Luminous Landscape Forum
Equipment & Techniques => Medium Format / Film / Digital Backs – and Large Sensor Photography => Topic started by: MattBeardsley on December 21, 2010, 05:25:25 pm
-
Hello LL,
I had a chance to cruise around Oakland CA's China Town with a Leica S2 and 70mm f2.5... It's a remarkable camera. I can't wait to get my hands on one for more extensive testing to really see how it stacks up against other 40 MP competition. My initial impression of image quality is very favorable, and the ergonomics are almost perfect... check it out:
http://photoartsmonthly.com/blog/2010/12/21/first-impressions-the-awesome-leica-s2/
thoughts? Also, please let me know if you catch any errors or weak points... I look forward to hearing from you!
Matt
[Jan. 17, 2011 update. I've scaled my blog up to a more formal web-based magazine and have updated the above link. Thanks for taking a look! ~matt]
-
Matt, the detail in those is insane. All hand held?
- N.
-
Matt, the detail in those is insane.
Yea. I had to use the F word when I clicked on the 100% crops.
-
The Achilles Heel is the terrible Chromatic Abberation in the 70mm (+others?) at wider apertures.
try it on a sunny day at F2.5-3.5 and see what there is around light:dark edges. great rings of red and green fringing.
Its one of the worst lenses in digital MF for this.
I could list other reasons why I didn't buy it, but this is a major one, and makes it near unusable with wide apertures.
Leica out did themselves: too complicated a lens design with multiple aspheric surfaces, creates problems.
Leica need to bring out a short barrel 70mm, smaller, lighter, cheaper, with ~f3.5 aperture, and concentrate on fixing the CA.
-
I thought CA is easily fixable in Raw?
Edmund
The Achilles Heel is the terrible Chromatic Abberation in the 70mm (+others?) at wider apertures.
try it on a sunny day at F2.5-3.5 and see what there is around light:dark edges. great rings of red and green fringing.
Its one of the worst lenses in digital MF for this.
I could list other reasons why I didn't buy it, but this is a major one, and makes it near unusable with wide apertures.
Leica out did themselves: too complicated a lens design with multiple aspheric surfaces, creates problems.
Leica need to bring out a short barrel 70mm, smaller, lighter, cheaper, with ~f3.5 aperture, and concentrate on fixing the CA.
-
my expert friend (and he won a MacArthur Genius award for this kind of thing) says that only symmetrical designs are easily fixed with RAW CA tool. Asymmetric not so good.
and anyways, didn't Leica boast that their lens don't need any post production fixing - that it was a sign of poor design?!
-
my expert friend (and he won a MacArthur Genius award for this kind of thing) says that only symmetrical designs are easily fixed with RAW CA tool. Asymmetric not so good.
and anyways, didn't Leica boast that their lens don't need any post production fixing - that it was a sign of poor design?!
What CA are you referring too? I have this exact lens on an S2 and see no CA. maybe you had a duff lens.
Branches of trees against bright sky, corners of frames, classic place to look for red/cyan fringing. On my images, none at all.
-
You are very lucky if that is the case.
However I am certainly not the only one to notice this. Two high end digital camera stores I know spotted it in their testing, and try the report at Digilloyds DAP site for more info (paid site).
If you are happy/lucky with your camera and lens, then great - end of story.
-
You are very lucky if that is the case.
However I am certainly not the only one to notice this. Two high end digital camera stores I know spotted it in their testing, and try the report at Digilloyds DAP site for more info (paid site).
If you are happy/lucky with your camera and lens, then great - end of story.
You were referring to Chromatic Aberration which I wrongly took to mean the kind of aberration that can be corrected in Lightroom etc. I read DigiLloyds references to Purple Fringing and Color Bokeh and it seems it is normal for fast lenses and disappears by f8. I'd not go so far as to call it the 'worst of all medium format lenses' however, but regardless of that, it's an astonishing lens in so many other characteristics and I really like having the extra light gathering of the f2.5 specification. It makes it very easy to focus and may have a lot to do with why the AF is so accurate.
-
N,
Good question; yes all the examples are hand held. The S2 feels a little smoother than big-box MF cameras in terms of mirror slap, which is nice for hand holding. Though I wouldn't get rid of the tripods just yet!
Also, I'm surprised others have experienced fringing or CA with this lens/camera combination. I didn't notice any, though the day of our test shoot was overcast and not especially high contrast. I'll take a very close second look and report back..
Since there isn't a "Phocus" or "Capture One" for Leica, does anyone know if Lens profiles are yet available for the S2 and Lightroom?
~Matt
-
OK... After some extreme pixel peeping, I found only one occurrence of any noticeable fringing and have attached an example. The first attachment is a 100% crop of the second attachment and shows the only area of fringe I found any any of the test images. The are is harshly backlit and a dark color surface, so it's the ideal extreme contrast to lead to problems. For what it's worth, Lightroom's Defringe > Highlight Edges tool virtually eliminated what is visible here.
