Resolution Limits of 35mm Photography

[Sirius Glass]

This is getting weirdly heavy.

Diffraction imposes an absolute limit on resolution at a given contrast at any given aperture. Up to the point (the aperture) where the lens is diffraction ... <snip>....see a diference. Over about 20 lp/mm, you can't see a real difference in a real print, though you might see a difference with a vernier discontinuity at 30+ lp/mm. Somewhere between 10 and 20 lp/mm is, as far as I can see, where most people lose the ability to see the difference.

Cheers,

R.

The only thing that you missed is that due to choices of glass and lens coatings some lens have better contrast than others which helps the resolution.

Steve

 

[Michel Hardy-Vallée]

Perhaps another detail that bear consideration is that it would be possible to obtain very high resolution with a custom 35mm system if for example the lens is optimized for a specific magnification ratio. Process lenses are such kind of design: their field flatness and resolution is maximized like crazy for a narrow range of ratio (usually 1:1).

Even more extreme are the lenses used in integrated circuit production (like the Ultra-Micro-Nikkor), which resolve a humongous number of lp/mm but only at ONE specific magnification ratio and using light of ONE wavelength. Someone correct me if I'm wrong, but I think those might be the lenses that push the resolution of details to the furthest extent. Yet if you use these lenses outside of their optimal range, you won't achieve those mindbending figures.

So if we want to bicker about theoretical limits, we could go way way beyond what our general-purpose lenses can accomplish.

 

[Roger Hicks]

The only thing that you missed is that due to choices of glass and lens coatings some lens have better contrast than others which helps the resolution.

Steve

Wouldn't argue with that, except to say that this is exactly what diffraction resolution means -- the point where diffraction forms the limiting factor, not Ole's 'other stuff' (including contrast resulting from choice of glass, type of coating, lens aberrations, number of lens-air surfaces...).

Cheers,

R.

 

[Roger Hicks]

...a humongous number of lp/mm but only at ONE specific magnification ratio and using light of ONE wavelength....

Absolutely true, which is why a range is more meaningful than an absolute number. Also, there's a degree of subjectivity in reading lp/mm charts on the film with a microscope: what one person sees an an easy 100 lp/mm, another will see as a bare 125 lp/mm.

Cheers,

R.

 

[Dan Fromm]

Roger, many years ago (10/78) the Modern Photography magazine that used to be published an article on the question "Is 100 lp/mm on film attainable?" Their answer was a highly qualified "Yes, under ideal conditions. In general, no."

The ideal conditions included: taking great pains to eliminate camera and subject movement; bracketing focus; using the highest resolution films available to them (High Contrast Copy developed for continuous tone; TP; SO 2483); still and clear air; and, oh yes, using a high contrast target.

They found that f/1.7 - f/2 "normal" lense for 35 mm SLRs could just do it under the ideal conditions sketched above, and typically only at f/4 or f/5.6.

Given the technique and emulsions most of us use, not to mention the conditions we work under, e.g., low constrast subjects, and the apertures we typically use, 100 lp/mm on film is fantasy. Every time I see Zeiss' results and the claims made for, e.g., Agfa Copex developed for continuous tone (=, I suspect, H&W Control, Gigabit film, Bluefire Police), I start doubting the sanity of the people who make the claims. I'm not convinced that the results, if true at all, are easily repeatable or that in practice we can come close to them.

There's another point about diffraction limits that many of us slide over. Since some of the major lens aberrations are off-axis and are somewhat controlled by aperture, the distance off-axis at which diffraction limits the resolution attainable (Strehl ratio > 0.85 is, I think, the conventional definition) is somewhat dependent on aperture. Brian Caldwell has made the point that the 55/2.8 MicroNikkor AIS is diffraction limited at f/4 in the central ~ 8 mm (that's diameter) of the image. Outside of that very sharp little circle, aberrations swamp diffraction. That, incidentally, is why its such a fine performer reversed at magnifications above 1:1.

Cheers,

Dan

 

[DrPablo]

Something I've not seen (I may have missed it) is that the practical limit of resolution on the film is often imposed by film location.

Ignoring all other differences, this suggests an advantage (by design) of the fixed position of digital sensors. Based on the variability of film position, how much resolution is practically lost by deviation of film position?

