kinoglass
Established
Perhaps this will help understanding the nature of the beast:
http://www.vanwalree.com/optics/spherical.html
http://www.vanwalree.com/optics/spherical.html
50 Sonnar and the focus shift
- Thread starter tbarker13
- Start date
- Latest activity Latest activity:
- Replies 118
- Views 11K
Krosya
Konicaze
kinoglass said:
kinoglass
Established
Lets say for a moment that there is no range finder (RF) in your camera and that you are using a ground glass (GG) to focus your SONNAR at f1.5 So, using a magnifying loupe you focus your lens sharp on the GG. Then you decide that the exposure will require to close the lens down to f5.6, now with the aid of the magnifying loupe you check your focus again on the GG and find that it has shifted to objects further a distance from the object where you had before focused the lens at full aperture. So, the focus has shifted (POSITIVE) by an amount which depends on the aperture and on previous distance where the lens was focused at first.
Now, if you wanted to ad a RANGE FINDER, a way to measure distances, to your camera and couple it to the lens focusing mount, you will wonder what aperture to set the lens at to couple it to the RF. If you choose f1.5 the RF will be off when the lens is closed down at any aperture other than f1.5 , and if you choose f2.8 the RF will be off at any other aperture than f2.8
ZEISS had originally, quite arbitrarly choosen f2.8 to calibrate the RF to the lens only to find that many users were more interested in critical focusing at f1.5 and now they are producing their lenses with ZERO focusing error at f1.5 so that what you see in focus in the RF is what the lens is focusing at.
How does the lens behaves at other apertures can be addressed by using the Depth Of Field (DOF) scale marked on the lens barrel.
Now, if you wanted to ad a RANGE FINDER, a way to measure distances, to your camera and couple it to the lens focusing mount, you will wonder what aperture to set the lens at to couple it to the RF. If you choose f1.5 the RF will be off when the lens is closed down at any aperture other than f1.5 , and if you choose f2.8 the RF will be off at any other aperture than f2.8
ZEISS had originally, quite arbitrarly choosen f2.8 to calibrate the RF to the lens only to find that many users were more interested in critical focusing at f1.5 and now they are producing their lenses with ZERO focusing error at f1.5 so that what you see in focus in the RF is what the lens is focusing at.
How does the lens behaves at other apertures can be addressed by using the Depth Of Field (DOF) scale marked on the lens barrel.
kinoglass
Established
I think using this lens requires a clever and creative management of all its parameters. Is it better to have perfect focus reference at f1.5 or f2.8? I will say that in either case I will be bracketing the focusing index, since the real beauty of the lens is not so much its plane of focus but the image rendered in the out of focus area. Both combined creates a very three dimensional image, an effect always wanted since it conveys the spatial atmosphere of real scenes and not easy to fit in a flat surface.
Krosya
Konicaze
Follow up question, if I may, so if I use Sonnar lens which has auto mode to close the lens down at the point of exposure, when I focus it's wide open, yet camera will close it down to a needed f stop - it means even with SLR I wouldn't really focus correctly, just hope that DOF cover that same error?
Did I undersand you correctly?
Did I undersand you correctly?
Rico
Well-known
Correct. The Canon EF 50/1.2L is a case in point. It has just be released, has brand new design, and suffers focus shift! With everything electronic, the EOS system can compensate for this optical defect, although no fix is currently available. In the meantime, exact TTL focus can be achieved by stopping down and focussing manually - tedious, but better than the RF options.Krosya said:
kinoglass
Established
That is correct. A fast lens in a reflex system will shift focus as the lens is stopped down. Ansel Adams (and many others) always re-focused their gigantic view cameras at the lens shooting aperture, often f64!
Any large format photographer will do the same.
Another example. Darkroom printers do refocus their enlargers at the actual lens aperture, many lenses can hardly give you a flat focus plane. Ages ago Leica used to sell adapters to fit the 50/2 Summicron optical head to an enlarger! Sharp? in the center of the image yes, the rest was pretty bad!
There is no PERFECT lens, only compromises, can't bend the LAWS of PHYSICS.
Any large format photographer will do the same.
