BobYIL
Well-known
It is not difficult to detect if a lens focus-shifts; we can also measure how far it does. First place two rulers perpendicular to each other as shown in the picture, the vertical one at 45 degrees to the horizontal axis. The number "20" is our reference point to focus and the testing distance is 1m exact from lens flange.
If your lens has no focus-shift then starting from wide open it will show such a performance (Nokton 50/1.5 Asph. @f1.5, 2, 2.8, 4 and 5.6)
The C-Sonnar is known of shifting focus. Here @f1.5, 2, 2.8, 4 and 5.6. This is a typical focus shift case:
Finally the reference class Summicron 50/2, current formula (!), @f2, 2.8,4, 5.6. Would you believe?!?
Conclusion:
If you like to use your lenses at wide apertures, then better be informed of how much they focus-shift with different apertures. Such optical discrepancies may become more apparent on digital bodies than film. Also sometimes they may deceive us by making us think our rangefinder needs recalibrating. Once you know how they "behave" then it's not difficult to compensate for such shifts.
If your lens has no focus-shift then starting from wide open it will show such a performance (Nokton 50/1.5 Asph. @f1.5, 2, 2.8, 4 and 5.6)
The C-Sonnar is known of shifting focus. Here @f1.5, 2, 2.8, 4 and 5.6. This is a typical focus shift case:
Finally the reference class Summicron 50/2, current formula (!), @f2, 2.8,4, 5.6. Would you believe?!?
Conclusion:
If you like to use your lenses at wide apertures, then better be informed of how much they focus-shift with different apertures. Such optical discrepancies may become more apparent on digital bodies than film. Also sometimes they may deceive us by making us think our rangefinder needs recalibrating. Once you know how they "behave" then it's not difficult to compensate for such shifts.
Robert Lai
Well-known
Was this on film or digital?
Popflash have a note on their C-sonnar listing that states that the lens is optimized for f/1.5 on film, but on digital, it may be slightly different. Apparently the digital M have a very slightly different lens flange to sensor distance than film. With film, the manufacturer of the camera has to guess at how much bowing in of the film will occur towards the lens.
Norman Goldberg (father of DAG), has a detailed explanation of film flatness issues in his book on camera design.
Popflash have a note on their C-sonnar listing that states that the lens is optimized for f/1.5 on film, but on digital, it may be slightly different. Apparently the digital M have a very slightly different lens flange to sensor distance than film. With film, the manufacturer of the camera has to guess at how much bowing in of the film will occur towards the lens.
Norman Goldberg (father of DAG), has a detailed explanation of film flatness issues in his book on camera design.
Chriscrawfordphoto
Real Men Shoot Film.
What Summicron? 50mm, 35mm, 75mm, 90mm? Which version of each focal length?
My 50mm tabbed (version IV) does not show any shift, and I have tested it using a method similar to what you describe.
My 50mm tabbed (version IV) does not show any shift, and I have tested it using a method similar to what you describe.
BobYIL
Well-known
The C-Sonnar shifts exactly the same on film and digital. Both M-flanges, digital or film have exactly the same register distances of 27.8mm. The digital, due to higher resolution, enhances the OOF-in focus differences more than what film does.
BobYIL
Well-known
"Finally the reference class Summicron 50/2, current formula (!),"
Mine is too the tabbed one and it focus-shifts.
Focus-shift is not a discrepancy based on sample.. it's a deficiency (or fault) of the optical construction.
taskoni
Well-known
I don't get it even illustrated so nicely, but my DR Summicron focus where I want him to focus to and rarely misses 
Regards,
Boris
Regards,
Boris
BobYIL
Well-known
I don't get it even illustrated so nicely, but my DR Summicron focus where I want him to focus to and rarely misses
Regards,
Boris
True.. no shift is observable; here's the Summicron DR, from left to right f2, 2.8, 4 and 5.6.
BobYIL
Well-known
This is Summicron V3. Another faultless focusing performance (f2, 2.8, 4 and 5.6)
BobYIL
Well-known
And finally Summicron Collapsible. Focus shift is noticable.
BobYIL
Well-known
And here are the versions tested:
Pastor Chris
Well-known
I just dropped off my version IV 50mm Summicron at Leica in Nj for an issue that they identified as "back focusing". They are going to make some adjustments with the focusing helix, among others, and change the mounting flange. They tested it with test charts set up similar to your rulers, although the center focus target is a red Leica dot from a camera. It will be interesting to see if it focus shifts when I pick it up. Although, from your description it seems as though a lens that focus shifts cannot be adjusted, it is a design fault. Is that correct? One advantage to being able to drive to Allendal is that they guys in the back can answer your questions and play with things while you are there. I'll let you all know.
BobYIL
Well-known
Are you sure it is not due to the lack of rangefinder calibration; i.e. you focus on the rangefinder to 100cm but on film/sensor plane it focuses to 105cm for example? (Naturally to be tested wide open only..)
Pastor Chris
Well-known
Oh yes, they checked the camera at the same time; quite a nice operation really, they also cleaned the sensor for me no charge. I know this is not a big deal on and M8 but it is nice thorough customer service.
njh411
slimnics
i would like to try this, but the images are all gone.
Ko.Fe.
Lenses 35/21 Gears 46/20
I check lenses after CLA. With similar to this:
http://www.firstchoicephoto.ca/FocusTestChart.pdf
Minimum focus distance, by 45 angle. I make zero line thicker with marker and use camera in portrait orientation.
I calibrate lenses on maximum aperture. My mint Cron 50 v4 focus shifts more than LNIB Nokton 50 1.5 and Planar 50/2. Rigid in user condition was also shifting, but less than v4.
http://www.firstchoicephoto.ca/FocusTestChart.pdf
Minimum focus distance, by 45 angle. I make zero line thicker with marker and use camera in portrait orientation.
I calibrate lenses on maximum aperture. My mint Cron 50 v4 focus shifts more than LNIB Nokton 50 1.5 and Planar 50/2. Rigid in user condition was also shifting, but less than v4.