I have since grown to appreciate Ctein's argument that the finest detail visible as a discontinuity in a straight line (the 'vernier acuity') is actually considerably higher, so my criterion for "pretty damn' sharp" is now 10-20 lp/mm. Under 10 lp/mm, you really can see a diference. Over about 20 lp/mm, you can't see a real difference in a real print, though you might see a difference with a vernier discontinuity at 30+ lp/mm. Somewhere between 10 and 20 lp/mm is, as far as I can see, where most people lose the ability to see the difference.

I'm assuming this is from a close viewing distance, like 10 or 12 inches, right? That's a pretty tall standard for a 35mm camera to meet unless you're really operating at high system resolution. A sharp 8x10 would require at least an 80 or 85 lpm resolution negative, and that's discounting losses from enlargement. That basically rules out most color photography.

Other than drum scanners, I'm not sure scanners really constitute an especially lossless system themselves, because Bayer interpolation is a heavy handed means of image acquisition. Oversampling obviously helps this in a way impossible in cameras with the same type of sensor. Still, I'm only so taken by scanning as an alternative to enlargement.

 

[Ed Sukach]

Did someone once say that discussions of a technical nature were out of place ... inappropriate... on a "Forum", such as this?

Let me restate my positon:

Me have camera.

Me take pictures.

Me no CARE about diffraction.

Once in a while ... I get good ones.

If anyone cares - and I'm quite sure most DON'T - Most of my Theorectical Optics training was accomplished at ITEK corporation.
After that, I joined the MIT Instrumentation Laboratory ... Eventually becoming Manager of the Optical Tooling Laboratory. From there ... Quality Assurance/ Supervision at American Science and Engineering (Computer Assisted Tomography - CAT Scanners)....

All that - and $5.00 - will get me a cup of coffee at Starbucks.

I LOVE optics -and even more - Photography.

 

[DrPablo]

Did someone once say that discussions of a technical nature were out of place ... inappropriate... on a "Forum", such as this?

If you're referring to my statement, that's hardly what I said or meant. My point was that your depth of inquiry into this subject was beyond what could be answered on a forum. It's completely in place and appropriate, but if you expect to do any more than scratch the surface you'll probably be disappointed.

 

[Roger Hicks]

Dear Dan,

Your point about where the image is diffraction limited is very important indeed, and as you say, all too often ignored.

My views on what is attainable are based on testing films for various magazines using several typical medium-contrast targets, plus a correspondence some years ago with Ctein on the same subject, plus conversations with Zeiss lens designers (including practical demonstrations on their part).

My own view is that the Modern statement is too conservative, but not much too conservative. As I say, I've seen 100 lp/mm centrally reasonably often, on Delta 100 and Pan F, but both exposure and development must be kept to a minimum, quite possibly less than would be tonally acceptable on a 'real' picture.

Exposure and development are two more points that are normally glossed over, because an extra stop can easily wipe 10-15 lp/mm off resolution, as can over-development. The many people who prefer slight overexposure in the interests of tonality, to say nothing of those who habitually overexpose grossly, will certainly never see 100 lp/mm.

I half agree with those who say that such figures are meaningless in the real world, but I don't think that's the same as saying they're worthless. A lens/film combination that delivers 80-100 lp/mm under test conditions will generally give crisper, sharper pictures in the real world than one that delivers 70-80, and MUCH better pictures than a nasty zoom that gives 50.

Then of course there's the question of 'sparkle', which both Zeiss and Ilford have quantified as very high MTFs at surprisingly low spatial frequencies. This is arguably more important than resolution.

Cheers,

R.

 

[Roger Hicks]

Dear Paul,

From all I have learned, if film could be in exactly the right place at all times, 125 lp/mm would be reasonably readily attainable on the sharpest and finest grained films. This is assuming that 100 lp/mm is reasonably readily attainable under the same ideal conditions with a real-world camera. So yes, a properly placed sensor has an inherent advantage over 'floating' film. But it has to be properly placed, which apparently isn't always the case with cheap SLRs, especially after allowing for discrepancies between the focusing screen and the sensor.

I fully take your point about scanning, and was indeed referring to drum scanning as being superior to most or possibly all enlarger lenses. This was based on my own experiences with published books in the last 25 years or so plus access to a drum scanner that was owned and operated by a friend who was also a lecturer in printing (photomechanical reproduction).

As for vernier acuity, I wasn't suggesting that this was readily attainable with 35mm; sorry for the confusion. And yes, 10-12 inches/25-30cm was what I was talking about.

