Aperture placement

[swmcl]

Hi,

Is it important where an aperture iris sits along the barrel of a lens? I presume it should be placed at a 'nodal' point - which I assume is located precisely between the two innermost elements ...

Is it as easy as that ? Or is this an ideal - in which case there is some compromise if it is not centrally mounted ...

Information on any compromise and / or placement is much appreciated.

Thanks,

[Mark Woods]

The nodal point of the lens may or may not be the center of the center of the lens design. Nodal point relates to the relationship of the taking lens to the world where the pan/tilt don't shift the image in relationship the foreground to the background. The center of the lens relates to the design of the lens. Regarding about where the aperture is placed is a decision of the lens designer. I know at the turn of the century, those iris fade outs done on screen were with an iris placed in front of the lens. A little funky, but it worked.

[Carsten Wolff]

Mark has already pointed out that aperture placement depends on the lens design and does not have to sit in the nodal point plane. Placement is usu. more critical with closely spaced cells and wide-angles as to where the aperture sits.
Many lenses can be mounted in different shutters, e.g. either a Copal, OR Copal-Press without modification or shims; looking at these two shutters for example, you'll know that their aperture diaphragms are in different positions: Front to iris 10.2mm vs. 11.7mm, respectively.

I'm sure you're thinking about a particular lens though; feel free to enlighten us... ;)

Carsten

[Peter K]

At first we have to ask why an (iris-)aperture at all? First to regulate the brightness in the image plane, second to cut outer rays that decrease the perfomance of the lens and third to decrease the "numerical aperture" and so to increase depth of field. All this without decrease the eveness of the image brightness. Or with certain lenses like WA-lenses to increase eveness of the image brightness.

If the aperture plane is shiftet from the "right" plane, the brightness of the image will not decrease evenly but the outer areas will show a shadow of the aperture, a vignetting.

Every lens, also a simple biconvex lens has a front- and a rear-nodal plane. This are not the same planes used as no-parallax-planes the lens (and also the camera) has to be mounted above the pan/tilt axis of the tripod head, so during moving the camera no image point moves relative to one other. This is the plane of either the front- or rear-entrance-pupill.

Peter

[Mark Sawyer]

The nodal point of the lens may or may not be the center of the center of the lens design. Nodal point relates to the relationship of the taking lens to the world where the pan/tilt don't shift the image in relationship the foreground to the background. The center of the lens relates to the design of the lens. Regarding about where the aperture is placed is a decision of the lens designer. I know at the turn of the century, those iris fade outs done on screen were with an iris placed in front of the lens. A little funky, but it worked.

Are you sure about this? I believe the center of the entrance pupil is the point where the background and foreground stayed aligned during rotation. Lenses have two nodal points, a front nodal point and a rear nodal point, (where a ray enters and where it apperars to exit.)

But yup, the correct aperture placement moves around sightly on different lens designs, (a "degree of freedom" for the designer), and its placement affects, and is used to correct, field curvature and coma in that particular design.

When an aperture is used to vignette an image, then it's a field stop rather than an f/stop, and conventional f/stops are in play as well. F/stops can be placed close in front of a lens without causing vignetting, as with the old landscape lenses or the Kodak Portrait Lens or Imagon.

[Mark Woods]

I think the discussion of front and rear nodal points refer to the center of the respective elements. The nodal point of the lens has to do with perspective and movement. We encounter this issue all the time in visual effects when shooting miniatures. The aperture can be placed in any number of positions determined by the lens designer. The vignetting caused by an aperture I would think relates to the image circle, and if the lens vignettes the image circle is too small for the format. I hope this helps. :-)

[Len Middleton]

Steve,

For some lens designs (telephoto, retro-focus wide angles), the nodal point may be outside the body of the lens all together (e.g. in front of telephoto lens, behind at retro-focus lens).

Therefore it may be impossible to locate the iris at that point without some sort of extension that defeats the rationale of the lens design e.g. retro-focus wide angle lens for a SLR so it does not interfer with the mirror.