I'd venture to say the S2 (at least the copy I tested) is very good at controlling fringing, especially as the subject's coat shows no fringes, nor any of the extreme highlights in the background. It's not an ideal shot to demonstrate S2 image quality, as it was a hasty capture and is not ideally sharp, exposed, etc. It is, though, the only image I found with any noticeable fringing. I'd certainly want to see further tests in more high-contrast lighting, like the backlit tree mentioned in an above post... can anyone share examples of problematic images?
Thanks for the question... I hope that helps!
~Matt
-
maybe you had a duff lens.
so sample variation? On a Leica lens at this price tag?
Interessting...
-
OK... After some extreme pixel peeping, I found only one occurrence of any noticeable fringing and have attached an example. The first attachment is a 100% crop of the second attachment and shows the only area of fringe I found any any of the test images. The are is harshly backlit and a dark color surface, so it's the ideal extreme contrast to lead to problems. For what it's worth, Lightroom's Defringe > Highlight Edges tool virtually eliminated what is visible here.
I'd venture to say the S2 (at least the copy I tested) is very good at controlling fringing, especially as the subject's coat shows no fringes, nor any of the extreme highlights in the background. It's not an ideal shot to demonstrate S2 image quality, as it was a hasty capture and is not ideally sharp, exposed, etc. It is, though, the only image I found with any noticeable fringing. I'd certainly want to see further tests in more high-contrast lighting, like the backlit tree mentioned in an above post... can anyone share examples of problematic images?
Thanks for the question... I hope that helps!
~Matt
I saw fringing at the first 100% crop you showed. The white van around the headlights. Besides that the details of the images look stunning.
-
Well, the images look very good, but I don't think that the 100% crops look any better than similar 100% crops from my CFV-39 and the 80mm Planar. Perhaps it's just more about the 'look' of 40MP MF files in general.
John
-
Well, the images look very good, but I don't think that the 100% crops look any better than similar 100% crops from my CFV-39 and the 80mm Planar. Perhaps it's just more about the 'look' of 40MP MF files in general.
maybe.
Based on the few samples I've seen from the S2 and the 70mm lens my impression is it is much better wide open and at close and mid distances than the 80 Planar. Whereas almost any shot I've seen from the S2/70mm at wide distances and stopped down was quite soft at the edges (center still very sharp)... I think here the Planar is even better. Maybe the 70mm is somehow optimzied for close and mid distances...?
Again... these are just impressions based on few samples of the 70mm available on the web and not backed up by own experience.
-
The ONLY really damaging criticsms levelled at the S2 have been that the AF focus works badly at large distances, and that there is no fine-focus adjustment so if a lens is not ok for the body you have to send both in, and sometimes lenses really are out of tolerance. Apart from that, it would seem that the camera is pretty much as described, and decently modern and integrated.
Note that some smartass with too much maths training on this forum worked out that a 10cm focus error at 5m is equivalent to a much lower rez camera, at some real-world aperture, which confirms it is worth taking a hard look of the real-life focus performance of any camera before buying it.
Edmund
-
maybe.
Based on the few samples I've seen from the S2 and the 70mm lens my impression is it is much better wide open and at close and mid distances than the 80 Planar. Whereas almost any shot I've seen from the S2/70mm at wide distances and stopped down was quite soft at the edges (center still very sharp)... I think here the Planar is even better. Maybe the 70mm is somehow optimzied for close and mid distances...?
Again... these are just impressions based on few samples of the 70mm available on the web and not backed up by own experience.
Leica certainly have a tradition of producing lenses which work best at or near maximum aperture. Whereas the 80mm Planar is best at f8.
John
-
I would expect that the "defringe" control in LR3 would remove virtually all of this 'defect'. There is loads of purple/green fringing in RAW shots from the 645D, and they vanish in LR.
Besides, if you're working in B&W, like a real man, it hardly matters. ;D
- N.
-
Besides, if you're working in B&W, like a real man, it hardly matters. ;D
Actually chromatic aberration / purple fringing removal (or using a system which does not have any) can be a very important step in B+W workflow.
If you're using sophisticated RGB B+W conversion, like a real man, then purple fringing / sensor bloom or chromatic aberration can create unwanted dark or light halos around your subject matter. For instance if you are using a red-biased black and white conversion to make more brighter/creamier skin tones you'll find that any dark subject matter with a hard edge has a (sometimes very strong) white halo where the red chromatic aberration was pushed up in tonality. Even worse is when you got a dark outline on the interior side of a bright edge - very unnatural.