This is one reason why I seldom print very big. I like super-sharp pictures of some subjects, so I generally contact print 5x7 inch and above; prefer not to go above 3x for roll-film; and limit myself to about 5x for 35mm.

Cheers,

R.

 

[Ed Sukach]

If you're referring to my statement, that's hardly what I said or meant. My point was that your depth of inquiry into this subject was beyond what could be answered on a forum. It's completely in place and appropriate, but if you expect to do any more than scratch the surface you'll probably be disappointed.

I am at the point where I no longer keep track - or CARE - about who wrote what, or where.

All this SEEMS to me to be on the same plane as those endless, unresolved, and unresolvable discussions among those who would like to be thought of as cognoscente - keepers of the "inside secrets" of photography. I can remember one group - closed to the profane - constantly arguing the merits and absolute necessity of using a lens of one particular design. If one eavesdropped, these conversations were filled with "Dagor", "Planar", "Petzval", "Arton" ... in what seemed to be a totally random sequence.
They seemed to spend the greatest part of their time in these discussions, and produced very few images... possibly because someone could use the critique of those images as ammunition in the next fiery debate.

I've come to the place where I have one overriding question: What is the relevance of all this? Will the final answer/s really be of use in my work, or will they simply be filed away in my memory and notes?

It is tempting to reminisce about the past - and my times at the Optical Bench. And yes ... there CERTAINLY was a great deal of influence on the results caused by the perception of the operator.

Now ... as I've written, I try to simplify - to direct my attention to the relevant "stuff" ... and I've found that keeping track - analyzing - evaluating - considering my emotional involvement (how would THAT be as a theme for discussion!!) is far more useful than ... I can feel the air being sucked out the room, now ... technical minutia.

Long live f/63!!!

 

[Ed Sukach]

On further thought - I feel that I may have been a bit hasty in my remarks regarding describing the quality of a lens through the use of its "Diffraction Limit".

It has been written here that manufacturers do not include that information ... because ... well - it is too easy to determine that from its Modulation Transfer Function graphs.

I have been using one lens - the Zeiss Sonnar CF f/4 150mm - on my Hasselblads for, lo, these many moons, blissfully ignorant of this characteristic - information that has been deemed critically important by a few participants here.
I am attaching the MTF data, directly from the Hasselblad catalog. With the hope that no one expires from sheer boredom, could anyone help me to determine that "Diffraction Limit" for this lens, so that I may, armed with this vital information, go afield and utterly destroy the competition?

 

[Sirius Glass]

On further thought - I feel that I may have been a bit hasty in my remarks regarding describing the quality of a lens through the use of its "Diffraction Limit".

It has been written here that manufacturers do not include that information ... because ... well - it is too easy to determine that from its Modulation Transfer Function graphs.

I have been using one lens - the Zeiss Sonnar CF f/4 150mm - on my Hasselblads for, lo, these many moons, blissfully ignorant of this characteristic - information that has been deemed critically important by a few participants here.
I am attaching the MTF data, directly from the Hasselblad catalog. With the hope that no one expires from sheer boredom, could anyone help me to determine that "Diffraction Limit" for this lens, so that I may, armed with this vital information, go afield and utterly destroy the competition?

Not enough posted information, but generally f/8 to f/11 for 35mm, f/11 to f/16 for MF are the aperatures that provide the best resolution. As the f/number increases beyond these points the diffraction starts to dominate. How much deteriation you will tolerate is up to you. For the physics related to the increased dominance of diffraction, I suggest that you read the following:
Seeing the Light Falk, Brill, Stork
Remote Sensing Schott
Modern Optical Engineering Smith
Optics Hecht
Principles of Optics Born, Wolf
Research is left to the reader as an assignment.

This thread got beaten to death a while ago, I am going back to sleep.

Steve

 

[Ed Sukach]

Not enough posted information, but generally f/8 to f/11 for 35mm, ...

Hmmm ... in answer to what has been labeled a "simple thing to do", I get reference to an entire library that I have to study ... Could it be ... possibly... not so simple after all?

Generally, most lenses perform best in the center of their aperture range - not a startling revelation - but I thought there was specific information to be derived from the MTF.

Granted, I was challe ... uh ... giving the example of a MF lens.

Attached is the Data Sheet for a more modern Zeiss f/2 50mm lens for the 35mm format, with more information... here goes...

 

[Mick Fagan]

All that - and $5.00 - will get me a cup of coffee at Starbucks.