Hope that provides some examples that it is not necessarily where we think it should be,

Len

[Mark Woods]

With the use of computer design, there are somethings in lens design that don't seem rational to what we experience on a day to day basis, but the lenses work and do the job. For example, I had an 8mm lens on a 35mm motion picture camera last night. The lens was about 24" long. Worked great on the miniature we shot.

[Peter K]

Are you sure about this? I believe the center of the entrance pupil is the point where the background and foreground stayed aligned during rotation. Lenses have two nodal points, a front nodal point and a rear nodal point, (where a ray enters and where it apperars to exit.)
The cardinal (http://en.wikipedia.org/wiki/Nodal_point#Nodal_points) or nodal points of a lens are definite as the point a ray which enters the lens at a small angle strikes as well the subject sided nodal point as the image sided nodal point.

To determinate the nodals points one has to focus the lens at a distant subject. Now the lens - not the camera - will be panned, moved perpendicular to the optical axis of the lens, and moved alonge the optical axis up to the point the image on the ground-glass doesn't moves during panning. (The position of the ground-glass has to be realigned to compensate the moving of the lens along the optical axis.) Now the plane perpendicular to the panning axis is the front nodal point.

For determinating the rear nodal point the same procedure has to repeated with the rear cell of the lens.

To mount camera and lens nodaly, the panning axis has to cut the plane of the entrance pupil.

When an aperture is used to vignette an image, then it's a field stop rather than an f/stop, and conventional f/stops are in play as well. F/stops can be placed close in front of a lens without causing vignetting, as with the old landscape lenses or the Kodak Portrait Lens or Imagon.
The Wollaston landscape lens was the first photographic lens with a front diaphragm, later used in many box-cameras. This has the advantage over a rear diaphragm of a better spherical correction. This is also the reason the single cell of a convertible lens like the Symmar has to be mounted with the diaphragm in front of the cell.

Of course an Imagon mounted with the sieve behind the shutter isn't so easy to use too. ;)

Peter

[Mark Sawyer]

The cardinal (http://en.wikipedia.org/wiki/Nodal_point#Nodal_points) or nodal points of a lens...

Well, the nodal points are among the cardinal points, but there are othe cardinal points, so the terms aren't quite interchangeable.

I'm not sure I followed your description on how to find the nodal points by moving the lens elements around, but trial and error are for approximates at best. The diagram of the simple thick lens at the wikipedia page shows where the nodal points are. The formula for finding them in a single element lens is 1/n-th the way from the surface to the center, where n is the index of refraction. I'm sure the formula for finding them in a complex lens would be way over my head! (My trig isn't what it used to be...)

But the Wiki page did confirm about where to rotate a lens:

"The nodal points are widely misunderstood in photography, where it is commonly asserted that the light rays "intersect" at "the nodal point", that the iris diaphragm of the lens is located there, and that this is the correct pivot point for panoramic photography, so as to avoid parallax error. These claims are all false, and generally arise from confusion about the optics of camera lenses, as well as confusion between the nodal points and the other cardinal points of the system. The correct pivot point for panoramic photography can be shown to be the centre of the system's entrance pupil."

[Mark Sawyer]

I think the discussion of front and rear nodal points refer to the center of the respective elements. The nodal point of the lens has to do with perspective and movement. We encounter this issue all the time in visual effects when shooting miniatures. The aperture can be placed in any number of positions determined by the lens designer. The vignetting caused by an aperture I would think relates to the image circle, and if the lens vignettes the image circle is too small for the format. I hope this helps. :-)

I'd beg to differ... each individual cell may have its own nodal points, but the entire system has a front and rear nodal point as well. And as Len noted, they can be outside the physical body of the lens on some designs, and inside on others. Regardless of the design or where the nodal points are, the correct pivot point should be the front entrace pupil at the optical axis.

