off center focusing

greenspun.com : LUSENET : Leica Photography : One Thread

I've become more confortable with the rangefinder viewfinder over the past few years,(coming from an extensive SLR backround) and find myself mixing up my compositions more these days. This often leads to off center main subjects. I've noticed on some of my recent images with my 50mm lens shot at wider apertures up close, that I'm slightly missing the focus on some of these off center images. I've been aiming the center of the camera at the main subject to bring the double image together(usually a face)and than quickly re-composing and tripping the shutter button. It sems like my son's shirt has been more in focus than his face on a number of images done this way. I am wondering if it is just a rangfinder limitation that changing the angle by re-composing is throwing off the focus distance slightly--enough to be noticeable at wider apertures (f2.0 & 2.8) and fairly close distances. I can't really think of any way around this. Anyone else experience this or found a way to foucs more accuratly off center?

-- Andrew Schank (aschank@flash.net), March 25, 2001

Answers

My background seems to be similar to yours and I go the same way you do with your RF camera nor knowing of any better way either. But the results I have gotten so far are satisfactory enough. At least not worst than the ones I get with my R4 or FM2n. I asume some technical issue on the hardware side could be involved but I'm not the great Leica M technician as to go any further this way. By the way: I own two M3's and focusing is a bit harder to get sure about with the newer one (a single stroke) but I have discovered that it is the same as accurate than the DS's but needs a little more attention while doing. Regards.

-- Iván Barrientos (ivanbarrientos@simltda.tie.cl), March 25, 2001.

I have had similar experiences as you, Andrew. More so with distances less than 2m and aperture wider than f2, but as subject distance increases and aperture less than f4 the problem is gone. This is a physical limitation of centre focussing, including AF with only center focusing spot. The design of the lens is such that idealy all subjects on the same plane parellel to the film plane will be in focus, as such as you swing the film plane one way or the other for reason of composition the subjects are no longer in the same plane that is parellel to the film plane, and therefore will be out of focus. I wish I could do better with words here as I could with a diagram, but I really don't know how to put a diagram on the site, so forgive me. Knowing this limitation, I tend to focus, recompose and then touch the focusing just a little so it is now just very slightly less (ie nearer) and then hope for the best. Of course, if the lens is stopped down, or the subject is further than 2.5m away, or if I am just street (candid) shooting then I don't bother. You keep some, you loss some. If you follow the discussion on the Leica camera site on Photo.net, there was a gentleman by the name of Ian Cruikshank (spelling?) awhile ago who mentioned this phenomenon as a drawback to RF cameras (and many others which I will not repeat here) and therefore illustrate the superiority of groundglass focusing on the SLR. Horses for courses. Hope this helps.

-- Steven Fong (steven@ima.org.sg), March 26, 2001.

Is this with the 50 Summicron, Andrew? I've not noticed the problem myself. When I photograph paintings in museums, I use either the 50 or 35 cron or 35 lux. I often focus on the inside edge of the frame, especially if it's an impressionist painting with few clean edges in it. Then I center the finder on the painting. I could be 3 feet to 15 feet away, but usually 4 to 6 feet. The pictures aren't necessarly tack-sharp, but I figured that was because of the fast film and low shutter speeds, often down to 1/30, sometimes lower. And of course, impressionist paintings aren't always so "sharp" themselves. I think I'll make a greater effort to focus in the center and see how it changes things.

I'd have guessed that curvature of field might be the culprit, but Steven Fong's explanation makes just as much sense. The boundary of the plane would be a little farther away than the center area. I wonder if a slight compensatory focus shift would help? The slightest possible shift to the left with the focus tab.

I'm dubious about things being any better with a reflex. The only place you can do *critical* focusing is in the center, where the microprism, fine grain collar, or what have you is. Focusing on the plain groundglass area at the edges, I find pretty frustrating (less so with a long fast lens like the 90 Summicron).

-- Bob Fleischman (RFXMAIL@prodigy.net), March 26, 2001.


