Related Links:
Canon Diffractive Optics Lenses [9/2000]
Is Medium Format Best? - Film Flatness, Diffraction...
[Ed. note: page was at http://members.home.com/penate/zoneplate.html
Zone Plate Photography (Guillermo Penate) [8/2000] [broken link 2/2003]
(zone plates use a series of concentric circles
to produce pinhole style photos..)
[zone plate article local archive)

Contrary to general belief, stopping down beyond f/11 may not improve sharpness of your images on-film. While your depth-of-field may continue to go up, your image sharpness may drop off. The reason is diffraction. Diffraction is a physical law related to the nature of light.

Diffraction effects explain why most 35mm lenses have upper f/stops of f/16 or f/22. By contrast, 6x6cm cameras may have f/32, and 4x5 inch camera users may employ f/45 routinely, even f/64 and f/128 for 8x10 inch camera fans. However, if you factor in the longer lenses used in these larger format cameras, you quickly realize that the limiting hole size is about the same in each format. That should make sense, since light doesn't know or care about film formats!

The only exception most 35mm users encounter are likely to be long telephoto lenses and macro-lenses. The long telephotos are similar to the view camera lenses in focal length, so they really aren't surprising. The various macro-lenses such as the 55mm f/3.5 micronikkor can reach f/32. But these macro-lenses are specially formulated for close-up work, where maximal depth of field is needed. In these cases, you accept the lower diffraction related limits as a trade-off for increased depth of field.

Astronomers and others use a rule of thumb known as Rayleigh's Law to estimate the diffraction limits of their optical systems. Unfortunately, you will see various numbers or factors used for this rule of thumb, depending on how liberal or conservative you are.

Regardless of which factor or rule of thumb you use, you can readily see how diffraction becomes quite limiting for resolution on 35mm lenses for f/stops beyond f/16!

Diffraction's Impact on Corner Resolution - Radial and Tangential

The above table of ideal (maximum obtainable aerial) resolution versus f/stop is interesting chiefly for its extension to tangential and radial angles of 10 degrees and 20 degrees from the center values. At f/2, our ideal lens may be capable of (aerial) center resolution of 713 lines per mm. But tangential resolution at 20 degree angles falls to 590 lpmm, or about 82% of the center value. Similar losses can be seen for other f/stops in the corners versus the center resolution values. Now you have an insight into another reason why edge resolution is less than center values for even ideal lenses.

RO Poirot ropoirot@aol.com writes:

>I am confused about the optimum aperture to use for landscape  photography with 
>a 4x5 view camera. Paul Hansma says that the optimum aperture for minimizing
>diffraction and defocusing effects (see PHOTO Techniques (M/A 1996), is a
>function of the spread in bellows extension between the near and far objects
>be in focus.  Thus for a bellows travel of 0.7 mm, he says f/16 is the  optimum
>f-stop to get maximum depth of field.  For  2.7 mm of travel it goes to  f/32,
>and for 11 mm travel it is f/64.

I haven't seen the article or their derivation. One can use a simple geometric model and assume that the "cone of rays" from the aperture converges to form its vertex at the focal plane, and then diverges. Points in front of or behind the plane of sharp focus will focus their "rays" behind or in front of the focal plane, and the plane will cut the conic section. Since the bellows spread includes nearly equivalent sections on either side of the focal plane, the relationship between the f/number (N), circle of confusion (c) and bellows spread can be expressed as

1)      N = d / (2 * c)
or
2)      c = d / (2 * N)

To include diffraction, one can use the Rayleigh estimate (given in mm),

3)       c = N / 1600,

and then set c in the first equation equal to the value in #3, implying that the aperture is stopped down to decrease c arising from defocus until to equals the limit imposed by diffraction.

4)        N / 1600 = d / (2 * N)
or
5)        N = sqrt(d * 800)             

So the optimum aperture value for near-far bellows draws of 0.7, 2.1 and 11 mm would be f/24, f/41 and f/94.

This "optimum" value is the aperture at which diffraction and defocus balance each other at the limits of the bellows draw distances for the near-far points chosen.

>On the other hand, John Fielder, in his book PHOTOGRAPHING THE  LANDSCAPE, say
>that even though extremely small apertures can ultimately compromise  resolution
>it won't happen at f/64 and f/90!  He also points out that Ansel Adams  create
>the F/64 club, who had a goal that everything should be in focus.

With respect to Adams, f/64 for 8x10 contact printed gives the same depth of field as f/32 in 4x5 or f/8 for 35 mm printed to 8x10.

Using the Rayleigh criteria, f/45 will give approximately 40 lpmm in the negative, enough for a very sharp 16x20 or 20x24 prints from 4x5 (10 and 8 lpmm in the print respectively). Since the magnification from 8x10 is less, less negative resolution is required, so comparable prints may be made from negatives exposed at f/90.

On the ground glass, on axis resolution of a sharply imaged point will decrease when observed with a 10X loupe as one stops down.

>If Fielder is right, why not always use the smallest aperture available,
>assuming one wants maximum depth of field.  Can anyone resolve this apparent
>discrepancy or is one of these gentlemen simply wrong?

There's at least one person who recommends f/45 for 4x5 as the "key" aperture for sharpness and depth of field. There are worse alternatives, such as using f/11 to avoid diffraction, which results in near nil depth of field.

Generally, one uses swings and tilts if possible, focuses, and stops down observing the near and far objects with a 10 X loupe. When they are sharp, it's the "correct" aperture. If they aren't sharp, one may consider making a smaller print.

Notes:

Source: Hicks, Roger.. Medium and Large Format Photography, p.16

Diffraction Limits (by %contrast and formula)
% contrast	f/5.6	f/11	f/22	formula
0%	275	140	70	1500/f
50%	180	90	45	1000/f

Related Postings:

rec.photo.equipment.medium-format
From: "Andy Shaw" andyshaw@erols.com
[1] Re: Diffraction & focal length, f# (Was: Why MF lenses aren't good)
Date: Mon Nov 30 1998

The size of the ideal point spread function is dependant upon:

the diameter of the aperture
the wavelength of the light
the numerical aperture

But this takes no account of aberations, all of which will increase the effective size of the point spread function

defocus (0W20): depends on the square of the aperture (not strictly an aberation but slight defocus will actually counteract the effect of some aberations and hence the reason for slight focus shift upon stopping down on "bad" lenses)

spherical aberation (first order 0W40): depends on the 4th power of the aperture

coma (first order 1W31): depends on the cube of the aperture

etc.

There are many more terms especially once you go off axis ("faster" lens are harder to correct than "slow" lenses).

In addition there are manufacturing problems - when a lens is ground under pressure the material actually flows, and some distortion will occur, so good lenses are often made over size and then cut down. The cutting down to size also introduces errors (centering for example).

The net result of all this is that is it not a simple thing to relate optical performance to the size or speed of the lens and real world performance is heavily dependant upon the lens designer comming up with something that can be manufactured on a factory production line at an acceptable cost.

From: Bob Wheeler bwheeler@echip.com
Newsgroups: rec.photo.equipment.large-format
Subject: Re: Diffraction At Small Apertures
Date: Sun, 15 Nov 1998

Follow Fielder's advice. Minimum lens apertures are specifically chosen so that you will not be effected by diffraction -- they all result in about the same size hole for all formats, and the size of the hole is the only thing light knows about.

I don't think Bob Solomon's comments are pertinent to your question.

Resolution is a competition between diffraction and depth of focus, d, (Hansma's defocus). The resolution due to diffraction decreases as the f-stop increases, but resolution due to d increases as the f-stop increases. Thus one has two opposite curves which combine an produce a maximum at some f-stop, depending on d.

The standard DOF equation draws a line that balances these two competing effects. For 4x5, the balance is about 20 l/mm on film, which means a 16x20 print will resolve all that your eye can resolve. This balance point is not at the maximum of the combination of the two curves. Hansma attempted to give a curve through the maxima (I think he made errors, but that is another story.) The result is that if you follow Hansma's advice you will achieve more than 20 l/mm on the film. Since you cannot see this difference in a print, you should think hard about why you want it, and balance this with the effort that it will take to attain it.

--
Bob Wheeler --- (Reply to: bwheeler@echip.com) ECHIP, Inc.

From: dickburk@ix.netcom.com (Richard Knoppow)
Newsgroups: rec.photo.equipment.large-format
Subject: Re: Diffraction At Small Apertures
Date: Sun, 15 Nov 1998

....

The optimum aperture for a lens is dependent on the lens design. Most lenses are best at about two stops down from maximum opening. But, I think the advise given in the first reference is really about depth of field. There is no optimum for depth, it depends on the scene and whether you can use camera movements to extend the depth without using small stops.

The second reference, if quoted accurately, is nonsense. There is very definite bluring from diffraction at f/64 unless you are strictly contact printing from 8x10 negatives. I can see some diffraction blurring at f/90 (I have one lens which stops down this far) even on an 8x10. Very small stops are sometimes necessary in order to obtain extrmeme depth of field but should be avoided when not necessary.

As far as the f/64 group, it was a reaction to the style of "pictorial" photography which had a vogue in the 1920's though the forties. Much of the "Pictorial" school used extrmeme diffusion, soft-focus lenses, paper negatives, and alternative printing methods to suppress most of the detail in photographs. This was probably an attempt to copy a certain style of illustrative art of a somewhat earlier period. Weston, Adams, and others thought this was worse than silly and started a counter school of work which showed sharp detail.

One advantage of working with a view camera is that you can see the effect of stopping down on the ground glass. The grain of the glass will distort the effect a little but you can still get a very good idea of depth of field and the effect on sharpness by examining the g-g image with a low power magnifier.

---
Richard Knoppow
Los Angeles, Ca.
dickburk@ix.netcom.com

From: sparks@col.hp.com (John Sparks)
Newsgroups: rec.photo.equipment.large-format
Subject: Re: Diffraction At Small Apertures
Date: 16 Nov 1998

RO Poirot (ropoirot@aol.com) wrote:

>On the other hand, John Fielder, in his book PHOTOGRAPHING THE  LANDSCAPE, says
>that even though extremely small apertures can ultimately compromise resolution
>it won't happen at f/64 and f/90!  He also points out that Ansel Adams  created
>the F/64 club, who had a goal that everything should be in focus.

You will definatly see a loss of sharpness at f/64 or smaller with 4x5. Most large format lenses have the highest resolution overall at f/22. A few are slightly better at f/16. Any smaller f/stop will show some loss of resolution. I am generally willing to accept the loss of sharpness I get with f/32 on 4x5, but try real hard not to stop down more than that. I have made a few negatives at f/45 when really necessary, but I find those negatives noticably unsharp (compared to other negatives) in a 16x20 print. Notice that these are effective aperatures so if you are making a macro photograph at 1:1, a marked f/32 is the same as f/64. You will see a loss of sharpness in this case even in an 8x10 print.

The f/64 club were mostly users of 8x10. In contact prints, you can get away with even f/128 and even with enlargments, because the enlargment required is so much less, f/64 is fine (the same diffraction as f/32 with 4x5) and f/90 is usable (but like f/45 on 4x5, I try not to use it if possible though I think the diffractions effects of f/90 on 8x10 is less visible than f/45 on 4x5).

>If Fielder is right, why not always use the smallest aperture available,
>assuming one wants maximum depth of field.  Can anyone resolve this apparent
>discrepancy or is one of these gentlemen simply wrong?

When I first got started with 4x5, I sort of did this but stopping down to f/32 even when a wider aperature was ok. I think this is a perfectly acceptable approach (but wouldn't use anything smaller than f/32 on 4x5 because the diffraction is noticable). When first learning to operate the view camera, this gives a greater number of fully sharp negatives than trying to get away with wider aperatures. The loss of sharpness at f/32 is minimal are rarely noticed. If f/22 is the sharpest aperature with many lenses, there is little reason to open up wider than f/22 unless you are trying to stop motion or want limited DOF. With 8x10, I find that I often use wider aperatures than f/64 (which gives the same DOF and diffraction as f/32 on 4x5). I think I am more confident in my focusing ability now and would probably end up using f/16-f/22 more on 4x5 if I started using it again.

John Sparks

From: brownt@ase.com
Newsgroups: rec.photo.equipment.medium-format
Subject: Re: Diffraction & focal length, f# (Was: Why MF lenses aren't good)
Date: Wed, 25 Nov 1998

> for example, if I tried to
> use f/64 on my 50mm normal lens for 35mm format, I would get horrible
> diffraction effects (that's why 35mm normal lenses only stop down to f/16
> or f/32). Yet, as the famous name f/64 indicates, using f/64 on large
> format normal or even wide angle lenses simply does not give this problem.
> So, it would appear that the objection to my objection does not in fact
> hold.

As a ball park figure, diffraction limited resolution is about 1600/f lp/mm, regardless of format or lens length. Some say more, some say less. I chose 1600 to round out the numbers. So if you set your 50mm lens on your 35mm camera to f/64 you'll be limited to 25 lp/mm: a muddy mess when enlarged 8X to make an 8x10 print, only about 3 lp/mm on the print. Then you set the 300mm lens on your 8x10 view camera to f/64 and you're also limited to 25 lp/mm. After you make an 8x10 contact print you'll get an incredibly sharp print with about 3 times the resolution needed for excellent sharpness (~7 lp/mm).

TB

From: hemi4268@aol.com (Hemi4268)
Newsgroups: rec.photo.equipment.large-format
Subject: Re: Q: Diffraction related to focal length?
Date: 27 Nov 1998

Hi

The best way to think about diffraction is ask yourself what is the highest resolution a lens will give at any given f stop and any given wavelength.

At .500nm wavelength or yellow green light so maximum resolution a lens will give is as follows:

f1= 2000 lines per millimeter
f2=1000 l/mm
f4=500 l/mm
f8=250 l/mm
f16=125 l/mm
f32=64 l/mm

If you use blue light, the resolution is higher and red light it's lower. Focal length does not effect anything.

So a 200mm f4 lens =500 l/mm
or a 50mm f4 lens =500 l/mm

Larry

Subject: Re: 35mm vs. medium format...?
From: jalbert@nyx10.nyx.net (Joseph Albert)
Date: Fri, 04 Dec 1998

...

>You have to be careful how and what you read in these tests.
>
>* 35mm is limited most by grain and resolution of the film, larger
>formats more by DOF/diffraction due to smaller apertures. TMAX100 is

sort of, but not really with respect to diffraction. with a larger format you shoot at a smaller f-stop and get more diffraction effects on-film, but you enlarge less so that if you set up the exact same shot with same DOF for two different formats and enlarge to identical size prints you will get the same loss of sharpness due to diffraction *on the print*. you mainly have to worry about diffraction when you try to enlarge past what has any hope of being down in 35mm, and since 35mm is no help in these situations, it makes no sense to say that diffraction is less of a problem with 35mm.

>how many lp/mm you can resolve and the quality of enlargement are not
>one in the same. Yes, I can still see the lines, but the grain is the
>texture, the larger formats are smoother. Look again.

also, what makes a print look crisp in sharp usually has alot more to do with high acutance than high resolution, and graininess limits acutance, so larger formats win out here. If you don't believe this, you can read all about it in Ansel's book, "The Print".

>* higher resolution and finer grain films do tip the scales in
>favor of smaller formats for the issue of DOF/diffraction limits alone.

not really. medium format optics are, so far, still sharper than film, so medium format benefits from those excellent quality film emulsions too.

use a (medium format or other) camera with tilt and swing, and you can shoot with wider apertures while getting enough DOF and diffraction isn't a problem-- a medium (or large) format camera with tilt blows away 35mm for images of scenes with high DOF requirements.

j. albert

From: lemon@lime.org (lemonade)
Newsgroups: rec.photo.equipment.medium-format
Subject: Re: Diffraction & focal length, f# (Was: Why MF lenses aren't good)
Date: Wed, 02 Dec 1998

sofjan@aol.com (Sofjan) wrote:

>   Plus i think MR Lemonade forgot to take the different Focal length of
> defferent format into account and also forgot what  f# really means.

On the contrary, you seem to have only gotten one post of mine from a fairly extensive discussion of a couple of weeks ago. With the help of other posters, the issue was resolved fairly satisfactorily.

To summarize again:

-diffraction per se is related to wavelength and absolute size of aperture;
-effect of this diffraction on the film is related to distance of diaphragm to film plane, which is roughly equal to the focal length;
-thus the longer focal length of MF lenses for comparable angles of view, which increases the absolute aperture size and thus reduces diffraction, is compensated for by the greater distance of the diaphragm from the film plane, which magnifies the effects of this lesser diffraction, to a degree which almost exactly compensates for the previous reduction;
-thus the f-number, or ratio of focal length to diameter, is a good characterization of diffraction at the film plane;
-the lesser degree of enlargement of larger negatives then reduces this effect, i.e. f/64 is usable for 8x10 prints from 8x10 negatives, although not for 8x10 prints from 35mm negatives;
-all this hinges on the approximate linearity of the relations involved;
-no one seemed to know how the position of the diaphragm, and thus the compensating effect, is affected by retrofocus vs wide-angle, or long focus vs telephoto, designs.

From: "Frank" red735i@earthlink.net
Newsgroups: rec.photo.equipment.large-format
Subject: Re: Diffraction
Date: Thu, 17 Dec 1998

I certainly do not want to sound like an expert in optics, nor do I want to argue with Bob Monaghan or the results of his home page,

but my understanding is that the diffractiion effect occurs more from the intersection of the iris blades than it does form the size of the iris hole....

.... the smaller the Fstop, the more acute the angle is at the intersection.... a sharper edge.....causing more diffraction effect...... Certainly, the more iris blades, the less the effect......because the angle that the blades intersect is less acute.

.You can see this by trying some experiments with an optical slit and an optical edge....you can see the results of diffraction very clearly with a very sharp edge ( razor blade )

So my rule of thumb is more blades, F22-F32 ..... certainly not scientific, but it works for me....

Arguments ? Opinions?

--
Frank Filippone

>I understand that too small an aperture can cause loss of image sharpness
>due to diffraction - light bending around the small edges of the aperture.

Date: Fri, 18 Dec 1998
From: Todd Maurer konabea@ibm.net
Newsgroups: rec.photo.equipment.large-format
Subject: Re: Diffraction

I think the number of blades plays more of a role in the ghost images of poin light sources caused by reflections within the lens. The more blades, the closer to a circle the ghost image tends to be. I do believe the star effect some very strong light sources (like the sun) develop is a defraction issue, but then the number of blades would effect the nunber of points to the star.

The defraction experiements from freshman physics that I remember did not involve any "corners" at all. A laser was pointed through a mask of closely spaced parallel lines. The light then got bent in a pattern perpendicular to the lines. Now what I could believe is that the closer to round the hole is the more the pattern would be a set of concentric circles rather than star pattern. At some small apperture the concentric circle pattern would become noticeable, however this would be long after image degradation becomes a problem.

Todd

From Nikon Mailing List:
Date: Thu, 27 May 1999
From: Andrew Koenig ark@research.att.com
Subject: [NIKON] 35mm going digital?

Tyler Harris says:

> In a previous discussion some suggested that in the future it may be
> possible to have something like a Gigapixel (or more) CCD in a digital
> camera with something like a 20mm (35 equivalent) lens and 'zoom digitally'
> when necessary, thus requiring only one high quality fixed lens.   However no
> one mentioned that if you do this even though you may end up with the angle
> of view of say a 300mm lens (after digitally zooming) you will have the
> depth of field of a 20mm lens.  Now in reality the CCD may be much smaller
> than 24x36mm meaning that you might end up with a 4mm lens being equivalent
> to 20mm lens which would be great for landscapes but terrible for  portraits.
> For shallow depth of field (especially telephoto shallow DOF) you will  still
> need 'big glass' (preferably Nikon).  My guess is that Nikon will have
> ~24x36mm format cameras with an F mount for quite some time (digital or
> otherwise).
>
> Please correct me if I'm missing something.

I remember seeing somewhere (perhaps on Nikon's website) that the hardest technical problem they faced on the 950 was designing a lens good enough to match the resolution of the CCD.

I believe it, but not for the obvious reason: A back-of-envelope calculation reveals that present digital cameras are very close to the resolution limits imposed by the laws of physics (specifically, the Dawes diffraction limit).

Here's the argument. If you look at the resolution figures for just about any lens, you will discover that you lose resolution if you stop it down too far. I think, for example, that the optimum aperture for the Nikon 50mm f/1.4 is either f/8 or f/11, and if you stop it down to f/16, you definitely lose resolution. You've hit the Dawes limit.

I won't bother you with the details, but this f/16 limit applies to any lens, of any focal length, as long as the physical image fits on a 35mm frame. In particular, it means that if you stop down a long telephoto to f/16, you will also have lost resolution compared to what the film can do, which means that you pay a heavy price for depth of field on those long lenses.

