Off center focussing with Leica Mgreenspun.com : LUSENET : Leica Photography : One Thread |
I was told in the Leica Forum that my question was already treated in this forum but I could not find it :I was wondering if there is a systematic error when focusing using the rangefinder and then recomposing the image. My concerns were about closeups made at full apperture with the 75mm f1.4 or the 90mm at f2.0, or even the 35mm at 1.4 or f2.0 with very off center subjects. Does anybody had found some problems? Are there some solutions? Alain M.
-- Alain Maestrini (alain.maestrini@att.net), March 05, 2002
It was answered here.
-- JM Woo (wooismyid@deletethispart.yahoo.com), March 05, 2002.
No different to any other camera with focus and recompossing. though focusing is more accurate witht the Leica is done correctly, being slightly more difficult with a lens over 50mm
-- Kristian (leicashot@hotmail.com), March 05, 2002.
At last, another puzzling mystery solved. Although having made my head spin reading all those formulas I am not sure of the best solution.Have people tried the theories out? In the common situation of say an enviromental portrait at 1m and the 35mm lens at F2 with subject in the side of the frame is it simply enough to focus on the chest (shirt) rather than face?
-- Giles Poilu (giles@monpoilu.icom43.net), March 06, 2002.
But here's a question: if even a 35mm lens has some field curvature, then focussing with the patch and then recomposing to put the subject at the frame edge may leave the subject at the same distance from the film plane but out of focus, right?
-- Douglas Kinnear (douglas.kinnear@colostate.edu), March 06, 2002.
Douglas, The focal error caused by off center focusing is more than twenty times that of residue film plane curvature.For example, for a Summicron 50/2 lens, off center focusing error is about 6%, while the residue film plane curvature is less then 0.5%
-- martin tai (martin.tai@capcanada.com), March 06, 2002.
Douglas: essentially, you are right. The image will be out of focus to some slight degree. The subject is still at the same distance from the camera, but the flat plane in which the subject lies is no no longer parallel to the film plane, once the camera is moved. The lens is corrected for a flat field. Once you move the camera, the subject still lies on the same radius; but the established point of focus is on the flat-field plane, a bit farther away.That's the best I can do at this time of night. Hope it helps.
-- Bob Fleischman (RFXMAIL@prodigy.net), March 06, 2002.
"I was wondering if there is a systematic error when focusing using the rangefinder and then recomposing the image. My concerns were about closeups made at full apperture with the 75mm f1.4 or the 90mm at f2.0, or even the 35mm at 1.4 or f2.0 with very off center subjects. Does anybody had found some problems? Are there some solutions? Alain M. " Alain, yes there is systemic error caused by geometry I have solve this problem using the DOF marks on the lens as a scale for correcting this off center focus shift
Suppose you focus offcenter with your M6 rangefinder. Now take a look at the lens barrel, the off center distance is suppose to be 1 meter hence the 1 m mark aligned with the center marker of the DOF scale like this: 0.8 1m 1.5m 2m 5m infinity ^ 16 11 5.6 2.8 | 2.8 5.6 11 16 To make correction, rotate the lens (as if focus the lens closer), and align the 1 meter mark with "5.6" instead of center marker: 0.8 1m 1.5m 2m 5m inf ^ 16 11 5.6 2.8 | 2.8 5.6 11 16 That is the exact location you lens must focus, when you pull the lens back to center, and you can read off the distance scale where the corrected distance is ( less then 1 meter) Case 2: distance = 2 meter Offcenter focus result: 0.5 1m 1.5m 2m 5m in ^ 16 11 5.6 2.8 | 2.8 5.6 11 16 Make correction by aligning 2m with "2.8" 0.8 1m 1.5m 2m 5m infinity ^ 16 11 5.6 2.8 | 2.8 5.6 11 16 Case 3: distance = 0.5 m after offcenter focusing: 0.5 1m 1.5m 2m 5m inf ^ 16 11 5.6 2.8 | 2.8 5.6 11 16 Make correction by aligning 0.5 with "11" 0.5 1m 1.5m 5m in ^ 16 11 5.6 2.8 | 2.8 5.6 11 16 In other words, the realignment should be such that the distance x the dof scale = 5.6 Note: the exact mathematic calculation calls for 5.2, however there is no "5.2" marker, so we can make do with '5.6"
-- martin tai (martin.tai@capcanada.com), March 07, 2002.
Martin,Your example is based on Your great rule of thumb "f1 @ 5,2m" which is meant for 1/2 frame recomposing (object at the very edge of frame). I like Your idea of using the DOF scale as a compensation dial, which is based on this rule, very much.
Am I correct that for 1/4 frame recomposing (object in the middle between edge and center of frame) the modified rule is "F1 @ 1,3m"?
For example: Measured distance (at center of frame) to object is 1 meter, recompose (object at 1/4 frame) and turn focus ring to the right until the 1m index on the focus ring is at the right 1,4 (should be exactly 1,3) engraving of the DOF scale.
If so the use of the "dial" at distances over 1,3 m is not easy because there are no engravings of f-stops bigger than 1 (Noctilux) or 1,4 (Summilux). The compensation must be somewhere between the index and the last available f-stop engraving.
