Does a lens bring a plane of points into focus or the points at a given distance into focus?

Sorry if this has been mentioned elsewhere. Jim

-- Jim Shields (jim.shields@tasis.ch), June 03, 2001

Answers

Is that lawn out front of your house a single object, "lawn", or is it individual blades of grass?

-- Michael Darnton (mdarnton@hotmail.com), June 03, 2001.

Jim:

The answer to your question is "Yes", depending on the lens, so I'll try a simple explanation... A lens with field "curvature" will bring a series of points at a given distance, usually defined by a spherical or elliptical radius from a given point somewhere behind the lens, all into focus on the film plane -- a spherical section of focus if you will. A "flat-field" lens brings all points on a plane that is perpendicular to the lens and whose center is at a given distance from the lens in focus on the film plane -- what we think of as a true "plane of focus".

-- Jack Flesher (jbflesher@msn.com), June 03, 2001.

Thanks, good answer. Now what do we know about Leica lenses, flat field or curved. The elliptical curvature surprises me by the way.

-- Jim Shields (jim.shields@tasis.ch), June 03, 2001.

The answer is again "yes". While I do not consider myself a Leica lens guru, I understand that some Leica lenses exhibit curvature, and some are very well corrected for field flatness, however when compared to the "norm" Leica lenses tend to be very well corrected. Speaking more from my general experience with lenses, I have found when shooting three-dimensional subjects, curvature rarely generates any detectable problem in the final image. However for copy work, or when shooting any large, flat surface, curvature can become more noticable, and hence detract from the intended image. Again, generally speaking, most macro lenses are well corrected for curvature as well as barrel and pincushion distortion, while faster lenses are designed to perform best when focussed from about 20X their focal length to infinity and generally exhibit more of these characteristics. (And we have not even begun discussing chromatic abberation!) So, as is usual with anything regarding optics, there are trade-offs one must weigh and then base their decisions on their own particular intended use.

-- Jack Flesher (jbflesher@msn.com), June 03, 2001.

Usually the wider aperture lenses have more problems with flat field, and also shorter focal lengths. I know the 3.5 50mm Elmar is a good flat field lens, as they were also used as enlarging lenses.

-- Andrew Schank (aschank@flash.net), June 03, 2001.

Field Curvature and Petzval Sum

The image of a plane from any lens, even a thin lens has a natural tendency to curved inward, because oblique object is further from the center of lens then axial object, hence naturally focus closer, resulting in curved film plane

Hungarian mathemantican Joseph Petzval (1807-1891) found a mathematical formula to represent this field curvature, the Petzval sum.

                     
                                  (n'-n) 
 Petzval  curvature = sum of  --------------
                                      nn'r

Where n' and n' are the refractive index on both sides of a lens surface, r is the curvature of lens surface. Various methods are used by lens designer to reduce or elimiate this field curvature, such as combine high reflective crown glass with low reflective flint glass.

Petzval sum is independent upon thickness of lens nor airspace, however there are many other abberations beside field curvature, changing sum parameters may reduce Petzval sum and increase others therefore, most lens still has some residue field curvature, except flat field lenses.

Some camera delibrately to curved the film plane, for example the famous Complan lens designed by Arthur Seibert, ex-Leitz lens designer, for the Minox camera used curved film plane to compenstate for curvature of field, and with the field curvature of field taken cared of, all the lens parameter can be used solely to reduce other abberations, resulting in a ultra sharp lens

-- martin tai (martin.tai@capcanada.com), June 04, 2001.

While I still have you, these wide voigtlanders are said to be not fisheye. Is the fisheyeness of lens related to this flat field business?

-- Jim Shields (jim.shields@tasis.ch), June 04, 2001.

Jim, fisheye is ultra wide angle lens with barrel distortion, it is not flat field lens. Macro lenses are flat field lenses, because these lenses is often used for photography very flat objects, such as documents, drawings etc and must have edge to edge sharp focus, therefore the image of the documents must all be focused on the film plane, with negligible field cuvature

-- martin tai (martin.tai@capcanada.com), June 04, 2001.

A Summicron 50mm lens has a typical Petzval curvature of about 860 mm

-- martin tai (martin.tai@capcanada.com), June 04, 2001.

Martin:

To add a bit more confusion, as I understand it the "fisheye" has no barrel distortion -- it is actually rendering the 3d wide-angle image correctly onto the smaller 2d film. (And was originally used for atmospheric study, as it has essentially no light falloff.) The "rectilinear" wide angles actually have pincushion distortion designed into them to correct for this optical phenomenon.

