The GetDPI Photography Forum

Great to see you here. Join our insightful photographic forum today and start tapping into a huge wealth of photographic knowledge. Completing our simple registration process will allow you to gain access to exclusive content, add your own topics and posts, share your work and connect with other members through your own private inbox! And don’t forget to say hi!

Diffraction vs DOF for MF wide angle lenses

torger

Active member
This is the definition of Hyperfocal distance, and is, technically, not disputable... But Merklinger's book "The ins and outs of Focus" suggest that we should focus nearer to infinity.
I do that myself too, since even if I overshoot hyperfocal distance by 50% or something the exponential behavior of DoF leads to that I still gain quite much in near limit than compared to focusing at infinity.

Example: 50mm f/11 hyperfocal distance is 14.8 meters, if focused exactly there we get DoF from 7.4 to infinity. If we overshoot by as much as 50% to 22 meters we still get sharp result from 8.8m (we lost 1.4 meters in near limit), while if focusing at infinity we get from 14.8m (lost 7.4 meters).

Overshooting a bit also guarantees a really sharp infinity, and the typical scene where hyperfocal applies we have fine structures at infinity (a forest-covered hill, a mountain etc) while up close the structures are generally larger and it hurts less if those are a bit soft.
 

Shashin

Well-known member
So, lets walk the walk. This is a test I actually did to find out how far I can push the system. I took an image under very bad light as the systemic resolving power is related to target contrast--so the image is not going to be as sharp as under direct light. Not only is this an overcast day, it is in a rather dark forest. This is also ISO400--I do a lot of handheld work and push my ISO as well. although this image was taken on a tripod. (Note: curves were applied because the original is flat and unsharp masking as well--no noise reduction.)



On the left is the entire image. It was shot at the diffraction limit of the system or close to it--f/22. The image on the right is a 100% crop and most likely not at the focal plane. This represents a 4.9mm x 4.6mm section of the 44x33mm sensor. If I made this into a 4x3 ft. print (and I printed this personally to 3 ft.), you are viewing a 5"x5" section of that print.

Digital photography has not rewritten imaging science. CoC is still a factor of format and not pixel pitch--you can check Zeiss's wonderful online pdf on DoF and Bokeh and you can visit the Cambridge Color web site. I also have a shelf full of books saying the same thing. While pixel peeping is a really neat thing to do and provides useful information, it is not a great way to definitively judge sharpness nor DoF. And the more pixels a sensor has, the less relevant the 100% view gets.

BTW, like I stated in my original post in this thread, I personally avoid the diffraction limit as I personally do not like it. But I have done great work at f/16 and do not hesitate to use it. I do on occasion shoot at f/22 when I need to.

My best suggestion is try and test (and in the field it is easy to take DoF field brackets--you don't need to stick to one). There are a lot of factors that influence sharpness--there are conditions, like the one above, where I would be hesitant to use small apertures--the test was just confirming for me where my limits were. I do not recommend playing by numbers which seems to be the fashion--"f/8 is the optimal aperture so stick to only that" kind of thing. There are really good reasons to give up a little at the plane of focus to gain something because of DoF--and this goes both with opening up or stopping down.
 
Last edited:

dick

New member
CoC is still a factor of format
Saying this again will not make it true...

You say:
you can print large for a trade show and if you know that you are photographing for 1/4 viewing distance condition, you can plan your shoot accordingly,
¿So CoC is a factor of the cropped format that corresponds to the smallest section of the image that a viewer will look at?

...and I am aware that people (and organisations, and books) have, for decades, assumed that the observer of the picture will always want to see the whole picture, and not move in closer to see more detail... Would it be nice if we could move on?

¿When people use a Canon lens on MFD, do they re-calculate their CoC so that they get no more res on MFD than they would have had using the same lens on 35mm?

¿If I want a high res image, and shift-and stitch with a 120mm lens instead of using a 60mm lens, do I re-calculate the CoC for the larger effective format, so that I get no more res? ...or do I do it so that I can print big at 360 original camera pixels per print inch, so that viewers can get close to the image to appreciate (the fine detail in) small sections of the image?

Perhaps you could try to see my logic instead of believing what you have read in books?
 

torger

Active member
Again it is about what the purpose of DoF calculations are. Traditional CoC and DoF theory is all about "acceptable sharpness" for relaxed viewing conditions. That thinking also leads up to that more than a 4 megapixel camera is a waste, because if you stand a bit away from the print, have average vision and don't care that much it will be acceptably sharp. We've all seen tests where prints from compact cameras are indistinguishable from ones made with medium format, and sure this happens when viewing conditions are relaxed. That "image science" has indeed not changed.