Probably a compensation is not really necessary in these cases but I just like to know if there is an (practical or theoretical) solution. Can You give any hint?
Thank You very much!
-- J. Hildebrand (j.hildebrand@hildebrand.de), March 07, 2002.
Martin--Regarding your immediate answer to my posting above: does that mean that the focussing error caused simply by recomposing (swinging the camera around) for off-center composition is much greater than any problem caused by lens field curvature? And so this problem would apply to RF camera as well as AF SLRs which have focussing sensors in dead center of the frame? Just want to make sure I understand.
-- Douglas Kinnear (douglas.kinnear@colostate.edu), March 07, 2002.
J."Your example is based on Your great rule of thumb "f1 @ 5,2m" which is meant for 1/2 frame recomposing (object at the very edge of frame). I like Your idea of using the DOF scale as a compensation dial, which is based on this rule, very much.
Am I correct that for 1/4 frame recomposing (object in the middle between edge and center of frame) the modified rule is "F1 @ 1,3m"?
" Proportinally, ie for 1/4 frame compensation use
1m @ F2.8 marker 1.4m @ F2 marker 2m @ F1.4 marker ...................... 2.8m @ F1 marker
AS long as distance x Fnumber = 2.8
Think of the two marker as a proxy of the viewfinder
| 2.8 5.6
center focus 1/4 focus 1/2 frame focus
-- martin tai (martin.tai@capcanada.com), March 07, 2002.
Douglas wrote: " Regarding your immediate answer to my posting above: does that mean that the focussing error caused simply by recomposing (swinging the camera around) for off-center composition is much greater than any problem caused by lens field curvature?Yes And so this problem would apply to RF camera as well as AF SLRs which have focussing sensors in dead center of the frame? Just want to make sure I understand. "
This rule apply to any 35 camera (RF,SLR, compact) as long as it has a DOF scale and allow manual adjusting the lens
-- martin tai (martin.tai@capcanada.com), March 07, 2002.
J, the answer did not come out correctly formated
" Proportinally, ie for 1/4 frame compensation use 1m @ F2.8 marker 1.4m @ F2 marker 2m @ F1.4 marker 2.8m @ F1 marker (AS long as distance x Fnumber = 2.8 ) Think of the two marker as a proxy of the viewfinder | 2.8 5.6 center focus 1/4 frame focus 1/2 frame focus
-- martin tai (martin.tai@capcanada.com), March 07, 2002.
Martin,This was really quick!
When I understand You correctly You are saying the rule of thumb for 1/4 frame recomposing is: "F1 @ 2,6 m" and not "F1 @ 1,3 m"
Did I misunderstand Your original posts?
You wrote back then:
(May 24, 2001) "How to get EXACT focus in Off Center Focusing The rule of thumb I provided is a simple method of handling off center focusing. However, the subject will not be exactly in focus, only within the limit of depth of field. However, if you want exact focus, you need to do one more very simple step: That is look at the DOF marks on your lens, and rotate the lens slightly to the left for a amount equals to the depth of field mark on the lens
For example, 50mm/1.4 Summilux, at about 1 meter,
1) Focus
2) Shift lens to left or right, by 1/4 of a screen, such that the object is now half way between one edge and the center At this point you may press the shutter
Do the following to get exact focus
3) Shift the Summilux to the right such that 1 meter aligns with the left DOF scale at F1.4
Focus 1m _________________| ^ |________________
1m _________________| ^ |________________ 1.4 1.4
Turn the Summilux barrel TO THE RIGHT to align 1m with the right side mark now the lens is focus at less then 1 m ( 1 m minus dof ) and put the subject back on EXACT FOCUS"
AND:
"At this stage, the Rule of Thumb for Off Center Focusing becomes:
"f1 and 5.2 meter " From here, proportionality comes in:
Stop down from f1 to f2, divide 5.2 in half = 2.6 m Stop down from f1 to f4, divide 5.2 by 4 to get 1.3m With this " F1 @ 5.2m " rule of thumb for move-object-from-center- to- edge, you can handle cases where you need only quarter frame move
Rotate camera by half as much (1/4 frame vs 1/2 frame ), you can either use 4 times bigger aperture, or move in 4 times as close at same aperture.
You can recite:
F1 @ 5.2 m half frame And do the following mental calculation:
So, F2 @ 2.6m half frame
Now I want only quarter frame move, <========
So, F1 @ 1.3m <========
and F2 @ 0.65m
......
There is only one line to remember
F1 @ 5.2m half a frame "
Did I misunderstand You or is/was something wrong (today or back then)?
-- J. Hildebrand (j.hildebrand@hildebrand.de), March 07, 2002.
J. F1 @ 5.2 m half frame (ie, at the edge, 18mm from center) And do the following mental calculation: So, F2 @ 2.6m half frame Now I want only quarter frame move, ( 9 mm from center ?) So, F1 @ 1.3m <======== No, F1 @ half of 5.6 == F1 @ 2.8 ...... There is only one line to remember F1 @ 5.2m half a frame " or f5.6 @ 1 m Fnumber and meter are interchangeable:Thus "1 meter @ F5.6, half frame "
I think remember this line is better.
-- martin tai (martin.tai@capcanada.com), March 07, 2002.