OK, now let's look at digital cameras. The typical megapixel camera nowadays seems to have about a 6mm lens, which gets about the same field of view as a 35mm lens on a 35mm camera. To get the same amount of information onto the CCD as a 35mm frame, the Dawes limit says that you need a physical lens aperture of about the same size. On a 35mm camera, a 35mm lens at f/16 is about 2mm in diameter, so the corresponding aperture for a typical CCD camera is about f/3.

In other words:

IF YOU WANT A DIGITAL CAMERA WITH A PRESENT-DAY-SIZED CCD TO HAVE THE SAME RESOLUTION AS A SLIGHTLY DEGRADED 35mm IMAGE, YOU MUST NOT STOP THE LENS DOWN BEYOND APPROXIMATELY f/3. DOING SO WILL DEGRADE THE IMAGE AS MUCH AS STOPPING DOWN TO f/16 ON A 35MM CAMERA, AND STOPPING DOWN FURTHER WILL DEGRADE IT FURTHER. THERE IS NO WAY AROUND THIS.

The only way to improve resolution while being able to stop down the lens further is to build a physically larger CCD, or to find some other way of taking a physical image the same size as a 35mm frame and shrinking it down *after* the image has been formed, perhaps using a fiber-optic bundle or something like that.

Unless I'm missing something fundamental, this reasoning implies that Nikon's already-announced approach of building a digital-camera body that will take F-series lenses is a necessary step towards obtaining 35mm quality in a digital camera. Obtaining such quality with a smaller image will require designing extremely fast lenses, which are well known to be extremely hard to design, and then requiring them to be used at large apertures, to avoid diffraction.

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

Date: Thu, 26 Aug 1999
From: h.nareid@nareid.demon.co.uk (Helge Nareid)
Newsgroups: rec.photo.equipment.medium-format
Subject: Re: Shutter Placement (was Re: why aren't leaf-shutter lenses obsolete yet?)

stafford@a.winona.msus.edu (John J. Stafford) wrote:

>  Steve Grimes says that it doesn't matter where a shutter
>  is placed, but I wonder if he would qualify that statement.
>  Doesn't a shutter which passes close to the film plane introduce
>  some diffraction? Wouldn't it be better (ideal?) to design
>  a lens so that the shutter is placed between certain clusters?

Diffraction can (at least in theory) be a problem with virtually any type of shutter. In practice, however, it is not a problem. Diffraction from the slit in a focal-plane shutter has no real effect in practical life, precisely because the distance from the shutter to the film is so short - diffraction gives an angular spread of the light, and an angular spread over a short distance has negligible effect. Also, appreciable diffraction effects can only be observed with very narrow slits - even at the shortest exposure times, the effect is very small.

For leaf shutters, there can be very appreciable diffraction effects when the shutter opening is at a minimum, but once again this forms only a very small proportion of the total exposure even at the shortest exposure times. I'd be more worried at vignetting effect from leaf shutters placed far from the aperture blades, but experience has shown me that this can't be seen with my Sinar behind the lens shutter, which is not in its theoretically optimum position.

In theory, the best locations for a shutter is either close to the focal plane (assuming that the shutter is designed to give equal exposure to all parts of the image - such as the slit-type focal plane shutter), or at the location of the aperture diaphragm (where most lens shutters are located). In real life, I have never seen any noticable effects from shutter placement.

-- - Helge Nareid
Nordmann i utlendighet, Aberdeen, Scotland

From Leica Mailing List:
Date: Sat, 22 Jul 2000
From: "Erwin Puts" imxputs@knoware.nl
Subject: [Leica] Diffraction limited; bo-ke

Someone suggested that I might have said that the Summicron 50 and other lenses are diffraction limited at f/8 and smaller. And therefore you should use smaller apertures to get the best performance. I did not state anything close to this. The Summicron is not diffraction limited at any aperture and so are many Leica lenses. Only a few lenses in the Leica stable are really diffraction limited, like the R- 4/280. The R-2/180 is for all intents and purposes diffraction limited at f/5.6 and should not be stopped down further if best performance is required. Diffraction limited means that the optical aberrations in a lens are so small that the physical limit of the Airy disc is approached. This means also that the lens automatically performs better at larger apertures as the diffraction effects (loss of contrast, loss of resolution, loss of encircled energy) increase when stopping down. It is always best to use the widest aperture that is feasible in a given situation. (when using modern leica lenses).

... [on bo-ke or bokeh]

Erwin.

From Pentax Mailing List:
Date: Mon, 24 Jul 2000
From: Bob Blakely Bob@Blakely.com
Subject: Re: (Fwd) Diffraction and all that - Long winded, but right

There is some misleading information in Erwin's post (attached below). I don't own a Leica (most excellent cameras) and would therefore be uncomfortable posting to their list. Further, it's not generally good form to post something this long as a first post anyway. Feel free, if you wish, to forward this to them. I would have interest in their comments but, not being a member of their list, I don't know if they have any real interest in such detail.

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

Diffraction is wave phenomenon observed in *all* electromagnet energy where *any* obstruction of *any* size exists. The character of the diffraction observed is dependent on the geometry of the obstruction. In cameras, this obstruction is the diaphragm or iris (or the lens element mounts if the lens is wide open).

Let's try to understand what's happening here. For the sake of simplification, let's consider a point light source such as a star centered on the axis of the lens. (Yes, they are point light sources for ALL practical purposes.) The photons from the star radiate in all directions, but a portion reaches the camera lens and enters. Some of the light is stopped by the lenses diaphragm and is absorbed by the black diaphragm blades (transduction to heat), some is reflected back out the front of the lens. Neither reach the film. A portion of the starlight falls *across* the aperture and is passed on to the film. So now we have a uniform intensity disk of light across the aperture. At this point, we stop thinking of the light as particles traveling in straight lines from the star, and consider it as a wave. The equations of a man named Maxwell take over here. The following will be hard to grasp for most. This is not an intellectual failing. Everyone I know has some difficulty with Maxwell, but I assure you that his equations are part of the cornerstone of modern physics and are Law. I can't provide an engineering course here, so I must ask you to bear with me. Please ask on optical engineer or an antenna engineer for confirmation of what I'm about to say.

After some serious review of Maxwell's equations as applied to this aperture, it is evident that the existence of the star is no longer relevant beyond the aperture. All that matters is the characteristics of the disk of electromagnetic energy across the aperture. After some considerable mathematical mastication, we discover that the disk of light produces a pattern radiating from the disk over the *entire* backside of the aperture. Further, the pattern in our case is radially symmetric if the disk is circular. Moreover, the amount of energy as a function of the angle from the center can be described in terms of a Bessel function. This function is both calculable and tabulated. This particular function has energy lobes, one large central lobe (aligned with the lens axis in our case) and a multiple of sidelobes decreasing in energy as a function of angle from the main lobe. between the lobes is a null.

In a perfect system, the nulls will have zero light energy. These sidelobes, separated by nulls, are the source of Erwin's aforementioned "rings". At this point, I point out that the main lobe has finite width, even though the star was a "point source". In addition, the amplitude of the first sidelobe is about 17 dB down from the main (central) lobe, or approximately a factor of 50 (more than 5 stops) lower in amplitude. Since the first sidelobe is more than 5 stops down from the main lobe, it and the rings Erwin discusses, are *not* the source of the unsharpness due to diffraction. Well then, what is?

Back to the main (central) lobe! As I said, it has finite width (beamwidth), even though the star is a point source of light. The width of this lobe is a function of wavelength, but most important for us, it is inversely proportional to the diameter of the aperture or light disk. Large aperture, small (narrow) lobe, sharp picture, small aperture large (wide) lobe unsharp picture. The size of this main lobe is the governing factor and so therefore the size of the aperture. There exists *no* way around this resolution limiting phenomena, and so the diffraction limit is the ultimate limit for optical (or any other electromagnetic, e.g. dish radio antenna) device.

Now to make the example more general. If the star is off the lens axis, then the light strikes the aperture at an angle. There is a delay between the arrival time between light reaching the one side of the aperture and light reaching the other side of the aperture. This leads to a constant phase shift of the light energy across the aperture and much more hairy mathematics will show that the main lobe and sidelobes will be displaced angularly to strike a different portion of the film in the expected position. Generalizing further, we can view a scene as an infinite number of point light sources radiating toward our lens and superimposing to form a continuous image.

A note on resolution. This main lobe beamwidth can be described as to size in terms of the two points either side of center (maximum energy) where the energy has fallen to 1/2 maximum energy. For radar folks, these are the -3dB points, for photographers, this is one stop down. Imagine photographing an image consisting of two parallel lines. If the lines spaced closely such that the 1 stop down portion of one side of the lobe from one line coincides with the 1 stop down portion of the other side of the lobe from the other line, then the energy adds and the energy at this point (between the lines) is the same as the maximum energy of either line's lobe. At this point we can no longer resolve the difference between the two lines and it has become impossible, even with the most sensitive and accurate of instruments to know if there were two lines or one fatter line. This *is* the theoretical limit of resolution.

A optical instrument is said to be "diffraction limited" if it's resolution is sufficiently close to the diffraction limit that no amount of optical wizardry will significantly improve it's resolution performance. A lens may be described to be diffraction limited at a certain f-stop, for example, the Pentax A* 85mm/1.4 can be said to be diffraction limited at f/4 (or at nearly f/2.8 if one believes two independent studies).

Do you want your lenses to be diffraction limited? Well, that depends...

Let's say that a particular manufacture were to make an 85mm/f2 lens that *is* diffraction limited. Excuse me. The mount on my camera is capable of f/1.2. Why can't I have a little more glass for the extra nudge of speed I need at night? If I think I need more resolution than available on any film (except to governments), I'll stop it down.

Keep in mind that other things besides lens resolution are important to the photographer.

Regards,
Bob...

From Leica Mailing List:
Date: Sun, 23 Jul 2000
From: imx imxputs@knoware.nl
Subject: [Leica] Diffraction and all that

There is some confusion when discussing diffraction and the optimum aperture of a lens. Diffraction is a physical phenomenon, that bends light rays at small openings, when geometrical optics would like to have straight lines. Diffraction limits the maximum resolution. An image point is never a true point, but a patch of light with a certain intensity distribution. For an aberration free system, the distribution pattern is a small central disc of high intensity and a series of annular rings of ever diminishing intensity (or amplitude). As diffraction is related to the diameter of the aperture, it is best to use a wide aperture when you wish to have the highest possible resolution.

A lens that is described as diffraction limited, has a very high level of optical correction such that the normal aberrations are reduced to such a small level that the size of the spot approaches the theoretically calculated (predicted) size and resolution is at its best. The notion of a diffraction limited lens applies always to the maximum aperture of course.

In most photographic lenses, the geometrical aberrations are quite lage (relatively speaking). These optical aberrations generate a "point" that is much larger than the one you get when only diffraction is effective, and optical aberrations become less effective when stopping down. When you stop down, you reduce the effect of the optical aberrations (get a smaller patch of light with a more even illumination distribution), but the effects of diffraction (due to the small opening) increase. if the lens is moderately well corrected, the size of the patch is so large, that the effects of diffraction due to a small aparture will not be detected. Older Leica lenses can be stopped down to 8 or 11 and you will see at every smaller aperture a better image quality, "better' that is related to the maximum aperture. More recent leica lenses have a higher level of abertion correction and now the size of the spot at medium apertures (like 5.6) is so small, that diffraction effects will be detected (loss of resolution, loss of contrast).

Most recent leica lenses, specifically the apochromatically corrected ones, like the R-2/180, and 4/280 are at full aperture already corrected so well that they approach the theoretical limit, when stopping down a bit and so may be described as diffraction limited at 5.6, (that is: geometrical aberrations are reduced the a very small level). With such a lens, stopping down always degrades the image because of diffraction effects. So we have three categories of lenses: moderately well corrected lenses, that only show diffraction effects at the smallest apertures. Well corrected lenses that show diffraction effects at medium to small apertures (5.6 and smaller). Such a lens is not to be described as diffraction limited at 5.6 or smaller. At these apertures the lens is not even close to the maximum (theoretical) resolution, required for the diffraction limit. True diffraction limited lenses, that show diffraction effects when stopping down a bit already. Here the resolution is close to the diffraction limited case.

To give an example. The R-2/180 at 5.6 resolves 540 lines/mm (270 lp/mm) and is diffraction limited as the theoretial resolution is close to this value.

The Summicron 2/35 at 5.6 resolves around 100 lp/mm and so is not even remotely close to the theoretical value and cannot be described as diffraction limited at 5.6. (but the diffraction effects are visible in a lower contrast). So do not confuse diffraction limited at 5.6. with the fact that diffraction effects are visible at 5.6.

Erwin

[Ed. note: new diffractive optics reduces long lenses weight and size 1/3rd..]
Date: 08 Sep 2000
From: Ville Voipio vvoipio@kosh.hut.fi
Newsgroups: rec.photo.equipment.35mm,sci.optics
Subject: Re: Here's the bomb Canon will drop at Photokina

w.j.markerink@a1.nl (Willem-Jan Markerink) writes:

> I have tried to inspire folks on sci.optics to shed a few expert words on
> this new Canon lens, but so far there's only silence (but my news-feed
> might be groggy, must check other server too).

(Dunno about experts, but...)

Without seeing anything more specific about the Canon objective I think it is possible that Canon really has a diffractive element in the system.

Diffractive optical elements are based on very small (in the order of one wavelength) bumps and pits on the element surface. These bumps and pits change the phase of the light coming through the element. As the phase is changed, the direction of the wavefronts coming through the element changes, i.e. the light changes its direction.

Diffractive optics offers many useful features. Diffractive optical elements can be manufactured with the same methods as CD or DVD discs. The mastering process is much more difficult, but the pressing and molding remains the same. This makes it possible to manufacture very low-cost elements with rather complicated functions. DOEs may be manufactured on glass by introducing a plastic (or otherwise softer) coating and then pressing, so the process is much easier than grinding and polishing.

DOEs are not limited to simple spherical optics functions. They may perform several corrections in one step.

Unfortunately, there are some bad news, as well. The first one is that calculating the correct surface profile for an optical element is very tedious. A lot of research is carried out on how to make the calculations more accurate and quicker. This, however is only a technical problem. The real problem is that DOEs have a devastating chromatic aberration.

So, DOEs are only good for monochromatic (usually laser) light. But there is a spot of light in this problem: the chromatic aberration of a DOE is opposite to that of a glass lens. So, by combining these two it is possible to manufacture a single element with very little chromatic aberration.

This idea has been around for several years, but I think the Canon lens is the first consumer application of this idea. There are some manufacturing considerations, and even though the technology is less expensive than other possibilities (using dublets and triplets, etc.), the lens might not be very cheap at first.

---

What's then the difference between a Fresnel lens and a diffractive lens? Both look the same.

Diffractive lens is based on the wave nature of light. The surface features are very small, and diffraction and interference are required. In a traditional Fresnel lens diffraction and interference are very much avoided, and the surface profile features have to be in the order of millimeters.

It is also possible to make amplitude-modulating diffractive optics. There the surface of the element is patterned with non-transparent stripes. This can be done with photographic emulsion or equivalent. The problem is the poor efficiency of these elements, so most DOEs are phase-modulating.

Fresnel lenses are usually rather low-quality. They are not used in imaging optics, as there are a lot of unwanted reflections. Most Fresnel lenses seem to be used in light steering optics, such as in overhead projectors, where a non-Fresnel lens would be very thick, very heavy, and very expensive.

---

It is rather difficult to explain the working principle of DOEs in a few words. I think the best way to get started with is to take som ephysics or optics book and start looking for an explanation about diffractive gratings. Same principle, but usually much more complex surface patterns.

---

> (or have all experts been issued a silence-plight?....:))

I hope not. And I am sure they will correct all the mistakes I've made above. I am only a user of DOEs, not a designer.

- Ville

--
Ville Voipio, M.Sc. (EE)

Date: Fri, 08 Sep 2000
From: Tony Spadaro t__spadaro@my-deja.com
Newsgroups: rec.photo.equipment.35mm,sci.optics
Subject: Re: Here's the bomb Canon will drop at Photokina

MIght as well add this to the mix

http://www.takeuchi.mce.uec.ac.jp/papers/upmm/1999-08e/

and I hope I've typed it right.

This is the company that may or may not be involved in the design/manufacture of the DO lens. It sure looks like it's related.

Date: Sat, 09 Sep 2000
From: Don Stauffer stauffer@gte.net
Newsgroups: rec.photo.technique.nature,rec.photo.digital
Subject: Re: lens, film resolution, MTF limits

The diffraction blur is actually sought after for portraiture because it softens the picture without really blurring it. That is, the diffraction spot has 'wings' that scatter some of the energy all over the picture (to some degree, as in flare).

Many soft focus portrait attachments are actually a screen to artificially bring about diffraction.

So indeed one wonders how far past diffraction limit should we strive for in a lens.

Roger N Clark wrote:

> snip
>
> This "rule of thumb" is simply using an MTF level that is at
> a slightly higher MTF than the Rayleigh criterion
> (approximately at the 10% MTF of the diffraction
> limited system).   
>
snip
>
> These are simply different definitions on where to cut off the
> limit of the MTF system.  They are not deriving the 0 point of the
> MTF curve.  But by stating the limit is X LPM means nothing without
> specifying the MTF level.
>
> It is clear that from my tests that the 4x5 system at f/45 shows
> much more detail than the 35mm at f/11, and that the information showing
> perceived image sharpness is significant at MTF levels below
> 50% MTF.  How much below is a little fuzzy (pun intended) and
> open to interpretation, but in my opinion, close to the Rayleigh
> limit.  I say this because there is more detail on the 4x5 slide
> than is scanned at 3300 dpi, significantly more in my opinion.
> At f/45, the Rayleigh limit is 3700 dpi.
>                                            
>
> Roger Clark

--
Don Stauffer in Minneapolis
stauffer@gte.net
http://home1.gte.net/stauffer/

From: hemi4268@aol.com (Hemi4268)
Newsgroups: rec.photo.equipment.medium-format
Date: 15 Oct 2000
Subject: Re: 700 lpmm film.. Re: The Medium Format Sweet Spot

Hi

All lenses follow the same diffraction rules. That is, with noon summer daylight sun above 75 degrees, the resolution values for a perfect lens is as follows.

f-1 =2000 l/mm
f-2 =1000
f-4 = 500
f-8 = 250
f-16= 125

Since most lenses are not perfect a typical lens over $200 will do the following

f-1 = 50 l/mm
f-2 = 300
f-4 = 500
f-8 = 250
f-16 = 125

Any readings in the 50 to 100 l/mm range is not lens resolution but rather resolution of the film image produced by that lens. Since most photographic systems opperate in the 18% of lens resolution rate (about 500l/mm lens resolution), 80 l/mm image is about right for say Tmax400 in just about any good camera. To go higher in resolution, one only needs to change film to tmax 100.

Larry

From Nikon Mailing List:
Date: Fri, 26 Jan 2001
From: ken rockwell kenrockwell@kenrockwell.com
Subject: [NIKON] Diffraction and sharpness

Hi Joe and all,

I've researched this extensively back when I worked at TRW. I intend to write it up, but for now you'll want to read the article in the March/April 1996 issue of Photo Techniques magazine

http://www.phototechmag.com/

To make a very long story short, for practical photography (8x or 22x loupe on Velvia) diffraction is invisible at up to f/16. It starts having a minor effect at f/22, and at f/32 it is obvious when carefully compared to the same image made at wider apertures. Thankfully this means that 35mm photographers can ignore most of this: I researched this for use with my large format cameras that go to f/128 and give visible blurred images with the naked eye at those apertures.

The diffraction limit of resolution is 100 lpmm at f/16, 70 lpmm at f/22, and 50 lpmm at f/32.

The real question you need to ask is: Well, if I need to stop down to get depth of field, but can't stop down too much because of diffraction, than what is the best aperture to use for a given scene?

Taking way too much math into account to arrive at this elegant solution, you need to redraw the DOF indicators on Nikkor lenses this way:

Where the lens says use f/45, use f/32
Where the lens says use f/32, use f/22 1/2
Where the lens says use f/22, use f/22
Where the lens says use f/16, use f/16 1/2
Where the lens says use f/11, use f/16
Where the lens says use f/8, use f/11 1/2
Where the lens says use f/5.6 use f/11
Where the lens says use f/4, use f/8 1/2
Where the lens says use f/2.8, use f/8

You can see an example of the replacement scale I drew for my 20mm at

http://kenrockwell.com/nikon/20af.htm

When you've done this, use the scale the way you were trained to use a traditional DOF scale, and use the f/stop indicated by the new scale to get the sharpest possible result taking into account both the contridictory effects of defocus (depth of field) and diffraction.

Ken R

Date: Wed, 17 Jan 2001
From: Jerry Coffin jcoffin@taeus.com
Newsgroups: rec.photo.equipment.35mm
Subject: Re: Lens Sharpness Review? (Lens Qualities...)

anders.-.eivor.svensson@swipnet.se says...