-- Jack Flesher (jbflesher@msn.com), June 04, 2001.

While on this subject (which has nothing to do with the original post): If you project your non-fisheye slides into a screen deeply curved across its width, which you can make by curving a piece of poster board in front of the projector, you will see a semi-fisheye effect in that horizontal lines will have a strong curvature, bulging toward the top and bottom edges. Now, if you were to project into the interior of a portion of a sphere (a dome), you would get the full fisheye effect. Of course, you wouldn't be able to hold focus. Now enter Omnimax, which does project a rectilinear image into a dome. As I understand it, they correct the projected image by projecting through a fisheye lens. Why a fisheye? I think because light rays traveling backwards (from the film to the screen) would deconstruct a fisheye image, if there were one on the film. But since the image on film is rectilinear, the result (I think) is a projected anti-fisheye effect. When this lands on the dome screen, the lens thus apparently cancels the fisheye effect of the dome (?). When sitting in the center of the theater, though, I see outward bending of verticals, like trees, at the left and right sides of the screen. So I think the correction isn't perfect.

Oh, I almost forgot. I understand that the fisheye Omnimax lens was made by E. Leitz, Canada.

So what? Nothing. It just seemed like the best chance I would get to tell about this.

-- Bob Fleischman (RFXMAIL@prodigy.net), June 04, 2001.

Thanks for your responses. I feel better about my confusion in these matters. These are not matters that normally concern me when taking pictures but this nonfisheye business about the new ultra wides interested (still interests) me. Jim

-- Jim Shields (jim.shields@tasis.ch), June 04, 2001.

Jack, fisheye lens has tremendous amount of barrel distorotion.

A 150 degree type I fisheye lens has 65 % barrel distortion.

A 150 degree type II fisheye lens has even more distortion, 74 % barrel distortion

In contrast, a standard lens has less then 1% distortion

-- martin tai (martin.tai@capcanada.com), June 04, 2001.

Jim,

Some lenses - especially macros, as Martin has mentioned - are designed to have a flat field and to bring a plane of points into focus. The Leica macro lenses work equally well at close range or infinity.

On the other hand, long telephotos with relatively few elements, such as the 400mm f/6.8 Telyt-R, have significant curvature of field. For the uses to which these lenses are typically put, however, such as a sharply focused subject against an out-of-focus background, this does not matter. Also, some lenses that are very well corrected for infinity can exhibit curvature of field at close range, e.g. the 180mm f/3.4 APO-Telyt-R.

Regards, Ray

-- Ray Moth (ray_moth@yahoo.com), June 06, 2001.

35mm camera's film surface is not abosolutely flat, because there is a standard 50 micron deep "film channel", thus 35mm film has a natural tendency to curve inwards. The curl of film has a curvature of approximately 3 m. This slight curvature compensates some what field, curvature of lenses. Even macro lens has no need to absolutely eliminate field curvature.

Chrome mounted in slide also has a natural curvature.

-- martin tai (martin.tai@capcanada.com), June 07, 2001.

>> 35mm camera's film surface is not abosolutely flat, because there is a standard 50 micron deep "film channel", thus 35mm film has a natural tendency to curve inwards. The curl of film has a curvature of approximately 3 m. >>

Martin, you're obviously a very technically minded person, but I find this statement pretty hard to swallow. The pressure plate pushing the film against the film rails is machined flat and the film curves in such a way that its centre line pushes out against this plate, so it seems unlikely to me - given also the small size of the negative - that it can have any residual curvature when held in position during exposure. As I recall, Contax made a vacuum pressure plate to hold the film absolutely flat - does this mean they had to recompute their lenses for use with this camera?

I'm not saying you're wrong, just that it seems pretty counter- intuitive to my limited intuition.

-- rob (rob@robertappleby.com), June 07, 2001.

Rob, film plane architecture is quite standard in most cameras, it consts of three parts

• A pair of inner film rails, each about 1.5mm wide, the rails are 24 mm apart, this film rail pair sits UNDER the film sprokect
• A pari of outer guard rails, 36 mm apart, they prevent the film from wobbling sideways. The height of the guard rails is 50 micron more than the height of the film rail
• A pressure plate

The presure plate is in contact with the outer guard rails but never touches the inner film rails, because they are 50 micron lower.

The pressure plate, the outer guard rails and the inner film rails form a 35mm x 50 micron tunnel, in which the film slides through. The film has a 50 micron leeway of moving in the optical axis direction inside the film channel

                         pressure plate

                 ======================================
                 ||  ------------------------------ || 
                 ||  ||         film             || || 
 
   The pressure plate does not presses on the film, but on the outer
guard rails.
   Film channel architecture provides much smoother film advance, 
otherwise, if the pressure plate presses the film on to the inner
film rails, then the film would be very stiff to move.