Ed, I will have to disagree with the above statement. As a result of reading your line and walking past one of those stores today, I went in to a Starbucks for the first time ever, and ordered a coffee.

Starbucks may call it coffee, but as a person who has lived in Melbourne Australia, for the last 34 years, and enjoyed a plethora of Italian coffee shops, that isn't what I would call coffee.

I now know, why Starbucks doesn't appear to be gaining ground in this city, used paper developer would be stronger than the stuff they served!

Mick.

 

[Dan Fromm]

Ed Sukach wrote "Generally, most lenses perform best in the center of their aperture range - not a startling revelation - but I thought there was specific information to be derived from the MTF."

Ed, find a library that has Modern Photography, either on paper or on microfilm, and look at the article "How Sharp Can You Get" in the October, 1978 issue. They reported that the f/1.7 - f/2 50 mm or so lenses they tried were best at f/4 or f/5.6, depending on the lens. That isn't the middle.

 

[Ed Sukach]

Ed Sukach wrote "Generally, most lenses perform best in the center of their aperture range...
Ed, find a library that has Modern Photography, either on paper or on microfilm, and look at the article "How Sharp Can You Get" in the October, 1978 issue. They reported that the f/1.7 - f/2 50 mm or so lenses they tried were best at f/4 or f/5.6, depending on the lens. That isn't the middle.

Dan, Instead of MY travelling to some odd place that still might have a microfiche reader (remember them? I once worked for a company that made them!), you could scroll back, right here, to the message I've posted on 03-06-2007 - (#26 Permalink.)

I've been thinking of an example where the quality of a lens decreases immediately when "stopped down"... and diffraction has NOTHING to do with that worsening in quality....

Occasionally, when testing a series of lenses from a certain production run... one sees an individual lens where ONE element has been assembled incorrectly - "reversed", so that the wrong surface faces front. That, sometimes, results in a image where the center of the field is one indistinct blur and it sharpens as the extremes of the field are approached. If one limits the ray bundle by introducing an aperture, the sharp areas will be eliminated and the overall quality decreased.

That, simply, is an error in the geometry of the lens. In practice, no geometry is "perfect"; most of the attention is given to the center of the image field; so it is a rare lens that conforms to the intended design that will exhibit that characteristic.

Where does one obtain these "MTF" (???) diagrams that allow one to determining the point where diffraction limits the quality of a SPECIFIC lens?

I've googled extensively - and haven't found one example.

 

[Jim Jones]

Although much science can be reduced to mathematical certainity, diffraction limited photography approaches voodoo. One can cite a formula for diffraction limited resolving power: 2000/f-number = lines per mm. (Lenses in Photography. Rudolf Kingslake, 1951) Other experts give conflicting formulae. Even in pinhole photography, where diffraction limiting is critical and by far the major factor in image sharpness, there is much disagreement about optimum pinhole diameters. In practical photography many other factors should be considered, such as film characteristics, exposure, development, subject contrast, lens abberations, film plane accuracy, focusing errors, etc. Since negatives must usually be enlarged before diffraction becomes apparent, most of the same factors apply to the printing. Dwarfing all of these factors is the way individuals interpret sharpness. One can narrow their choices of lenses and technique by considering the criteria that works for other photographers with other equipment, but that's just a starting point. Quibbling over the minute details of diffraction limited photography is as pointless for photographers as arguing religion. It can lead to stimulating (and acrimonious) debates, but it does little to promote photography.

 

[Dan Fromm]

Ed Sukach wrote "I've been thinking of an example where the quality of a lens decreases immediately when "stopped down"... and diffraction has NOTHING to do with that worsening in quality...."

Ed, all Zeiss Luminars and Mikrotars, Leitz Photars, Nikon Macro Nikkors, Reichert Neupolars are all less sharp one stop down from wide open than they are wide open. Diffraction has EVERYTHING to do with it.

 

[Ed Sukach]

Ed Sukach wrote "I've been thinking of an example where the quality of a lens decreases immediately when "stopped down"... and diffraction has NOTHING to do with that worsening in quality...."

Ed, all Zeiss Luminars and Mikrotars, Leitz Photars, Nikon Macro Nikkors, Reichert Neupolars are all less sharp one stop down from wide open than they are wide open. Diffraction has EVERYTHING to do with it.