[ ... ]

> Is there any confusion of what diffraction really is ?
>
>From what I've seen, there's an awful lot of confusion on what it is.
> My knowledge about this phenomena is limited to fairly
> simple textbooks, but AFAIK, there is no way diffraction can
> be avoided if the *physical aperture* is very small. For 35
> mm lenses, this is usually at f8-f11 or so.

Diffraction is _always_ present, for any finite aperture size. Diffraction takes place as light passes close to the aperture. As the aperture becomes larger, a smaller percentage of the light passes close enough to the aperture to be diffracted. As the aperture decreases, a larger percentage of the light passes close enough to the aperture to be diffracted.

Therefore, at any finite aperture, diffraction places a limit on the resolution that can be achieved. Above a certain (fairly small) size, however, other factors in the design of a lens (such as chromatic aberration) will normally cause more unsharpness than diffraction. The unsharpness due to diffraction becomes much smaller than that due to other factors, so the diffraction has little visible effect on the outcome.

FWIW, in a few specialized cases, diffraction becomes meaningful at considerably larger apertures than most of us are accustomed to seeing it rear its ugly head. For example, some aerial photography is normally done with extremely good lenses and (crucial point) uses band-pass filters so you only ever see light very close to a specific wavelength. This can drastically reduce chromatic aberration, to the point that some of the lenses can be diffraction limited even at relatively large f/stops (e.g. f/2).

> All information I have about optics (I am by no means a
> optical expert, however) points in the same direction - good
> lenses lenses become worse at some point after f8 due to
> diffraction, not due to lens design.

This is true. It is possible, however, for a lens to be WORSE than the diffraction limit at something like f/11. At f/11, the usual formula for the diffraction limit gives a maximum resolution of about 145 lpmm. If the lens design is poor enough that it can only resolve (say) 120 lpmm without diffraction, then it will still only be able to resolve 120 lpmm with it as well.

There's no question that as you reduce the aperture, any lens (no matter how bad) would become diffraction limited. The only real question is exactly when you'll hit the limit -- at f8, 11, 16 or what.

Also keep in mind that as I mentioned above, _every_ picture combines both some light that was focused with some light that was diffracted. As you approach the diffraction limit, you'll see a higher percentage of diffracted light even though it isn't necessarily enough to complete obscure detail. This is at least part of the reason why you tend to see lower resolution as the target contrast is reduced.

Assume you can distinguish a 1% difference between "light" and "dark". If the difference between the light and dark areas is 100%, then you can distinguish detail when diffraction is covering the dark part up with 99% of the level of light from the light part. If you reduce the contrast to only 20% difference, then you can only distinguish dark from light when up to 19% of the light is diffracted into the dark part.

--
Later,
Jerry.

sci.optics
From: "A. E. Siegman" siegman@stanford.edu
[1] Re: Diffraction Simulation Software
Date: Mon Mar 26 2001

Diffraction 2.00, Mac shareware, copyright by John Lindal 1992-93, acknowledgements to Absoft Corp 1988, is a demo-type "draw an aperture in one window, see the diffraction pattern in the second window" type demo program that still runs (for me anyway) under OS 8.6. No idea if the author is still around.

DAFRAC 3.0.2, similar but rougher demo program by Greg Forbes and Pete Hoch, Univ of Rochester, 1986-88, also runs.

Date: Sat, 09 Jun 2001
From: "Roger N. Clark" rnclark@qwest.net
Newsgroups: rec.photo.equipment.35mm,rec.photo.digital
Subject: Re: A diffraction test posted

John Stafford wrote:

> With respect to your initiative, Tony, and not to detract from your
> assertion, I think your test is invalid. A lens properly designed
> specifically for a small CCD format should perform _ much better_ than
> you can show using a lens designed specifically for the much larger
> 35mm format.

Wrong. You can't do better than diffraction limited. (and things like phase contrast microcopy doesn't apply in this case, and it too obeys the same laws of physics.)

This is a good and simple test and very valid.

More information on diffraction effects at:

http://www.users.qwest.net/~rnclark/scandetail.htm

Roger Clark

Date: Mon, 9 Jul 2001
From: "Q.G. de Bakker" qnu@worldonline.nl
Newsgroups: rec.photo.equipment.35mm
Subject: Re: diffraction limited Re: Lens testing at f8 or f11?

Michael Gudzinowicz wrote:

> >> instance, at f/8 the Rayleigh limit is approximately 1600/8 =
> >> 200 lpmm. A very good lens might resolve close to 200 lpmm if the
> >> aerial image is examined, and a "dog", 100 lpmm. Both lenses are
> >> optimal at f/8, and both show worse performance f/11 or
> >> smaller apertures. In other words, they are both "diffraction  limited"
> >> at f/8.
> >
> >No, no. "Optimal" performance is in no way the same as "diffraction limited"
> >performance.
> >The very good lens may be diffraction limited at f/8, the "dog"  certainly
> >isn't. It is limited by aberrations, even at f/8.
>
> And the very good lens is limited by aberrations at f/8, or it would
> perform better at wider apertures. You have been misled by the
> marketing departments and magazines.

The very good lens in your example above is diffraction limited at f/8, the "dog" clearly isn't. If the very good lens would show worse resolution at larger apertures it only means that aberrations at those larger apertures are not sufficiently reduced, so this lens quite simply isn't diffraction limited at these apertures. That however does not mean in any way that it is not at f/8. Not at all. Your "or it would perform better at wider apertures" is false.

And what marketing departments and magazines? All that is said is that if a lens is "diffraction limited" its performance, as far as resolution is concerned, is limited by diffraction only. That really is what the term means.

Marketing departments and magazines may try and convince us that a particular lens is diffraction limited (i.e. has all aberrations reduced so far that they contribute less to limiting resolution than diffraction does) and if we believe them when they tell a lie, yes, then we would be misled.

But that's not what we are discussing here, is it?

Again: "diffraction limited" means resolution is only limited by diffraction, and not by aberrations. No more no less.

> The -> only <- criteria required to make the claim of diffraction
> limited performance at a certain aperture is that further stopping
> down harms or limits performance. That is the definition used in the
> optical design literature.

Not so. If stopping down does indeed result in worse performance, it not necessarily follows that the good performance at the larger aperture was limited by diffraction only.

The converse however is true: a lens that really is diffraction limited will show less resolution when stopped down. I don't think optical design literature would use the definition you give.

> It doesn't matter if the lens is excellent
> or a "dog" at that aperture and never has mattered.

Of course it does matter. If at a given aperture aberrations are reduced enough so they don't limit resolution, the lens is diffraction limited, at that aperture. If aberrations do still limit resolution, the lens clearly isn't diffraction limited. Simple.

> The reason one lens is better than another with the same diffraction
> limited aperture is often due to aberrations not affected by reducing
> the aperture and quality criteria such as uniformity over the entire
> lens surface, lack of decentration, mechanical tolerances of the
> barrel and short/long term alignment, group positioning in zoom
> lenses, etc.

Your use of "diffraction limited aperture" is incorrect. You clearly use it to mean the aperture at which the lens has its best performance. But best performance by no means equals to being diffraction limited. Yes, there are many things that can degrade image quality, but unless diffraction is the only thing that does so, you musn't use the term "diffraction limited".

> Every book on optical design will state that as soon as a "real" lens
> is introduced into the system, it is limited by aberrations, and is
> no longer the "perfect" or "ideal" lens which was assumed by Rayleigh
> as a condition for his analysis. "Diffraction limited" does not and
> never has implied that the lens suddenly becomes "perfect" or
> "ideal" at an aperture at which further stopping down harms
> performance. Only a salesperson would like you to beleive that.

Your again mixing two things up: what it means for a lens to be diffraction limited (and yes, that would indeed mean being a perfect lens), and whether or not a specific lens indeed is diffraction limited. Two different matters. As you say, it is nigh impossible to actually build a perfect, diffraction limited lens. But that fact doesn't change what "difraction limited" means.

> All that the term "diffraction limited" means is that the aberrations
> which may be "corrected" or "covered up" by stopping down have
> corrected as much as possible, and further reduction of the aperture
> diameter diminishes performance due to diffraction. That is the
> textbook definition used by designers, not the misinterpretation used
> by salespeople and marketers. In other words, diffraction can limit
> the performance of the "good lens" as well as the "dog" at the same
> aperture - the "dog" doesn't improve as it is stopped down either.

I don't know your textbook definition. The one i know is (again) very simple: diffraction limited means that resolution is limited by (guess what?) diffraction. If any aberration has an effect limiting resolution more than diffraction would, resolution (again, guess what?) simply is not diffraction limited. It would be a misinterpretation to call a lens "diffraction limited" when in fact its performance (as far as resolution is concerned) is indeed limited by aberrations. That's your interpretation, isn't it?

> When the manufacturer or designer following accepted optical
> definitions and criteria, states that a lens is diffraction limited
> at f/5.6, all that means is diffraction harms performance if it is
> stopped down more.

No, no! It means that *at that precise aperture*, resolution is *equal to what diffraction allows it to be*.

Of course it can only get worse when stopping down further, yet this particular lens might very well be diffraction limited at *all* apertures. Diffraction limited resolution will roughly be halved when going from f/5.6 to f/11, but if the lens still manages to resolve as good as diffraction would allow, it still is diffraction limited at the smaller aperture, despite the reduction in resolution compared to wider apertures.

> The performance is still limited by aberrations or
> the lens would have been better at wider apertures.

Not so. If a lens manages to resolve the amount of detail diffraction will allow at a given aperture, it really is diffraction limited. Does that mean it *has to* perform better at wider apertures? No, it doesn't. It all depends on whether or not other resolution limiting factors, i.e. aberrations, still are reduced to the same low level at the wider aperture. If they indeed are would depend on the particular lens design alone. But if they are not, the lens quite simply isn't diffraction limited at that wider aperture, and that could easily account for the worse performance.

There is no conclusion you can draw from that regarding the intial, smaller aperture, and whether or not the lens was indeed diffraction limited at that aperture.

> Some may have
> been reduced by stopping down, and others are not affected.

And if uncorrected aberrations limit resolution, the lens isn't diffraction limited. As simple as that.

> In addition, the term says nothing about aberrations and defects in
> _every_ lens which are not affected by aperture.

Would you expect it to do?

> The people selling lenses would like you to believe that a lens which
> is advertised as "diffraction limited" at f/5.6 or whatever, is
> Rayleigh's hypothetical "perfect" or "ideal" lens which has no
> aberrations affecting performance. They can sell a worn out piece of
> plastic junk which is decentered; has image wander, on-axis
> astigmatism, distortion, and misalingment of groups tilting the focal
> plane; and shows rainbows instead of tangential lines off-axis due
> to lateral color. On axis, it can demonstrate poor, acceptable or
> even high resolution.

Yes. We can all pretend to be more than we really are. So what? Does that have any bearing on the actual meaning of terms we abuse to describe what we are not? I don't think so.

> The manufacturer can properly and legally call the dog "diffraction
> limited" at f/5.6 if stopping down further harms performance. That is
> the _only_ criteria required to make the claim of "diffraction
> limited" performance at a certain aperture as defined by the optical
> industry.

I never heard it defined this way. Until now.

> The marketing department, hoping that a consumer will misinterpret
> that statement as meaning that the lens is "perfect" or that _all_
> aberrations and defects are insignificant compared to diffraction.

That would be the correct interpretation. Sad if the lens doesn't perform that way: the consumer will have been lied to...

> In fact, in the "dog" example above, the residual aberrations and
> defects are not insignificant,

True. So it isn't diffraction limited.

> however, diffraction simply adds to
> the problems which are not corrected by reducing the aperture. The
> "dog's" poor performance often tails off a bit slower as a percentage
> of optimal (suboptimal) performance, since the disk due to residual
> aberrations not corrected by aperture reduction is often larger than
> that of an excellent lens. However, the size of the sums increase
> past the "diffraction limiting" aperture.

So once a "dog" always a "dog".

> The term is more misleading that the "new" definition of
> "apochromatic" used by the manufacturers and standards
> organizations discussed at length in the LF group.

I can honestly say that i have never seen or heard this term abused the way you describe it. But that could be just because of me not reading enough brochures ;-) But i haven't seen anyone use the term "diffraction limited" the way you define it either, and that, i'm afraid, i fully attribute to you not using it correctly ;-)

Date: 8 Jul 2001
From: bg174@FreeNet.Carleton.CA (Michael Gudzinowicz)
Newsgroups: rec.photo.equipment.35mm
Subject: Re: diffraction limited Re: Lens testing at f8 or f11?

Kennedy McEwen rkm@kennedym.demon.co.uk wrote:

>>Michael Gudzinowicz wrote
>>
>>It is possible to have two lenses - one top notch and another,
>>a dog, which are diffraction limited on axis at f/5.6. The better
>>lens may have all aberrations corrected by stopping down to f/5.6,
>>and exhibit high resolution. The "dog" may show aberrations, lower
>>contrast and mediocre resolution which does not improve when
>>stopped down to f/8, since it is diffraction limited at f/5.6.
>>
>If its resolution is diffraction limited at f/5.6, NO lens will show
>improved resolution if stopped down to f/8 - neither a "top notch" lens
>nor a "dog"!  Diffraction limited resolution is inversely proportional
>to f/# : the resolution cut-off for a diffraction limited lens is
>1/(W x f/#), W is the wavelength of the light.  So your "top notch" lens
>will show just the same degradation of on-axis resolution as the "dog"
>in stopping down to f/8 if both are DL at f/5.6.

Actually, you are citing the Rayleigh criteria which sets an upper limit on resolution based on aperture size and distance from the aperture, if it isn't expressed as an angle.

That criteria assumes that there is _no_ lens in the system, or that the lens is "perfect" or "ideal" = has no aberrations.

As I stated, "diffraction limited" does not imply in any way that the lens is "perfect" and shows no aberrations. It simply is defined as the aperture at which one does not improve performance by stopping down further. There are aberrations which are not affected by decreasing the aperture, and it is possible to have poor and good lenses with the same limiting aperture with very different characteristics.

>Certainly, the poorer lens may be diffraction limited at f/5.6 on axis
>(ie. on axis all other aberrations are insignificant compared to
>diffraction) but only diffraction limited at the edge at f/11 (ie. at
>the edge of field other aberrations only become insignificant compared
>to diffraction at f/11).

Other aberrations are _not_ insignificant compared to diffraction, simply because the diffraction limiting aperture is used. For instance, at f/8 the Rayleigh limit is approximately 1600/8 = 200 lpmm. A very good lens might resolve close to 200 lpmm if the aerial image is examined, and a "dog", 100 lpmm. Both lenses are optimal at f/8, and both show worse performance f/11 or smaller apertures. In other words, they are both "diffraction limited" at f/8.

"Diffraction limited" implies simply that there is no additional benefit in stopping down the lens. It does not imply, and never has implied in the optical literature, that performance is equivalent or that the lens is "perfect" at its diffraction limit. That is an invention of marketing groups and salespeople.

With large format lenses, I've tested lenses which were "best" at the same aperture, but the resolution of one was 1/3 that of another. Off-axis, peformance was worse, and never peaked. It was an old Schneider lens, and undoubtedly the problems were related to poor quality control and manufacturing defects, since the design was sound.

>A lens which is DL at f/5.6 is BETTER than a lens which is DL at f/8,
>since the sum of other aberrations must be 1.4 times less.

Generally, that is the case when comparing lenses of the same or similar design. However, some aberrations (primarily, lateral chromatic aberration and distortion) are not corrected by stopping down, and the "diffraction limit" aperture says _nothing_ about them.

Likewise, decentration of elements during manufacture or use (edge finishing, cementing alignment, cell construction, and cell alignment) generally are more important that the aberrations corrected by stopping down when testing lenses. The result might be marked one-directional coma across the field, tilting of the focal plane, or less consequential (for pictorial usage) tangential distortion/shift or image run-out. A large portion of the cost of high quality lenses is related to the prevention of decentration during manufacture.

All I am pointing out is that lenses with the same diffraction limited aperture are not necessarily equivalent in performance, nor are they "perfect".

In the newsgroups people are often given the impression that an inexpensive lens (cheap in construction and design) which is diffraction limited at f/8 has the same performance as a well made state of the art lens stopped down to f/8 (diffraction limited at f/4). That is not the case. Likewise, lenses with the same diffraction limited aperture may have marked differences in performance on axis and across the field.

Date: 7 Jul 2001
From: bg174@FreeNet.Carleton.CA (Michael Gudzinowicz)
Newsgroups: rec.photo.equipment.35mm
Subject: Re: diffraction limited Re: Lens testing at f8 or f11?

Robert Monaghan rmonagha@smu.edu wrote:

>the interesting aspect of lens resolution testing at f/8 to f/11 is that
>many lenses are diffraction limited at these f/stops, so even a  relatively
>low cost lens will often perform at its best around f/5.6 to f/11, and
>moreover, if diffraction limited (as many 50mm normal lenses are, for
>example), then no lens regardless of mfger or cost will outperform the
>resolution - that's what diffraction limited means ;-)

That is not what it means...

All "diffraction limited" means is that performance will not improve if the lens is stopped down past the diffraction limiting aperture, since diffraction will reduce resolution.

It is possible to have two lenses - one top notch and another, a dog, which are diffraction limited on axis at f/5.6. The better lens may have all aberrations corrected by stopping down to f/5.6, and exhibit high resolution. The "dog" may show aberrations, lower contrast and mediocre resolution which does not improve when stopped down to f/8, since it is diffraction limited at f/5.6.

...

Date: Sun, 8 Jul 2001
From: Kennedy McEwen rkm@kennedym.demon.co.uk
Newsgroups: rec.photo.equipment.35mm
Subject: Re: diffraction limited Re: Lens testing at f8 or f11?

Michael Gudzinowicz bg174@FreeNet.Carleton.CA wrote

>It is possible to have two lenses - one top notch and another,
>a dog, which are diffraction limited on axis at f/5.6. The better
>lens may have all aberrations corrected by stopping down to f/5.6,
>and exhibit high resolution. The "dog" may show aberrations, lower
>contrast and mediocre resolution which does not improve when
>stopped down to f/8, since it is diffraction limited at f/5.6.

If its resolution is diffraction limited at f/5.6, NO lens will show improved resolution if stopped down to f/8 - neither a "top notch" lens nor a "dog"! Diffraction limited resolution is inversely proportional to f/# : the resolution cut-off for a diffraction limited lens is 1/(W x f/#), W is the wavelength of the light. So your "top notch" lens will show just the same degradation of on-axis resolution as the "dog" in stopping down to f/8 if both are DL at f/5.6.

Certainly, the poorer lens may be diffraction limited at f/5.6 on axis (ie. on axis all other aberrations are insignificant compared to diffraction) but only diffraction limited at the edge at f/11 (ie. at the edge of field other aberrations only become insignificant compared to diffraction at f/11).

A lens which is DL at f/5.6 is BETTER than a lens which is DL at f/8, since the sum of other aberrations must be 1.4 times less.

--
Kennedy

From Nikon MF Mailing List;
Date: Sat, 21 Jul 2001
From: "Dale Cotton" dale-cotton@home.com
Subject: Re: Extreme DOF

> I presume, therefore, I'm misunderstanding what he means
> by "diffraction limited spot size"? I would like to understand
> these fundamentals eventually, so further input would be much
> appreciated ;)

In case anyone eventually tries to respond to this plea, I've found the source of my misunderstanding in a large format thread:

"Actually, in photography, it IS correct to say that diffraction only depends on f stop. Diffraction IS a function of the physical size of the aperture. However, diffraction patterns are angular patterns and are thus also a function of how far the aperture is from the screen used to view the diffraction pattern (i.e., the film). Note that a shorter lens is closer to the film than a longer lens. So even though the shorter lens has a smaller physical size of the aperture (greater diffraction), it is closer to the film and the diffraction pattern spreads less than it would with a longer lens which is further away from the film. So, in photography, it is absolutely accurate to say that the diffraction pattern is determined only by the f stop.

Cheers, DJ."

-- N Dhananjay (ndhanu@umich.edu), January 12, 2001.

Thus, as everyone else but I understood, DOF improves as f/stop becomes smaller, but this varies with the absolute size of the aperture. In contrast, diffraction degradation increases as f/stop gets smaller, regardless of the size of the aperture.

From Nikon MF Mailing List;
Date: Sat, 4 Aug 2001
From: "shepherdlen" Shepherdlen@btinternet.com
Subject: Re: Re: Whole F Stops

From: "John Owlett" owl@postmaster.co.uk

> The practical upshot of diffraction is that, in 35mm photography, you
> shouldn't normally use apertures smaller than f/22.  (Some purists
> would limit us to f/16, but I find I often need f/22 to get the
> foreground sharp.)  With some close-ups, the only way to get any
> depth-of-field at all is to use f/32 and risk the diffraction.