Also not all programs with a "defringe" or "chromatic aberration tool" perform at the same level for any given camera/lens.
(http://www.captureintegration.com/wp-content/uploads/2_C16tweaked-vs-LRdefault-2.jpg)
I was not able to remove the CA of this image with LR (right) without damaging other parts of the image but it was easy in C1 with no deleterious effects (left) - more about this image (http://www.captureintegration.com/2010/12/03/noise-reduction-in-c1-6/)
-----
@Edmund: CA and blooming can be very easy to remove in post. But it can also be a nightmare. It depends very much on which raw processor, and what camera/lens. The most difficult chromatic aberration is often CA in slightly out of focus areas where the algorithms have a hard time telling the difference between magenta/green subject matter and magenta/green chromatic aberration.
Doug Peterson (e-mail Me) (doug@captureintegration.com)
__________________
Head of Technical Services, Capture Integration
Phase One Partner of the Year
Leaf, Leica, Cambo, Arca Swiss, Canon, Apple, Profoto, Broncolor, Eizo & More
National: 877.217.9870 | Cell: 740.707.2183
Newsletter (http://"http://www.captureintegration.com/our-company/newsletters/") | RSS Feed (http://"http://www.captureintegration.com/2008/08/11/rss-feeds/")
Buy Capture One 6 at 10% off (http://"http://www.captureintegration.com/phase-one/buy-capture-one/")
-
maybe.
Based on the few samples I've seen from the S2 and the 70mm lens my impression is it is much better wide open and at close and mid distances than the 80 Planar. Whereas almost any shot I've seen from the S2/70mm at wide distances and stopped down was quite soft at the edges (center still very sharp)... I think here the Planar is even better. Maybe the 70mm is somehow optimzied for close and mid distances...?
Again... these are just impressions based on few samples of the 70mm available on the web and not backed up by own experience.
I find I use the 35mm more for landscape shots, the 70mm is less useful for 'front to back sharp' images. Here is a screengrab from the 35mm displayed at 100%, it's the extreme bottom left corner of the frame, a hand held grab shot whilst in the city a few months back. I think you will agree there is zero corner sharpness fall-off. f8@1/250.
What I do see with this lens is modest curvature of field, like so may wide lenses. I thought this was a problem until I realised that its a bit like having built in 'tilt'. Shooting vertically, the foreground is sharper than it would be if it was a flat field lens. Lloyd Chambers makes the point that lenses are rarely designed to be flat field, and to work with the lenses' characteristics rather than against them. This is also where MTF charts fall down as they are made from flat plane targets. Resolution does not necessarily drop off in corners, it can easily be simply out of focus whilst being still just as sharp in a different plane of focus.
The 70mm is as sharp at f2.5 as it is at f8, no question. (Same with the 35mm, although the curvature of field is more pronounced wide open) No other medium format lens is fully sharp wide open, even the amazing Digitars need to be used at f5.6 or f8 for max IQ. On the S2 you don't choose an aperture for quality reasons, just for DOF reasons.
For your entertainment I have attached the (measured) MTF from the new 120mm macro - the contrast at 40lpm at the edge is better than most lenses in the centre!
-
I find I use the 35mm more for landscape shots, the 70mm is less useful for 'front to back sharp' images.
the question was if it is useful for left to right sharp images at wide distances :-)
-
the question was if it is useful for left to right sharp images at wide distances :-)
OK, then the answer is yes, no problem.
No fall-off of sharpness at edges at middle to far distance. Slight curvature of field wide open but for the sorts of landscapes we are talking about you'd be shooting for a bit of DOF anyway. This is a shot from the Barrier Reef, 100% crop from right side of frame. The old pontoon is about 200m away. f8/500th, handheld since the platform I was on was not rock solid. This image is 1100px wide out of 7500px for the full frame.
-
This business of CA is all a bit odd. As Doug says, it really does matter even if you are working in B/W, because it smudges the edges of things like tree branches against the sky. When I was working with film on the 'Blad, I never really saw any CA at all, even when shooting colour. Then, when I got the 39MP back, to my horror I had loads of CA with the 50mm and to a lesser extent the 60mm. Really bad, in certain circumstances. The 80mm and upwards were fine. I just thought that this was inevitable with old W/A lenses and digital sensors, and resigned myself to dealing with it in software.
Then, I discovered I had a defect on my brand-new CFV back, a vertical magenta line which showed up at 100% about one third in from the left of the frame (a "column fault", they called it). To cut a very long story short, the DB went back to Denmark under warranty, they tried to fix it by re-calibrating but could not, and in the end they fitted a brand-new Kodak sensor, checked it through and returned it.