I cited one example from experience where diffraction had NOTHING to do with image quality. Those lenses happened to be for use in microfiche readers and had NO diaphrams. When disassembled and reassembled correctly, they were fine. It was not meant as a blanket, universal FACT

Shouting now, are we? I say the effects of diffraction at large apertures are INSIGNIFICANT. Is the one who shouts loudest going to be RIGHT?

I'm not familiar with ALL the Luminars, Photars, Macro Nikkors, and ... who...? - Reichert Neupolars ...
Let's see their MTFs ... and possibly we can determine if that statement seems reasonable.

,,,

 

[Roger Hicks]

I cited one example from experience where diffraction had NOTHING to do with image quality. ,,,

Dear Ed,

Well, it did, sort of. At full (and indeed, the only) aperture, it might well be that the usual run of optical shortcomings (to say nothing of film location questions) were more important than diffraction. On the other hand, at (let us say) a full aperture of f/4, there was no way the resolution could exceed about 450 lp/mm for green light. Or at f/2, more than about 900 lp/mm...

I fully take your point about the irrelevance of diffraction limits in real pictures (or microfilm readers) in the real world, at least below about f/11 to f/16 for 35mm, but that still doesn't mean that the theoretical considerations are invalid in a limiting case.

Have you signed up for f/63 in the Lounge yet?

Cheers,

R.

 

[Dan Fromm]

Ed, you wanted (emphasis yours) an example of a specific lens that lost image quality when stopped down from wide open. I named four lens families all of whose members have that property. So don't yell at me, just accept reality.

I suggest an exercise for you. Buy a 100/6.3 Luminar. Or a 120/6.3 Macro Nikkor. Test it at 1:1 at f/6.3 and f/8. Report back.

 

[Ed Sukach]

Ed, you wanted (emphasis yours) an example of a specific lens that lost image quality when stopped down from wide open. I named four lens families all of whose members have that property. So don't yell at me, just accept reality.

I am trying to narrow the research to a specific example ... I don't consider "families" to be SPECIFIC. Is there MTF information universal to all the members of the "family"? ... if so, where is it? ... or if it is not, where is ONE MTF chart that can be used to support your theory?

Or are you bent out of shape that I have the insolence associated with the failure to accept YOUR word as ultimate truth - on sheer faith alone - a' la "Lemming Syndrome"?

I suggest an exercise for you. Buy a 100/6.3 Luminar. Or a 120/6.3 Macro Nikkor. Test it at 1:1 at f/6.3 and f/8. Report back.

There is some reason for me to spend money and jump through hoops to satisfy your argument? What is this ... a demand that I report back to you? - or anyone else? Some here seem to be hung up on getting others to "report". It won't happen - I left the Consulting Business a few moons ago.
Can I, with all the respect appropriate here ... ask a question: Who the hell made YOU my boss?

Hmmm ... you admit there WAS shouting. I can't "yell" at you in retaliation, but the reverse - you screaming at me - should be absolutely OK.

Nah!

I love the idea of "accepting reality." I'll do that, and I invite anyone else around here - present company NOT excluded - to do the same.

Vive le f/63!!!

 

[Dan Fromm]

Ed, I've done the experiment. At 1:1 a known good 100/6.3 Luminar is sharper at f/6.3 than at f/8. I've got the same result with a 100/6.3 Neupolar. If you can't accept my results, the best thing for you to do is try to duplicate them.

As is, you won't accept anyone's word for anything, you won't do your own testing, you won't open a book or go to a library. If I had a 2x4 and could reach you I'd address you smartly with it. This is said to work on mules, might work on you.

 

[HerrBremerhaven]

. . . . .

I've come to the place where I have one overriding question: What is the relevance of all this? Will the final answer/s really be of use in my work, or will they simply be filed away in my memory and notes?

. . . . .

One of the things that comes to mind, and this refers to the recent Zeiss tests, was the conclusion Zeiss gave about the high resolution results. Quite simply, they stated that it was assurance that their (newest) lenses would not be the limiting factor in your photography. I suppose if someone wanted to read more into that, then it might be assumed that the (newest) Zeiss lenses outperformed most films.

I also recall a recent print from a 6x9 transparency. The original shot was done hand held using a $10 AGFA folder with a simple triplet design lens. Judging by tests done by Christopher Perez and others, I would guess the lens and my hand held technique probably resulted in a resolution capability well under 30 lp/mm. Maybe under ideal conditions, using 35mm gear, I could get a really similar, or technically close result, though I don't have any desire to test that.

Ciao!

Gordon Moat
Dead Link Removed