As far as I know, if you focus on a point, then every point in the plane running through it and that is parallel to the film is also in focus. So, if you swing around as you recompose, you will be out of focus. This is definitely an area where a multi-point AF system is a bit better (and so also a groundglass), but not enough to notice, except when you are shooting wide open and close up (portraits are especially troublesome as the eyes get less than perfectly sharp).

I've never really noticed the problem, except at f2 and more to the point f1.4, and especially at f1.8 with a Nikkor 85/1.8 when I've prefocused using the center, and not the groundglass.

The only solution is to get a better sense of where your subject will be in the final composition, and nudge the focus a bit... Its not perfect but it will improve your percentage of in focus shots in these rare situations.

-- Mani Sitaraman (bindumani@pacific.net.sg), March 26, 2001.


I have never had this problem in spite of using my lenses wide open and close all the time. I was beginning to feel left out :-) so I tried a few tests to see what was up. It turns out that I do not focus and then rotate the camera; I focus and then "slide" the camera sideways in the direction I need to go. I do not have your problems so I guess it works.

Cheers,

-- John Collier (jbcollier@home.com), March 26, 2001.



If you are rotating the camera as you recompose, then you are not measuring the same distance (it is longer by the cosine of the angle of rotation. There are two solutions to this problem (which exists in all cameras with center-frame focusing). One; use a small enough f-stop that the difference in distance is subsumed in the depth of filed. Two; move laterally as you recompose, rather than rotating the camera about its vertical axis.

-- Dan Brown (brpatent@swbell.net), March 26, 2001.

This out-of-focus issue pronounces everywhere when at first focussing a subject in a center of VF (whether RF or SLR cameras), and then composing it out of a center, especially with wide-open super fast mid-tele lens at mid-distances. So named a curvature of field does not solve this problem. We change an angle of view, and therefore we change a distance to a subject. A camera doesn’t know where our subject is gone when composing and shoots what it sees in a center where it has been focused to before. It’s easy to get a sharper picture when at first composing a subject being out of a VF center, and then focussing it and shooting. I use a swing or tilt on LF- camera or pure ground glass on 35mm/MF SLR. But RF-camera does not give us such a possibility. I got use to use a zone of DOF which shows RF-spot in M3’s VF/RF. However when shooting with M4-P I “nudge the focus a bit” as Mr. Mani Sitaraman recommends.

-- Victor Randin (www.ved@enran.com.ua), March 26, 2001.

I'll have to see if I can get a feel for John's "sliding technique" instead of the rotating style that is less than efficient undere these circumstances. I also noticed that without exeption, I seem to be slightly back focused using the technique I was using, and never front focused. My subjects ears are usually in focus, but the eyes are slightly soft and wouldn't make a great enlargement. I also realized only recently have I been using the ground glass on my Nikon N90S to focus off center directly (with very good results), as I think I used to use the split image on my Nikon FM and rotate the camera causing the same problem up close at wide apertures I have with the M3. I probably notice it more on the M3 because the lens is so sharp wide open, and the fact that I use it more for low light no flash shots than I did my SLR's. Thanks for the input.

-- Andrew Schank (aschank@flash.net), March 26, 2001.

The amount of "rotation" necessary to recompose a head-shoulders shot off center with a 50mm at 1.5-2m just isn't enough to overshoot the DOF of the lens, even at f2. I've used the central focusing spot on an SLR to focus off-center and recompose macro shots at 1:1 with a 200mm lens. By contrast, I've gotten many soft images with longer lenses at wider apertures with the Leica even without recomposing, this is half the reason I sold my 75/1.4. Any chance the problem is in the initial focus acquisition, not something that happens as you recompose? Are the shots all sharp where you don't recompose?

-- Jay (infinitydt@aol.com), March 26, 2001.

Jay, the depth of field is only inches in those situations, and I'm afraid the geomety of rotating the camera actually is making the focus shift. I use an M3 that is in perfect calibration and has a terrific effective base magnification. I have had no trouble when the image is in the center for pictures taken even at f2.0-you can count eyelashes. The fact that when I shift off center always gives me a slight back focus confirms in my mind that the plane of focus was aquired correctly and moved a few inches behind the subject when rotating the camera. This is likely why the new AF cameras have those irritating blinking focus points in the finder so you can chose a spot off center and focus the subject at the actual place you want them in the composition. I'm going to play around later today to see if I understand what John means by sliding the camera instead of rotating it. I'm all for learning a better technique if one is available.