But when it comes to shooting technique, would you strive for making the sharpest possible image for the given situation, or just "sharp enough" for a relaxed viewing condition? The recommendations for tripods etc would look quite different if we were into "sharp enough" -- of course we want "as sharp as possible"!

In that case DoF tables and hyperfocal distances based on traditional CoC are not helping us making the proper decisions. Traditional tables will indicate that we can shoot at large apertures and will lead to an unnecessarily sharp plane of focus and blurry DoF edge. It is a waste! The logical error with this is that when we choose CoC as the smallest spot we can see it is meaningless to have anything sharper than that, so why don't we just stop down until the plane of focus is as blurry as the DoF edge?

This is what CoC = Airy disk is all about. These tables will show DoF where the plane of focus in no sharper than the DoF edge even when nosing the print, and it takes diffraction into account. This will cause us to stop down more not less than if using tables with a large CoC.

With film there were reasons to not stop down much, 1) you could not do it at all (large format cameras rather limited in minimal aperture), 2) overly long shutter speeds (large film formats require very small apertures and thus long shutter speeds) 3) the small aperture effects contrast loss and diffraction could not be compensated for in post-processing as effectively as we can digitally. So in that case it made sense to have a relaxed DoF since we needed a tradeoff with less diffraction and larger apertures.

With digital we are not limited in the same way so we can find a much better balance between diffraction blur (and shutter speed) and CoC blur to give a more detailed overall image. As a comparison, for many 8x10" lenses the smallest aperture is f/64, the equivalent aperture for my 48x36mm sensor is f/11 (64/5.5 crop factor) which actually is the largest aperture I shoot landscapes at!

I may buy into that shooting at a larger aperture and have blurry DoF edges can give the image a more layered 3D look, a subtle depth, but I think that only makes sense when you actually have a main subject to focus at, so the plane of focus it at some logical position for the composition (i e not hyperfocal and generally not when tilted). In those cases I often let some stuff be outside the stringent DoF and blur slightly.
 
Last edited:

Shashin

Well-known member
Saying this again will not make it true...

You say:
¿So CoC is a factor of the cropped format that corresponds to the smallest section of the image that a viewer will look at?

...and I am aware that people (and organisations, and books) have, for decades, assumed that the observer of the picture will always want to see the whole picture, and not move in closer to see more detail... Would it be nice if we could move on?

¿When people use a Canon lens on MFD, do they re-calculate their CoC so that they get no more res on MFD than they would have had using the same lens on 35mm?

¿If I want a high res image, and shift-and stitch with a 120mm lens instead of using a 60mm lens, do I re-calculate the CoC for the larger effective format, so that I get no more res? ...or do I do it so that I can print big at 360 original camera pixels per print inch, so that viewers can get close to the image to appreciate (the fine detail in) small sections of the image?

Perhaps you could try to see my logic instead of believing what you have read in books?
The format is the image area--cropping and stitch impact that. The trouble with amateurs is you get little pieces of information and really don't understand how they go together. Perhaps you can make a little effort to study this and come back when you understand the fundamentals.
 

Shashin

Well-known member
Again it is about what the purpose of DoF calculations are. Traditional CoC and DoF theory is all about "acceptable sharpness" for relaxed viewing conditions. That thinking also leads up to that more than a 4 megapixel camera is a waste, because if you stand a bit away from the print, have average vision and don't care that much it will be acceptably sharp. We've all seen tests where prints from compact cameras are indistinguishable from ones made with medium format, and sure this happens when viewing conditions are relaxed. That "image science" has indeed not changed.
You are misrepresenting what the science does. To say there is no point on having a high-resolution system mean you have missed the point and also don't understand the image is a collection of resolutions and details over the area. The science is a model to predict the perception of certain qualities in an image. Nor is it a fixed model. Your visual system can actually perceive details smaller than what it can resolve--this is called detection. So a system exceeding the basic limit of the defined CoC is still impacting the image and in ways that can be perceived. But sharpness is not a definition of detail.

But when it comes to shooting technique, would you strive for making the sharpest possible image for the given situation, or just "sharp enough" for a relaxed viewing condition? The recommendations for tripods etc would look quite different if we were into "sharp enough" -- of course we want "as sharp as possible"!
But "sharp as possible" is a subjective call and there are two sides to that. If you don't have enough DoF, then the image is not as sharp as possible and you will have to pay for getting that with the effect of diffraction.

