if you need an LCC when shifted. Sharpness in the outer part of the image circle probably not as good as with the HR 32. Shift range reduced when compared with their use on 35mm.
5. Tilt-shift lenses for 35mm on HCam or Alpa FPS with the Sony 44x33 CMOS chip: as above, but with liveview. Could be great if crosstalk issues are not serious (which we do not yet know)?
6. Tilt-shift lenses on 35mm: image quality in some areas is not as high as I wish.
Did I forget something? A mirrorless camera with a 44x33 CMOS chip and an adapter for canon mount shift lenses would be a great addition to these possibilities, but that could happen in the next week or in a few years. And I have lost conficence that very much will happen concerning lenses. While everybody seems to work on fast lenses for full frame - virtually nothing happened with shift lenses. Nikon is behind canon. No one makes a new 35mm shift lens. The new Schneider 28mm is very heavy and counter-intuitive to handle.
So for those who made their choice for a system on the basis of architecture as their main subject: what did you choose and why ?
-
Torger - thank you for your detailed response.
I'm really curious what will happen when I use an LCC - but you said that P1 will correct it well, so I'm hopefull. When the rumors of an even higher resolution Sony 35mm chip come true - maybe full frame (at least with shift lenses) will run into the same problems as medium format needing an LCC for every shot.
Lightroom, RawTherapee and Capture One and my own Lumariver HDR all have LCC correction (best in Lumariver HDR of course ;) ), but none of them can correct crosstalk. I've experimented with an crosstalk cancellation algorithm but it's not stable enough for production, and I'm not sure if it ever will be. So if the dull colors in the corners indeed are crosstalk they will still be dull after LCC correction.
Sooner or later Sony will launch a 50 megapixel 35mm chip, it will likely be more problematic yes as height/width ratio of the pixels will be larger. I still think that it will only be visible on extreme shifts of wide angle tilt-shift lenses like the TS-E 24 and TS-E 17. Mass-market users won't accept that you get color cast on regular wide angle lenses so I don't think Sony would release such a sensor. Some APS-C sensors have already tiny pixels and there are wide angles for them that work fine, which means that it will work for a high mp 35mm chip.
Tech camera lenses have much more extreme angles. If the TS-E 17mm has 20 degrees you need to go up to 120mm or so before a tech cam lens delivers such a low angle. My Schneider 60XL has 45 degrees angle at the image circle edge, and thus push even MF CCDs into crosstalk, but you can handle mild levels of crosstalk without much image degradation.
-
So for those who made their choice for a system on the basis of architecture as their main subject: what did you choose and why ?
I'm a landscape photographer and only shoot architecture occasionally, but if I pretend I was architecture pro, I would choose an Alpa MAX with Hasselblad H5D-50 (or H5D-200MS) and Schneider Digitar lenses 28mm and up.
For my landscape work I use a Linhof Techno, and I like view camera design better for that purpose (I do tilt/swing etc quite often), but for pro architecture I'd rarely use tilt/swing, mostly wide angles, and having a high precision focusing ring on a pancake camera would be more comfortable in the often weak light conditions with dark wide angles. When I shoot architecture I quite often shift both horizontally and vertically simultaneously, thus the MAX body from Alpa. The RM3Di has too much overkill on the focusing ring precision, I think Alpa has the ideal balance.
Back/lens choice is kind of philosophical. With an IQ260 and Rodenstock lenses like the 32HR I would get a little bit better image quality (and more expensive!). But I care about photo gear, and the H5D-50+Schneider Digitar series is designed more how I think "large format digital" should be, ie lenses should be symmetric and distortion free as far as possible, and the sensors should support that. The Kodak sensor is not exactly class-leading in DR, but it's adequate and unlike the Dalsa in the IQ series it's much better of suppressing crosstalk, meaning that you can use the full image circles of the Schneider 28 and 35 without significant crosstalk issues. I also think that the 49x37mm sensor size is better balanced in terms of movement range on 90mm image circles. The distortion of the Schneider range is low enough so I wouldn't need to correct it in post, which I think is how it should work. The LCC is a necessary evil though as there is today no sensor that can handle this type of optics fully :-\
-
Marc:
That's a very good summary of the current situation.