-- Andrew Schank (aschank@flash.net), March 26, 2001.


As Dan said, the problem is cosine error; basic trigonometry... An easy fix I have used is to lean forward about 6 or 8 inches after recomposing -- this usually cancells out the bulk of the cosine error to an acceptable level. (Sliding, as the others have metioned will work as well; both accomplish the same thing.)

Cheers,

Jack

-- Jack Flesher (jbflesher@msn.com), March 26, 2001.


I belive what you do is very common in any leica M user, focus and recompose, to be honest I have never had that problem, and specialy with a summicron (what ever it´s age), probably and for sure with a summilux wide open; have you make any test with a tripod, try photographing a flat wall wide open, or a flat news paper, so you can see what are the posibilities of your lenses, well you have probably done this already, please let us know of your results, and of the lenses you´re using.The best of luck.

-- R. Watson (al1231234@hotmail.com), March 26, 2001.

Sorry!

I meant to convey I leaned forward, or into the subject when focusing, and back to my normal stance for re-composing and shooting!!! (It has been a hectic morning!)

Cheers,

Jack

-- Jack Flesher (jbflesher@msn.com), March 26, 2001.


John, by sliding, I assume you mean moving the camera in a plane parallel to the intended plane of sharp focus. If that's what you mean, then how far does the camera typically have to be slid? I am imagining an off-center portrait situation that might require maybe a two-foot slide. Is that about right? Can this be done without introducing about as much error, in this presumably hand-held movement, as the cosine error itself, if the lens is swung to recompose?

-- Bob Fleischman (RFXMAIL@prodigy.net), March 26, 2001.

We are talking about wide open and up close so I think two feet is a bit excessive. If I moved sideways two feet I would be taking a picture of something else! :-) It depends on the lens and where you want the subject in the frame. In my case it is 50/2, 35/1.4 or a 21/2.8 and usually involves people so the eyes are what I am focusing on. With the 50/2 I focus on one eye and then probably shift sideways not more than six inches or so. The 21/2.8 requires more movement but has substancially greater DOF so no problems there. Try it out and let me know what you think.

Cheers,

-- John Collier (jbcollier@home.com), March 27, 2001.



How good it is to be in the same club with so many generous people we have on board here. I asume I should have thought of the "cosine" problem myself since I'm an engineer. I'm ashamed that I didn't. But I didn't. Yes, it is basic trigonometry. Yes, there are ranges where it is more noticeable in relation with the DOF you use. Yes, the "sliding" technique should work (isn't it what the large format cameras allow you to do in order to get rid of the problem ?) Yes, I began trying "sliding" myself and will report as soon as I have results to share. Thanks everybody. You all are such nice persons.

Best regards

Iván

-- Ivan Barrientos (ingenieria@simltda.tie.cl), May 12, 2001.


This is caused by the geometry of the situation

A person's face is about 8 to 9 cm in front of his shirt

Suppose the distance of the person's face to lens is 120 cm

Then the distance of shirt to lens is 128 cm

Now suppose you shift your camera by 20 degree to recompose

The lens is still focused at 120 cm

However, now the face and the shirt are at an angle to the optical axis of the lens, and the perpendicular distance of face to lens becomes 112 cm( = 120 x cos 20 ), and the perpendicular distance of shirt to lens becomes 120 cm( = 128 * cos 20 ). Since the lens was focused at 120 cm, that is why the shirt is sharply in focus, and the face is not.

There for when using wide open lens with shallow DOF, if you want to recompose, focus on the shirt, and not the face to get the face sharp after the camera is shifted

-- martin tai (martin.tai@capcanada.com), May 13, 2001.


Mama mia, my propousals with the 35 were peanuts, you´re talking 75/1.4 or 50/1, at 1 mt., interesting the trigonometri solution, I´ll stay on to read solutions.

-- R Watson (al1231234@hotmail.com), May 14, 2001.