[Mark Woods]

Hello Mark,

I believe that's what I said in the first sentence. And the second sentence also agrees with what you posted. :-)

[Mark Sawyer]

Hello Mark,

I believe that's what I said in the first sentence. And the second sentence also agrees with what you posted. :-)

My fault; my brain was fried today! (End of the third quarter grading day...) When you said:


Nodal point relates to the relationship of the taking lens to the world where the pan/tilt don't shift the image in relationship the foreground to the background.

...I somehow jumped to panning and stitching, as some of my students have been doing that lately. That's where things rotate around the front entrance pupil. Duh, I is stoopid! :o

[Peter K]

Well, the nodal points are among the cardinal points, but there are othe cardinal points, so the terms aren't quite interchangeable.
There are also the front and rear focal points, the vertexes of front and rear lens and the focal plane. But if the lens is used in air the principal planes coincidences with the nodal points.

To illustrate the determination of the nodal points after Moessard I will made a drawing, but later this day, it's nearly 3 a. m. here. Sorry :o

Peter

[Mark Woods]

Mark,

I'm grading papers today too. That's why I have so much time for these discussions. :-) That Oslo paper is killer. Too bad I'm Mac based or I'd download the educational software. It would come in handy when I do my lecture on lenses at the American Film Institute. Might be worth getting a cheap Dell to get it. Oh well....

[Mark Woods]

Mark, where do you teach?

[Mark Sawyer]

I'm just a crappy old high school teacher at Marana High, in the backwoods of Arizona!

It's kinda fun though... I have some really dedicated kids who are doing 4x5 and 8x10, even a bit of wet plate work is starting. Others, I have to threaten and cajole to get a crummy cell-phone pic out of...

Glad you liked that Oslo pdf. It's one of my favorites!

[Mark Woods]

Hey Mark, with all the digital stuff I have to teach, the Fellows want to shoot on film. They love it and we teach both captures. They're totally blown out with my 8x10 contact prints and love to shoot 35mm Motion Pic capture. I'm very proud to teach at the Conservatory that is considered #1 for Cinematographers on the planet. (And our faculty is modest too. ;-))

[swmcl]

So! Its not as easy as I thought.

Firstly, I think I'm sorry I mentioned the technical term 'nodal' ... it seems it was a cardinal sin !!!

But seriously, thanks guys. Yes I have a lens in mind, an old magic lantern lens for a start. It may be a Petzval I don't know.

Where do I go from here? Must I get accurate measurements and then learn a computer simulation program ? It is beginning to be very complicated. I was hoping to cut it in two and be done ...

Rgds,

[Peter K]

Where do I go from here? Must I get accurate measurements and then learn a computer simulation program ? It is beginning to be very complicated. I was hoping to cut it in two and be done ...
Steve, forget the computer simulation program. For this you would need not only to know the curvatures and thickness of ever lens but also refractive indexes and Abbe's numbers of the glass used. Also this old glasses are not in modern glassmaker's catalogues and also not in the data-bank of the programs.

So measure and note the distances of the lenses, diaphragms etc. Than make the threads a little bit longer to get the possibility of alignment by turning the barrels against each other.

Peter

[Peter K]

Here is the schematicaly drawing.
37920
To determinate the nodals points one has to focus the lens at a distant subject. Now the lens - not the camera - will be panned, moved perpendicular to the optical axis of the lens, and moved alonge the optical axis up to the point the image on the ground-glass doesn't moves during panning. (The position of the ground-glass has to be realigned to compensate the moving of the lens along the optical axis.) Now the plane perpendicular to the pan axis is the front nodal p point.

For determinating the rear nodal point p' the same procedure has to repeated with the rear cell of the lens.

Cheers

Peter

[swmcl]

Thanks Peter,

Your help is appreciated. I'll need a little while to digest this.

Cheers,

[Emmanuel BIGLER]

Hello all from France !

I agree with most that has been said above about nodal = principal cardinal points and the role and importance of the entrance pupil.
Cheeke in tongue : European solidarity here implies that I agree at 200% with Peter K, natürlich ! H und H' come from Hauptpunkte, Hauptebene ! ;)

Some additional remarks to the original questions
Is it important where an aperture iris sits along the barrel of a lens?