My my work, I am using the 60MP Dalsa, with the 40mm Rodenstock. I can easily get 15mm of shift and in a push 18mm. The Rodenstock lenses have a internal market for the edge of the image circle, which on a 60MP Dalsa starts to give a hard vignette at 16mm or beyond. This is true with all the Rodenstock HR HR-W and I believe the latest HR-SW (90mm). If I use the 40mm at 18mm of shift many times this hard vignette will have detrimental effects on the image, mainly if the image has blue sky. So I have to consider the shot.
As Torger mentioned the 60XL is the wonder lens. I can take this lens to 25mm of shift (yes 25mm). I use the CF on mine which helps on the light fall off and I have started to use a CF on my 40mm Rodenstock.
I don't see the permanent damage from crosstalk on these types of shifts. There is color/sat fall off, but I feel that it's correctable. This is on a CCD full frame Dalsa chip. The downside is that I am pretty much limited to iso 50 and 140 (IQ260), so there are times that I cannot consider using this setup.
For my needs the CMOS in the 250 is showing great results all the way to around iso 1600, which in the field for me is plenty. However the current chip @ 50MP with a 1.3 crop and crosstalk issues is not a solution. So I still bracket exposures.
I would love to see a shutter solution for the tech market that allows 1/3 or even 1/2 stops since the copols are so rigid in their stops. These extra stops can make a difference (I feel) with the CCD chips since they don't have anywhere as much push available as the CMOS does.
I will also say that if Sony makes a 50MP 35mm chip, I strongly doubt they will consider crosstalk issues. They didn't on the current 36MP chip. The effect of extreme shifts on the 36MP 35mm chip are pretty harsh and do create some non recoverable color issues IMO. Sony has no shift lenses, none planned, and tend to plan their chips around what they offer in lenses. Those who shift I believe are a minority of total users so more than likely their needs may not get considered. Sony has bigger issues with front curtain vibration to worry about.
I do hope Phase One does figure out a solution for CMOS full frame that is tech camera friendly. I don't think we will see it next week in Germany, but it would be nice if it did get mentioned. But even on the MF market, I feel the tech camera use is still a minority of the total users.
Paul
-
To state the obvious: Stitching will allow to use the center of the image circle, and perhaps even use a longer focal length for additional resolution and makes flexible aspect ratios possible. The penalty of additional work after the shot is only modest when compared to the gain in flexibility. The image quality can easily be superior to wide angle alternatives.
Of course there are shooting situations which benefit from shooting single frames, so one should make informed trade-offs and choices.
Cheers,
Bart
-
The widest Solution for the HCam with anIQ250 is our new cut 2,8/14-24mm Nikon AF-S G-ED .
You will be able to see this on Photokina on the Novoflex booth in hall 2.1 Booth A40
Greetings from Germany
Stefan
-
"I would love to see a shutter solution for the tech market that allows 1/3 or even 1/2 stops since the copols are so rigid in their stops. These extra stops can make a difference (I feel) with the CCD chips since they don't have anywhere as much push available as the CMOS does. "
Paul - the Hcam allows 1/4 Fstops with all Canon Lenses.
Actually we could do 1/8 stops, but this is beyond needed accuracy.
Regards
Stefan
-
The widest Solution for the HCam with anIQ250 is our new cut 2,8/14-24mm Nikon AF-S G-ED .
Stefan, any images you could share? Any data on the image circle of the 14-24? Could you shift the lens on an IQ250 or is the image circle to small for this?
Thanks,
mjon
-
mjon
The Nikon with an EOS adapter on the HCam is unable to use movements, because there is no shift mechanism !
The lens will fully cover from 14-24mm for an IQ250 !
The movements on a Sony A7/R/S with the HCam Mirex EOS To Emount is about 4-5mm at 14mm, 10mm at 18mm and abou 6-7mm at 24mm.
I will do plenty of images when I am in Cologne, I will have a Sony A7S for testing and I am planning to do a lot of night shots with this combo.
Greetings from Germany
Stefan
-
Paul - the Hcam allows 1/4 Fstops with all Canon Lenses.