RULE OF THUMB FOR MAXIMUM ROTATION ANGLE FOR SUMMILUX LENSES

Following this discussion, we know that when recomp after focusing, the object distance is reduced.

The question now is, what is the maximum angle one is allow to shift a Summilux at F1.4 such that the originally focused object is still within the near limit of depth of field of a Summilux ? After some calculation, I find, for exmaple, for Summilux 50mm/f1.4 lens at 1 meter, you are not allow to move more than 11 degree at 5 meter you cannot move lens away from the object by more than 24 degree. For Summilux 75mm /f1.4, the permitted turn angle while recomposing is even smaller: 1 meter, turn no more than 7.5 degree ! at 5 meter, turn no mroe than 16 degree. Note that the angle of view of a Summilux 50/f1.4 lens is 45 degree, divided by 4 I got 11 degree, which is the limit of recompositon angle for this lens at 1 meter.

Note also the angle of view of Summilux 75/1.4 lens is 31 degree, divide 31 by 4, we get 7.5 degree, which happens to be the permittable recomposition move angle for such a lens at one meter !

Here is my rule of thumb for Summilux: 1) The maximum allowable recomposition move angle of a Summilux lens at one meter is the angle of view of this lens divided by 4 2) The maxium allowalbe recomposition move angle of a Summilux lens at 5 meter is half of the angle of view of such a lens.

Here, all Summilux is used at wide open F1.4

Therefore, a simple rule of thumb is formulated, which provide a guide in off center focusing- recomposition

When a Summilux is used at f1.4, as long as the move angle is within the allowable limited, the original object, will still be considered sharp, because it is just within the limit of depth of field

-- martin tai (martin.tai@capcanada.com), May 14, 2001.


My formulated rule of thumb works for both Summilux M and Summilux R It was calculated based on a circle of confusion of 1/30 mm.

It works also for any other brand of lenses, for example the rule for Carl Zeiss Planar 50/f1.4 is exactly the same as Summilux 50/1.4

I don't think I have read anywhere( in books or magazines ) any similar explicit rule of thumb which links up the geometry of recomposition focusing with DOF theory and the angle of view of lens.

If you have seen it anywhere, please let me know

-- martin tai (martin.tai@capcanada.com), May 14, 2001.


Off Center Focusing Rule of Thumb for Summilux Simplified

The rule of thumb for off center focusing for Summilux used at wide open made used of the angle of view of a lens, which you can get from many Leica books or brochures. However, how are you going to measure 7.5 degree ( For Summilux 75mm at 1 M ) or 11 degree etc

Is there any simpler method, without looking of "angle of view" and make use of a angle measure

Yes.

The viewfinder frame on M6 and the viewfinder of any SLR camera carries the angle of view information.

Off Center Focusing Maximum Rotation for Summilux

For any Summilux used at F1.4, when the object distance is one meter, you may rotate your camera up to a quarter of a screen wide, measured horinontally, for object at 5 meter, you may shift up to half a screen.

Angle of view is measured from diagonal to diagonal. By using the screen wide horizontally I achieved to goals 1) Greatly simplified the rule of thumb 2) Build in a guarantee. For example for object at 5 meter, you may shift your camera as much as you like as long as the object is still in the viewfinder-- even if the object is now at the edge of the horizontal frame line ,or viewfinder frame.

                        _________________________
                        |                        |
                        |                        |
                        |                        |
                        | 5m   1m  [ ]           |
                        |                        |
                        |                        |
                        |                        |
                        --------------------------

[ ] indicates the rangefinder superimpose window, or center colar of SLR

1M and 5m indicates the maximum shift permitted at 1m and 5 m.



-- martin tai (martin.tai@capcanada.com), May 15, 2001.


Off Center Focusing rule of thumb for Summicron

In the above rangefinder frame line or SLR viewfinder if you use any Summicro or any lens at F2, replace 5m with 3 M replace 1 M with 0.6 M you get the rule of thumb for off center focusing.

For Summicron at f2, for object at 0.6 meter move camera no more than quarter of screen horizontally; for object at 3M, rotate caemra no more than half a screen

Copyright: Martin Tai 2001

-- martin tai (martin.tai@capcanada.com), May 15, 2001.