Hi Owl

Good to see you are still around and, as ever, being practical.

The deterioration due to diffraction between f16 and f22 exists in mathmatics but in the real world is below 10% and between f16 and f32 is under 30%. As modern lenses out resolve any film faster than 100ASA it is difficult to see any quality fall off in most pictures.

Len Shepherd.

From Nikon Mailing List;
Date: Sun, 05 Aug 2001
From: bryce hashizume bryceh@cpsc.ucalgary.ca
Subject: Re: Whole F Stops

> The deterioration due to diffraction between f16 and f22 exists in
> mathmatics but in the real world is below 10% and between f16 and f32 is
> under 30%.

This is a bit of a generalisation. The Rayleigh diffraction formula says that the maximum possible resolution of a lens is 102 lp/mm at f/16; 74 lp/mm at f/22; and 51 lp/mm at f/32 (small variantions in these numbers are possible depending on what values you use in the formula). Of course this assumes that the lenses are diffraction limited at these apertures, which isn't purely true for most lenses. On the other hand, many lenses are capable of resolving well over 100 lp/mm at f/8 or f/ll, but NO lens is capable of resolving more than 51 lp/mm at f/32. So diffraction can have a very definite effect on the final image.

> As modern lenses out resolve any film faster than 100ASA it is
> difficult to see any quality fall off in most pictures.

This is another common misconception. Just because a lens has an aerial resolution of 80 lp/mm and the film also has a resolution of 80 lp/mm, it certainly doesn't mean that you can achieve a resolution of 80 lp/mm with that film & lens combination.

The resolution of film/lenses is usually measured as the target resolution which achieves an MTF contrast of 0.1, or 10%. A lens with a resolution of 80 lp/mm can render an 80 lp/mm target with 10% contrast. The same thing goes for an 80 lp/mm film. But when you use both film and a lens, you have to multiply the contrast levels of the film and lens. So if you take a picture of an 80 lp/mm target with an 80 lp/mm lens and 80 lp/mm film, you get an MTF of only 0.1 x 0.1 = 0.01, or 1%. Our eyes can't discern 1% contrast, so the lens/film combination definitely does not yield a system resolution of 80 lp/mm.

Noted Leica-phile Erwin Putz suggests that if you want to achieve a resolution on the film equivalent to the resolution of the film alone, then you should use a lens that has 3 times the resolution of the film. For example, if you want to get something close to 80 lp/mm on your 80 lp/mm film (maybe 78 or 79 lp/mm... you'll never get 80 lp/mm unless you have some sort of wonder lens that has an MTF of 100% at 80 lp/mm), then you should use a lens with a resolution of at least 240 lp/mm.

At any rate, it doesn't really matter if you do have a lens with a resolution of 240 lp/mm if you stop down to f/32 or f/22. That's because the lens is effectively diffraction limited at these apertures, and you can't get a resolution of more than 51 or 74 lp/mm (respectively), anyway.

bryce

From: ramarren@bayarea.net (Godfrey DiGiorgi)
Newsgroups: rec.photo.equipment.medium-format
Subject: Re: Highest "F" stop in MF
Date: Wed, 02 Jan 2002 

Yes. 

What Robert's comment fails to state is the reason why smaller formats
have more DoF than larger formats. Depth of Field is parametrized by the
physical size of the aperture and the magnification of the negative, not
directly to focal length. That physical aperture size is normalized to the
focal length by the use of f/numbers (f/number == focal length divided by
aperture size). So smaller format cameras have more Depth of Field for a
given f/number because they use shorter focal length lenses for the same
coverage (field of view) and the resulting aperture dimensions are
smaller.

The bounding point in lens resolution is when the aperture reaches the
point where it by itself is diffracting a sufficient amount of the
incoming light to perturb the lens' ability to image cleanly ... the
diffraction limit. This happens at an opening around 2.0-2.5mm in
diameter, so you want to keep your lens' minimum aperture above that point
in most cases. For a "normal" 75mm lens on 6x6 format, f/32 is a 2.35mm
aperture ... at the diffraction limit (approximately) where for a "normal"
50mm lens on 35mm format, f/22 is past the diffraction limit at 2.27mm
opening. 

Returning to DoF, consider a Minox camera with a format size of 8x11mm.
It's 15mm lens is fixed aperture at f/3.5 so it has an opening of 4.2mm,
well above the diffraction limit, and its hyperfocal range produces a DoF
of 6' to infinity in focus with coverage of a 43mm lens in 35mm. A 35mm
format lens of 43mm focal length would need to be set to f/11 to achieve
similar DoF, a 6x6 lens of 75mm focal length would need to be set to f/19
or so for similar DoF.

Godfrey

steven.sawyer@banet.net wrote:

> Can I follow that logic and say that micro format (16mm/Minox) would have even
> greater depth of field?
> 
> Robert Monaghan wrote:
> 
> > actually, 35mm > med fmt > LF  as far as usable DOF goes; the reason LF
> > cameras seem to have more DOF is that they cheat by using movements to put
> > plane of focus where needed. But if you try to shoot a complex landscape
> > with flowers in front, rocky outcrops in mid-range, and a mountain at
> > infinity you all want in focus, you had better be using something wide and
> > on 35mm ;-)


From: "Q.G. de Bakker" qnu@worldonline.nl>
Newsgroups: rec.photo.equipment.medium-format
Subject: Re: Highest "F" stop in MF
Date: Thu, 3 Jan 2002 

Godfrey DiGiorgi wrote:

> The bounding point in lens resolution is when the aperture reaches the
> point where it by itself is diffracting a sufficient amount of the
> incoming light to perturb the lens' ability to image cleanly ... the
> diffraction limit. This happens at an opening around 2.0-2.5mm in
> diameter, so you want to keep your lens' minimum aperture above that point
> in most cases. For a "normal" 75mm lens on 6x6 format, f/32 is a 2.35mm
> aperture ... at the diffraction limit (approximately) where for a "normal"
> 50mm lens on 35mm format, f/22 is past the diffraction limit at 2.27mm
> opening.

True, but even far before the "bounding point" is reached resolution will
suffer. Roughly, every 2 stops a lens is closed down maximum attainable
resolution is halved. So image degradation by diffraction does not only
occur at the "dark" end of the aperture range, but also when going from
f/1.4 to f/2, i.e. when changing aperture size from 53.6 mm to 35 mm. It
will perhaps only be visible in a perfect lens, but the effect is there.

> Returning to DoF, consider a Minox camera with a format size of 8x11mm.
> It's 15mm lens is fixed aperture at f/3.5 so it has an opening of 4.2mm,
> well above the diffraction limit, and its hyperfocal range produces a DoF
> of 6' to infinity in focus with coverage of a 43mm lens in 35mm. A 35mm
> format lens of 43mm focal length would need to be set to f/11 to achieve
> similar DoF, a 6x6 lens of 75mm focal length would need to be set to f/19
> or so for similar DoF.

And what if we enlarge the negatives produced by these cameras so they all
end up the same size, and then viewed these from the same distance?




From: David Littlewood david@nospam.demon.co.uk>
Newsgroups: rec.photo.equipment.35mm
Subject: Re: diffraction limited optics was Re: Why is Leica so expensive?
Date: Fri, 5 Oct 2001 

ArtKramr artkramr@aol.com writes
>> have the 50mm lens (in several M42/K variants) and it is a top performer
>>(esp. for $25 US used). By f/8, I'd expect aberrations to be well
>>controlled and performance diffraction limited in both lenses.  So as far
>>as resolution goes, I would be surprised if the Leica lens delivered
>>noticeably higher resolution, given it is presumably diffraction limited
>>in these ranges around f/8 and the pentax 50mm is diffraction limited too. 
>>Right? 
>>
>>Personally, I think this test is probably accurate as Keppler reported it; 
>>namely, that both the Pentax 50mm and Leica 50mm lenses were diffraction 
>>limited by f/8 and performed roughly similarly in on-film resolution. I 
>>would have been surprised at any OTHER result, given the nature of 
>>diffraction limits. But it does come as a surprise to many folks that a 
>>good cheapy lens can equal performance of much pricier lens at these 
>>diffrction limited f/stops, at least in terms of resolution. 
>
>from what you have said I don't think you own a diffraction limited lens or
>anything even close.
>
Every lens is diffraction limited if it is stopped down far enough.
-- 
David Littlewood


From: T P please.reply@newsgroup>
Newsgroups: rec.photo.equipment.35mm
Subject: Re: diffraction limited optics was Re: Why is Leica so expensive?
Date: Sat, 06 Oct 2001 

"Mxsmanic" mxsmanic@hotmail.com> wrote:

> "David Littlewood" david@nospam.demon.co.uk> wrote 
> 
> > Every lens is diffraction limited if it is
> > stopped down far enough.
> 
> Yes, but for 35mm cameras, that means around f/16, not f/8.


I never before thought I would agree with Manic on anything.  
I hate to admit it, but this statement is bang on the button.

I know David didn't state the idea that a 35mm lens could be 
diffraction limited at f/8, but Photodo make that claim.

It is of course completely laughable.  It might be true in a Minox
submin, but *never* in a lens for the 35mm format unless its optical
design were *wholly* incompetent.

Not even Cosina lenses approach that.

-- 
Best regards,
TP


From: artkramr@aol.com (ArtKramr)
Newsgroups: rec.photo.equipment.35mm
Date: 06 Oct 2001 
Subject: Re: diffraction limited optics was Re: Why is Leica so expensive?

>ArtKramr wrote:
>
>> from what you have said I don't think you own a diffraction limited lens
>or
>> anything even close.
>
>You cannot own a lens with a diaphragma and not own a diffraction limited
>lens.


When we talk of diffraction limited lenses we are talking about lenses that
produce their best performance wide open.  The introduction of a diaphragm 
immediately causes diffraction and lowers the performance. In these cases the
wide open performance is so high  that introducing diffraction by stopping down
does not give performance improvements comensurate with the smaller aperture.
To call every lens diffraction limited is to confuse the issue. What determines
the precise limits? How do you  measure the degree of diffraction to  call it a
limit?.  To  choose the aperture at which the lens performs best and call that
a diffration limit is to totally miss the point. Examples of diffratrion
limited lenses are the Apo-El Nikkor 105mm F/5.6 and the 80mm F.5,6 S-Planar
which according to their makers produce best performance wide open and lose
peroformance  even stopped down even 1/10th stop. These are difraction limited
lenses in the true sense of the term. Diffraction limited means the
introduction of any diffraction at all limits their performance. Don't confuse 
"a diffraction limited lens" with the idea of the point at which any lens is
limited by diffraction.



Arthur Kramer
Visit my WW II B-26 website at:
http://www.coastcomp.com/artkramer



From: "Q.G. de Bakker" qnu@worldonline.nl
Newsgroups: rec.photo.equipment.35mm
Subject: Re: diffraction limited optics was Re: Why is Leica so expensive?
Date: Sat, 6 Oct 2001 

ArtKramr wrote:

> When we talk of diffraction limited lenses we are talking about lenses that
> produce their best performance wide open.

That is a wrong way of using the term. Diffraction limited means that
residual geometrical aberrations have
been reduced to such a low level that the point spread function is
essentially the same size as if the geometrical aberrations were all zero.
All lenses are diffraction limited at a certain f-stop, as diffraction
increases the point spread relentlessly when stopping down, while
aberrations are reduced.

>  The introduction of a diaphragm
> immediately causes diffraction and lowers the performance.

Even wide open the lens has a diaphragm, and has diffraction. But (most)
lenses show aberrations too. While diffraction increases when stopping down,
most aberrations decrease. So the net result of stopping down moderately
usually is an increase in performance. Until the lens is diffraction
limited, i.e. until the lens is stopped down so far that diffraction is
worse than residual aberrations.

> In these cases the
> wide open performance is so high  that introducing diffraction by stopping down
> does not give performance improvements comensurate with the smaller aperture.
> To call every lens diffraction limited is to confuse the issue. What determines
> the precise limits? How do you  measure the degree of diffraction to  call it a
> limit?.  To  choose the aperture at which the lens performs best and call that
> a diffration limit is to totally miss the point.

So does using the term "diffraction limited" to say a lens is good. You will
have to mention the aperture at which it is diffraction limited as well to
make sense.

> Examples of diffraction
> limited lenses are the Apo-El Nikkor 105mm F/5.6 and the 80mm F.5,6 S-Planar
> which according to their makers produce best performance wide open and lose
> peroformance  even stopped down even 1/10th stop. These are difraction limited
> lenses in the true sense of the term.

"True" sense?
The "true" sense is that they, unlike other lenses, are diffraction limited
at their largest aperture.

> Diffraction limited means the
> introduction of any diffraction at all limits their performance. Don't confuse
> "a diffraction limited lens" with the idea of the point at which any lens is
> limited by diffraction.

You don't need to "introduce" diffraction. You can't have a beam of light
with a limited size and not have diffraction. Diffraction's influence
increases with decreasing aperture sizes, while residual aberrations
decrease, so at one point all lenses are diffraction limited. That point may
well be at a lens' largest aperture, usually it isn't. But that does not
alter the fact that all (!!!) lenses are diffraction limited at some point.
I don't find that confusing at all...



From: artkramr@aol.com (ArtKramr)
Newsgroups: rec.photo.equipment.35mm
Date: 06 Oct 2001 
Subject: Re: diffraction limited optics was Re: Why is Leica so expensive?

>When we talk of diffraction limited lenses we are talking about lenses
>that
>> produce their best performance wide open.

>sidual geometrical aberrations have
>been reduced to such a low level that the point spread function is
>essentially the same size as if the geometrical aberrations were all zero.
>All lenses are diffraction limited at a certain f-stop, as diffraction
>increases the point spread relentlessly when stopping down, while
>aberrations are reduced.

We are talking about lenses that are at their best performance wide open. And
stopping  down introduces diffraction that is GREATER than any of the
advantages of stopping down. Therefore best performance is without stopping 
down (introducing additional diffraction) and is therefore diffraction limited
at full aperture. It means that  a diffraction limited lens is  aways used wide
open for best performance. 

>Even wide open the lens has a diaphragm, and has diffraction. But (most)
>lenses show aberrations too. While diffraction increases when stopping down,
>most aberrations decrease. 

No. You we are talking about lenses at their best wide open and stopping down
at all reduces performance due to diffraction such as the lenses I mentioned
which are termed diffraction limited by their makers.

>So does using the term "diffraction limited" to say a lens is good. You will
>have to mention the aperture at which it is diffraction limited as well to
>make sense.

Yes. And That is wide open.

>True" sense?
>The "true" sense is that they, unlike other lenses, are diffraction limited
>at their largest aperture.

Now you got it.



Arthur Kramer
Visit my WW II B-26 website at:
http://www.coastcomp.com/artkramer


From: Charles Richmond cmr@iisc.com>
Newsgroups: rec.photo.equipment.35mm
Subject: Re: diffraction limited optics was Re: Why is Leica so expensive?
Date: Sun, 07 Oct 2001 


a "diffraction limited lens" refers to a lens whose optics are so
good that it is diffraction limited fully open. This is true in
every branch of optics including photography. The statement that
a lens is "diffraction limited at f-whatever" is an expression of
how badly that lens falls short of its potentially best performance.
Lens marketeers and people justifying their purchases try to make
the two phrases somehow synonymous. 

-- 
*  Charles Richmond    Integrated International Systems Corporation   *
 


From: artkramr@aol.com (ArtKramr)
Newsgroups: rec.photo.equipment.35mm
Date: 07 Oct 2001 
Subject: Re: diffraction limited optics was Re: Why is Leica so expensive?

>Who is this "we"? Most of the posters here (including me) agree with
>QG's usage: a lens is diffraction limited when used at an aperture at
>which its resolution is limited only (or only significantly) by the
>effects of diffraction.

This is not an election where the majority rules. In this case the majority is
wrong.

>To use the term as you suggest (A lens whose performance wide open is at
>the limit set by diffraction) leads to the absurd conclusion that any
>lens can be made diffraction limited by adding an artificial aperture
>restriction in its construction.

I'm afraid you don't  even begin to have an understanding and  your last
statement is what is absurd.


Arthur Kramer
Visit my WW II B-26 website at:
http://www.coastcomp.com/artkramer


From: brianc1959@aol.com (brian)
Newsgroups: rec.photo.equipment.35mm
Subject: Re: diffraction limited optics was Re: Why is Leica so expensive?
Date: 7 Oct 2001 

You're more or less correct, but the real situation is more
complicated.  Some 35mm camera lenses really are diffraction limited
at f/8.  For example, the 50mm f/1.4 Nikkor when stopped to f/8 will
perform better than the Rayleigh limit out to the long side of a
24x36mm frame.  Performance dips slightly from there out to the
extreme corner, but most people would agree that this lens is
diffraction-limited at f/8.  Other lenses perform as well as this, and
some perform better.  The original Vivitar Series I 90mm f/2.5 macro
is diffraction-limited at f/5.6 near the center of the image even when
focussed at infinity.  Most wide angles will be diffraction-limited at
f/8 very close to the center of the field, but fall off rapidly due to
color fringing.  The Nikkor 20mm f/2.8 for example is diffraction
limited at f/8 only over a 3mm diameter circle at the center of the
image.  Color fringing can prevent diffraction-limited performance
over a wide field even at very small stops, which is one reason why
this aberration is so harmful.  For example the 20mm Nikkor has a
diffraction-limited image circle about 15mm in diameter at f/22. 
Compare this to the full 43mm image circle required to cover the
format.  In fact this lens would not be strictly diffraction limited
over the entire format until you stop it down to f/32, which is not
possible without installing a different diaphragm.

Brian




T P please.reply@newsgroup> wrote ...
> "Mxsmanic" mxsmanic@hotmail.com> wrote:
> 
> > "David Littlewood" david@nospam.demon.co.uk> wrote 
> > 
> > > Every lens is diffraction limited if it is
> > > stopped down far enough.
> > 
> > Yes, but for 35mm cameras, that means around f/16, not f/8.
> 
> 
> I never before thought I would agree with Manic on anything.  
> I hate to admit it, but this statement is bang on the button.
> 
> I know David didn't state the idea that a 35mm lens could be 
> diffraction limited at f/8, but Photodo make that claim.
> 
> It is of course completely laughable.  It might be true in a Minox
> submin, but *never* in a lens for the 35mm format unless its optical
> design were *wholly* incompetent.
> 
> Not even Cosina lenses approach that.


From: artkramr@aol.com (ArtKramr)
Newsgroups: rec.photo.equipment.35mm
Date: 07 Oct 2001 
Subject: Re: diffraction limited optics was Re: Why is Leica so expensive?

>You're more or less correct, but the real situation is more
>complicated.  Some 35mm camera lenses really are diffraction limited
>at f/8.  For example, the 50mm f/1.4 Nikkor when stopped to f/8 will

No. Let's review the bidding. A diffraction limited lens was always known  as a
lens that gives best performance wide open. Of course this definition got a lot
of people upset when they realised that their favorite Leitz, Nikon, Zeiss
camera lenses perform nowhere hear that good wide open. So when it is asked why
wasn't the 50mm Summicron made to be diffraction limited and proivide its best
perofmance wide open? This embarassed the companies that made these relatively
low cost optics, which includes all consumer optics, the hedging and marketing
hype began to muddy up the waters. This was done by clouding the issue by
saying all lenses are diffraction limited because they have stops where
diffraction degrades the image. Above all they must wipe out the idea that
diffraction limited means best performance wide open. Lenses can be made that
way. But not F/1.4 or even F/2.0 lenses. And look to pay about $4,000 and up
for lenses of this type. But I own two lenses that are truly diffraction
limited and classed as such by thee makers. But they cost very big bucks, but
no other lenses I own, and I own many, compare to these diffraction limited
designs. So when someone tells you that all lenses are diffraction limited, go
get a  second opinion. (grin)


Arthur Kramer
Visit my WW II B-26 website at:
http://www.coastcomp.com/artkramer


From: "Q.G. de Bakker" qnu@worldonline.nl>
Newsgroups: rec.photo.equipment.35mm
Subject: Re: diffraction limited optics was Re: Why is Leica so expensive?
Date: Mon, 8 Oct 2001 

DBaker9128 wrote:

> Maybe I can pipe in. In Rudolf Kingslake's book entitled A History of the
> Photographic Lens, diffraction is defined (pg 317) as:
> "A narrow-angle scattering of light at an edge, such as at the blades of
an
> iris diaphragm. This often causes streaks of light to radiate from any
bright
> light in the scene, each streak being perpendicular to the edge causing
it. If
> the aperture of a lens is very small, say on the order of one millimeter,
> diffraction may cause a general loss of definition in a photograph."
> I would submit for the group that in the context of Kingslake,  Q.G.
deBakker's
> definition of diffraction limited optics (small aperture limiting
resolution)
> holds sway.