Actually we could do 1/8 stops, but this is beyond needed accuracy.
Regards
Stefan
Stefan:
Thanks for the info. Will the Hcam work with a lens in the Arca R mount? or Cambo, Alpa etc? I was interested in the ability to use say a Rodenstock 40mm HR-W. The current line up of electronic shutters, is pretty limited to what you can get in regards to shutter speeds, i.e. most max at 1/500 and don't go longer than 1". Plus they require a USB attached controller or something similar.
Paul
-
Paul
we make a version called HCam-B1v where the canon lensmount is replaced by a 120x120mm large "Hole".
Ontop of that we can mount about ANY adapter frames from the mentioned makers allowing adaptations of accoring lensboards.
This will work without much hassles from 60mm onwards and longer, with a recessed lensboard we could bring the lenses more backwards
so I guess we could reach maybe 32mm. But then we don´t have any movements.
Further: all the aperture settings are to be done from a copal shutter , the electronics that we use for our Canon adaptation is of course not working.
The problem with all these lenses is and stays: their flange focal distance is too short.
This limits their usability severly. In some cases I would say beyond any practical use for photography. This is definitely valid for the short Schneiders starting from 24mm which are
real wideangles, you will not find much usage for them, even Alpa has stopped offering them.
Rodenstock has understood and builds retrofocus now, But I fear their flange focal distance is still very short using a factor of only 2x.
Regards
Stefan
-
The Nikon with an EOS adapter on the HCam is unable to use movements, because there is no shift mechanism !
I see. I thought, there was an EOS adapter including a shift mechanism :-(
Thanks!
mjon
-
As Torger mentioned the 60XL is the wonder lens. I can take this lens to 25mm of shift (yes 25mm). I use the CF on mine which helps on the light fall off and I have started to use a CF on my 40mm Rodenstock.
Tried the 60XL on my Alpa setup and thought it just ok in its resolution, certainly inferior in the center to the Roddy 60mm f4 Digaron-S, but then that is not a stitching/shifting lens. I want it to be good, so believe me, am not a 'knocker' of other people's gear, but it has problems for me - sharpness I just mentioned, and that vignetting. Having to put a CF on a back working at ISO50, means incredibly slow effective ISO, and therefore long exposures. I need to freeze sidewalk action sometimes, and even distant people will blur at slower shutter speeds. That CF is not without cost.
I would love to see a shutter solution for the tech market that allows 1/3 or even 1/2 stops since the copols are so rigid in their stops.
Paul
The Alpa FPS has been out for nearly 2 years now! 1/3, 1/2, 1/4, or whole stops. whatever you wish.
-
Good points. Who knows, you might have a bad copy of the 60XL, it can happen. Most find it very sharp on 60 or 80MP backs. Mine is.
I didn't really overlook the FPS, as it's limited to Alpa mount only, as an Arca user I guess I am looking for a more universal solution. Something like the current electronics on the market, but with more range in speeds. What little that was per-released on the DHW version seemed to show promise, but the 2K cost per shutter, was not going to work for many.
Also I know the FPS has a a lot of lens mount options, but with the 28 Rod and 35 SK can you still have tilt? I just don't know enough about it.
No doubt right now for more variety in shutter speed options is seems to currently about the only game in town, and if I had already made the investment in Arca R mount, I would be definitely looking at one. I could have kept my 17mm and 24mm TS-E lenses!
Hopefully Arca and Cambo are working on something similar (frankly I am surprised neither has), but with maybe a more realist cost of entry, i.e. less than 8.5K. (that was the FPS price last time I checked).
Paul
-
Thanks everybody for the input.
Bart - I forgot mentioning stitching as an option. And I know some people get great results with that technique. But for me it is not useful. I work as a professional architecture photographer and go through lot's of images. Often I combine more than one exposure for interiors because of harsh contrasts - even with the high dynamic range of the A7R. I just can't stitch with the amount of images I need. And I like to see the final composition in what ever kind of viewfinder.