Off Center Focusing with Leica Minilux

Focusing on a subject then lightly press down on shutter release to lock the exposure and autofocused distance, move the focal point away from the subject and recompose the picture is a technique often used in compact cameras.

Mininlux has a 40mm/f2.4 lens.

For subject about 1 meter distance, camera can be rotated about a quarter of a screen

For subject of 3 meter, the camera can be rotated half a screen (ie, move the subject from center to edge ) in order to fall within the depth of field at widest aperture of f2.4 p> For outdoor pix, the automatic program of Minilux usually select a smaller aperture, with resulting more depth of field.

For indoor pix the camera selects f2.4, then it is necessary to take into account the degree of camera rotation

For example, in indoor shot, if the subject is only 1 meter away you focused it then shift the subject to the edge of the viewfinder frame, that is half a screem move, the subject will definitely be out of focus, as it falls outside the DOF

-- martin tai (martin.tai@capcanada.com), May 15, 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



-- martin tai (martin.tai@capcanada.com), May 15, 2001.

A Simple Formula for 35mm Off Center Focusing

I have derived a very simple formula for cover how much an object can be shift before it it outside the limit of depth of field


                   ---------------------------------
                  D  \               D' |
                       \                |
                         \              |
                           \            |
                             \          |
                               \        |
                                 \      |
                                   \    |
                                     \  |
                                       \|
                                        |\
                                        |  \
                                        |    \
                                     d' |      \d
                       -----------------------------------
                        FILM            |<- L -->   PLANE
In the diagram, M6 (or other rangefinder, including AF ) focused initially at D, the move the lens to D' direction to recopose

In the film plane, the image of D is at D, and d' is the center of the film plane.

L is the lateral shift of the object on the film plane.

How much lateral shift is permitted such that object D is still considered sharp within the depth of field ?

SQRT(FSTOP * D ) L = ---------------------- 4 Where D is the object distance in mm

The advantage of expressing allowable lateral shift in the film plane is the independence of quantiy L from focal length, in other words, how much one can shift depends only on object distance and fstop

  
                                 '
                     DISTANCE mm   1.4     2    2.8
                      1000         9.4mm  11.2  13.2
                      2000         13.2   15.8  18.7
                      3000         16.2   19.3   23

For example for 50, 80 Summilux at f1.4, when the object is at 1 meter, the lateral image shift about 9 mm on the film plane, or about a quarter of a frameline, then the image will still be within DEPTH OF FIELD

Example 2 Summilux 50,80 used at f2 or Summicron 50,90 at F2, for object at about 1 m, you can shift almost 1/3 screen

For any lens used at f2.8, you can shift right to the edge of a frame.

Because this formula uses the lateral image shift on film plane instead of using the "angle of view" of lens which involves diagonal , therefore the resulting formula is much more elegant.

-- martin tai (martin.tai@capcanada.com), May 23, 2001.


Opps The fomula for allowable lateral image shfit is

                SQRT(FSTOP *  D )
          L = -------------------------
                      4

          Where L is the lateral shift of image, whos circle of 
cofusion  is  <= 1/30 mm
          D is the straightline distance of the object(in mm ) when 
originally focused with  rangefinder






-- martin tai (martin.tai@capcanada.com), May 23, 2001.

Martin; for years I though the only formula to make me feel dizzy was Tequila and beer, let me put my feet on ground to tell you how interesting your last posts was. I just need to read it slowly.

-- R Watson (al1231234@hotmail.com), May 23, 2001.

It happens a lot with the 35 summilux non asph,

having the feel of how sphereical aberration develops through focusing range,this may be a fast way to react before passing moments, my experience with f/2 and faster lenses is soo pour yet, but my summaron works beautifuly in this aspect, you know it only goes to 2.8; this may not be the best lens for arquitecture, but with people is good.

-- R Watson (al1231234@hotmail.com), May 24, 2001.


Roberto You can use your M6 to find out how much the camera can be rotated at a chosen distance and aperture.

For example, stand in front of a wall papered wall, chose a mark, set the lens at 1 m, and move the camera back or forth, such that the double images becomes one-- ie, to measure 1 m with rangefinder.