The only correct definition of diffraction limited is that diffraction,
caused by limited size of beam of light (and it is limited at all (!!!)
apertures, not just small ones), causes more point spread than any residual
geometric aberration. So resolution is not limited by the effect of faults
but by diffraction, and by diffraction alone.

Both aberrations and diffraction change with aperture size, so a lens that
is not diffraction limited wide open will unavoidably be at a certain
smaller stop. There is no escaping the increase of image degradation caused
by diffraction, unless you prevent a lens being stopped down. Aberrations
however do get less when stopping down. So there will be a particular
aperture size/f-stop in any lens at which the lens' performance is
diffraction limited. It's one of those laws of nature. No escape.



From: "Q.G. de Bakker" qnu@worldonline.nl>
Newsgroups: rec.photo.equipment.35mm
Subject: Re: diffraction limited optics was Re: Why is Leica so expensive?
Date: Mon, 8 Oct 2001 

Charles Richmond wrote:

> Ok, you clearly understand what is meant by "diffraction limited".
> If I inferred or said otherwise then I apologise. So then why the
> seemingly stubborn refusal to recognise that the phrase "a diffraction
> limited lens" refers to the special case of a lens that meets the
> standard when not stopped down.

So tell me this: What is it that is being "limited" by diffraction? Indeed,
resolution. Now do lenses have resolution only wide open? No?
So would it perhaps be possible that diffraction limits resolution at any
other aperture than wide open as well?
So why not use the term wherever it applies, i.e. does it lose it's meaning
when applied to an f/8 lens whose resolution is limited by diffraction? Does
it change when this lens in fact is a f/4 lens stopped down to f/8?
No it doesn't, it is still diffraction that is limiting the lens'
resolution. The meaning of the term is not dependent on whether or not the
lens is stopped down or not. It only deals with the question what it is that
is limiting resolution.

So the question i would like to see answered is why you so stubbornly refuse
to
accept that this criterion can be met by any lens? Yes, most need to be
stopped down. And true, most of them do not show their best resolution when
they eventually are diffraction limited. But it is not implied in the
criterion in any way that they should. Is it? If you think it is, please
explain.

> From a post of mine from last December about stopping down and
> then claiming to be diffraction limited:
>
> > Only because the diffraction distortion is increased by stopping
> > down. It's kind of like shooting a picture through the air above
> > a fire and claiming that the lens performance is only limited by
> > atmospheric turbulence.

Which would be correct, but for the word "only". "Mostly" would be better.


From: "Q.G. de Bakker" qnu@worldonline.nl>
Newsgroups: rec.photo.equipment.35mm
Subject: Re: diffraction limited optics was Re: Why is Leica so expensive?
Date: Tue, 9 Oct 2001 

Charles Richmond wrote:

> a "diffraction limited lens" refers to a lens whose optics are so
> good that it is diffraction limited fully open. This is true in
> every branch of optics including photography. The statement that
> a lens is "diffraction limited at f-whatever" is an expression of
> how badly that lens falls short of its potentially best performance.
> Lens marketeers and people justifying their purchases try to make
> the two phrases somehow synonymous.

I can't help but noticing how there is no difference between the two.
First you mention the term is used as a quality criterion stating that at a
certain aperture (which happens to be wide open) the resolution of the lens
is limited by the amount of diffraction at that aperture.
Then you say linking the term to a particular f-stop is a different use and
meaning of the term...



From: rcochran@lanset.com (Richard Cochran)
Newsgroups: rec.photo.equipment.35mm
Subject: Re: which diff. limit rule? ;-) was Re: diffraction limited optics
Date: 9 Oct 2001 

artkramr@aol.com (ArtKramr) wrote 
> >Hi Arthur:
> >Perhaps this is nitpicking, but it is possible to design and build a
> >lens that performs best wide open, but is nevertheless not
> >diffraction-limited until you stop it down.  T
> 
> I don't see how.  If it is best wide open, it is diffraction limited at that 
> wide open aperture. If it improves when stopped down, then it wasn't at its
> best wide open therefore it was not a diffraction limited lens.

Note that he didn't say the lens would improve when stopped down, he
merely said it would become diffraction-limited when stopped down.

In all the references I've found, a diffraction limited lens is one 
which has all aberrations (optical errors) corrected to the point that 
residual wavefront errors are substantially less than 1/4 wavelength 
of light at the focal point.  This is what is necessary for the
lens to achieve the Dawes limit of resolution, that is, the
ability to separate two closely-spaced point sources of light into 
two distinct images, when the angular separation of the point
sources measured in arc-seconds is the aperture (in mm) divided
by 116.

Another, more intuitive, way of saying it is that it a diffraction
limited lens has a resolving power that is as good as any lens of 
its focal length and aperture could possibly have.  All aberrations
are insignificant.  The physics of diffraction is all that's limiting 
the lens.

It's true that a diffraction-limited lens will perform worse when
stopped down, but it's not necessarily true that a lens which performs
worse when stopped down is diffraction-limited.

A lens might perform better wide-open than at any other aperture, but 
it still might not perform as well as another lens of that same focal 
length and aperture.  Such a lens is limited by aberrations, not
diffraction.

--Rich


From: artkramr@aol.com (ArtKramr)
Newsgroups: rec.photo.equipment.35mm
Date: 09 Oct 2001 
Subject: Re: which diff. limit rule? ;-) was Re: diffraction limited optics

>Another, more intuitive, way of saying it is that it a diffraction
>limited lens has a resolving power that is as good as any lens of 
>its focal length and aperture could possibly have.  All aberrations
>are insignificant.  The physics of diffraction is all that's limiting 
>the lens.
>

That's true.

>It's true that a diffraction-limited lens will perform worse when
>stopped down, but it's not necessarily true that a lens which performs
>worse when stopped down is diffraction-limited.

If the lens is at it's best wide open and degrades when diffraction is
introduced  I would say it is indeed a diffraction limited lens.

>A lens might perform better wide-open than at any other aperture, but 
>it still might not perform as well as another lens of that same focal 
>length and aperture.  Such a lens is limited by aberrations, not
>diffraction.
>
>--Rich

I think you have just provided a description of a non existant lens. If a lens
suffers from abberations it is very unlikely it would be at it's sharpest wide
open and none of the abberations would  be lessened by stopping down. At least
I have never seen such a lens nor have I ever in examining  hundreds of lenses
on an optical bench seen a lens that was bad wide open that wasn't improved to
some degree by stopping down. Have you ever seen such a lens?

Arthur Kramer
Visit my WW II B-26 website at:
http://www.coastcomp.com/artkramer


From: brianc1959@aol.com (brian)
Newsgroups: rec.photo.equipment.35mm
Subject: Re: which diff. limit rule? ;-) was Re: diffraction limited optics
Date: 9 Oct 2001 

Arthur:
I'm a professional lens designer, and I actually have designed and
tested a mass-production lens that was sharpest at its widest
aperture, but was not rigorously diffraction-limited until it was
stopped down a couple of stops.  Your definition of
diffraction-limited performance is interesting and perhaps useful, but
it is not the definition used in the optical industry. Think of it
this way:  in a purely diffraction limited system the performance
increases in a predictable way as the aperture is enlarged.  In this
case, the best performance is achieved at the largest aperture.  If
small aberrations are introduced, then it is still possible to
maintain the best performance at the largest aperture, but the
performance at that aperture nontheless falls short of the diffraction
limit.

Brian
 
artkramr@aol.com (ArtKramr) wrote...
I think you have just provided a description of a non existant lens. If a lens
suffers from abberations it is very unlikely it would be at it's sharpest wide
open and none of the abberations would  be lessened by stopping down. At least
I have never seen such a lens nor have I ever in examining  hundreds of lenses
on an optical bench seen a lens that was bad wide open that wasn't improved to
some degree by stopping down. Have you ever seen such a lens?

Arthur Kramer
Visit my WW II B-26 website at:
http://www.coastcomp.com/artkramer


From: artkramr@aol.com (ArtKramr)
Newsgroups: rec.photo.equipment.35mm
Date: 10 Oct 2001 
Subject: Re: which diff. limit rule? ;-) was Re: diffraction limited optics

>Your definition of
>diffraction-limited performance is interesting and perhaps useful, but
>it is not the definition used in the optical industry. Think of it
>this way:  in a purely diffraction limited system the performance
>increases in a predictable way as the aperture is enlarged.  In this
>case, the best performance is achieved at the largest aperture.  If
>small aberrations are introduced, then it is still possible to
>maintain the best performance at the largest aperture, but the
>performance at that aperture nontheless falls short of the diffraction
>limit.
>
>Brian
>
 But let's assume that small  abberations are not  introduced, then this lens
will not fall short of the  diffraction limit. At least that is the position
taken by Nikon in the 10 years I worked with them The 105mm F/5.6 Apo El Nikkor
is such a lens and Nikon includes instructions packed with the lens  warning
that stopping down even one stop from F/5.6 to F/8.0 will degrade performance.
Carl Zeiss makes the same statements regarding their S-Planar 80mm F/5..6
lenses which came out of their scientific division, not their photo division. I
guess it is a case of exactly which lenses we are talking about and who made
them to what standards. By the way, the above lenses were about $4,000 each in
barrel mount so we are not talking about the run of the mill consumer lenses
that most people use..


Arthur Kramer
Visit my WW II B-26 website at:
http://www.coastcomp.com/artkramer


From: artkramr@aol.com (ArtKramr)
Newsgroups: rec.photo.equipment.35mm
Date: 09 Oct 2001 
Subject: Re: which diff. limit rule? ;-) was Re: diffraction limited optics

>Hi Arthur:
>Perhaps this is nitpicking, but it is possible to design and build a
>lens that performs best wide open, but is nevertheless not
>diffraction-limited until you stop it down.  T

I don't see how.  If it is best wide open, it is diffraction limited at that 
wide open aperture. If it improves when stopped down, then it wasn't at its
best wide open therefore it was not a diffraction limited lens.

>Also, there is an additional complication in that diffraction-limited
>performance is often reached at a relatively wide aperture on-axis but
>at a much smaller aperture at the corner of the field.
>
>Brian

Not really. difraction limited lenses hav eno edges. Let me explain. The Apo El
Nikkor 105mm F.5.6 is diffraction limited at F/5.6. But  it only covers 35mm,
yet had a focal length of 105mm It only uses the center of a highly corrected
field  which is why we say it has no edges, the edges are never used  and it's
covering power is deliberately limited for that reason. The same is true of
apochromats made for photo mechanical work such as Apo Artars and Apo Ronars.
Only the center is used so the usual problems of  working with  edges is
avoided.


Arthur Kramer
Visit my WW II B-26 website at:
http://www.coastcomp.com/artkramer


From: artkramr@aol.com (ArtKramr)
Newsgroups: rec.photo.equipment.35mm
Date: 08 Oct 2001 
Subject: Re: which diff. limit rule? ;-) was Re: diffraction limited optics

>Perhaps Mr. Kramer or others here will know if this is also true for the
>use of the term "diffraction limited" for wide open lenses, in its
>(initial?) sense?  Or is there a more precise definition of where
>diffraction's impact becomes the limit for such optics? Which of the rules
>of thumb is closest to such a value? Do different photo/lens manufacturers
>use different limits for classifying their optics as "diffraction limited"  
>wide open, or do they agree on one definition?
>definitions in the 1950s prime refers to any lens 


When we talk of a true diffraction limited lens we  are talking about a lens
that gives it optimum performance wide open, The moment we stop it down, it
loses performance.  It is that simple and that absolute. But the "popular" and
invalid  use of diffraction limited as applied to all lenses and the stop where
diffraction  is unacceptable is a subjective target in constant motion.. There
is no standard for that. It is in the eye of the beholder and is based on what
any individual is willing to accept and a hopelessly subjective standard of
what anyone considers acceepetable or not acceptable. It is in fact a non
defiition definition. 


Arthur Kramer
Visit my WW II B-26 website at:
http://www.coastcomp.com/artkramer


From: brianc1959@aol.com (brian)
Newsgroups: rec.photo.equipment.35mm
Subject: Re: which diff. limit rule? ;-) was Re: diffraction limited optics
Date: 8 Oct 2001 

Hi Arthur:

Perhaps this is nitpicking, but it is possible to design and build a
lens that performs best wide open, but is nevertheless not
diffraction-limited until you stop it down.  This is actually a fairly
common occurence in the short focal length zooms used on small sensor
consumer digicams.  "Diffraction-limted" means a fairly specific thing
that can be expressed in several ways:  1)less than 1/4 wave
peak-to-valley wavefront error, or 2)less than 0.07 waves RMS
wavefront error, or 3)more than 0.8 Strehl ratio, and so on.  I think
your definition is a good rule of thumb, but I would never apply it to
work I do for paying clients.

Also, there is an additional complication in that diffraction-limited
performance is often reached at a relatively wide aperture on-axis but
at a much smaller aperture at the corner of the field.

Brian


artkramr@aol.com (ArtKramr) wrote 
> >Perhaps Mr. Kramer or others here will know if this is also true for the
> >use of the term "diffraction limited" for wide open lenses, in its
> >(initial?) sense?  Or is there a more precise definition of where
> >diffraction's impact becomes the limit for such optics? Which of the rules
> >of thumb is closest to such a value? Do different photo/lens manufacturers
> >use different limits for classifying their optics as "diffraction limited"  
> >wide open, or do they agree on one definition?
> >definitions in the 1950s prime refers to any lens 
> 
> 
> When we talk of a true diffraction limited lens we  are talking about a lens
> that gives it optimum performance wide open, The moment we stop it down, it
> loses performance.  It is that simple and that absolute. But the "popular" and
> invalid  use of diffraction limited as applied to all lenses and the stop where
> diffraction  is unacceptable is a subjective target in constant motion.. There
> is no standard for that. It is in the eye of the beholder and is based on what
> any individual is willing to accept and a hopelessly subjective standard of
> what anyone considers acceepetable or not acceptable. It is in fact a non
> defiition definition. 
> 
> 
> Arthur Kramer
> Visit my WW II B-26 website at:
> http://www.coastcomp.com/artkramer



From: brianc1959@aol.com (brian)
Newsgroups: rec.photo.equipment.35mm
Subject: Re: which diff. limit rule? ;-) was Re: diffraction limited optics
Date: 10 Oct 2001 

Hi Arthur:

The lenses you mention may actually have been diffraction limited, but
they certainly could not have had zero aberration.  No non-trivial
lens has zero aberration, even $1,000,000+ photolithographic
objectives.  It is a fact that diffraction can actually hide optical
faults in many instances.

My only point earlier was that the diffraction limit is a reasonably
well defined thing, and it is not defined as the point at which
stopping down degrades performance.  To repeat, the commonly accepted
standards for a diffraction limited system are:  1)peak-to valley
wavefront error of 1/4 wave or less, or 2)RMS wavefront error of 0.07
waves or less, or 3)Strehl ratio of 0.8 or higher.  There are other
criteria involving the area under the measured MTF curve, but these
are not really standard.  A lens having a peak-to-valley wavefront
error of 1/2 wave might very well perform best wide-open, but no
optical designer would ever call it diffraction-limited.

All of these measures are easy to calculate using lens design
software, but the most meaningful way to test a lens to determine
whether it is actually diffraction limited or not is with an
interferometer or with a Ronchii or knife-edge test.  Using a star
test on a bench can detect small aberrations, but it is very difficult
to determine whether or not one of the rigorous criteria for
diffraction-limited performance has been met.  Resolution testing is
almost completely useless if you want to determine whether a lens is
truly diffraction limited or not.

Brian


> >
>  But let's assume that small  abberations are not  introduced, then this lens
> will not fall short of the  diffraction limit. At least that is the position
> taken by Nikon in the 10 years I worked with them The 105mm F/5.6 Apo El Nikkor
> is such a lens and Nikon includes instructions packed with the lens  warning
> that stopping down even one stop from F/5.6 to F/8.0 will degrade performance.
> Carl Zeiss makes the same statements regarding their S-Planar 80mm F/5..6
> lenses which came out of their scientific division, not their photo division. I
> guess it is a case of exactly which lenses we are talking about and who made
> them to what standards. By the way, the above lenses were about $4,000 each in
> barrel mount so we are not talking about the run of the mill consumer lenses
> that most people use..
> 
> 
> Arthur Kramer
> Visit my WW II B-26 website at:
> http://www.coastcomp.com/artkramer


From: brianc1959@aol.com (brian)
Newsgroups: rec.photo.equipment.35mm
Subject: Re: which diff. limit rule? ;-) was Re: diffraction limited optics
Date: 10 Oct 2001 

Hi Robert:

To my mind, a completely diffraction limited lens will have a Strehl
ratio greater than 0.8 all the way to the corner of the field.

Most 35mm optics are not diffraction limited when used wide open even
in a narrow zone around the center of the field.  But then again, do
you really need 300+ cycles per millimeter performance on 35mm? 
Stopping down a little makes a big improvement in most cases.  I once
used a 50mm f/1.2 Nikkor for a project involving a high resolution CCD
because the lens is actually fully diffraction-limited at f/5.6 over a
small image area, and provided resolution well in excess of 200 cycles
per millimeter.  So, you may have been exaggerating a bit, but you
weren't far off.

Medium format optics with conservative specs are no doubt quite good,
especially near the center.  But I doubt that they are truly
diffraction-limited when used wide open.  There is no reason for them
to be.  After all, 85cycles/mm on film is really good, and there is no
need to reach 200-300 cycles.

Brian


rmonagha@smu.edu (Robert Monaghan) wrote 
> So who sets these coverage and edge effect criteria? Maybe my 35mm SLR OEM 
> optics are diffraction limited (wide open) on APS? ;-) ;-)
> 
> I suspect that many slower medium format lenses would have central aerial
> resolution values that would easily reach the Dawes criterion (circa
> 1200/f#, or even 1600/f#) when used to cover only 35mm too, and so qualify
> as "diffraction limited" by these standards, even wide open ;_) I mean, my
> zeiss 250mm or even Kowa 250mm at f/5.6, wide open, hit 85 lpmm on film,
> being 6x6cm medium format optics, and have central aerial resolutions that
> should easily equal the relaxed Dawes criterion or even Kingslake's
> 1600/f# factor. Lots of lesser $ med fmt lenses probably do too, if they
> are slow enough wide open ;-) If you aren't holding the corners of the
> format at these diffraction limits, but only using the central area, it
> seems a lot easier to reach goal optically. 
> 
> And as for $4,000 Zeiss lenses, that seems like a relative bargain for a 
> specialty lens from Zeiss. Bet those prices have gone up. Even the regular 
> hassy lenses are in that range nowadays, drat! ;-) grins bobm
> 
> e.g. fisheye 30mm is $6,335 from mail order B&H;
> 110mm f/2 normal lens is $3,918 " "
> 40mm f/4 is $4,197
> 250 f/5.6 SA is $5,279
> phew! ;-)


From: "LABourdillon" labour@NOSPAM.ix.netcom.com>
Newsgroups: rec.photo.equipment.35mm
Subject: Re: which diff. limit rule? ;-) was Re: diffraction limited optics
Date: Wed, 10 Oct 2001 

Most posters here seem to be saying the same thing - yet don't agree?

As has been mentioned by others, diffraction limited optics refers to lens
systems where all aberrations have been reduced to a minimum. Thus the len's
resolving 'power' (ability to separate two closely spaced point sources of
light - e.g. stars) is then limited *only* by the aperture of the lens. Any
diffraction limited (circular) lens system viewing a point source of light
(at increased magnifications) will produce a concentric pattern of closely
spaced light-dark rings with a bright spot in the middle. The bright spot,
the Airy disk, should contain about 80% of the light. The pattern is given
by a set of Bessel functions originally developed by Sir George Airy
(England ~1871). There's a lot more to the mathematics, but I'll leave it
here.

A diffraction limited lens at full aperture will therefore still (usually)
be diffraction limited when it is stopped down. This does not include
diffraction effects from the iris edges that become noticeable at very small
f-stops. Some lenses that are not diffraction limited at full aperture may
become diffraction limited (improve) at reduced apertures. This is simply
due to the fact that by stopping down the other lens aberrations may be
reduced.

