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

Landscapelover

Senior Subscriber 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
 
Last edited:

Shashin

Well-known member
By diffraction, I assume you are talking about sharpness of the image. DoF is another aspect of sharpness.

First, photography as we think of it as a creative pursuit is about creating pleasing images. Pleasing is a technical term and basically states that the output is subjective--it is the observer's perception and reaction to an image. Another important point which is often lost is the photography is not comparative--the image is judged on its own merit. Comparisons are done to show changes, but images are presented individually. So when we look at an image, we want to perceive it as sharp. It does not matter if another image or set of conditions is sharper.

Sharpness is a subjective quality as it lies in the observation of the image. It is not an absolute quality in the object. A meter is an object quality--it is either 100cm or it is not. The observation and outside conditions do not change that. Sharpness is a perceived quality based on many outside conditions like viewing conditions and the eyesight and taste of the observer. So the concept of viewing conditions arise when talking about sharpness. The standard viewing condition in the simplest form is to view an image from a distance equal to the diagonal of the area. This is a nice definition as the actual size of the image is irrelevant--the larger the image, the greater the viewing distance and so the perception of the image is always the same--a 300dpi 8x10 viewed at 12" looks the same as a 150dpi 16x20 viewed at 24" (you notice the number of pixels do not change).

A simple definition of sharpness and DoF is to do with the smallest permissible circle of confusion that is perceived as sharp. This value is also subjective and does vary both with camera/lens manufacturers and photographers--if you don't like one set of numbers, you can make your own. Zeiss defines the permissible circle of confusion as 1/1500th of the format diagonal. DoF is related to format, not pixel pitch, as it is in relation to viewing conditions that assume a particular distance an image is viewed and in relation to the human visual system.

Now folks will say they don't always look at an image at standard viewing distance. But here is the neat thing, the illusion of sharpness is robust and has a built-in error. The definition of a photo-quality print is 300dpi, assuming an 8x10 print viewed at about 12". To keep the math simple, than means the diagonal of the print is divided into pixels that are about 1/3000th of the image diagonal. The definition Zeiss gives for the permissible circle of confusion is 1/1500th of the diagonal. This means if you view an image at half viewing distance, the plane of focus should be acceptably sharp--a 16x20 viewed at 12" is still sharp even at 150dpi.

MFD easily exceeds these numbers.

So, the CoC depends on the format diagonal. Sharpness is a function of the format size and viewing condition.

I believe you are wanting to know how far you can stop down to maximize DoF without diffraction becoming obvious. With a 6x4.5 format, f/22 or f/32 should be the diffraction limit. Your personal diffraction limit may be different. I routinely stop my 645D and P25+ down to f/16. I sometimes use f/22, but that is getting o my threshold. But it is not a simple thing, high-contrast scenes work better and unsharp masking makes a huge difference. When I got my camera I did some tests. I shot some usual condition all the way down to f/22 and made 36" prints form them. The benefit of the added DoF certainly outweighed the impact of diffraction--sharpness is not what you see at 100% when comparing it to other conditions. I would urge you to test yourself to see the results, and by results, I mean going to a print.

Now, as much as I respect my fellow photographers, we all suffer from too much experience. We can see the defects in an Ansel Adams print from 200m. We have come to learn to recognize very small changes in the image. We obsess over individual pixels. But the audience does not see nor care about the minutia. What they see is a great, sharp image. I have never had someone walk up to me and ask about the diffraction in my print.

I hope that answers your question.
 

Landscapelover

Senior Subscriber Member
Much appreciate my friend!
It is a very thoughtful comment from someone who really knows about photography/publishing.
Again, my question is about the "print", not the "pixel peep".

Thanks
 
Last edited:

dick

New member
The definition Zeiss gives for the permissible circle of confusion is 1/1500th of the diagonal. This means if you view an image at half viewing distance, the plane of focus should be acceptably sharp--a 16x20 viewed at 12" is still sharp even at 150dpi.

MFD easily exceeds these numbers.

So, the CoC depends on the format diagonal. Sharpness is a function of the format size and viewing condition.
I see that several people agree with you... but some of us buy MFD to produce pictures bigger than 10 * 8 " inches, and which can be observed from distances closer than the diagonal... if not then why not use an iPhone?

Were big pictures, like 40 foot frescos, made to be observed from the diagonal?
 

Shashin

Well-known member
I see that several people agree with you... but some of us buy MFD to produce pictures bigger than 10 * 8 " inches, and which can be observed from distances closer than the diagonal... if not then why not use an iPhone?

Were big pictures, like 40 foot frescos, made to be observed from the diagonal?
I never said anything about limiting print size to 8x10. Basically, print size is not a factor and viewing distance closer than standard will work very well.

Comparing a fresco to a photograph is not very meaningful and really has nothing to do with the topic.
 

dick

New member
I never said anything about limiting print size to 8x10. Basically, print size is not a factor and viewing distance closer than standard will work very well.

Comparing a fresco to a photograph is not very meaningful and really has nothing to do with the topic.
You said:
So, the CoC depends on the format diagonal.
...and who could try to make us believe that... ¿except for a pro who is still trying to do professional photography with a 6Mpx camera?