Then move rotate the lens slightly such that the index now aligns with the longer end of the f2 dof marks ( there are two marks, one on each side of the index, one closer, the ohter further )

Now check again with your rangefinder, now the one meter object will not be off on your rangefinder.

Stand on the same sport, but move the camera left or right, to find the mark on wall paper now looks in focus in the rangefinder

Then note down how much the new infocus object deviates from the old object-- in the rangefidner, you will find out it is approximately a quarter of frame shift.

Do the same with 1 meter/f2; 2 meter at f1.4, f2 etc. and make notes

The simplest method, is to use your customary distance and aperture in head-and-shoulder shots, then stand such distance from wall paper, set the lens at exactly such a distance, make sure the rangefinder do superimpose, then find a object on wall paper which is at the edge of the viewfinder frame (of that particular lens ) then measure with rangefinder again, check to see whether it is out side the DOF mark.

From these experiments, you get a table for off center focusing tailored to your own application

-- martin tai (martin.tai@capcanada.com), May 24, 2001.


Steven has some numbers, and this correspond well with the table


                                 '
                     DISTANCE mm   1.4     2    2.8
                      1000         9.4mm  11.2  13.2
                      2000         13.2   15.8  18.7
                      3000         16.2   19.3   23



At 2m or more, f2.8 will cover object moved from center of focus right up to edge ( 18.7mm is more than a half of 36mm frame width )

At 3M , one can almost open up to f1.4

The hard part is up close, at 1 m, f4 can cover center to edge rotation of camera. to circustances

-- martin tai (martin.tai@capcanada.com), May 24, 2001.


The Simplest Formula for Off Center Focusing

The formula

                SQRT(FSTOP *  D )
          L = -------------------------
                      4

Can be further simplified by subsitution of L = 18mm (half of a frame width at film plane ) :

                         5.2  
                       D = --------------- meter
                              Fstop 


From which we get a simplified table

Example, with any lens at F2, if the object is not closer than 2.6 m, you can focus object at center then recompose it up to the edge.

-- martin tai (martin.tai@capcanada.com), May 24, 2001.


At this stage, the Rule of Thumb for Off Center Focusing becomes:

"f1 and 5.2 meter "

From here, proportionality comes in:

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 


-- martin tai (martin.tai@capcanada.com), May 24, 2001.

The government work, Martin! Two kopecks from me, please. You know there are universally adopted understandings in the LF-photography for moves of lens axle or film plane, such as *tilt* (up and down), *swing* (to the right, left), and *shift* (parallel move). Maybe it would be better to use these instead of *rotating* camera or *shift* meaning the same?

Is it necessary to calculate for an R-lens or for any other SLR-lens? We can easy focus a subject at any point of an SLR-screen at its center (split image) or at its edge (groundglass), except for a Leicaflex standard, of course, that has not a groundglass in the screen design. The difference in distances is very visible when using an SLR in your examples. There is no the off center focussing problem when shooting with an SLR, and of course with an LF-camera, which allows us to focus simultaneously even any two subjects at any points of groundglass when using the Schoimpflug’s rule. Unfortunately we need calculating or guessmatic-focussing to solve off center focussing problem, when shooting with RF-cameras :^{

-- Victor Randin (ved@enran.com.ua), May 25, 2001.


Biktop,

Yes, "swing" is a better word then "rotate" "shift".

Camera (LF, MF reflex ,most SLR )using groud glass has the advantage of whole screen focusing instead of center focusing. Some SLR camera has no groud glass, only a splitprism, for example, Zeiss Ikon Contaflex has no ground glass the view is always clear, except the central splitprism; with this kind of SLR, off center focusing is much like RF.

-- martin tai (martin.tai@capcanada.com), May 25, 2001.


A Novel Use of F1 @ 5.2 meter Rule of Thumb

     The formula
                         5.2  
                       D = --------------- meter
                              Fstop 

Or F1 @ 5.2 METER rule can be used to allow for adjustment of focusing in off center focus

That is how it works.