Stefan - I like the idea of your shift adapter for the Nikon 14-24mm a lot. I am curious how the image quality of that lens is when you use it outside of the "official" image circle (it is great within) on the A7R. A problem I see for my kind of shooting with this solution: how can I correct images for distortion? You have a lot of variables: the varying distortion depending on the focal length you use. And the different amount of rise / fall / shift you use. As far as I know none of this information can be recorded and saved in exif. So I do not see a chance to create profiles for correction of distortion. Maybe not a problem in a lot of cases for landscapes, but for architecture it is.
The new Alpa Adapters with movements for Hasselblad H and Contax 645 look interesting to me. That could be a way of using wide retrofocus MF-lenses on a 44x33mm-chip. No LCCs and eventually (hopefully) no problems with crosstalk too. Does anybody know the image circles of Hasselblad H or Contax 645 lenses. How much shift would lenses with 28mm or wider allow on a 44x33mm-chip? Would it be more than just a few millimeters? Hopefully they will create profiles for correction of distortion for their Alpa plugin software. But it looks as if they don't record the amount of used movements in exif which would be great.
Marc
-
Stefan - I like the idea of your shift adapter for the Nikon 14-24mm a lot. I am curious how the image quality of that lens is when you use it outside of the "official" image circle (it is great within) on the A7R. A problem I see for my kind of shooting with this solution: how can I correct images for distortion? You have a lot of variables: the varying distortion depending on the focal length you use. And the different amount of rise / fall / shift you use. As far as I know none of this information can be recorded and saved in exif. So I do not see a chance to create profiles for correction of distortion. Maybe not a problem in a lot of cases for landscapes, but for architecture it is.
With PT-Lens it is possible to correct for example the distortion of my great Pentax 645 A 35mm lens, shifted to any position.
So, in theory, you "only" need a lens profile and the knowledge of the focal lens and the horizon in your image.
-
Update on the performance of the TS-E 24mm L II in the shifted corner:
Since in another thread people refer to this test I wanted to give a short update: when I did this test I did not know that my copy of the TS-E 24 needed adjustment. The sharpness in the shifted corners in my original test is much worse than usual with the lens. Here are crops and full image on a different copy on the 5DsR with 12mm of shift. Sorry it is not on the same subject and for the grey low-contrast weather conditions, but I think one can clearly see the difference.
Marc
-
Thanks for the info!
Best regards
Erik
Update on the performance of the TS-E 24mm L II in the shifted corner:
Since in another thread people refer to this test I wanted to give a short update: when I did this test I did not know that my copy of the TS-E 24 needed adjustment. The sharpness in the shifted corners in my original test is much worse than usual with the lens. Here are crops and full image on a different copy on the 5DsR with 12mm of shift. Sorry it is not on the same subject and for the grey low-contrast weather conditions, but I think one can clearly see the difference.
Marc
-
The new Alpa Adapters with movements for Hasselblad H and Contax 645 look interesting to me. That could be a way of using wide retrofocus MF-lenses on a 44x33mm-chip. No LCCs and eventually (hopefully) no problems with crosstalk too. Does anybody know the image circles of Hasselblad H or Contax 645 lenses. How much shift would lenses with 28mm or wider allow on a 44x33mm-chip? Would it be more than just a few millimeters? Hopefully they will create profiles for correction of distortion for their Alpa plugin software. But it looks as if they don't record the amount of used movements in exif which would be great.
Marc
Without having any further research on the subject than posts on the web, it seems that Contax lenses have significantly larger image circles than the respective Hasselblad H ones, because H lenses where designed as to cope with the maximum of 37x49mm sensors that where the largest ones when the line first appeared.
On the Alpa site one can find examples of the Contax 35mm and 140mm lenses, which should be the two ones with the narrower image circle since the 35mm is the wider one and the 140, the one with the longest minimum focusing distance (1.3m!!!), the 35mm (according to Alpa) allows 6mm of shifts on an IQ-180 and the 140 is able to to do full 12mm shifts (on ALPA FPS) with the same back... It seems that Contax lenses will prove able to cope well not only with 33x44 sensors, but with up to 37X49mm ones too...
The matter is of particular interest to me too as I own the whole series (but the 350) of the C645 lenses, but it seems to me (by examining the lenses mechanically) that the 45-90mm vario-sonnar zoom, should prove a pleasant surprise if used with a view camera and the same should apply to the 120mm APO micro lens...