In that case DoF tables and hyperfocal distances based on traditional CoC are not helping us making the proper decisions. Traditional tables will indicate that we can shoot at large apertures and will lead to an unnecessarily sharp plane of focus and blurry DoF edge. It is a waste! The logical error with this is that when we choose CoC as the smallest spot we can see it is meaningless to have anything sharper than that, so why don't we just stop down until the plane of focus is as blurry as the DoF edge?
That is a personal evaluation. DoF is simply a model. CoC values can be chosen to help you visualize what your condition produce. But these CoC values are not absolute. If a set of values does not seem good to you, you can change them, this is why difference lens manufacturers have used different values.

You yourself have redefined the values based on your personal work. The trouble is you are taking your subjective choice and saying this is now true for everyone. The normal DoF model assumes an average human response and so is the best starting point. Just because you have refined the number for your self does not undermine the theory. And certainly folks still zone focus with lenses and so the theory works regardless of your personal frame.

This is what CoC = Airy disk is all about. These tables will show DoF where the plane of focus in no sharper than the DoF edge even when nosing the print, and it takes diffraction into account. This will cause us to stop down more not less than if using tables with a large CoC.
If that work for you, fine. But I am not talking about personal preferences or advice. To say the size of the Airy disk, and really it is not the Airy disk unless you image point sources, is the defining attribute make no sense in relation to format. The spot size in reference to the entire image area defines the point--it does make a difference if it is 1/1500th, 1/3000th, 1/7000th of the image size. Absolute frames are not helpful. And this is why simply adding pixels to an image does not softer nor does it change the DoF.

With film there were reasons to not stop down much, 1) you could not do it at all (large format cameras rather limited in minimal aperture), 2) overly long shutter speeds (large film formats require very small apertures and thus long shutter speeds) 3) the small aperture effects contrast loss and diffraction could not be compensated for in post-processing as effectively as we can digitally. So in that case it made sense to have a relaxed DoF since we needed a tradeoff with less diffraction and larger apertures.

With digital we are not limited in the same way so we can find a much better balance between diffraction blur (and shutter speed) and CoC blur to give a more detailed overall image.
There are reasons not to stop down much with digital. Will will also notice that the minimum aperture for film camera lenses were different based on format.

Exposure time is irrelevant here. BTW, unsharp masking is originally a film technique.

I think you have completely missed my the point of my posts. I am not say you stop to any particular aperture. I am simply showing how the attributes of DoF and diffraction work together and their limits. I am not doing this from a personal opinion based on my working method. But I am working from the definitions of DoF and my experience in seeing it in real terms. It works.

I have also stated if you want to redine it for your personal work, as you have, you can.
 

dick

New member
The format is the image area--cropping and stitch impact that. The trouble with amateurs is you get little pieces of information and really don't understand how they go together. Perhaps you can make a little effort to study this and come back when you understand the fundamentals.
...and I thought that it is you who is quoting clichés out-of-context without understanding what you are talking about?

I have the advantage of being an engineer, and understanding technology... and there are many amateurs here who know more than most pros - but that is not saying much.
 

torger

Active member
Shashin, we probably agree more than we think. I think we both end up in using about the same apertures :). I'm not sure I'm arguing against you in some way even, I just try to explain my view on DoF in a little bit more detail. I surely agree that we can define CoC to whatever we like depending on what our purpose with it is.

What I try to say is that the traditional DoF with fixed large CoC (0.2mm for 8x10, 0.1 for 4x5, 0.03 for 35mm etc) which often end up in tables here and there and for hyperfocal advice etc do not give the results most digital users expect.

Digital users don't expect when they pixel peep their image that the infinity is blurry when they have used hyperfocal technique. A digital user would then have preferred to stop down more and focus farther ahead, which happens to provide a sharper image (or equally sharp) for all viewing conditions than the original one.

The traditional DoF tables are more about describing what is seen in a typical relaxed viewing condition after the print is made, rather than helping the digital user to achieve optimal result in the field which provides best possible input to make prints for various viewing conditions.

Nearly always someone asks about how to achieve the sharpest image. Then I think talking about the fuzzy subjective viewing condition-based DoF, "it depends on what you want etc" does not really help much.

I say there is another way :), a way to look at DoF that helps anyone optimize their shooting technique out in the field so that regardless of viewing condition of the final print there's the best input for it.