                     

The purpose of Contax RTS III vacumm sucker is to reduce this 50 micron film channel leeway, thus reduce focusing error. However I do not own RTS III, so I don't know the details.

-- martin tai (martin.tai@capcanada.com), June 07, 2001.

AFAIK, there is one camera which does not adopt 50 micron film channel architecture, that is Minox GT-E 35mm caemra

Minox GTE has four focal plane rails of equal height its pressure plate presses directly on the film to four focal plane rails.

            
                     pressure plate
                ============================
                 ---------------------------
                 ||    film               ||

              GT-E:  four equal height focal plane rails

                    

Minox GT-E's direct pressure plate film plane architecture elimiates this built in 50 micron focusing error in most cameras, thus enhance the overall sharpness the the lens-camera combination.

-- martin tai (martin.tai@capcanada.com), June 07, 2001.

My Leica R5 has such a film channel architecture.

I don't know Leica M or LTM has similar film channel.

Come to think of the difficulty in loading my IIIf from the bottom, I guess, it would be hard to have such kind of fours rails it would make film loading even more difficult, and in case there is need to remove film in the middle (without rewind ), the "guard rails" would make such removal imposssible. I remembe when loading film, I used to push in then pull out the film several times, if there was gaurd rails, I would not be able to pull the film out and retry.

So my guess is, LTM may not have film channel, instead it may have GT-E style film plane architecture. If so, it may explain why my Elmar 50/3.5 was so sharp

-- martin tai (martin.tai@capcanada.com), June 07, 2001.

In the previous posts, 50 micron film leeway has already taking in account of film thickness. Without the film, the the raw height difference of "guard rails" vs "film rails" is 200 micro; average color print film or chrome is about 150 micron thick, leaving net 50 microm for the film to move.

For thin base film such as Kodak Technical Pan film, which has a thickness of 113 micron, leaving a film channel leeway of as much as 87 micron !

-- martin tai (martin.tai@capcanada.com), June 07, 2001.

Another factor which affects the film plane flatness is more common with auto wind camera or camera equiped with motor drive.

This phenomenon is known as film undulation. When motor drive pull film quickly then stop suddenly, the center of the film may bulge toward the lens( some time even two bulges ).

Contax RTS III has an elaborate vacuum system for keeping film flat. However, IMO, it is rather complicated.

Only the other hand, the Minox spy camera uses a simple and very effective method to keep the film plane flat: It has a flat pressure plate, which presses directly on the film against a oblong opening (8 mm x 11 mm), and since the film has such a small dimension, there is practically no film curling. During film advance cycle, the film pressure plate automatically retracts, thus the film can move freely with little friction, when the film moved to next frame, the pressue plate then press again on the film, keeping it flat. Since there is no such kind of 50 micron to 90 micron deviation of film from "film plane" there is no focusing error introduced by film plane deviation, that explain why Minox camera can produce sharp image out of such tiny negative. It is indeed a high precision camera.

-- martin tai (martin.tai@capcanada.com), June 07, 2001.

But I thought the Minox used a curved gate to correct field curvature. (?)

-- Bob Fleischman (RFXMAIL@prodigy.net), June 08, 2001.

Bob, in Minox A and B, the COMPLAN lens indeed has a curved film plane, and these camera has a slightly curved pressure plate. In later models, from Minox C to LX and the current model TLX (titanium) CLX (chrome ), the lens(name MINOX ) was recomputed to yield a flat film plane.

My own lens test with Chasseur d'image lens test chart and observation with Olympus microscope, I found that the 15mm/f3.5 COMPLAN lens of my Minox B is sharper ( 177 line pair/mm ) vs 163 lpmm of the later MINOX lens.

-- martin tai (martin.tai@capcanada.com), June 08, 2001.

In large format photography, a plane object forms a plane image, and the plane does not necessily perpendicular to the optica axis of lens. The rules of lens tilt and back tilt is governed by the Scheimpflug rule and Scheimpflug Hinge rule.

In large format camera lens, Petzval curvature is less a problem then 35mm, because Petzval curvature is proportional to the focal lens.

-- martin tai (martin.tai@capcanada.com), June 09, 2001.

There are various kind of pressure plate: plain plate, perforated pressure plate, and pressure plate made from scratch proof and ultra flat ceramic (Contax RTS III, Contax T2 )

-- martin tai (martin.tai@capcanada.com), June 09, 2001.