I'm now heavily converting a Rollei X-Act 2 camera which will end up having Contax 645 lens interface control and one will be able to use both Sony α7 cameras and multishot MFDBs on it without having to convert the rear standard (there is "special" care taken to avoid the recess that the α7 handgrip causes and rotation is retained with both the α7 or an MFDB). I'll soon be able to examine the image circle of all C645 lenses and will post the results...
-
Hi Theodoros, any particular reason for choosing the Rollei-Xact for this purpose?
-
I find many shift lenses really horrible in their edge and corner performance. Wouldn’t it be much better if we don’t apply shift lenses anymore?
Why not use normal lenses, tilt the complete camera, leave sufficient space for cropping and make our corrections in post? It’s what I do in C1.
These days with high megapixel sensors the resolution after correction in post seems still more than sufficient, and it is at least more than what we get with shift lenses.
Am I missing something here?
-
I find many shift lenses really horrible in their edge and corner performance. Wouldn’t it be much better if we don’t apply shift lenses anymore?
Why not use normal lenses, tilt the complete camera, leave sufficient space for cropping and make our corrections in post? It’s what I do in C1.
These days with high megapixel sensors the resolution after correction in post seems still more than sufficient, and it is at least more than what we get with shift lenses.
Am I missing something here?
I'm quite sure the good shift lenses still outperform cropping and keystone correction. But indeed it would be interesting to make a side-by-side comparison of the two methods.
-
Hi Theodoros, any particular reason for choosing the Rollei-Xact for this purpose?
Yes there is.... The camera is used for testing purposes only, as the end product will have different mechanisms design that will lead to even smaller overall size and greater range of movements, so the important thing for testing purposes was to use a camera as a base that would be easy to convert the plates of the standards and additionally would allow (again by relatively simple modification) for the front and rear standard plates to come as close as possible to each other... The Rollei X-act2 allows the front and rear standard plates to collide (with out a bellows on and after a relatively simple conversion of the rear standard only) yet retaining the yaw free design of the original design. With a WA bellows on, the total of the rear standard plate width + (sqeezed) bellows + front standard plate width adds up to only 24mm. Additionally, there are plenty of lens mounts available from Novoflex which are easily interchangeable to each other.
EDIT: An other requirement that the x-act-2 can be modified to fullfill, is that after the rear standard modification the rear standard's surface is completely flat...
-
I find many shift lenses really horrible in their edge and corner performance. Wouldn’t it be much better if we don’t apply shift lenses anymore?
Why not use normal lenses, tilt the complete camera, leave sufficient space for cropping and make our corrections in post? It’s what I do in C1.
These days with high megapixel sensors the resolution after correction in post seems still more than sufficient, and it is at least more than what we get with shift lenses.
Am I missing something here?
There is more to a shift lens than just the question of resolution. I prefer to compose my image in the viewfinder.
But if you only have the question of resolution and sharpness in mind:
For longer focal lengths the loss introduced by keystone correction is not too much. I don't see too much difference between my TS-E 90mm at full shift and my 85mm 1.8 with keystone correction.
But that is not true fur the wide lenses - especially as wide as 35, 24 or even 17mm. The resolution loss through keystone correction can be quite extreme - especially when you shift much.
Another effect reduces sharpness when you do not use a tilt-shift lens: if you want to take a photo of a facade and use a normal lens you have to tilt camera and lens to have everything in your frame. That means that your plane of focus is tilted too, parts of the facade are out of focus - and depth of field is not always deep enough to compensate for this.
A lens as good as the TS-E 24mm II will win against a keystone corrected normal lens.
-
The widest Solution for the HCam with anIQ250 is our new cut 2,8/14-24mm Nikon AF-S G-ED .
You will be able to see this on Photokina on the Novoflex booth in hall 2.1 Booth A40
Greetings from Germany
Stefan
Hi Stefan, have you tried the Nikkor 17-35mm f2.8 instead of the 14-24mm one? How the image circle behaves at respective focal lengths with this lens? I think it may have wider image circle at 24mm? ...how does it behave at 28mm and 35mm?