I actually use DoF/hyperfocal/tilt tables out in the field to make actual decisions about aperture, if it is better to tilt than to use hyperfocal, etc. CoC = Airy disk works. The resulting DoF is easy to understand what it means, a good reference point for deep DoF photography that helps you optimize your technique. It is also well-balanced, it does not lead to insanely small apertures, at the "emergency apertures" you really have lots of DoF.

Why Airy disk? Because landscape apertures are always diffraction-impacted in modern cameras so one does not need to care about pixel pitch, and it makes no sense to have a smaller CoC that the Airy disk since it defines the resolution limit (sort of, CoC = 0.8xAiry disk is probably more exact but at this level, the pixel level, it is not important). Why not larger than the Airy disk? Because if our goal is to get deep DoF and we have important stuff within it it makes no sense to give more sharpness in the plane of focus than at the DoF edges, then it is better to stop down more increase diffraction and get more DoF.

There are some subtle things around this of course but this DoF definition is very understandable and good to use as reference to make modifications from. But of course, if someone finds using a fixed size 0.04mm CoC more useful that's ok.

I don't want to claim that "my method" (I was certainly not first to use it) is some sort of ultimate truth, I just try to explain why it is really useful to base DoF calculations on that if you intend to use it out in the field when you do landscape photography, and why focusing based on traditional DoF tables do not give as good results.
 

Shashin

Well-known member
...and I thought that it is you who is quoting clichés out-of-context without understanding what you are talking about?

I have the advantage of being an engineer, and understanding technology... and there are many amateurs here who know more than most pros - but that is not saying much.
Sorry Dick, but bullying has never really worked on me. You can shoot off as many personal attacks as you like. I really don't care if you are an engineer--one field does not translate into another.

I really feel sorry for you.
 

Jack

Sr. Administrator
Staff member
Thanks everyone for your thoughtful comments. I really appreciate them.
Jack..What f-stop you will use for IQ 180 with SK 55mm to get the best DOF from few feet to infinity (No focus stacking or T/S)?
Thanks.
Pramote
You can't do 2 feet to infinity with a 55 over the IQ180 without tilts. You can barely get 10 feet to infinity without tilts at f16, and f16 is sub-optimal due to excessive (at least for my tastes) diffraction.
 

Bob

Administrator
Staff member
I would like to hear opinions from friends at GetDPI regarding diffraction vs DOF for MF wide-angle lenses for landscape photography (without using T/S lens or focus stacking). What is the best compromise for final print?

Thank you,
Pramote
back to the original question.
"Best" compromise on DOF vs diffraction.
I doubt that there is one "Best".
DOF in a print viewed at a distance of its diagonal measure depends on the diagonal measure of the sensor as cropped, the focal length of the lens, and the aperture used.
Diffraction has only to do with the numerical aperture. When diffraction is extreme, it will soften the whole image, so that the sharpest areas will begin to visibly blur, but those areas that already are OOF may blur, but I doubt one would notice a difference. As a general practice I try to shoot at f/8 or f/11 maybe 16 at the small end just so I do not limit my cropping options. At "normal" viewing distances for a full frame image you can probably lose a bit of sharpness because folk's eyes are not that good.
If on the other hand, your customers are photographers primarily, then all prints are viewed at nose-length distance and through a hand glass. In that case your COC=about 1.414 times your pixel pitch and you need to shoot at f/8.
In reality, there is a lot more flexibility for the usual crowd.
A case in point, I sold a 17 by 24 image of a 6 Mpix file awhile ago. It was not even particularly sharp or that well focused.
It is hanging in an office lobby over a sofa.
I almost refused to print it, but the client is thrilled and as far as she is concerned it is "perfect" and even better, she paid.
So the answer to your question is a big "depends"
-bob
 

GrahamWelland

Subscriber & Workshop Member
:thumbs: What Bob said ... It depends. For mortal viewers I think Shashin and the vast majority of material on the CoC/DoF seem correct. For pixel peepers it's different rules applying.

What I do know is that I've yet to read one of these threads without someone getting their whiteys in a twist at some point. Not one - yet.
 

torger

Active member
A case in point, I sold a 17 by 24 image of a 6 Mpix file awhile ago. It was not even particularly sharp or that well focused.
It is hanging in an office lobby over a sofa.
I almost refused to print it, but the client is thrilled and as far as she is concerned it is "perfect" and even better, she paid.
So the answer to your question is a big "depends"
Great story :). I've sometimes hopelessly tried to point out quality differences between print A and print B which to me are obvious and significant, but the non-photographer (and even photographers too) do not see it or do not care.

