Calibrating Summilux 50/1.4 pre-asph for M9?

[Alex Krasotkin]

Guys, I have noticed focus shift, when using my Summilux 50/1.4 pre-asph (late silver version) at f/1.4 and even at f/2 sometimes with Leica M9. I had never experienced such problems using this lens on M7 and M6 wide open.

Whait is a solution for this issue? i guess that I need to send it to Solms for calibration and 6bit coding? Will it cure the problem?

many thanks,
Alex

 

[marcoinhawaii]

I sent my 35mm summilux to be adjusted for this reason. Hopefully Leica can adjust to work properly with M9. I hope to post a proper answer soon.

 

[Doug]

The ideal adjustment will be to put the best focus at the aperture you want. It's just a fact that the digital sensor is pickier about focus than film. I've sent several lenses (including the focus-shifting C-Sonnar) to DAG for 6-bit coding and focus check/adjustment. Servicing like this can put the focus at a chosen aperture, but does not otherwise solve the "focus problem." The focus will still shift as you change apertures. For example, the C-Sonnar focus was adjusted to be "on" at about f/2, and this minimized the shift at f/1.5, while at smaller apertures like f/4 depth-of-field enclosed the shift (the best focus is in the part of DOF nearer the camera).

 

[jcm0]

Alex,

I am surprised to read your post. I too have a 50mm summilux pre-asph (v3) model, and have seen some (limited) focus error at f1.4 althought I did very recently send my lens to Leica Solms for focus alignment (mine was way off at infinity).

I am not at home to verify, but there is a difference between focus shift and a rangefinder/lens calibration problem. What I have understood, is that the focus shift is related to the lens optical design and cannot be cured.

There is a focus shift when the focus behavior change while changing the lens diaphragm. That means the real focus point at a given distance is different when you change the lens diaphragm.

I would be surprised if this is the case with the 50mm summilux pre-asph lens as I didn't have seen any post on the subject. I would think this is rather a M9 mis-alignment problem (assuming your lens hasn't been modified by any preceeding user as was mine), which can be fixed.

Here are the measurements I made at f1.4 with a ruler and taking pictures at an angle of 45° ; I will make the measurements at other apertures as soon as I have a little time:

0.7m -> +1.5cm (post focus, ie the focus point on the final image was at 71.5cm whereas I did focus at another point which was nearer, at 0.7m - value indicated by the lens),
0.8m -> +1cm,
0.9m -> +0.5cm,
1m -> -1cm,
1.2m -> -2cm,
1.5m -> -2.5cm,
2m -> -3cm

In my case where my lens was modified by a preceeding user, these limited focus errors can be a mix between my M9 rangefinder misalignment ant a not perfect focus alignment on the lens by Leica Solms. The most inconvenient focus error is - with my usage of the lens - at short range.

Let me know what test you did conduct to conclude to a focus shift on your lens. Please also note that the focus errors may have been masked on your M7 due to either one or a mix of the following reasons: you rather use grainy films, and/or you made shots at lower speed = shake (with the M9 one rather use higher iso hence higher shooting speed), and/or you changed your shooting style (ex: more portrait/close-up, more light in your pictures, ...)

For Marco: All the 35mm summilux except the latest FLE version are known to be prone to focus shift (that's the reason why Leica built the FLE version).


Regards,

JC

 

[icebear]

Your M9 should be checked for sensor position and fine tuned if out of spec. Then your lens should be matched to the calibrated camera. I had this done for my M9 , 2/90 'cron and Noct' . Perfect now.

 

[jcm0]

Hi Klaus,

Did you remember the price for the your M9 sensor position fine tuning and the price for your 2 lens tweaking (repair made by Leica Solms ?) ?

Thx

JC

 

[jaapv]

Your M9 should be checked for sensor position and fine tuned if out of spec. Then your lens should be matched to the calibrated camera. I had this done for my M9 , 2/90 'cron and Noct' . Perfect now.

No - it should not be matched to the camera. Both lens and camera are adjusted to a fixed standard. Then if that is done there can be a bit of tolerance matching between critical lenses and the camera.

 

[V-12]

Yes I agree, it is about time that it was more widely understood that Leica nor any repair technician worth using would calibrate a lens to a particular body, or visa versa. Each is calibrated separately to a standard, anything else is a bodge job even if it is possible to make a lens work only with one particular body.

Earlier lenses were not calibrated to the tighter tolerances necessary to work on a digital body, so I would imagine this is the main problem, focus shift aside. So I'd say the lens needs to be calibrated again.

 

[Alex Krasotkin]

Guys, many thanks for your replies. So I assume, if I check my M9 with demo 50/1.4 asph in the local Leica shop and will find no focusing issue, I need to send only my lens to Solms?