The correct placement of the iris determines
- residual aberrations
- and has a prominent importance for minimizing distorsion

In a symmetrical lens design if you do not place the iris at the centre of symmetry of the lens, you usually get distorsion.
In the Wollaston "landscape lens" meniscus single lens design, the shape of the meniscus (its radii of curvature) and the placement of the iris are determined by several trade-off between minimizing aberrations and distorsion.
http://www.optics.arizona.edu/Nofziger/UNVR195a/Class9/fig22.jpg
The Wollaston lens is a purely academic case of little interest, except, may be, to show the kind of trade-off required in optical engineering, however it can be entirely computed by hand with a table of logarithms and a slide rule ; I suspect that some of us here only swear by the slide rule for their environmentall-friendly, zero carbon-dioxide routine computations

---------------

For example look at the optical diagram of a classical dialyte lens (4/4) used in process cameras and still widely used by LF photographers today.

http://www.prograf.ru/rodenstock/images/apo-ronar_scheme.jpg

I presume it should be placed at a 'nodal' point - which I assume is located precisely between the two innermost elements ...

Well, before speaking in public about notions in optical engineering, it is always good to precisely understand their meaning, otherwise the discussion can go in the wrong direction. I assume that after reading the previous posts, all our readers know what the nodal points N and N' are ( = H and H' principal points for a photographic lens in air), but I'm afraid that there is still a confusion floating around ... in air or glass ;) about the various optical spaces encountered when light travels across a compound photographic lens.
The entrance space is air where the subject is located ; yes it can be water but let us start simply ;)
Each time a ray crosses a diopter, you change from one optical space to another.
The iris is physically located in air but in a certain intermediate optical space different from the entrance optical space and different from the exit optical space.
The entrance pupil is the image of the iris as seen from the front of the lens. The entrance pupil is located in the entrance optical space. This is not the same optical space as where the iris itself is located.
Hence by definition of the entrance pupil, and taking into account that the proper rotation point for parallax-free panoramic stitching is the entrance pupil, and nowhere else, the iris can hardly ever been located at the same place as the entrance pupil, except in theory for an ultra-thin single lens element, or some very special lens design.

About parallax effects in panoramic stiching and the entrance pupil I had posted here an explanatory diagram...
http://www.largeformatphotography.info/forum/attachment.php?attachmentid=90&d=1149151905

The relevant discussion started by (and starred by, of course) Leonard Evens is here
http://www.largeformatphotography.info/forum/showthread.php?t=17775&highlight=entrance+pupil

However in quasi-symmetric lenses like those in use in LF photography, the iris being located very close to the quasi-centre of the lens, it can be shown by simple ray-tracing considerations, that its images, the pupils, are located very close to the nodal points. Only very asymmetric designs like telephotos with long focal lengths have the entrance pupil located far from the nodal points.
Hence the entrance pupil is very often located very close to the entrance nodal or principal point !!!!!

Is it as easy as that ? Or is this an ideal - in which case there is some compromise if it is not centrally mounted ...

No, things are very, very complicated indeed.
For a perfectly symmetric lens design used at the 1:1 ratio, you should place the iris at the centre of symmetry otherwise you get distorsion. But even this rule of symmetry does not guarantee that you get a good design.
For example take a single lens element and try to use it for large angular field images of distant objects : in this case a symmetrical bi-convex lens element with the iris close to it is not the best design, one of the best designs for one single element is the strange Wollaston meniscus, a concave-convex single lens element, the concave surface facing the subject and with an iris placed in front, in air !

Now try to focus a parallel beam of light (like a laser), you do not care for a wide angular field at all, the optimization differs totally, you can place the iris wherever you like in front of the lens if the beam is really parallel to the optical axis, and it can be shown in classical textbooks that the best shape is an asymmetric bi-convex lens element, not at all Wollaston's meniscus.
http://www.thorlabs.com/newgrouppage2.cfm?objectGroup_ID=900

In the 1:1 ratio with a symmetrical lens whatever design, distorsion is : zero, nil, nada, totally cancelled by symmetry. Plus : zero coma and zero lateral colour, which are nice bonuses for the engineer, and explain why quasi-symmetrical designs are so good and so often encountered in LF photography.