One must face it that there is really a small amount of people that can even see a difference if a print was shot with a medium format system with excellent technique rather than a hand-held compact camera, or care about it if they happen to see a difference.

Fine print quality is a connaisseur thing. Usually the most picky connaisseur is the photographer that shot the image and made the print, certainly not the customers, but sometimes the customer might be a picky photographer...
 

fotografz

Well-known member
Huh? Can you help us non-native speakers? Even the urban dictionary failed to help me :)
"Tightie-Whities" are slang for men's white underwear briefs ... "in a bunch" or "twisted" refers to uncomfortable and irritating ...making the wearer "cranky" ... which in turn means ill-tempered.

Somehow when you have to explain it, it loses a lot :ROTFL:

-Marc
 

Landscapelover

Senior Subscriber Member
back to the original question.
"Best" compromise on DOF vs diffraction.
I doubt that there is one "Best".
DOF in a print viewed at a distance of its diagonal measure depends on the diagonal measure of the sensor as cropped, the focal length of the lens, and the aperture used.
Diffraction has only to do with the numerical aperture. When diffraction is extreme, it will soften the whole image, so that the sharpest areas will begin to visibly blur, but those areas that already are OOF may blur, but I doubt one would notice a difference. As a general practice I try to shoot at f/8 or f/11 maybe 16 at the small end just so I do not limit my cropping options. At "normal" viewing distances for a full frame image you can probably lose a bit of sharpness because folk's eyes are not that good.
If on the other hand, your customers are photographers primarily, then all prints are viewed at nose-length distance and through a hand glass. In that case your COC=about 1.414 times your pixel pitch and you need to shoot at f/8.
In reality, there is a lot more flexibility for the usual crowd.
A case in point, I sold a 17 by 24 image of a 6 Mpix file awhile ago. It was not even particularly sharp or that well focused.
It is hanging in an office lobby over a sofa.
I almost refused to print it, but the client is thrilled and as far as she is concerned it is "perfect" and even better, she paid.
So the answer to your question is a big "depends"
-bob
You can't do 2 feet to infinity with a 55 over the IQ180 without tilts. You can barely get 10 feet to infinity without tilts at f16, and f16 is sub-optimal due to excessive (at least for my tastes) diffraction.
Jack & Bob..Thanks very much.
Does it mean that the only ways I can achieve DOF from few feet to infinity for the IQ 180 with insignificant diffraction are T/S (probably from a Tech Camera) and focus stacking?

Tim Fitzharris, an award winning nature & landscape photographer, published a lovely book "National Audubon Society Guide to Landscape Photography" in 2007. He used Mamiya with non-digital (even manual ones) Mamiya lenses and Phase One P25 throughout his book with f/16 and f/22. His pictures looked fantastic. He also claimed they looked great on his big prints.
His approach was quite different from more recent comments concerning diffraction. Does a smaller MP DB (P25) make difference?

Pramote
 

GrahamWelland

Subscriber & Workshop Member
His approach was quite different from more recent comments concerning diffraction. Does a smaller MP DB (P25) make difference?

Pramote
yes - the 9u sensor works better with stopped down lenses in my experience. I also have a P25+ and it gives you a lot more leeway with f-stop usage for pixel peeping images. As Bob alluded, 1.414 x sensor pitch (9u) gets you to pixel peeping CoC twice as large as you'll get with the IQ180 (5.5u).

Btw, I also have that book and remember that the high gloss print sharpened images look super sharp at the small size. I'm sure his larger prints look fine too but (ducks for cover) it depends on how big they are and how close you are when viewing them.
 

dick

New member
...Phase One P25 throughout his book with f/16 and f/22. His pictures looked fantastic. He also claimed they looked great on his big prints.
His approach was quite different from more recent comments concerning diffraction. Does a smaller MP DB (P25) make difference?

Pramote
yes - the 9u sensor works better with stopped down lenses in my experience. I also have a P25+ and it gives you a lot more leeway with f-stop usage for pixel peeping images. As Bob alluded, 1.414 x sensor pitch (9u) gets you to pixel peeping CoC twice as large as you'll get with the IQ180 (5.5u).
Yes... but the physical (format) size is the problem, as DoF depends on magnification, which depends on format size... so, if anybody made a quality, high-res 35mm or micro-4/3 camera ¿and made the sensor available in a digiback? it would be great for landscapes and macro... my GH2 is a good camera for the money, but the files fall apart when you try to do any adjustments in Photoshop, especially if you use high ISO.
 
Top