 

[ferider]

Yes I agree, it is about time that it was more widely understood that Leica nor any repair technician worth using would calibrate a lens to a particular body, or visa versa. Each is calibrated separately to a standard, anything else is a bodge job even if it is possible to make a lens work only with one particular body.

Earlier lenses were not calibrated to the tighter tolerances necessary to work on a digital body, so I would imagine this is the main problem, focus shift aside. So I'd say the lens needs to be calibrated again.

Right.

Then what is the "standard" optimal aperture and focus distances of the v2 Summilux that Leica should calibrate for ? Obviously in the 60s, Mandler must have known about the M9

Also, assume Leica adjusts Alex's lens to focus a single inch closer on his M9 at minimum focus. You don't think he will notice the difference with his film bodies ?


Alex still hasn't told us how much front or back focus we are talking about. If it is an inch or two, it is indeed in the range of the 'lux shift, and whoever calibrates the camera will need to know from Alex which aperture and focus distance he wants to be optimal.

If it is more than that, his camera is kaput and needs to be fixed and adjusted to his 'lux (since it works on his film cameras). Nothing to do with "tighter tolerances", "standard" calibration of a lens from the 60s, bla, bla, bla.

Roland.

 

[jcm0]

No - it should not be matched to the camera. Both lens and camera are adjusted to a fixed standard. Then if that is done there can be a bit of tolerance matching between critical lenses and the camera.

That's interesting to know. Thank you Jaapv.


Roland, you're talking of the 'lux focus shift. Do you own this lens ?

I made a short hand-held test last night, and would say that I didn't see any obvious (more than 1cm) focus shift (ie : point of focus shifting when changing the lens aperture). That's a good news for me . I'll check that more carefully this week-end.

 

[douglasf13]

Even my 50 Summicron Collapsible has noticeable focus shift on my M9. These silicon sensors are a lot more finicky than film. At F2, the Cron slightly front focuses. At f2.8, it is dead on. From f4 on, the focus point slowly drifts toward the back a bit. I basically ignore the focus shift from f2.8 on, and, when shooting wide open, I know to slightly nudge focus to the back (or move the camera slightly forward.)

 

[CaptZoom]

Posting so that I can follow this thread as it develops.

 

[flash]

Right.

Then what is the "standard" optimal aperture and focus distances of the v2 Summilux that Leica should calibrate for ? Obviously in the 60s, Mandler must have known about the M9

Roland.

Since the flange to sensor measurement is exactly the same as the flange to film plane, why would he need to know about the M9? The only difference is the acceptable tolerance to this measurement which is more relaxed for film.

Factory calibration is wide open where there can be no focus shift. If you choose to have a lens recalibrated for a different aperture it should be done only after the body and rangefinder are calibrated to specification.

Gordon

 

[D&A]

The ideal adjustment will be to put the best focus at the aperture you want. It's just a fact that the digital sensor is pickier about focus than film. I've sent several lenses (including the focus-shifting C-Sonnar) to DAG for 6-bit coding and focus check/adjustment. Servicing like this can put the focus at a chosen aperture, but does not otherwise solve the "focus problem." The focus will still shift as you change apertures. For example, the C-Sonnar focus was adjusted to be "on" at about f/2, and this minimized the shift at f/1.5, while at smaller apertures like f/4 depth-of-field enclosed the shift (the best focus is in the part of DOF nearer the camera).

This is exactly what I've been recommending to people who have lenses that have focus shift like the C-Sonnar or leica's 35mm Lux asph (pe FLE)....namely to have their lens adjusted for best focus at f2...so that there is a small amount of front focuswide open and a smaller amount of back focus at f2.8, f4 f5.6...where the ever growing depth of field covers much of the shift.

Dave (D&A)

 

[brusby]

Summilux Ver2 focus test

Summilux Ver2 focus test

Hey Alex, I did some focus tests on several of my lenses on the M9 last night, including a silver version 2 Summilux that is one of my favorite lenses. I wanted to check them all under the same circumstances to see how they focused wide open and then to determine how much focus shift each had.

I'll try to upload some 100% zooms of the ver2 Summilux test.

The test was done with the camera on a tripod about 4 feet from the focus point (which is the close end of the toothpick). Each successive shot is closed down 1 F stop.

I actually have 2 versions of this lens, a chrome one and a black one. The black one just came back from Don at DAG so I'm pretty sure it's working correctly. I tested both lenses and the results were almost identical so I feel pretty confident these results should be representative of properly functioning Ver2 Summiluxes.

BTW Alex, I really like the photos I saw on your website. Very nice!!! Please keep up the great work!

F 1.4

F 2.0

F 2.8

F 4

 

[V-12]

Right.