Suppose you focus an object which is near the edge of a viewfinder frame, read it off the lens, and you find out the distance is 2.5 M

5.2 divided by 2.5 is about 2

Rotate the lens barrel so that the lens moves a distance about the the dof mark corresponding to F2

2.6 m -----------------------| * |---------------------

Off center focus result = 2.5 meter

2.6 m -----------------------| * |---------------------

f2 f2 slightly turn the lens to align the measure distance on the short end of DOF scale Now you are EXACTLY IN FOCUS. And, further, you can use any aperture you like, not necessarily F2, because, the lens is exactly in focus, no just within depth of field, therefore you can use even F1. Example 2: Offcenter focus result = 1.3 m

5.2/1.3 =4

Therefore, move the "1.3m" point to align with the right f4 DOF mark, and lo, you have exact focus ! The DOF scale on all your lens become a compensation dial for off center focusing ! Nothing can be simpler !



-- martin tai (martin.tai@capcanada.com), May 25, 2001.

This thread has been cleaned up slightly by straightening out some heading and list tags. If I have altered a post incorrectly, please let me know! -- Your Friendly Neighborhood Moderator

-- Tony Rowlett (rowlett@mail.com), May 25, 2001.

Tony thanks for the clean up. This thread evolves through many intermediate step, the final formula f1 @ 5.2 meter and how to use the DOF scale on 35mm lens as a off center focus compensation dial is the most useful

However, today, when I make a search at photo.net, and found there was a similar thread, and to my surprise, David Jacobson 's Lens FAQ http://www.photo.net/photo/optics/lensFAQ Q16 already had an answer, although in different forms and slightly different numerical value. I am glad that I did not know David's formula before hand, otherwise I would not dive deep into this theme and derive an answer on my own, and in the process get greater insight into this theme. It is worth the effort.

Particular the method of how to use the DOF scale on 35mm lens as a off center focus compensator dial is unique to this thread.

-- martin tai (martin.tai@capcanada.com), May 25, 2001.


How to Use DOF scale to make correction for 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 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

When you pull your lens back to center, that is the exact location you lens must focus, 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"

1m realign to 5.6 because 1x5.6 =5.6 2m realign to 2.8 because 2x 2.8=5.6 0.5 realign to 11 because 0.5 x 11 =5.6 1.4m realign to f4 because 1.4 x 4 = 5.6 4m: realign to f1.4 because 4x 1.4 =5.6 .......................



-- martin tai (martin.tai@capcanada.com), March 07, 2002.

Hmmmmmmmmmm.

Martin, it has to matter how far off center the focusing was done. Does this method assume a 1/2 width rotation, out to the edge?

-- Bob Fleischman (RFXMAIL@prodigy.net), March 07, 2002.


Bob, To the far edge

-- martin tai (martin.tai@capcanada.com), March 07, 2002.

Martin, I just had a look at this on my 35mm - at 1m re-foccusing to 5.6 changes the focussed distance to just under 0.9m (ie 4 inches closer) - this sounds like a lot.

And anyway, if as some have said we should focus on the shirt not the face shouldn't the focal distance be longer (ie turn the lens to align the 1m mark with the left hand 5.6) so we focus on 1.1m?

Help!

-- Giles Poilu (giles@monpoilu.icom43.net), March 07, 2002.


Giles

When you use the "realign 1m x f5.6" method, you just focus on where you want it sharp, ie the face What happens is that when you off center focus at the face, then turn the camera to center C, the lens now is focused at the shirt (further ) hence it is necessary to adjust the focusing of the lens to focus closer.


()------------------------- C . | . | | . | | . | | . | | . | | . face shirt



-- martin tai (martin.tai@capcanada.com), March 07, 2002.

Thanks Martin - I was muddled, your diagram is very helpful. My most common scenario is a portrait off centre at say 1m - but with the face between edge and centre (ie rule of thirds).

If your formula is for the extreme edge, then in the above case should I align at F4?

-- Giles Poilu (giles@monpoilu.icom43.net), March 07, 2002.


Giles Yes, adjust proportionally.

For off center half frame instead to edge, realign 1 m to F2.8, for rule of the third position, realign 1m to f4, yes

-- martin tai (martin.tai@capcanada.com), March 07, 2002.