...larry


brian brianc1959@aol.com> wrote...
> Hi Robert:
> To my mind, a completely diffraction limited lens will have a Strehl
> ratio greater than 0.8 all the way to the corner of the field.
>
> Most 35mm optics are not diffraction limited when used wide open even
> in a narrow zone around the center of the field.  But then again, do
> you really need 300+ cycles per millimeter performance on 35mm?
> Stopping down a little makes a big improvement in most cases.  I once
> used a 50mm f/1.2 Nikkor for a project involving a high resolution CCD
> because the lens is actually fully diffraction-limited at f/5.6 over a
> small image area, and provided resolution well in excess of 200 cycles
> per millimeter.  So, you may have been exaggerating a bit, but you
> weren't far off.
>
> Medium format optics with conservative specs are no doubt quite good,
> especially near the center.  But I doubt that they are truly
> diffraction-limited when used wide open.  There is no reason for them
> to be.  After all, 85cycles/mm on film is really good, and there is no
> need to reach 200-300 cycles.
>
> Brian
>
>
> rmonagha@smu.edu (Robert Monaghan) wrote 
> > So who sets these coverage and edge effect criteria? Maybe my 35mm SLR OEM
> > optics are diffraction limited (wide open) on APS? ;-) ;-)
> >
> > I suspect that many slower medium format lenses would have central aerial
> > resolution values that would easily reach the Dawes criterion (circa
> > 1200/f#, or even 1600/f#) when used to cover only 35mm too, and so qualify
> > as "diffraction limited" by these standards, even wide open ;_) I mean, my
> > zeiss 250mm or even Kowa 250mm at f/5.6, wide open, hit 85 lpmm on film,
> > being 6x6cm medium format optics, and have central aerial resolutions that
> > should easily equal the relaxed Dawes criterion or even Kingslake's
> > 1600/f# factor. Lots of lesser $ med fmt lenses probably do too, if they
> > are slow enough wide open ;-) If you aren't holding the corners of the
> > format at these diffraction limits, but only using the central area, it
> > seems a lot easier to reach goal optically.
> >
> > And as for $4,000 Zeiss lenses, that seems like a relative bargain for a
> > specialty lens from Zeiss. Bet those prices have gone up. Even the regular
> > hassy lenses are in that range nowadays, drat! ;-) grins bobm
> >
> > e.g. fisheye 30mm is $6,335 from mail order B&H;
> > 110mm f/2 normal lens is $3,918 " "
> > 40mm f/4 is $4,197
> > 250 f/5.6 SA is $5,279
> > phew! ;-)



To: camera-fix@yahoogroups.com>
From: Bob Shell bob@bobshell.com>
Date: Tue, 27 Nov 2001 
Subject: Re: [camera-fix] Home made pin hole lens.



> From: Ron Schwarz rs@clubvb.com>
* Date: Tue, 27 Nov 2001 
> To: camera-fix@yahoogroups.com
> Subject: Re: [camera-fix] Home made pin hole lens.
> 
> OK, sounds like you're leaving the glass in place.  IMO all you'd get is a
> severely-diffracted image.  Most medium format (and smaller) lenses suffer
> diffraction once you stop down past the mid range. Large format stuff is
> less susceptible to it, mainly because the longer the focal length, the
> larger the absolute "hole size" at any given f stop.

I have a set of precision cut pinholes which screw into the filter thread of
a lens.  Yes, they produce diffraction, but no more than a plain pinhole
camera (diffraction is the limiting factor in pinhole camera sharpness).
They produce incredible depth of field.

Bob



From nikon mf mailing list;
Date: Tue, 07 Aug 2001 
From: bryce hashizume bryceh@cpsc.ucalgary.ca>
Subject: Re: re: diffraction, fstops etc.


> re: large format lenses at f/64 - how do they do that? ;-)
> large format lenses are longer focal lengths than 35mm for same usage;
> normal lenses run about 135mm to 150mm to 165mm in 4x5" systems; call it
> 160mm / f16 ~ 10mm pupil/opening, 160/f32 ~ 5mm opening, f/64~ 2.5mm
> vs 50mm normal lens on 35mm at f/22 is about 2+mm too - same size  opening;
> logical since light doesn't know how far the film is back there ;-)

I'll have to disagree with you here, Bob.

The light does know how far back the film is. For example, a 50/1.8 Nikkor
and a Zeiss 80/2.8 may have the same rear coverage (measured in degrees),
but you can't use the 50 on a 'blad and expect to get a full frame 6x6
image. The point is that the cone of light gets wider the further it has
to travel (and the light from the 80 has to travel further than the light
from the 50).

Similarly, diffraction causes light to be "bent" outwards as it passes the
aperture blades, and the further the light travels, the more it is bent
from its original path (even though the deviation, as measured in degrees,
does not change). Think of a wedge. The further out the wedge you go, the
greater the distance between the two edges. The thin end of the wedge
would be where the aperture blades lie.

So we can see that the distance to the film, as well as the physical size
of the aperture, is important in determining the level of diffraction.
And, of course, an f-stop is a measure of just both these things, and that
is why the f-stop is used to determine diffraction leels, and not the
physical aperture size alone.

bryce


From rangefinder mailing list:
Date: Mon,  1 Oct 2001 
From: Michael Darnton mdarnton@hotmail.com>
Subject: RE: lens questions (lil' OT)


Why do lens openings stop when they do? Diffraction. The End. If you
notice with virtually any lens on a 35mm camera (for instance) after the
quality maxing out at about 2-stops down from wide open the quality
begins its slide into the toilet from that point smaller. A manufacturer
just has to decide at which point the potential damage to his reputation
isn't worth putting another number on at the end of the scale. On their
own, most diaphrams will shut down to almost a pinhole, but you wouldn't
like the results.

I believe diffraction damage is dependent on absolute hole size, rather
than relative size (f-stop), which is why longer lenses can stop down to
a smaller relative number, and still not reach the absolute minimum hole
where diffraction becomes a problem.

--Michael


From rangefinder mailing list:
Date: Mon,  1 Oct 2001 
From: Winfried Buechsenschuetz w-buechsenschuetz@gmx.de>
Subject: RE: lens questions (lil' OT)



Michael Darnton wrote:

> I believe diffraction damage is dependent on absolute hole size, rather
> than relative size (f-stop), which is why longer lenses can stop down to
>
> a smaller relative number, and still not reach the absolute minimum hole
>
> where diffraction becomes a problem.

That's absolutely right, and the same thing came to my mind when
re-reading my post. Diffraction depends on an absolute value. Since the
focal lengths of MF are approx. 1.5 times those of 24x36, their
diffraction will always be smaller. Also, if you run a 24x36 lens at,
say, f/2, its aberrations will appear. An MF lens with f/2.8 will have
similar diffraction, but much less aberration errors.

Winfried


From rangefinder mailing list:
Date: Mon, 1 Oct 2001 
From: Dante Stella dante@umich.edu>
Subject: RE: [RF List] lens questions (lil' OT)

Diffraction is not the reason.  You only need 6 lp/mm for a print of any
size at its normal viewing distance, and you never get anywhere near this
low with any lens in any format at any aperture.

Look at this page:

http://bobatkins.photo.net/info/dofdiff.htm

At f/32, you can still get 44+ lp/mm, which is more than you'll probably
ever need.  As it has been pointed out repeatedly, it is impossible to tease
more than 60lp/mm out of any complete optical path.  Medium format can get
away with far smaller apertures because the enlargement is not nearly as
great.

The numbers above are the *theoretical limits* with perfectly-designed
lenses.  If there is any optical reason to limit the maximum aperture, it
is because certain design-related lens aberrations creep back in as you
stop down (this has nothing at all to do with diffraction).  Sonnar-type
lenses, for example, are much worse at f/11 than they are at their peak
performance of f/2-5.6.  This is why Nikon made the minimum on their
50/1.5 rangefinder lens (predecessor to the 1.4) f/11.

Dante



From: "Bruce Wilson" b.e.wilson@usa.net
Newsgroups: rec.photo.equipment.35mm
Subject: Re: Diffraction and B&W;
Date: Tue, 19 Feb 2002


The effects of diffraction are a photo that is a bit fuzzier (unsharp) than
it can be. Most lenses have some aperture where it is sharpest, typically
f/16 or f/22. Stop down smaller than that and you increase the depth of
field (the hyperfocal distance) at the expense of overall image sharpness.

Typically only those who shoot on a tripod worry about it (handheld shooters
use the largest aperture possible and deal with camera shake as the biggest
contribution to unsharp photos). A critical eye can see the difference
between diffraction and shake easily (diffraction is unsharp in all
directions, shake is usually blurred in only one direction,usually the
vertical one).

It's easy to test your lens to find the sharpest aperture, just keep track
of your aperture for many shots of the same scene (use a tripod) and see
which ones are sharpest.


Colored fringes are what's known as chromatic aberration, and it is
typically seen in poorer-quality wide-angle zoom lenses, near the outer
edges when focused up close. All lenses are designed to focus properly using
green light. Chromatically-corrected lenses (most lenses you can buy) add
additional glass elements to correct for one other color of light.  These
work fine for most photography.  Lenses marked 'apo' or 'apochromatic' are
corrected for chromatic aberration of green plus two other colors. They are
the best lenses to use if you are going to play with focal distance (like
using macro lenses and tubes for close-up work). If you don't do macro work
an apo lens won't do you much good (in fact, an apo lens, because it has
more glass elements to generate reflections, can give poorer results than a
simpler lens when focused near infinity; I still carry a simple
chromatic-corrected four-element tessar-type lens for large-format shooting
even thought I have a nice apo planar-type lens because the tessar is
sharper when focused at infinity at f/16).

This might be more than you wanted to know.


--
Bruce Wilson
http://chem.dynu.com/photo
-
"Ken Durling" kdurling@earthlink.net wrote...
> Well, thanks folks.  I feel better!  ;-)   Actually I wouldn't know
> what diffratcion even looks like.  I was imagining something like
> color fringing, ergo my Q about B&W.;    Glad to hear it's not a real
> big concern.
>
>
> Ken


From: "Mxsmanic" mxsmanic@hotmail.com
Newsgroups: rec.photo.equipment.medium-format
Subject: Re: DOF "overrated"?
Date: Sun, 17 Mar 2002 


"Stephe" ms_stephe@excite.com wrote ...
> Is this actually diffraction?
Yes.

> Is that the correct term?
Yes.


Diffraction is an unavoidable consequence of physics.  No lens can be
designed to get around it.  It limits maximum resolution at all
apertures, but for all but the smallest apertures, the limitation is so
generous that defects in the lens usually limit resolution much more.  A
"diffraction-limited" lens is one of such high optical performance that
the only obstacle to perfection is the unavoidable fact of diffraction.

> But isn't there some set point/aperture size
> where actual 'difraction' sets in and a MAJOR
> loss of optical quality happens?

Diffraction is always present.  Its influence on the image becomes
increasingly obvious as a lens is stopped down.  At some point, it
limits resolution more than any defect in the lens, but the transition
is completely smooth, and so there is no obvious point at which the
effects of diffraction suddenly "set in."

> Like no matter how crappy a lens is, it gets worse
> beyond that point?


For any lens, no matter how crappy, there is some aperture that is so
small that diffraction becomes the overriding limitation on image
quality.  Most crappy lenses cannot stop down far enough to reach this
point, however.


From: "Q.G. de Bakker" qnu@worldonline.nl
Newsgroups: rec.photo.equipment.medium-format
Subject: Re: DOF "overrated"?
Date: Sun, 17 Mar 2002


Mxsmanic wrote:

> A "diffraction-limited" lens is one of such high optical performance that
> the only obstacle to perfection is the unavoidable fact of diffraction.

Not quite. A lens does not need to be perfect to be diffraction limited. All
that is required is that the effects of diffraction are worse than the
effects of uncorrected aberrations. Most if not all lenses are (or would be)
difraction limited, if stopped down far enough. It so happens that lens
manufacturers put a limit to how ar we can stop lenses down, and they, very
shrewdly, put that limit somewhere near the aperture at which the results
would start to be really bad. To make sense, the term "diffraction limited"
therefore should always be accompanied with an indication of the aperture at
which it is.


Stopping down is a contest between correcting aberrations and increasing
diffraction. In most lenses aberrations are best corrected when stopped down
two or three stops. Stopping the lens down further will only increase
diffraction.

There however are lenses in which aberrations do not improve when stopping
down (either because due tio the design there are very little aberrations at
maximum aperure already, or simply because stopping down will have no
effect, and aberrations remaining at the level they are). Stopping down will
thus only increase the effect of diffraction, limiting resolution. An
example of such a diffraction limited lens at maximum aperture is the
Hasselblad/Zeiss Sonnar f/5.6 250 mm.

> Diffraction is always present.  Its influence on the image becomes
> increasingly obvious as a lens is stopped down.  At some point, it
> limits resolution more than any defect in the lens, [...]

And at that point the lens becomes "diffraction limited".


From: "Leonard Evens" len@math.northwestern.edu
Newsgroups: rec.photo.equipment.medium-format
Subject: Re: DOF "overrated"?
Date: Mon, 01 Apr 2002

"Bogdan Karasek"
 wrote:

> Dilbertdroid2 wrote:
>>
>> <> diffraction effects at similar settings.  Diffraction is not a defect
>> in lens design, it is an inevitable consequence of physical laws
>> governing the behavior of light.<<
>>
>> Still copying things you don't understand out of books, I see.    What
>> a shame that you started out as a "disease" on the 35mm equipment board
>> and have turned into a plague on Leica and M.F. boards.  (and probably
>> a bunch of other ones I don't see.)   Just when do you have time to
>> take  pictures with all this newsgroup posting?
>
> Hi,
>
> You reply to the person who quoted above did not help me understand the
> diffraction phenomenon.  Rather than attacking the person, perhaps you
> could help me understand the quote. I am rather curious about this
> phenomenon.


What the quote says is that diffraction is a physical phenomenon which
would exist even if one had a perfect lens.   It is a result of the
wave nature of light.     What happens is that the image of a point
source is a central bright area surrounded by progressively fainter
rings.   That is called a diffraction pattern.   Usually, only the
central area is of significance.   It is brightest at the center and
drops off towards its edges.   There is a standard way of defining its
diameter.    This circular region is called the Airy disc after the
physicist who studied this phenomenon.

The diameter of the Airy disc is given by

C*l*N

where  l  is the wavelength of the light source,  N is the f-number of
the lens opening, and  C  is a proportionality factor depending on the
units.   Usually the diameter of the Airy disc is given for a specific
wavelength of green light since the eye is most sensitive in that range.
The important point is that it increases as the f-number increases.


As you stop down any reasonable photographic lens, aberrations due to
lens defects become less important, but because the f-number increases,
the diameter of the Airy disc increases.    The Airy disc would still be
there if there were no lens at all.    A circular pinhole will produce
a similar diffraction pattern.

I don't know who posted the original comment and why "Dilbertdroid2" is
so angry with him/her.   The anger may be based on other posts, but what
was quoted above is standard information which any serious photographer
should be aware of.   How important it is in practice is not entirely
clear since lens manufacturers usually limit the smallest aperture so as
to avoid significant diffraction effects.    That means they prevent you
from trying apertures which could seriously degrade your picture.  Even
so, when deciding which aperture to use, you should be aware that the
further you stop down, the more diffraction will limit the sharpness of
your picture.   Usually one finds that there is a best aperture for
sharpness where lens aberrations have been reduced significantly but
diffraction has not yet created any problems.    Just where this will be
will depend on what types of pictures you are taking and the quality of
your lenses.   The quality of the lens won't affect the Airy disc, but it
will determine how far you have to stop down to avoid defects in lens
design.

Because the formula for the diameter of the Airy disc is independent of
focal length, the more you have to enlarge the image, and hence the size
of the disc in the final image,  the more important
it becomes.    So typically large format lenses which aren't enlarged
very much can be stopped down to very small apertures while 35 mm lenses
typically are limited to f/22 or less.   Digital cameras with much smaller focal
lengths are diffraction limited at relatively large apertures because
such an image might be enlarged 40 times or more.

...

--
Leonard Evens      len@math.northwestern.edu      847-491-5537
Dept. of Mathematics, Northwestern Univ., Evanston, IL 60208



From: brianc1959@aol.com (brian)
Newsgroups: rec.photo.equipment.35mm
Subject: Re: Resolution: Digital vs Film
Date: 21 May 2002


"Q.G. de Bakker" qnu@worldonline.nl wrote 
> brian wrote:
>
> > [...]  Another way to think about this is
> > to consider angular resolution.  Even though the Airy disc size is
> > constant for a given f/# it subtends a larger angle when a smaller
> > focal length is used.  Thus, a longer lens will have more angular
> > resolution than a shorter lens when both are stopped down to the same
> > f/#.
>
> Yep. But that is because the actual size of the 'aperture', not the f/# is
> important for resolution. This is a misleading way of describing things. A
> longer lens will have a larger aperture size than a shorter lens at the same
> f/#. But you can forget about both focal lengths and focal length related
> descriptions of aperture size.
> When resolution is determined by diffraction (i.e. Airy disk size only)
> resolution can be described independent of focal length, as a simple
> relation between the wavelength of light used and the aperture diameter. No
> focal lengths, thus no f/#s are involved.

Not misleading at all.  Angular resolution (line pairs per degree)
depends only on the actual entrance pupil diameter (assuming
diffraction-limited performance and zero magnification).  However,
resolution expressed as a spatial frequency (line pairs per
millimeter) does depend on the f/#.  I imagine that most people in
this NG think in terms of spatial frequency, whereas most astronomers
think in terms of angular resolution.

For example, if you have a 10mm entrance pupil diameter a diffraction
limited 10mm focal length lens (f/1) will produce about 1700 line
pairs per millimeter at the image plane.  However, a 1000mm focal
length lens (f/100) will only produce 17 line pairs per millimeter.
Both lenses have exactly the same size opening, but the have very
different spatial resolution.  Angular resolution will be identical,
however.

Brian
www.caldwellphotographic.com




From: "Brian Ellis" bellis60@earthlink.net
Newsgroups: rec.photo.equipment.large-format
Subject: Re: Resolution vs Fstop
Date: Tue, 21 May 2002 


I make no claims to being an optics expert. However, for the last five years
I've been using the methodology outlined in the article "Image Sharpness and
Focusing the View Camera" that appeared in the March/April, 1996 issue of
Photo Techniques magazine to focus and determine the aperture setting for
any given photograph. FWIW, the conclusion in the message below, to the
effect that there is no visible difference in apertures of F 45 and faster
with 4x5 or larger film, is inconsistent with the conclusions reached by the
authors of this article. According to them, for any given situation there is
an optimum aperture. If an aperture smaller than, or greater than, the
optimum aperture is used, image sharpness will suffer to some extent
(according to the authors of this article). The article includes a simple
table that can be used to determine the optimum aperture. I don't have the
technical knowledge to know whether they are correct or not but I've been
using their methodology and obtaining results that are satisfactory to me.
However, I haven't tried to determine just how visible the use of apertures
smaller or larger than the optimum aperture might be in the final print.

"Roger N. Clark" rnclark@qwest.net wrote
> bpnp@sover.net wrote:
>
> > In a recent issue  B&W; magazine ran an feature on the outstanding work of
> > Clyde Butcher. In the B&W; interview with Mr. Butcher he is quoted as saying
> > " the more you stop down the more quality you lose because there less
> > information per millimeter as you stop down. If you could shoot at F16, then
> > you getting as much definition as you have available on the film"  Is he
> > referring to iris diffraction I wonder? I shot most of the time at F 32 & 45
> > with my 5x7, would I be better off at F 32 & 22?  90% of the time camera
> > movements take cae of any focus problems.
> > Regards
> > Nelson
> Here is my web page that lists information on diffraction
> size and diffraction limits.
>
> http://www.clarkvision.com/imagedetail/scandetail.html#diffraction
>
> In my opinion if movements can compensate for most of your
> depth of field issues, then stop down the minimum possible.
> However, I show on this same page that image detail on
> fine grained film (fujichrome velvia) at f/45 is still
> limited by film grain and not diffraction (although it
> is a pretty close match at this point).
>
> So, in practice at f/45 and faster and with 4x5
> and larger, I don't worry about it.
>
> Roger
> Photos, etc at:
> http://www.clarkvision.com



From: "Ralph W. Lambrecht" lambrecht@btinternet.com
Newsgroups: rec.photo.equipment.large-format
Subject: Re: Resolution vs Fstop
Date: Tue, 21 May 2002 


Nelson

Diffraction gradually limits lens resolution while stopping the lens down. Some
lens aberrations on the other hand, are being reduced while stopping down.
Consequently, there is an optimum lens aperture for every lens. This optimum
depends somewhat on the particular lens and its aberrations. In addition, the
magnification required and the acceptable circle of confusion considered
produce different optimum apertures for different film formats.

As a rule of thumb, optimum f/stops are as follows:

35mm     f/11-16
6x6         f/22
4x5         f/45

However, if depth of field has higher demands, don't hesitate to stop down
further.