Then what is the "standard" optimal aperture and focus distances of the v2 Summilux that Leica should calibrate for ? Obviously in the 60s, Mandler must have known about the M9

The 'standard' is what Leica calibrate the lens to before it leaves the factory , otherwise you would be in a camera shop checking each lens to see where it focuses, or prefering those lenses assembled by Hans instead of those put together by Wolfgang.

Since you don't seem to know the digital M's require greater focusing accuracy from the lens than film cameras. This means that while some older lenses will be OK and within the new tighter tolerance, some others will be outside, but still within the old tolerance. These are the ones that need calibrating to the new standard. Focus distance and aperture have nothing to do with it from a customers point of view because the aim is to make the lens work on all bodies, not just one. It is hardly sensible for a company known for high precsion products to have factory calibrated lenses circulating the secondhand markets all set to different focus points and 'best at' apertures

 

[jaapv]

I fear you have missed the meaning of the word "tolerance" in technical jargon. It means the limits of the spread of acceptable adjustments. Which is more narrow on a sensor than it was on film. So some - but not all- older lenses need to be adjusted within the new tolerance range to be accurate on a sensor. Obviously that will not affect their usability on film, as a matter of fact it may be improved.

Right.

Then what is the "standard" optimal aperture and focus distances of the v2 Summilux that Leica should calibrate for ? Obviously in the 60s, Mandler must have known about the M9

Also, assume Leica adjusts Alex's lens to focus a single inch closer on his M9 at minimum focus. You don't think he will notice the difference with his film bodies ?


Alex still hasn't told us how much front or back focus we are talking about. If it is an inch or two, it is indeed in the range of the 'lux shift, and whoever calibrates the camera will need to know from Alex which aperture and focus distance he wants to be optimal.

If it is more than that, his camera is kaput and needs to be fixed and adjusted to his 'lux (since it works on his film cameras). Nothing to do with "tighter tolerances", "standard" calibration of a lens from the 60s, bla, bla, bla.

Roland.

 

[jcm0]

Hey Alex, I did some focus tests on several of my lenses on the M9 last night, including a silver version 2 Summilux that is one of my favorite lenses. I wanted to check them all under the same circumstances to see how they focused wide open and then to determine how much focus shift each had.

I'll try to upload some 100% zooms of the ver2 Summilux test.

The test was done with the camera on a tripod about 4 feet from the focus point (which is the close end of the toothpick). Each successive shot is closed down 1 F stop.

I actually have 2 versions of this lens, a chrome one and a black one. The black one just came back from Don at DAG so I'm pretty sure it's working correctly. I tested both lenses and the results were almost identical so I feel pretty confident these results should be representative of properly functioning Ver2 Summiluxes.

BTW Alex, I really like the photos I saw on your website. Very nice!!! Please keep up the great work!

F 1.4
F 2.0
F 2.8
F 4

Hi Brusby,

Your test @4 feet do show a focus shift - about 2 inches at f4 which is a lot . By chance, thanks to the DOF, the point of focus is preserved at any aperture. The DOF seems to extend from about -1 inch (regarding the point of focus) to the end of the DOF region, which is very usable once this behavior is known.

Did you do made some tests at other distances @f1.4 ? It could be interesting to compare with my findings on another summilux lens.

Regards,

JC

 

[brusby]

Yes, JC, I noticed the focus shift too. And you're absolutely correct that the point of main focus stays pretty sharp as this lens is stopped down, although it's so close to being out of focus it doesn't give much room for error.

It's almost exactly the same on another V2 Summilux I tested.

With other lenses, like a Noctilux, the shift was so severe that the original focus point went completely out of focus as I stopped down, and refocusing would have been required to keep the subject sharp.

I haven't done the tests at any other distance.

Almost all the lenses I checked had a very narrow band of focus in front, and much more in focus behind the object I was focusing on. This might be good for landscape shots, but I've been doing more informal portraits, and for that I'd much prefer if the situation were reversed, i.e., with the thing I'm focusing on at the back of the focal field. That way, if I'm focusing on someone's eyes, they would be the last thing in focus which would throw the ears and the background more out of focus, kinda like the way traditional large format portraits looked.

I think a practical solution for someone like me might be to have the lenses optimized or collimated at one stop more closed down than the maximum aperture. Maybe then the subject would be at the rear of the focus field when shooting wide open and it would give a little more lattitude when stopping down.

Or maybe my rangefinder in the camera is off a bit and needs adjusting.

Hi Brusby,

Your test @4 feet do show a focus shift - about 2 inches at f4 which is a lot . By chance, thanks to the DOF, the point of focus is preserved at any aperture. The DOF seems to extend from about -1 inch (regarding the point of focus) to the end of the DOF region, which is very usable once this behavior is known.

Did you do made some tests at other distances @f1.4 ? It could be interesting to compare with my findings on another summilux lens.

Regards,

JC