Another example. Darkroom printers do refocus their enlargers at the actual lens aperture, many lenses can hardly give you a flat focus plane. Ages ago Leica used to sell adapters to fit the 50/2 Summicron optical head to an enlarger! Sharp? in the center of the image yes, the rest was pretty bad!
There is no PERFECT lens, only compromises, can't bend the LAWS of PHYSICS.
Dan States
Established
It seems likely that Zeiss chose F2.8 as a compromise between full aperture accuracy and the problem of back focus at smaller apertures and at infinity.
If a Sonnar is set to focus accurately at 1.5, it will then back focus at all other apertures. Since the shift of focus is as much as a meter when focused at 5 meters you will likely see the shift quite clearly at apertures 2.8-4. Is there anyone here who can't tell when they have misfocused by a meter at F4?
This will mean a lack of clarity when focused at infinity. M8 users who are already shocked by the visibility of focus shift on their 35mm Leica lenses will not like the answer on the 50mm Sonnar.
When I first noticed this tendancy months ago some posters who had not owned the lens acted like I was some kind of idiot who could not understand the "qualities" of this lens. The plain fact is this lens has focus shift that exceeds its depth of field and as such should go back to the drawing board if it's really meant to be used seriously.
I suspect that Zeiss cranked out this lens primarily for their biggest market: Japanese COLLECTORS, who could care less about focus shift since they rarely use their cameras. (not that there is anything wrong with that...it's their business).
If a Sonnar is set to focus accurately at 1.5, it will then back focus at all other apertures. Since the shift of focus is as much as a meter when focused at 5 meters you will likely see the shift quite clearly at apertures 2.8-4. Is there anyone here who can't tell when they have misfocused by a meter at F4?
This will mean a lack of clarity when focused at infinity. M8 users who are already shocked by the visibility of focus shift on their 35mm Leica lenses will not like the answer on the 50mm Sonnar.
When I first noticed this tendancy months ago some posters who had not owned the lens acted like I was some kind of idiot who could not understand the "qualities" of this lens. The plain fact is this lens has focus shift that exceeds its depth of field and as such should go back to the drawing board if it's really meant to be used seriously.
I suspect that Zeiss cranked out this lens primarily for their biggest market: Japanese COLLECTORS, who could care less about focus shift since they rarely use their cameras. (not that there is anything wrong with that...it's their business).
Huck Finn
Well-known
Dan States said:
Dan, I know that you tested this lens extensively. Did you determine the 1-meter shift from your own tests or are you quoting the marketing statement from Zeiss that was posted on another thread?
I had a more recent correspondence with one of the Zeiss engineers in Germany. He stated the following:
"As a rule of theb, you may assume that up to f/5.6 the focus shift roughly corresponds to half of the total depth-of-field."
Since depth of field at 5 meters is less than a meter an any of the apertures you've mentioned, this would mean a focus shift of less than half a meter & it would mean a distance that is covered by DOF.
Dan States
Established
Huck, I set up the camera on a tripod and shot multiple images at exactly 15' distance. (It was from one end of my sunroom to the window frame of the other).
I focused carefully on the window frame and shot at 1.5,2.0 and 2.8. I printed the images full frame at about 6x magnifcation (on 8x10 paper). At 1.5 the front focus seemed to be about a 2.5 feet ahead of the point of focus...It got closer at 2.0 and was pretty much ok at 2.8. (by the way, the Planar Tony Rose sent me to test at the same time was dead on at all apertures)
I have a stack of shots from a walk in the woods that all show front focus. The full aperture shots are trashed, but the f2.0-2.8 shots are mostly useable if you don't mind the unusual placement of focus plane.
I tested the lenses (yes, two of them, plus the Planar) on BOTH of my M6 cameras because I thought I was either loosing my mind or I had a hosed up camera.
I finish by saying that believe it or not I'm not a very picky sharpness freak. I am only looking for 8x enlargements or smaller that are reasonably in focus.
On a happier note...I still love that Planar! It's SO easy to print negs from that lens...no flare and contrast that you can manage.