However if you need even a slight departure from this very restricted situation : 1:1 magnification ratio plus symmetrical lens design, think about a dialyte lens design used for distant objects, in principle you have to make compromises.
For a lens optimized for long distances, you should depart from a purely symmetrical lens design.

However, it is a matter of fact that many view camera lenses designed in the XIX-st and XX-st century (except telephotos and except the beloved asymmetrical Tessar design ; I am not aware of retrofocus lenses for the large format ?) are only a slight modification of a pure symmetrical design.
The miracle is that even if you do not respect the 1:1 magnification ratio, a quasi-symmetrical lens design can nevertheless offer you outstanding performance.
If you consider a telephoto or a retrofocus design, which are so common in 35mm and medium format photography, the iris has to be located after balancing many, many parameters, including residual aberrations and distorsion.
This is the art of lens engineering and cannot be explained in a few words.

However, since we mostly deal with quasi-symmtrical lens designs, yes in this case the iris is located very close to the quasi-centre of quasi-symmetry of the lens.

And for almost all non-symmetric designs used for LF cameras, the iris is located inside the lens, unlike the Wollaston lens !
Look at the tessar design, the iris is located in the middle of the lens.
The only exception I know for LF lenses is the "odd" Rodenstock Imagon(TM) where the lens, a cemented doublet, is located behind a very strange aperture that in France we call : "the infamous Imagon(TM) mashed potatoe machine grid ;)"

[Mark Sawyer]

So! Its not as easy as I thought.

Firstly, I think I'm sorry I mentioned the technical term 'nodal' ... it seems it was a cardinal sin !!!

But seriously, thanks guys. Yes I have a lens in mind, an old magic lantern lens for a start. It may be a Petzval I don't know.

Where do I go from here? Must I get accurate measurements and then learn a computer simulation program ? It is beginning to be very complicated. I was hoping to cut it in two and be done ...

Rgds,

If you like the fabled "Petzval Signature" you're probably better putting it in the wrong place anyways! Putting it in the wrong place will just alter the field curvature (the Petzval isn't a flat-field lens anyways), and maybe add a little coma, which would just add to the charm for most photographers who like this sort of lens. It wouldn't hurt the resolution at the center, and wouldn't cause significant vignetting ulless you were waaaaaaay off.

Or if you want it "right", ask around about other Petzvals with a similar focal length, and ask the owners where their apertures are.

[Mark Sawyer]

Hey Mark, with all the digital stuff I have to teach, the Fellows want to shoot on film. They love it and we teach both captures. They're totally blown out with my 8x10 contact prints and love to shoot 35mm Motion Pic capture. I'm very proud to teach at the Conservatory that is considered #1 for Cinematographers on the planet. (And our faculty is modest too. ;-))

Sounds like a great place to teach! We do about 60% digital, 40% darkroom. Most of my kids want to do do digital when we're in the darkroom, and darkroom when we do digital, of course! A lot of the stuff is "high-school-ish", but fun if you can learn to appreciate it. With apologies for wandering off-topic, one of our shadows/lighting demos from last week:

http://i55.photobucket.com/albums/g139/Owen21k/DSC00498a.jpg

[Mark Woods]

Nicely done! Is this one of your shots or a student's?

My class runs me ragged. They are curious about everything, as I'm sure yours does you.

[Mark Sawyer]

That was a demonstration print, made "by committee" as it were. Three students and myself, with me mostly making suggestions after a few earlier near misses, and them saying, "be quiet, we know what we're doing!" :D

[Mark Woods]

Oh man! I know what you're saying. I had the same experience on Thursday while we were shooting a beer pour at high speed with strobes. I had some wine :-) at the dinner break with my fellow teacher (does a lab while I do the theory -- sometimes the theory goes on stage).