All was good, and then I realised after a few weeks of use that the CA problem with my W/A lenses had gone. Well, just a tiny bit of fringing if you went mad with TV aerials against the light, but nothing like what it had been. In fact I now have to really look for CA to find it at all. Now, I don't understand the tech stuff behind all of this, but it seems to me that CA is not necessarily just a lens problem, but can also be affected by sensor alignment and calibration too, somehow. Which might explain why we are getting conflicting reports on this with regard to the Leica S2.
John
Edit - having just re-read the above, I am struck by the thought that it all sounds pretty improbable, because how could the sensor calibration affect CA? And yet, that's what seemed to happen.
-
Nick, thanks for posting the crop.
-
This business of CA is all a bit odd. As Doug says, it really does matter even if you are working in B/W, because it smudges the edges of things like tree branches against the sky. When I was working with film on the 'Blad, I never really saw any CA at all, even when shooting colour. Then, when I got the 39MP back, to my horror I had loads of CA with the 50mm and to a lesser extent the 60mm. Really bad, in certain circumstances. The 80mm and upwards were fine. I just thought that this was inevitable with old W/A lenses and digital sensors, and resigned myself to dealing with it in software.
Then, I discovered I had a defect on my brand-new CFV back, a vertical magenta line which showed up at 100% about one third in from the left of the frame (a "column fault", they called it). To cut a very long story short, the DB went back to Denmark under warranty, they tried to fix it by re-calibrating but could not, and in the end they fitted a brand-new Kodak sensor, checked it through and returned it.
All was good, and then I realised after a few weeks of use that the CA problem with my W/A lenses had gone. Well, just a tiny bit of fringing if you went mad with TV aerials against the light, but nothing like what it had been. In fact I now have to really look for CA to find it at all. Now, I don't understand the tech stuff behind all of this, but it seems to me that CA is not necessarily just a lens problem, but can also be affected by sensor alignment and calibration too, somehow. Which might explain why we are getting conflicting reports on this with regard to the Leica S2.
John
Edit - having just re-read the above, I am struck by the thought that it all sounds pretty improbable, because how could the sensor calibration affect CA? And yet, that's what seemed to happen.
Maybe the new sensor had cells with a different design from the old one?
Maybe the cover glass creates prismatic effects?
Edmund
-
my expert friend (and he won a MacArthur Genius award for this kind of thing) says that only symmetrical designs are easily fixed with RAW CA tool. Asymmetric not so good.
Hmm. I would have thought it was the other way round.
Lateral chromatic aberration causes the sort of colour fringing where each spot is imaged as a stubby spectrum, always orientated radially from the centre of the image. There is basically a slight magnification difference as a function of wavelength, and it can be corrected to first order by re-dimensioning the Red and Blue channels to coincide spatially with the Green. This is easily done in digital post processing and there is a proper optical basis for doing it: each of the RGB channels contained a pretty sharp image; they were just out of register with each other.
Axial chromatic aberration, or secondary spectrum, is a lot messier. It does not vary with field position, only with f-number - it's the type that is worst at wide apertures - and axial distance from the plane of focus. Only one (usually) of the RGB channels contains a sharp image; the others are out of focus. There is no rigorous optical basis to post-processing algorithms for correcting this - from what I've seen, they seem to work by cosmetic fudges like desaturating the fringe colour, rather than by physics - although, from Doug's attachments, it's clear that some of the fudges are more successful than others. At the plane of focus for one colour, you cannot push the other defocused light/colours back into focus, post-capture. The closest you can come is with some sort of adaptive, channel-selective deconvolution process. But that is non-linear, unpredictable, locally variable, and stuff like the noise statistics then go out the window. Outside the plane of focus, where the axial chromatic is manifested as colour fringing on the bokeh 'blobs', even deconvolution alone won't work, because there is also a spatial offset between the defocused fringes of difference colours. That sounds somewhat like lateral chromatic, but it doesn't have its global invariance; rather, this effect varies locally with each bokeh point. To truly correct it would require knowledge of the degree of defocus of all points in the image. Talk about messy!
So what about the lens design? Symmetrical lenses exhibit minimal lateral chromatic aberration, by virtue of their symmetry. Since they lack the easily correctable type of fringing, I don't understand why they would be more "easily fixed with RAW CA tool" than asymmetric lenses, which can show both types of aberration, including dollops of the more easily correctable type.
Ray
PS. Do I now qualify for a MacArthur Genius award? ;D
-
How about more ELD glass ?
Frank
-
The manifestation of longitudinal/axial CA (i.e., how much it shows up in the image) is not only a property of the glass, but also of the spectral sensitivity of the sensor. In other words, the entire optical system must be considered.
-
The manifestation of longitudinal/axial CA (i.e., how much it shows up in the image) is not only a property of the glass, but also of the spectral sensitivity of the sensor. In other words, the entire optical system must be considered.
One needs to consider assymetries in the sensels, I would think.
Edmund