I focused carefully on the window frame and shot at 1.5,2.0 and 2.8. I printed the images full frame at about 6x magnifcation (on 8x10 paper). At 1.5 the front focus seemed to be about a 2.5 feet ahead of the point of focus...It got closer at 2.0 and was pretty much ok at 2.8. (by the way, the Planar Tony Rose sent me to test at the same time was dead on at all apertures)
I have a stack of shots from a walk in the woods that all show front focus. The full aperture shots are trashed, but the f2.0-2.8 shots are mostly useable if you don't mind the unusual placement of focus plane.
I tested the lenses (yes, two of them, plus the Planar) on BOTH of my M6 cameras because I thought I was either loosing my mind or I had a hosed up camera.
I finish by saying that believe it or not I'm not a very picky sharpness freak. I am only looking for 8x enlargements or smaller that are reasonably in focus.
On a happier note...I still love that Planar! It's SO easy to print negs from that lens...no flare and contrast that you can manage.
Huck Finn
Well-known
Thanks, Dan. I agree that 2.5 feet is a lot. You may note that earlier in this thread, I posted that in that same correspondence I received from Zeiss, they indicated that henceforth they will be adjusting the cam for f/1.5 at the factory. I hope that with the increased DOF will now cover the focus shift at the smaller apertures where it will now manifest. I'm looking forward to posts from those who have sent it back for this same readjustment.
Harry Lime
Practitioner
Boy Scout said:
The Noctilux also has a pretty serious focus shift.
Take a look at Erwin Puts' article on his site.
Oddly enough I never noticed it when I owned one.
;-)
Huck Finn
Well-known
According to Erwin, the CV 50/1.5 Nokton does as well. Zeiss says that its current version of the Sonnar is higher contrast than the vintage lenses & that this makes the focus shift stand out more. I'm betting that with some of these other lenses with focus shift, it show up as more of a softening of focus than a dramatic front focus as it soes with the C-Sonnar. Trading off contrast at wider apertures to preserve resolution was not uncommon with classic lens design.
john_s
Well-known
The Noctilux focuses perfectly at full aperture in my experience. Using the focus errors cited by Erwin, it looks as though the focus shift is just within DoF, using about 35um for the circle of confusion (i.e. not very stringent). All of the depth afforded by stopping down is behind the focused distance around the range from f1.4 to f4.Harry Lime said:
The error for the C Sonnar is more, which is why they chose to focus properly at f2.8.
Erwin has a page about DoF and it's worth reading, because there is more to it than the simple geometrical representation that is used to illustrate what happens, and from which the usual equation is derived. Here I would have to agree that the C Sonnar behaves better than simple calculations would suggest. (By simple calculations, I mean using the basic equation for DoF applied to the distance of optimum focus compared to distance set by the rangefinder).
It may be that most if not all wide aperture lenses have some focus shift with aperture changes. I suppose they're not usually noticed because (1) the lenses focus properly at full aperture and here the SLR has an advantage, and (2) the DoF more or less takes care of the rest.
Dan States
Established
I never noticed focus shift on my Noctilux until Puts told us all about it...Hmmm.
The Noctilux shift was not critical because it was within the depth of field at all apertures. I used my Noct a lot at 2.8-8 and it was always great. When Puts wrote about the shift I examined some negatives under a 25x microscope and yes, there is a bit of a shift.
As a black and white only shooter I spend a lot of effort managing contrast through development. The notion that the old Sonnar f1.5 had this problem but was so flat we didn't notice doesn't make sense to me. First of all because my old Sonnar has loads of contrast, and second because you can compensate for differences in development. I always at 15% to standard times when shooting older coated lenses, and 25% with uncoated lenses like the Summar. I can then print at grade 2 pretty consistently.
I DO own another lens with severe shift...The F2 Sonnar. I find it to shift focus dramatically on stopping down. Why it was worse than the 1.5 I don't know.
The Noctilux shift was not critical because it was within the depth of field at all apertures. I used my Noct a lot at 2.8-8 and it was always great. When Puts wrote about the shift I examined some negatives under a 25x microscope and yes, there is a bit of a shift.