Giles wrote:

"Martin, I just had a look at this on my 35mm - at 1m re-foccusing to 5.6 changes the focussed distance to just under 0.9m (ie 4 inches closer) - this sounds like a lot. "

I am glad you did the experiment on you 35mm lens

Here are the data: For a f=35mm lens, if the off center distance is 1 meter the centered distance is exactly 896.5 mm a difference of 103.5 mm or 4 inches !

Now, check this out with your M6 + 35mm lens, focus at an offcenter object (at the edge), walk closer such that the RF indicate 1 meter then refocus at the center , to check that now the 1 m mark aligns with F5.6



                              Lens
                              ()-------------------
                                .     center distance = 896.5 cm
                                    .
                                        . off center distance= 1m
                                            .
                                               .
.

-- martin tai (martin.tai@capcanada.com), March 07, 2002.

Revised Off Center Focus Guide Number

My previous formula for off center focus compensation was not close enough. The guide number 5.2, 5.6 were too loose.

After some more calculation, I arrived a much much better guide number:

Off center distance x Fstop = 4
I check this guide number for Within the focusing range of these lenses, this guide number works very well.

Not applicable for shorter distance ( ie with extension ring) ( For example, not good for 135mm lens offcenter focus distance =1 m, which is outside the 1.5 M near limit )

There are two ways to use this D* Fstop =4 rule of thumb

Method 1. Select Fstop according to D*Fstop =4 rule and let the depth of field takes care of the rest

For example, with any lens from 21mm to 135mm, if you off center focus = 1 M, the select Fstop >=4, then after recomposition, the object still remain withing depth of field; if off center distance=2 then any Fstop >=2 will be good

Method 2. Readjustment of focus

Still use the same D* Fstop =4 rule

Method 1, relies on depth of field, you do not adjust the distance, but select Fstop according to D*Fstop=4

Method 2, you can use ANY Fstop of your choice, however you must use D*Fstop =4 rule to find out the Fstop, to which you align the distance: purpose: you get object on sharp focus, not just within depth of field.

-- martin tai (martin.tai@capcanada.com), March 09, 2002.


Martin,

in method 2 is possibly an mistake. I think You wanted to say:

· If D= 1 Meter, align 1 M on focus scale with F4 on DOF scale · If D= 1.4 Meter, align 1.4 M on focus scale with F 2.8 on DOF scale · If D= 2 Meter, align 2 M on focus scale with F2 on DOF scale · If D= 4 Meter, align 4 M on focus scale with F1 on DOF scale

I assume that this revised rule is for half frame recomposing. Am I correct that for quarter frame recomposing the rule is as follows:

· If D= 1 Meter, align 1 M on focus scale with F2 on DOF scale · If D= 1.4 Meter, align 1.4 M on focus scale with F 1.4 on DOF scale · If D= 2 Meter, align 2 M on focus scale with F1 on DOF scale

I tried to check Your revised rule with a SLR (R6.2 with 2/50). It seemed to me when looking at the focusing screen that Your old rule (5.2/5.6) was more precise. I needed even more realignment than 5.6 at 1M. Is it because of curved field at short distances?

-- Joachim Hildebrand (j.hildebrand@hildebrand.de), March 26, 2002.


you people have waaaay too much time on your hands....

-- grant (lotusphotography@yahoo.com), April 25, 2002.

To make the rules simpler and more precise, I have made up a table, which I hope will make things much simpler, see my article

"Leica M & R Recompose Focus Guide " at

http://groups.yahoo.com/group/Leicafile/files/articles/recompos e.html

-- martin tai (martin.tai@capcanada.com), April 27, 2002.


Joachim, to be more precise, no one number covers all focal length, actually each focal length needs a different rule of thumb number, there are a dozen lenses in M and R, which makes "rule of thumb" complicated.

Hence I made a Leica M and R Recompose Focus Guide, from which the Fstop number can be read off directly from the table, given the center focused distance and focal length of lens

Hope the table method will be much simpler to use and more precise See the link in my preceding post

-- martin tai (martin.tai@capcanada.com), April 27, 2002.


Moderation questions? read the FAQ