Ralph W. Lambrecht

bpnp@sover.net wrote:

> In a recent issue  B&W; magazine ran an feature on the outstanding work of
> Clyde Butcher. In the B&W; interview with Mr. Butcher he is quoted as saying
> " the more you stop down the more quality you lose because there less
> information per millimeter as you stop down. If you could shoot at F16, then
> you getting as much definition as you have available on the film"  Is he
> referring to iris diffraction I wonder? I shot most of the time at F 32 & 45
> with my 5x7, would I be better off at F 32 & 22?  90% of the time camera
> movements take cae of any focus problems.
> Regards
> Nelson




From: brianc1959@aol.com (brian)
Newsgroups: rec.photo.equipment.35mm
Subject: Re: Resolution: Digital vs Film
Date: 21 May 2002 


> So confusing that i need to tidy that up that a bit... ;-)
>
> It was meant to say:
> But you went on juggling f/#s and focal lengths, to show that when the
> aperture (expressed as a ratio of the focal length) can (by the right choice
> of focal lengths) be shown to be the same size in two lenses (having
> different focal lengths and different f/#s), resolution will be the same.
>
> The same can be expressed by simply saying that only the aperture's size
> matter. All the stuff i put inside parentheses above only creates the
> impression that focal lengths and f/#s do indeed play some obscure role.
>
> I know, i know: you did show that two lenses having the same f/# but
> different focal lengths have different aperture sizes and different angular
> resolution.
> Same thing. Delete all reference to f/#s and focal lengths and it will still
> be true.


I doubt that we disagree on anything fundamental here. The f/# and
wavelength completely determine the maximum spatial frequency that can
be detected at the image plane of a diffraction-limited lens.  Armed
only with a knowledge of the aperture size (entrance pupil) you can
calculate the angular resolution but you cannot calculate the spatial
frequency.  Spatial frequency matters in most kinds of photography.
Astronomers and spies may care only about angular resolution because
that is what determines the smallest detail you can image.  However,
most photographers will care about spatial resolution because it is a
key ingredient of sharpness.  For example, everyone in this NG would
be floored by the image quality of a fully diffraction-limited 15mm
f/4 wide-angle lens for 35mm format.  On the other hand, a 1500mm
f/400 lens will produce an image on 35mm format that is no better than
a pinhole image in terms of sharpness.  Yet both lenses have the same
aperture size and angular resolution.

Brian


From: bachchaconne@my-deja.com
Newsgroups: rec.photo.equipment.35mm
Subject: Re: Superspeed lenses and "window effect".
Date: Thu, 23 May 2002


>bachchaconne@my-deja.com wrote 
>http://www.nikon.co.jp/main/eng/society/nikkor/n02_e.htm

>Andrew:
>Regarding the diffraction rays, they emanate in a direction
>perpendicular to the *flats* of the polygon, not the tips.  If you
>have an even-sided polygon this means that there are really two sets
>of rays that overlap and reinforce each other.  Thus, odd-sided
>polygons produce twice as many rays as there are sides, while
>even-sided polygons produce the same number of rays as there are
>sides.  Triangular and hexagonal aperture stops both produce 6 rays.
>A pentagonal aperture will produce 10 rays.

Brian: the site mentions more than once 'six rays' from the hexagonal
iris and that's what I see on the sample pic on the site, another from
a magazine, and my own Yashica 28mm.  Also it says 'Later Nikkor
lenses were designed with an odd number of diaphragm blades in order
to eliminate these "rays"...'

The angle of these rays do not correspond to your theory either.  If
each ray are really 2 that overlap and reinforce, they should 'spread
out' a lot more, and we would be able to see the 'orthogonality to the
side of the polygon' or at least something close.  I haven't
encountered this orthogonality in these situations.  Can you provide
any sample pic?

Andrew



From: brianc1959@aol.com (brian)
Newsgroups: rec.photo.equipment.35mm
Subject: Re: Superspeed lenses and "window effect".
Date: 24 May 2002 

...(quote above posting)

Andrew:
If two roads cross at right angles, then one person might say that
there are two roads, but another person might say that there are 4
roads emanating from the point of intersection.

Each flat side of a polygon aperture produces a single ray that goes
through the center of the image point.  You can also think of it as
two rays 180 degrees apart, having an intersection at the center.

I said earlier that a hexagonal aperture will produce 6 rays, although
you can also think of it as 12 rays in which half of them overlap.
This is why a pentagonal aperture can produce more rays (10) and a
triangular aperture will produce the same number (6).

If you look carefully at the Christmas tree picture at the link you
provide you will notice that two of the rays point along the
horizontal direction, and the others are 60 degrees from horizontal.
There are no vertical rays.  I own a sample of the 50mm f/2 that is
discussed in the article.  Its hexagonal iris is oriented like this:


          x
        x   x
        x   x
          x

In words, the iris has two vertical flat sides oriented in the
vertical direction when the camera is positioned in landscape
orientation.  This aperture produces two pairs of rays in the
horizontal direction, and they overlap.  The rays are perpendicular to
the flat sides.  If the rays emanated from the points of the aperture
then there would be vertical rays but no horizontal ones.  So, the
angle of the rays really does correspond to what I stated earlier.

By the way, its not my theory.  Its just basic diffraction theory that
started with a Dutch guy named Huygens, and it is more than 300 years
old.


Also, the article is completely incorrect when it suggests that going
to an odd-numbered polygon aperture will eliminate the rays.  All this
does is increase the number of rays and reduce their intensity.  Only
a circular aperture will be rayless.  The article is also incorrect in
suggesting that an odd-sided aperture will result in higher
resolution.  The article is interesting and informative, but you
shouldn't take every word too literally because of all the errors.

Brian
www.caldwellphotographic.com


From hasselblad mailing list:
Date: Tue, 23 Jul 2002 
From: Henry Posner/B&H; Photo-Video henryp@bhphotovideo.com
Subject: [HUG] Re: Auto Focus or Focus Confirmation

you wrote:
>after speaking with a friend who's
>a HB/Nikon kind of guy, he said my shooting it at f22 might've not been
>optimal for the lense!?! I realize most lenses have a sweet spot, but I'm
>not sure I buy his explanation.

Stopping down too far is bad for optimum image quality. You're asking the
light to bend a great deal, and bending = distortion.

--
regards,
Henry Posner
Director of Sales and Training


From hasselblad mailing list:
Date: Wed, 24 Jul 2002 
From: "Q.G. de Bakker" qnu@worldonline.nl
Subject: Re: [HUG] Re: Auto Focus or Focus Confirmation

Mark Rabiner wrote:

> And "Diffraction" which i hear lowers contrast and softens things up.
> I have however used the smallest F stop on my Hasselblad lenses for 25
> years before someone gave me the idea that maybe it's not such a great
> idea.
>
> My impression is that with what we have here: Zeiss Hasselblad Medium
> format lenes: we can stop down not just a lot but all the way with no
> visible detriment to the image; a quality image. I've got a whole body
> of work full of it
>
> I think this stopping down rule applies mostly to 35mm photography
> because large format photography I know people stop down with great
> voraciousness and great compulsion.
> Those lenses are apparently designed for it and so i think is the
> Hasselblad glass.

Nope.

You can't "design away" diffraction. It's an inescapable law of nature.
Maximum achievable approximately resolution halves with every two stops the
aperture is closed. And the efects are even more visible when the
lens involved is a high quality one.

At first (with most lenses), stopping down will increase performance due to
elimination of residual aberrations (though without fail with every stop
diffraction increases and limits maximum achievable resolution further).

When the detrimental effects of aberrations are reduced below those of
diffraction (which happens at the fabled "sweet spot"), any further stopping
down will brutally reduce resolution. The only gain is depth of field
(which, as we all know, is no more than "acceptable" unsharpness. "Gain"?).

The criteria set out DoF and the effects of diffraction however are at odds:
stopping down a lens all the way may well result in lower maximum resolution
than is demanded by DoF criteria. Everything will look about equally "sharp"
alright, but really sharp it isn't.

So if you can help it, don't stop down further then necessary.



From hasselblad mailing list:
Date: Wed, 24 Jul 2002 
From: Ken Martin kmartin@ventur.net
Subject: Re: [HUG] Re: Auto Focus or Focus Confirmation

Mark:

   My experience is would tend to validate your thoughts in that some lens
do better at small F/stops than others.  Over the years I have taken photos
with my Canon macro lens at F 32.  I have a series butterfly photographs
that are enlarged to 11"x14".  They were taken at F 32 with my Canon 100MM
micro.  People who view them almost always remark how sharp they are.  I
have a couple of after market lens that it is impossible to get critically
sharp photos  at other than mid F stops.  I just got back a couple of
16"X16" enlargements from the lab that were taken with my Hasselblad/Zeiss
250MM at F45.  These are some scenics that I recently took around Yosemite
and you can count the pine needles in the trees.  Acceptably sharp? most
people would think so.  Technically as sharp as the lens is capable of
producing at a larger F/stop, probably not.  Some of us at my age remember
hearing or reading about Ansel Adams and some of his friends that formed the
"F 64" club.  Those photographers were in the pursuit of sharp pictures
using very small F/stops.  I have viewed some of those photographs and  I
would consider them successful.

   In nature and outdoor photography we cannot control the conditions as we
can in the studio.  Factors such as atmospheric conditions, angle of light,
required DOF all can push lens to the limits and its ability to produce
sharp images.  It is an extremely rare nature photograph that only one
defined plane needs to be in focus.  In fact, that is true with almost all
photography, be it in the studio or in the great outdoors.  It is always a
question of how large the circle of confusion and degree of distortion we
can consider as "acceptably sharp".  With current lens technology and the
eyeballs that the creator has given us we will never achieve perfect
sharpness.

Ken Martin
...


From hasselblad mailing list:
Date: Wed, 24 Jul 2002 
From: Anthony Atkielski anthony@atkielski.com
Subject: Re: [HUG] Re: Auto Focus or Focus Confirmation

Henry writes:

> Stopping down too far is bad for optimum
> image quality. You're asking the light to
> bend a great deal, and bending = distortion.

It doesn't bend any more at small apertures than it does at large apertures.

The real enemy is diffraction.  The smaller the aperture, the more the light
rays diffract, producing a series of concentric circles for each point of
light, instead of a single point.  There is no way to avoid this, as it is a
fundamental physical limitation of light.  If it didn't exist, smaller
apertures would always produce the sharpest images.  But since it exists,
there is some point between the largest and smallest aperture of a lens that
usually produces the best image (stopping down reduces lens aberrations, but
at some point diffraction gets so bad that it negates the advantage of
reduced aberrations).

Most lenses are designed such that the smallest available aperture is the
smallest that can reasonably be used on that lens without causing too much
loss of sharpness due to diffraction.


From hasselblad mailing list:
Date: Wed, 24 Jul 2002 
From: Anthony Atkielski anthony@atkielski.com
Subject: Re: [HUG] Re: Auto Focus or Focus Confirmation

Mark writes:

> I think this stopping down rule applies mostly
> to 35mm photography because large format photography
> I know people stop down with great voraciousness
> and great compulsion.

The larger the format, the greater the focal length required.  And the
greater the focal length, the greater the actual size of the aperture for a
given f-stop.  And since diffraction depends on the actual physical size of
the aperture, this means that smaller f-stops can be used on larger formats
without excessive loss from diffraction.

> Those lenses are apparently designed for it and
> so i think is the Hasselblad glass.

There isn't any way to design around diffraction losses.  It's an absolute
limit.


From hasselblad mailing list:
Date: Wed, 24 Jul 2002 
From: Anthony Atkielski anthony@atkielski.com
Subject: Re: [HUG] Re: Auto Focus or Focus Confirmation - stopping down

Stuart writes:

> I think for any lens and any format the
> sharpest focus is always in the middle range.

For an ideal lens, the sharpest focus is always wide open, because that's
the aperture at which diffraction is least apparent.  However, since
real-world lenses usually show too much aberration wide open, the best
f-stop for a real-world lens is usually somewhere between wide open and
completely stopped down.  A few lenses are truly diffraction-limited at
certain f-stops (meaning that the loss from diffraction is greater than the
loss from aberration); a lens that is diffraction-limited wide open will
give the sharpest images when it is left wide open.


From hasselblad mailing list:
Date: Wed, 24 Jul 2002 
From: Anthony Atkielski anthony@atkielski.com
Subject: Re: [HUG] Re: Auto Focus or Focus Confirmation

Edo writes:

> The effects of diffraction aren't related to
> focal length.

They are a function of aperture size--but aperture size, when it is
expressed in terms of f-stops, _is_ a function of focal length, and so
diffraction in relation to f-stop is a function of focal length as well.

> At physically larger apertures, rays will bend
> at smaller angles ...

The light rays always bend at the same angles, irrespective of aperture.
Diffraction is not caused by the bending of light rays; it is a consequence
of quantum electrodynamics.



From hasselblad mailing list:
Date: Wed, 24 Jul 2002 
From: "Nelson L. Mark, SC001" phair1@jklsoftware.com
Subject: [HUG] RE: Stopping Down

According To Wildi in the newest edition, he states on page 207-208 that:

"At small apertures, the diffraction may affect the image sharpness. This
is true. But it is also true that manufacturers of high-quality lenses
usually limit the minimum aperture of their lenses to a point where the
quality loss is not objectionalble or even noticable. This explains why
some of the Zeiss lenses may only stop down to f/22 or even f/16. As a
results, the Hasselblad photographer need not be concerned about closing
the aperture too far. Stop down completely if the depth of field requires it."

I think f/64 and be there just about covers it... :)

I'm heading to Lost Wages on Thursday...hmmm, I'll have to try shots of the
top of the Paris hotel at f/22 on Velvia and see...


From hasselblad mailing list:
Date: Wed, 24 Jul 2002 
From: "John L. Odom" john@odomsite.com
Subject: [HUG] Diffraction, Focusing

Hi All:

The diffraction problem is a function of the lens' effective diameter,
not the f-stop per se.

For a given f-stop the diameter of a short focal length lens is much
smaller than for a longer lens.
because of this diffraction effects are much more severe with 35 mm than
with MF or LF. f22 on an 8 X 10 setup is probably larger in diameter
than the f8 on a 35 mm. diffraction also affects certain kinds of images
more than others.

In photomicrography we often use clear glass focusing screens, with a
set if scribed lines on the surface. The focusing magnifier is then
focussed on the line, and then the image brought to the same focus. This
permits seeing finer detail in the image than is possible with a ground
glass.

Although a newby to HB I am a microscopist and optical instrument
builder/modifier.

John Odom


from hasselblad mailing list:
Date: Wed, 24 Jul 2002 
From: Jim Brick jbrick@elesys.net
Subject: Re: [HUG] RE: Stopping Down

My photography colleague is a fine art photographer. She uses a Pentax
67II. Her project over the past year was/is Silicon Valley's urban dilemma.
She uses the SMC 55mm lens and everything she photographs is at f/22.
Maximum aperture. Lots of urbanscapes.

She just had a show in San Francisco, somewhere around 30 photographs, 75%
of them 30x40's, printed by her in a rental darkroom with who knows what
kind of enlarger and lens. But they had a 50" Kreonite RA4 processor which
is why she went there.

Anyway, all of her photographs were "dead sharp". Nothing gave a hint of
being degraded. And the show was in tight quarters so it was easy to view
the 30x40's up close and personal.

I suspect that manufacturers do limit the max aperture to a value that
won't show noticeable degradation. I also suspect that the small apertures
on Zeiss lenses can be used almost with impunity.

Jim

 Nelson L. Mark, SC001 wrote:
>According To Wildi in the newest edition, he states on page 207-208 that:
>
>"At small apertures, the diffraction may affect the image sharpness. This
>is true. But it is also true that manufacturers of high-quality lenses
>usually limit the minimum aperture of their lenses to a point where the
>quality loss is not objectionalble or even noticable. This explains why
>some of the Zeiss lenses may only stop down to f/22 or even f/16. As a
>results, the Hasselblad photographer need not be concerned about closing
>the aperture too far. Stop down completely if the depth of field requires it."
>
>I think f/64 and be there just about covers it... :)
>
>I'm heading to Lost Wages on Thursday...hmmm, I'll have to try shots of
>the top of the Paris hotel at f/22 on Velvia and see...



from hasselblad mailing list:
Date: Wed, 24 Jul 2002 
From: "Q.G. de Bakker" qnu@worldonline.nl
Subject: Re: [HUG] RE: Stopping Down

Nelson L. Mark, SC001 wrote:

> According To Wildi in the newest edition, he states on page 207-208 that:
>
> "At small apertures, the diffraction may affect the image sharpness. This
> is true. But it is also true that manufacturers of high-quality lenses
> usually limit the minimum aperture of their lenses to a point where the
> quality loss is not objectionalble or even noticable. This explains why
> some of the Zeiss lenses may only stop down to f/22 or even f/16. As a
> results, the Hasselblad photographer need not be concerned about closing
> the aperture too far. Stop down completely if the depth of field requires it."

I think that the number of stops a Zeiss/Hasselblad lens can be stopped down
is solely and wholy dictated by the mount they are put in.
Why, for instance, would a 110 mm lens not stop down further than f/16, and
a 250 mm lens go all the way down to f/45? That is, one of the two available
250 mm lenses goes all the way to f/45. The other one, starting at one stop
wider also ends one stop wider.

So it's a nice story Wildi tells, but personally, i don't see much value in
it.



from hasselblad mailing list:
Date: Wed, 24 Jul 2002 
From: Henry Posner/B&H; Photo-Video henryp@bhphotovideo.com
Subject: [HUG] Re: Auto Focus or Focus Confirmation

 you wrote:
>I always read about Ansel shooting almost everything at f22 and above - but
>he shot B&W; with a much bigger camera and obviously, I'm no Ansel!!

The old rule of thumb was that lenses operate best when stopped down 2 (or
perhaps 3) stops from wide open. if f/22 is the min aperture on your lens,
diffraction is all but inevirtable. When Ansel shot with a big ol' 8x10,
f/22 was right in the middle, and f/96 or f/128 was REALLY stopping down. :-)

--
regards,
Henry Posner
Director of Sales and Training
B&H; Photo-Video, and Pro-Audio Inc.
http://www.bhphotovideo.com


From: "Richard Knoppow" dickburk@ix.netcom.com
Newsgroups: rec.photo.equipment.large-format
Subject: Re: "best" f-stop question
Date: Fri, 27 Sep 2002

"Tourtelot" tourtelot1@attbi.com wrote
> I know!  "The only dumb question. . . ."
>
> I just started shooting LF (and BTW, my first 4x5 chromes came back
> perfectly exposed.  I don't know what I did wrong.) but I think I saw
> something on the web about an optical fault that occurs when the f-stop is
> set way down at the bottom of the range.  I can't find that information
> anywhere now, but to have a camera on sticks and f64 available and not be
> able to use it; what a shame that would be.  But I am thinking that
> different lenses have different "optimal" apertures.  Is it related to brand
> and/or formula, or simply focal length?  Where is this written down?
>
> Now to track down a light leak in a holder.
>
> BTW, what are the thoughts on Fuji Astia?  I have been shooting Provia but I
> came across some cheap Astia.
>
> Thanks.
>
> D.

  There are several factors which change with the lens stop.
The first is that some of the lens faults, or aberrations,
get smaller as the lens is stopped down making the image
sharper. Also, when most lenses are wide open the corners of
the image are partially shadowed, or vignetted, by the edges
of the lens mount. As the lens is stopped down a point is
reached where the aperture is no longer vignetted and the
illumination is more uniform.

  On the other side is diffraction, which is inversly
proportional to the size of the aperture. The edges of the
aperture spread the light, that's why a pin-hole acts as a
lens. As the aperture becomes smaller the light is spread
out more resulting in a bluring of the image. So, if we plot
image sharpness against aperture we will find that there is
some aperture where its best. The aberrations are reduced
but the diffraction is not bluring the image too much.

  The stop which is the optimum depends on how well the lens
is corrected _and_ the image angle. As we move further from
the center of the image (and optical axis) the image quality
in general becomes poorer. The optimum stop becomes smaller
off axis since some aberrations become larger.

  As a rule the largest stop should be the one at which the
lens mount no longer obscures part of the iris diaphragm.
Obviously, this stop becomes smaller the further off axis we
get.

  For lenses of "normal" focal length, i.e., FL about equal
to the diagonal of the format the optimum aperture will be
two to three stops below maximum. Most lenses must be
stopped down two stops to eliminate vignetting by the mount.

  As far as aberrations are concerned there is no real rule.
A few lenses will be sharpest near the center of the image
when wide open but even these will usually improve off axis
when stopped down.

  The diffraction effect depends on how much the negative
image is to be enlarged. For 35mm cameras the smallest stop
should probably not be more than f/11. Many lenses for 35mm
cameras do not stop down beyond f/16 for this reason. Medium
format cameras can tolerate f/22. 8x10 cameras can be used
at f/64 although careful observation will see diffraction
blurr
---
Richard Knoppow
Los Angeles, CA, USA
dickburk@ix.netcom.com


From: "Bob May" bobmay@nethere.com
Newsgroups: sci.optics
Subject: Re: airy disks
Date: Fri, 22 Nov 2002

This program calculates the diffreaction pattern for a pair of stars.  It
does run in DOS tho.
http://bobmay.astronomy.net/difdraf.zip  for the program.  It is freeware
from some long time ago.