As a black and white only shooter I spend a lot of effort managing contrast through development. The notion that the old Sonnar f1.5 had this problem but was so flat we didn't notice doesn't make sense to me. First of all because my old Sonnar has loads of contrast, and second because you can compensate for differences in development. I always at 15% to standard times when shooting older coated lenses, and 25% with uncoated lenses like the Summar. I can then print at grade 2 pretty consistently.
I DO own another lens with severe shift...The F2 Sonnar. I find it to shift focus dramatically on stopping down. Why it was worse than the 1.5 I don't know.
kinoglass
Established
Again, the Summilux does have less noticeable negligeable focus shift. Yet, the rendition of three dimensional space is not as good as compared to the Sonnar. That is the strong point of the Zeiss lens, that is what the designers wanted to give the photographers. Otherwise they suggest to use the Plannar design.
When the Sonnar came out back in 1932 Leitz (Leica) was in a pickle because they had nothing like it and existing patents kept them out of such a design. They ended up buying the rights from the English firm Taylor to make their 50mm f1.5 Obviously there wass no deal possible with Carl Zeiss. Duiring the Berlin Olympics Zeiss bet paid handsome dividends. The Sonnar was the baby not only for the photo journalists but for the cinematographers. Leni Rifenstahl one of them.
When the Sonnar came out back in 1932 Leitz (Leica) was in a pickle because they had nothing like it and existing patents kept them out of such a design. They ended up buying the rights from the English firm Taylor to make their 50mm f1.5 Obviously there wass no deal possible with Carl Zeiss. Duiring the Berlin Olympics Zeiss bet paid handsome dividends. The Sonnar was the baby not only for the photo journalists but for the cinematographers. Leni Rifenstahl one of them.
john_s
Well-known
kinoglass said:
Does anyone know what a clinical lens like the 50mm Summilux ASPH would be like with the mildest soft focus filter like a Softar 1? My understanding is that the Softar softens the image but it is still focused, which I think would be better than a C-Sonnar image that would be out of focus. Would the out of focus areas be ameliorated?
Mazurka
Well-known
john_s said:
The Softar (or its close copy Hoya Softener) does not change the focus.
The 50 ASPH's bokeh is not bad, if that's what you're saying. OTOH, the Softar does not improve the background with my 85/1.8 manual Nikkor. I would expect it to have minimal effect, if any, on any lens' bokeh also.
What the filter does is emphasizing ("bleeding") highlights and lowers the contrast a little. The effect is like the "Leica glow" of yesteryear.
kinoglass
Established
The problem of photography, painters had the same problem, is how to represent and render in a flat surface, wheter paper of a projection on a screen, a three dimensional scene. The problem of bringing every ray of light to coincide is resolved. How to deal with the out focus part of the image is the secret tool of the designer to make that three dimesional illusory touch as real as possible when the image is represented on a flat surface. And by the way a computer can't resolve this very subjective parameter. A sensitive human been is required. This is art not just science.
cellison
Member
What's the verdict?
What's the verdict?
I should have my first Leica sometime next week (M3 SS) and I'm looking for a 50mm lens for it. I like some example images I've seen from the C-Sonnar T* 1.5/50 ZM... beautiful bokeh but I'm somewhat concerned about the focus shift issue people seem to be having. I read that Zeiss have tweaked the lens so it's sharp wide open rather than at 2.8 but I'm wondering how far off it goes when stopped down a few notches... a meter? more?
Anyone here with a recent version of the lens have any sample images or comments?
It comes down to a new C-Sonnar T* 1.5/50 ZM or a used Summicron for roughly the same price.
Chris
What's the verdict?
I should have my first Leica sometime next week (M3 SS) and I'm looking for a 50mm lens for it. I like some example images I've seen from the C-Sonnar T* 1.5/50 ZM... beautiful bokeh but I'm somewhat concerned about the focus shift issue people seem to be having. I read that Zeiss have tweaked the lens so it's sharp wide open rather than at 2.8 but I'm wondering how far off it goes when stopped down a few notches... a meter? more?
Anyone here with a recent version of the lens have any sample images or comments?
It comes down to a new C-Sonnar T* 1.5/50 ZM or a used Summicron for roughly the same price.
Chris
Similar threads