(If you use a low res camera and look at the picture from far enough away, it looks not too bad,)
 

Shashin

Well-known member
Dick, you must be right! What an articulate argument--how did you know I was a Pro using a 6MP camera?
 

fotografz

Well-known 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
I think the specificity of "landscape photography" and wide angle lenses lends itself to Shashin's opinion to some degree ... where it is more about pictorial qualities of a viewed print ... the boundries of which may be somewhat subjective from photographer-to-photographer.

However, Dick has a point in general. In my experience, f/11 for MFD becomes the point of more careful scrutiny because of final usage. I strive for an optimal f/5.6 or f/8 and revert to T/S when I routinely push past it too frequently to get visual acuity front to back ... but I do not shoot landscapes.

I have done images that were enlarged 6 to 10 feet wide and used in trade-show booths where the images will be viewed closer than normal, even inspected for product detail. Or images that are massively cropped to show a detail.

Also have had images made into photo murals for corporate lobbies, although those aren't viewed at nose-close distances usually.

The other application of W/A optics is industrial inspection where details are important.

-Marc
 

Shashin

Well-known member
However, Dick has a point in general.
Really? Apart from his charm, he has not actually said anything. LOL

In my experience, f/11 for MFD becomes the point of more careful scrutiny because of final usage. I strive for an optimal f/5.6 or f/8 and revert to T/S when I routinely push past it too frequently to get visual acuity front to back ... but I do not shoot landscapes.
So, my point is it is subjective--determined by a viewer. You seem to be saying exactly the same thing.

And yes, I print big. I just did a 7x10 foot image for a museum.
 

fotografz

Well-known member
Really? Apart from his charm, he has not actually said anything. LOL



So, my point is it is subjective--determined by a viewer. You seem to be saying exactly the same thing.

And yes, I print big. I just did a 7x10 foot image for a museum.
Yes, I was agreeing with you, especially given the specificity of the OP's question ... but not dismissing Dick's opinion (blunt or not) because the applications he referenced are ones I have faced many times.

It is one thing to have a person gazing at a landscape at some normal distance where it can all be taken in, and a mechanical engineer inspecting a part on a 10' machine photo at a trade show, or a museum curator or art lecturer inspecting the brush strokes of a painting, or an art director cropping a 1" section of fabric from a 15" piece to show detail in a catalog, or some product shot with slightly fuzzy DOF when the retoucher goes to knock out the background, etc. etc. etc.

I also think it differs a bit lens to lens, but that is just an impression not a tested fact.

-Marc
 

Shashin

Well-known member
I really don't think people are reading what I wrote. First, I defined the area I am discussing--that is the creation of a pleasing image and in reference to maximizing DoF. I also do scientific and reproduction imaging and the underlying assumptions still hold true--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, nothing I wrote really changes except the emphasis on DoF.

Also, I am not simply saying this for standard viewing distance, but also for shorter than standard. The illusion of sharpness under standard models is robust, as I explain. You can walk up to a print and it holds together. If you are printing large, you certainly cannot maintain a 300dpi "native" resolution, if you follow some arguments about what CoC is required. Nor do you need to.

And more pixels do not make an image softer--the underlying assumption for setting the CoC to 100%. Take a sharp image and print it large. Resample the image to give it more pixels and print it the same size. The difference? Nothing. Putting more pixels on a sensor likewise does not make the resulting images softer.

Another thing that is being missed is the CoC defines an upper limit. An image will have CoCs smaller than that limit.

I don't think we are really disagreeing.
 

Jack

Sr. Administrator
Staff member
DoF is directly related to focal length -- the shorter the focal length, the deeper the DoF. Diffraction is directly related to aperture and nominal CoC, or in the case of digital imaging, the pixel pitch. Thus, like every other decision surrounding gear choices, there is a tradeoff of benefits and no free lunches... That said, generally speaking you will have greater apparent DoF with shorter lenses and smaller physical sensors, but other aspects of image quality will change/suffer in proportion. It's why MF looks different than FF DSLR, and FF DSLR looks different than M4/3rds when all are enlarged to the same size prints. Which looks best is largely subject and personal-preference dependent, though I have found most people's personal preferences seem to lean toward the "MF look" being "superior" -- even if they cannot clearly express the specifics of why they feel that way...

However, I don't think any of the above has anything to do with creating "the best" possible art. To my thinking, that is achieved whenever the artist clearly conveys their message with whichever gear they've chosen to use, period. In that, I have seen spectacular works produced with M4/3rds and equally crappy work produced with MF; I've also seen spectacular work produced with MF and crappy work produced with M4/3rds. It aint the arrow being used, it's the indian behind them putting them to use.

But that's me and I respect others opinions will vary.
 

Landscapelover

Senior Subscriber 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
 

torger

Active member
We love to debate about DoF definitions, which ends up on CoC size.

Originally DoF was defined something like this: "just sharp enough for a relaxed viewing condition for people with average vision". This is what Ansel Adams used and all other great photographers in the film era.

Then came digital. Tastes have changed.