--
Bob May


From: Nichols@OpticsNotes.Com (OpticsNotes)
Newsgroups: sci.optics
Subject: Re: airy disks
Date: 22 Nov 2002

Bob,
I took a quick look at opticsnotes.com and didn't find an airy disk
java applet, but you might find something helpful, if you browse
around yourself on the site. There are a lot of optics and physical
optics books that will probably take care of this inquiry, too.

Best of luck,

Bruce
http://www.opticsnotes.com
Optics and Photonics Tutorials, References and Resources

"Bob McKenzie" philip.smith@sympatico.ca wrote
> Hi,
> Does anyone know of any websites that have java applets or diagrams showing
> the overlapping diffraction patterns of light directed into 2 circular
> apertures on an opaque?
>
> or something even remotely similar?
> or an explaination of whats going on?
>
> basically its described above, and the centre to centre separation of the
> diffraction patterns is the same as the centre to centrer separation of the
> holes.   I want to determine what the separation of the two holes are when
> the airy disks observed on a screen overlap by one half of their radius..  I
> would just like to understand whats going on, im finding it hard to picture
> thanks
>
> bob


From: Diffractive.Optics@free.fr (Diffractive Optics)
Newsgroups: sci.optics
Subject: Free Diffractive Optics CAD Software
Date: 11 Jul 2003

Free software that handles all steps of Diffractive Optics design is
available at:

http://diffractive.optics.free.fr

Synthesis, simulation of performances, tolerancing and fabrication
files generation are fully supported.


From: brianc1959@aol.com (brian) 
Newsgroups: rec.photo.equipment.large-format 
Subject: Re: Diffraction again - frequency 
Date: 14 Sep 2003 

mikescarpitti@yahoo.com (Michael Scarpitti) wrote 
> > But about diffraction and color - I realize that diffraction degrades 
> > sharpness, but exactly what effect does color (frequency) have? 
> 
> Not relevant. The light is scattered at the edge of a surface. 
> Wavelength differences within visible spectrum irrelevant. 

Francis was exactly correct: diffraction is directly proportional to 
the wavelength. The visible spectrum extends from about 400nm 
(violet) to 700nm (deep red). So, for a diffraction-limited system 
you get about variation in linear resolution of about 1.75 depending 
on the color. The total information content would vary by as much as 
3x. In other words, an image taken in the deep visible violet can 
have 3 times as much information as an image taken in deep visible red 
light. This is a very significant variation IMO, similar to the 
change in image quality you would get in going from 35mm to 645 
format. 

Brian 
www.caldwellphotographic.com 


From: hemi4268@aol.com (Hemi4268) 
Newsgroups: rec.photo.equipment.large-format 
Date: 15 Sep 2003 
Subject: Re: Diffraction again - frequency 

 
Hi 

I always like to add a simple table of resolution and depth of focus for each f 
stop. This would be typicl reading in noon summer sun at about 10000 foot 
candles in brightness. 

F stop Lines per MM Depth of Focus MM 
f-1 2000 .0001 
f-2 1000 .0004 
f-4 500 .0016 
f-8 250 .0064 
f16 125 .0256 
f-32 64 .1028 
f-64 32 .4112 
f-128 16 1.6400 
f-256 8 6.5000 

Larry 	 


From: hemi4268@aol.com (Hemi4268) 
Newsgroups: rec.photo.equipment.large-format 
Date: 18 Sep 2003 
Subject: Re: Diffraction again - frequency 

> The 
>> diffraction-limited depth of focus is approximately the f/# squared, 
>> measured in microns. 

The true diffraction-limited depth of focus is ((f-stop 
squared)X(wavelength)x2). In most cases where peak wavelength is .5 microns, 
the calculation can be reduced to f-stop squared. 

Larry 


From: hemi4268@aol.com (Hemi4268) 
Newsgroups: rec.photo.equipment.large-format 
Date: 15 Sep 2003 
Subject: Re: Diffraction again - frequency 

>Stroebel's book probably uses an approximation to the airy disk formula, 
>that drops the 1.22 factor in favor of a 1 factor 
For ease of thinking, factors can be dropped. Thatis why I always use the 
simple table in telling how any lens might perform 
In noon summer sun with a peak wavelength of 500 nm, you can expect any 
perfect lens to do the following. 

f-1 2000 lp/mm 
f-2 1000 
f-4 500 
f-8 250 
f-16 125 
f-32 64 
f-64 32 
f-128 16 
f-256 8 
f-512 4 
f-1028 2 

A 500 micron pinhole with 2 lines per millimeter resolution sitting 500000 
microns or 19 inches from the film, produces it's system peak performance. 

Larry


From: "G.P" askme@my.email.pls 
Newsgroups: rec.photo.equipment.large-format 
Subject: Re: Diffraction again - frequency 
Date: Mon, 15 Sep 2003

"Francis A. Miniter" miniter@attglobal.net wrote 
> The size of an Airy Disk (the measure of diffraction) is directly 
> proportional to the wave length of light. I have seen the following 
> formula in Stroebel et al "Basic Photographic Materials and Processes" 
> (1990), p. 169: 
> 
> R = 10^6 / [wl * f-N], 
>
> where R is resolving power in lines per inch 
> wl is wavelength of light 
> f-N if f-stop.  
> 
> Curiously, Ctein gives almost the same formula, except that he uses a 
> constant of 820,000. I have tried in vain to find the source of these 
> alleged constants. Perhaps, someone here who has more experience in 
> optics can explain. 

Are you sure R is in lines per inch? As I recall, the diameter of Airy 
disk is given by 1.22 x wl * f-N , if you put 2 of those airy disks side by 
side, with its centers separated half the diameter of the airy disk, they 
could barely be resolved as 2 separate image points, the reciprocal of that 
minimum distance will give you the resolution in line pair per unit of 
distance (whatever unit of distance "wl" is expressed in). If you express 
"wl" in nanometers, you then have: 

R = 1 / [1.22 * wl * f-N] line pairs per nanometers (wl in nanometers) 
line pairs per nanometers does not really make sense, so lets multiply by 
10^6 to have the answer in line pair per millimeters and still keep "wl" 
expressed in nanometers. 

R = 10^6 / [1.22 * wl * f-N] line pairs per millimeter (wl in nanometers) 
But 10^6 / 1.22 = 819,672 that we could approximate to 820,000 

so we have: 

R = 820,000 / [wl * f-N] lp/mm (wl still in nanometers) 

Stroebel's book probably uses an approximation to the airy disk formula, 
that drops the 1.22 factor in favor of a 1 factor, as a results the formula 
has 10^6 instead of 820,000, but still the resolution has to be in lpp/mm 
and not lpp/inch. 

Corrections welcomed, 
Guillermo 


From: hemi4268@aol.com (Hemi4268) 
Newsgroups: rec.photo.equipment.large-format 
Date: 16 Sep 2003 
Subject: Re: Diffraction again - frequency 

> If you use the Rayleigh criterion, the resolution in 
>line pairs per millimeter is given by 1/(Airy disc radius) = 4.03 
>lp/mm for the 8" distance and 1.67 lp/mm for the 19" distance. 

It seems to me that a 500 micron hole would produce a 2 lp/mm image from zero 
focal length to about 19 inch focal length using 500nm light. 

Actually, the needed f-number would be 1028 for 2 lp/mm. Times that by 500 
microns and it's 514000 micron distance from the pinhole to the film . Divide 
that by 1000 and it's 514mm or 20.23622047 inches. 

So I guess the real distnaces examples actually should be 10.11811024 inches 
for f-256 and 20.23622047 inches for f-1028. 

I like ABOUT 16 inch focal length at ABOUT f-512 using a $650 gold plate 500 
micron scientific pinhole to give me overall excellent results. 

A negative images is placed on 10x16 Kodak Supra paper. Exposure is 45 seconds 
at f-512 processed in a RA 4 roller transport processer. Color negative print 
is then scanned, reversed, color corrected in photoshop and then printed out on 
8x12 dyesub paper using a Kodak 8650 printer.. 

Larry


From: hemi4268@aol.com (Hemi4268) 
Newsgroups: rec.photo.equipment.large-format 
Date: 18 Sep 2003 
Subject: Re: Diffraction again - frequency 

> In the case of your 0.5mm pinhole the 
>optimum "focal length" is about 8". At 19" the diffraction spot is 
>substantially larger (for green light) than the 0.5mm. 

And this is what I disagree with. that 0.5mm pinhole is good to about 20 inches 
of 1/2 meter. Somehow your or my calculations are off. All I know is, that 
0.5mm pinhole at 16 inches at f-512 gives excellent results biving about 2 
lp/mm resolution on 10x16 paper. After the color negative image is reversed, 
color corrected, reduced and printed out on Dyesub 8x12 paper the results are 
more then excellent. 

Larry 


From: brianc1959@aol.com (brian) 
Newsgroups: rec.photo.equipment.large-format 
Subject: Re: Diffraction again - frequency 
Date: 18 Sep 2003 

hemi4268@aol.com (Hemi4268) wrote
> > but the point here is that resolution of a pinhole 
> >image does not stay constant as you change the image plane distance. 
> 
> But I say yes it does. The resolution of a pinhole image is always the size of 
> the pinhole regardless of the focal length up to the diffraction limit of that 
> pinhole. 
> 
> Larry 

Larry: 

Its true that the resolution of a pinhole image is approximately 
constant up until the point where the Airy disc is about the same size 
as the pinhole. If you move the image plane further from this point 
then the resolution will drop. In the case of your 0.5mm pinhole the 
optimum "focal length" is about 8". At 19" the diffraction spot is 
substantially larger (for green light) than the 0.5mm. 

Brian 
www.caldwellphotographic.com




From: hemi4268@aol.com (Hemi4268) 
Newsgroups: rec.photo.equipment.large-format 
Date: 16 Sep 2003 
Subject: Re: Diffraction again - frequency 

> but the point here is that resolution of a pinhole 
>image does not stay constant as you change the image plane distance. 

But I say yes it does. The resolution of a pinhole image is always the size of 
the pinhole regardless of the focal length up to the diffraction limit of that 
pinhole. 

Larry


From: brianc1959@aol.com (brian) 
Newsgroups: rec.photo.equipment.large-format 
Subject: Re: Diffraction again - frequency 
Date: 16 Sep 2003 

hemi4268@aol.com (Hemi4268) wrote
> >The optimum "focal length" of a pinhole occurs when the diameter of 
> >the Airy disc is about the same size as the pinhole itself. 
> 
> To give an example this happens when a 500 micron pinhole is about 500000 
> microns or about 19 inches from the film. This give a focal length of 19 inches 
> at about 2 lp/mm resolution. 
> 
> Sure you could use this same size pinhole at say 8 inches but the resulting 
> image would be lower in scale since the focal length would now be now 8 inches. 
> Resolution would still be the same 2 lp/mm. 
> 
> Larry 

Larry: 

You've completely ignored diffraction, which will make a very big 
difference when you move from 8" (203mm) out to 19" (483mm) from the 
image plane. In both cases the far-field approximation will be valid, 
so the simpler Fraunhofer diffraction model will give accurate 
results. Accordingly, the Airy disk radius is given by: 

r = (1.22*R*wavelength)/(pinhole diameter), 

where r is the Airy disk radius, and R is the distance from the 
pinhole to the image. Assuming the wavelength is 0.0005mm 
(bluish-green light), the Airy disk radii for your two pinhole 
placments will be 0.248mm and 0.589mm for 8" and 19" distances, 
respectively. If you use the Rayleigh criterion, the resolution in 
line pairs per millimeter is given by 1/(Airy disc radius) = 4.03 
lp/mm for the 8" distance and 1.67 lp/mm for the 19" distance. You 
under-calculated resolution by a factor of two (unless you assumed 
near-infrared illumination for the 8" case and deep blue illumination 
for the 19" case), but the point here is that resolution of a pinhole 
image does not stay constant as you change the image plane distance. 

Brian 
www.caldwellphotographic.com



From: brianc1959@aol.com (brian) 
Newsgroups: rec.photo.equipment.large-format 
Subject: Re: Diffraction again - frequency 
Date: 15 Sep 2003

"Richard Knoppow" dickburk@ix.netcom.com wrote
> The effect of wavelength on a pinhole image will be the 
> size of the image since a pinhole has no definite focal 
> length. The red image will be larger than the blue image, in 
> other words, it has lateral chromatic aberration. The effect 
> of a very small stop in a lens may be somewhat different 
> since lens magnification is involved. 
> Another part of the original question had to do with the 
> location of the filter. For a lens one wants the filter at a 
> point of minimum vergence of the light. Converging or 
> diverging rays throught a plane parallel surface produce 
> chromatic aberration, spherical aberration, astigmatism, and 
> I think even some other aberrations. The thinner the plate 
> and the closer its index to that of the surrounding medium, 
> the better. For normal use in air thin gelatin filters have 
> the least effect. Since the light from very distant objects 
> entering a lens is very nearly collimated (plane waves) the 
> aberrations introduced by the filter are negligible. 

Hi Richard: 

The optimum "focal length" of a pinhole occurs when the diameter of 
the Airy disc is about the same size as the pinhole itself. This 
length does change with wavelength because diffraction is proportional 
to wavelength. So when you use a pinhole in white light the only 
downside is that the pinhole placement is slightly non-optimum for the 
short and long wavelengths. There is no lateral chromatic aberration 
because the chief ray can't change direction as it passes through the 
pinhole. If you use a lens with a tiny stop then you can get lateral 
color and distortion since the chief ray is refracted. 

The aberrations introduced by placing a plane parallel plate into a 
converging f/180 beam are entirely negligible unless the plate is more 
than about 4 inches thick. The main aberration introduced by a fat 
plate in a beam this slow is lateral color. If a narrow line color 
filter is used, then other aberrations like astigmatism don't get 
serious until the plate is about a meter thick! On the other hand, if 
the converging beam is fast, say f/1, then a parallel plate even 0.1mm 
thick can significantly degrade the image quality, although here the 
degradation is due to spherical aberration and coma with almost no 
chromatic aberrations at all. 

Brian 
www.caldwellphotographic.com



From: hemi4268@aol.com (Hemi4268) 
Newsgroups: rec.photo.equipment.large-format 
Date: 15 Sep 2003
Subject: Re: Diffraction again - frequency 

>since pinhole do not focus light 

Actually, the resolution or focus is the size of the pinhole. Example, a 500 
micron pinhole produces 500 microns of resolution or 2 lines per millmeter at 
all distances from zero focal length to about 16 inches. At 16 inches 
diffraction takes over and the image falls apart. 

So a 500 micron pinhole set at 16 inches is twice the focal length as that same 
pinhole set at 8 inches. Although, both images will be about 2 lines per 
millmeter. As a result, both images will have the same sharpness but one will 
be twice the scale. 

Larry


From: hemi4268@aol.com (Hemi4268) 
Newsgroups: rec.photo.equipment.large-format 
Date: 20 Sep 2003
Subject: Re: Diffraction again - frequency 

>The notion that the difference in diffraction between blue light and 
>red light at f/180 on a 180mm lens and a 4x5 negative is observable is 
>totally absurd.
 
The calculation is 180 squared equals 32400x.400nm blue light = 12960x2=25920nm 
or 25.920mm. 

Now 180 squared for 32400x.500 greensih light = 16200x2=32400nm or 32.400mm. 
32.400-25.920=6.48mm difference depth of focus. 

Larry


From: brianc1959@aol.com (brian) 
Newsgroups: rec.photo.equipment.large-format 
Subject: Re: Diffraction again - frequency 
Date: 18 Sep 2003

john@stafford.net (J Stafford) wrote 
> 
> I claimed nothing. I posed a query to Net Wisdom, and it is clear from the 
> excellent responses that some experiments are warranted. And I wrote that 
> I was using a deep green filter because green gave me the colors that 
> worked best with the subject. However, shooting toward the blue/uv 
> frequency range seems promising for another subject. 
> 
> But at this point I'm busy enough with the Day Job, and have found the 
> F180 aperture on the wide 5x4 to be a Bad Thing so it is back to F8 and 
> modest filtration - business as usual. 
> 
> An aside - to the Net - shooting into the UV requires a UV ('quartz') lens 
> (and extended UV film of course). No? I've only seen one such lens in my 
> life - the one made for the Hasselblad, and apparently they have become 
> outrageously expensive. 

Shooting toward the blue/UV will definitely help sharpness at f/180. 
The optical correction will be fine (still *very* 
diffraction-limited!), but transmission losses might start to be a 
problem. Even with a high-transmission UV lens the exposure time will 
be extremely long because of limited illumination. Special UV lenses 
aren't strictly required for UV shooting, but they do help. I 
recently tested a new 105mm fluorite/quartz objective against a 
conventional 105mm micro-Nikkor, and found that the conventional glass 
did transmit a small amount of UV. 

Brian 
www.caldwellphotographic.com



From: "f/256" askme@my.email.pls
Newsgroups: rec.photo.equipment.large-format
Subject: Re: Makes a sharp f/90
Date: Thu, 15 Jan 2004

"Collin Brendemuehl" dpcwilbur@excite.com wrote 
> aldenphoto@aol.com (Ken Smith) wrote
>
> Don't forget that f90 on 19" is a LARGER diameter tahn f45 on 240mm.
>
> f-stop
> focal length 90 64 45
> Inches mm aperture d(mm)
> 240 2.67 3.75 5.33
> 16.50 419.10 4.66 6.55 9.31
> 19.00 482.60 5.36 7.54 10.72
>
> Is this not correct?

Unfortunately diffraction is directly proportional to the f-stop and not
just inversely proportional to the aperture diameter, in other words, like
doesn't really care if the hole is big or small, it cares about the ratio
focal_length / aperture_diameter. All lenses at f/90 would have similar
diffraction, but as it has been already mentioned in this thread, for equal
size prints, shooting big formats has the advantage of requiring less
enlargement which in turn lessen the impact of diffraction, compared to
smaller formats.

Guillermo


From: dpcwilbur@excite.com (Collin Brendemuehl)
Newsgroups: rec.photo.equipment.large-format
Subject: Re: Makes a sharp f/90
Date: 15 Jan 2004
aldenphoto@aol.com (Ken Smith) wrote
> Every time I see a 16.5" or 19" APO-Artar up for sale, (and I haven't
> been able to snag one yet) the seller always likes to say, "Makes a
> sharp f/90" Is that possible with these or other process lenses? I
> don't dare go past f/45 with my sharpest lens, a 240 Apochromatic
> Fuji, which goes to f/90.
>
> I made a question a few years ago, (which was misconstrued) about
> whether older lenses used by the infamous f/64 folks were somehow more
> capable of sharp images at smaller stops. The question does not
> concern the aesthetic choice of the f/64 group for all over sharpness
> to counter the soft psuedo art trend that continued to linger.

Don't forget that f90 on 19" is a LARGER diameter tahn f45 on 240mm.

f-stop
focal length 90 64 45
Inches mm aperture d(mm)
240 2.67 3.75 5.33
16.50 419.10 4.66 6.55 9.31
19.00 482.60 5.36 7.54 10.72

Is this not correct?

Collin


From: "RLL" rlombardo3@cox.net
Newsgroups: sci.optics
Subject: Re: "Difraction Limited" question
Date: Sat, 17 Jul 2004


The diffraction limited spot diameter for a circular aperture is defined by:

Ds = 2.44 * lambda * F/no.

So for 550 nm visible light, the focused spot diameter for an F/1 lens is
1.34 microns; for an F/16, it's 21.5 microns.

The formula for the diffraction limited angle again for a circular aperture
is:

Theta = 2.44 * lambda / aperture diameter

This result will change as a function of aperture diameter change.

When one combines this formula with the spot diameter product formula, F *
theta, where F is defined as the focal length, you can derive the spot
diameter formula mentioned initially.

I hope this helps.

- Russ in Santa Barbara


"Charles" ckraft@SPAMTRAPwest.net wrote ...
> In photography there is a limit to increasing depth of field due to
> increased diffraction limiting due to small aperture stop. Not
> really, I guess, but the image quality degrades due to diffraction.
>
> Is the diffraction relating to the /f number being used, or the actual
> aperture opening? A 28 mm lens at /f 16 would have a much smaller
> opening than a 200 mm lens at the same /f 16 setting.
>
> I've wondered this for a long time, and I also wonder if I have asked
> the question in a way that can be understood.
>
> Oh, another question. In some optics books I see diagrams that
> represent the power density at small offsets from the center dot of a
> focused source, I don't know what the diagrams are called, and I
> can't find any on the web to refer to. In the microwave world I
> worked in they would be called the mail lobe and side lobes of an
> antenna.
>
> Anyway, my question, are these calculated or measured. I can sort of
> see how they could be calculated, although I couldn't do it with out
> learning a lot. I can't see how they could be measured, it seems that
> any measurements at this scale would be working beyond the limits of
> test equipment.
>
> Obviously I'm a dummy, but I like reading about smart things.
>
> thanks
> - Charles





End of Page