We now like even sharper images, and the gear can do it and the prints are possible. I prefer, and I see many others do the same a new DoF definition along the line "the edge of the DoF is so sharp it is virtually indistinguishable from the plane of focus even at close inspection, after post-processing sharpening has been applied". That is if you focus at the hyperfocal distance the infinity should not look blurry even at 100% pixel peep.

According to my own experiments I find a CoC set to the larger of 2x pixel pitch and airy disk diameter is a good choice to match this definition, and since we use small apertures in landscape airy disk diameter is always the larger one.

So how small apertures can we use? Smaller than we did with large format film (relative to the film size)! The thing is that deconvolution sharpening is rather effective, a tool we did not have back in the film days, so we can blur down the image with quite some diffraction and sharpen it back up again to fine results. So how small can we go? It is hard to say something absolute since it is a matter of taste. With my medium format system I shoot f/11 when I can but go all the way to f/32 if I have to. If I used 8x10" film instead the equivalent apertures would be about f/64 for f/11 and f/165 or so for f/32, and as we know many of large format lenses are actually limited to f/64... so DoF has become easier with digital (smaller apertures, better sharpening), so I think it is perfectly ok to adapt the newer more stringent definition of DoF with the smaller CoC.

How you relate to DoF is also very much a personal style and also scene dependent, some like to let things go slightly blurry to give the image more depth. Concerning DoF tables and hyperfocal distances etc which I use when out shooting I prefer the new tighter definition of DoF though, I think it is a better more well-defined reference to relate to. When it comes to the technical aspect of landscape photography most of us are interested in making the image "as sharp as the gear and scene allows" rather than "no sharper than required for a relaxed viewing condition".
 
Last edited:

dick

New member
If you are printing large, you certainly cannot maintain a 300dpi "native" resolution, .
I am not an advocate of Pan-and-stitch, but it can give you 300 original camera pixels per print inch on 10 foot pictures... and move-and-stitch (photographing in sections) could do for a 40 foot fresco!
Take a sharp image and print it large. Resample the image to give it more pixels and print it the same size. The difference? Nothing.
This is correct if the raster image processor makes the most of the file. With high quality Multi-Shot files up-sampling might help, but, again it depends how well the printer software would do the same job.
DoF is directly related to focal length -- the shorter the focal length, the deeper the DoF. , generally speaking you will have greater apparent DoF with shorter lenses and smaller physical sensors,
Yes, Jack, but I think, technically, DoF is dependent on magnification ¿Which is, I think, what you are saying? and if we use a lens of double the focal length from double the distance, we get the same magnification and very similar DoF.

You get better apparent DoF with low-res cameras and cheap lenses, because the CoC is relatively larger, and if nothing is very sharp then slightly out of focus looks relatively sharp.
 

torger

Active 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
If one would use my model it works like this - the more you stop down the more diffraction you get, and then you should increase the CoC to mirror that it can be larger without noticing a difference from the plane of focus. CoC = Airy disk diameter I think is an excellent choice. Then you just feed this into your hyperfocal formula and get the focus distance and near limit. The larger f-number the nearer limit but more diffraction to deal with in sharpening.

Example using CoC=Airy disk diameter for 55mm

f/11: 17.8 meters hyperfocal distance, near limit half of that 8.9 meters
f/16: 8.9, 4.5
f/22: 4.5, 2.3
f/32: 2.3, 1.2

unfortunately I did not have a table in feet, but multiply with 3 you get an approximate. As you can see for each stop you increase the hyperfocal distance is halved.

This will give you so sharp images that if you nose your print you cannot see any difference between the plane of focus and the DoF edge, which is usually want one will want if you have something important at infinity.
 

torger

Active member
You get better apparent DoF with low-res cameras and cheap lenses, because the CoC is relatively larger, and if nothing is very sharp then slightly out of focus looks relatively sharp.
Exactly. If the small camera would have equal number of megapixels and be able to fully resolve them there would be no difference. Larger sensor + longer focal length can in full be compensated with smaller aperture (unless we hit an upper limit of the lens which can be the case for large format film) at the cost of longer shutter speed. Yes we get more diffraction, but it is a zero-sum game, 20 megapixels on a large sensor can accept more diffraction than 20 megapixels on a small one.

Appearent DoF concerning viewing condition is dependent on sooo many factors. A worst case: detailed landscape scene printed big hanged in a gallery at comfortable viewing height, people will in addition to see the whole picture walk close to look at the details. A best case: a portrait hanged above a fireplace - you can't come close and you would not want to either, portraits are preferably viewed so you can see all at once.

Due to this I have dropped the traditional large CoC definition and instead use tables with small CoC (=Airy disk diameter), then I know where it is "critically sharp", and when the scene is such that I can or should relax it I do, but I think it is much easier to have the "critically sharp at close inspection"-DoF as reference. The basis for this to work though is a camera where you can stop down much and sharpen well, which applies to all digital cameras but not to large format film.
 

dick

New member
...if you focus at the hyperfocal distance the infinity should not look blurry even at 100% pixel peep.

According to my own experiments I find a CoC set to the larger of 2x pixel pitch and airy disk diameter is a good choice to match this definition, and since we use small apertures in landscape airy disk diameter is always the larger one.
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.
 
Top