the myth of DIFFRACTION

[Markvetnz]

What a condescending, pretentious wanker.

[MilT0s]

Lot of discussion on this topic. My thoughts have already been expressed by others.

First of all I find very very difficult for people with degrees in photography to have deep understanding of what diffraction is. Why? Because it is a complicated matter of science so unless you are a physicist or at least have some advanced mathematics knowledge it is very difficult to follow the mathematics formulas of diffraction. I am a PhD student in photonics and it is always very difficult to explain physics students what diffraction is without using the formulas. Obviously photography students learn many technical issues but only in the context of what is needed to improve their photography skills. They are not optic systems designers after all!

Now does diffraction affects the quality of the photograph? Of course it does but as said by others it is not a major factor (unless we have to do with microscopy photography!). We can see diffraction near the edges of any obstacle. When using a small aperture the edge/opening area ratio of the diaphragm becomes large and the problems occurred by diffraction more obvious. It also depends on the pixel size of the sensor. The CiC tutorials on this are really excellent. Both simple and advanced, clearly written and extremely informative.

But really who cares about the scientific details of diffraction? That knowledge won't make me a better photographer. DoF is very important for photography so I need to know how the aperture affects it. If I want to make very large prints is also useful to know the sweet spot of my lens, usually about f8-f11 as shown in the graph, because for various reasons (diffraction included) my lens performs better at these openings. I don't think that anyone considers the diffraction formulas when taking a photograph in the field. So, apart from my interest in science, that is just what I need to know to (try to) make some great photographs.

[Simon Garrett]

Mitos,

I agree with what you say. Diffraction definitely is an issue for those that obsess about sensor resolution (mega pixels). Now there's a whole new topic for debate. But those that think they can stop down a D800 to f/22 and still get an image resolution limited only by the 36M pixels of the sensor...

[rhadorn]

Even in that situation, I would suggest that it's not a big deal because as the print size increases so does the viewing distance - and of course the greater the viewing distance, the less detail that our eyes can resolve.

I have also seen wonderful pictures which were totally out of focus. This could also be a good argument against caring for limiting diffraction.

I don't share your viewing scheme. I don't look at big enlargments from a greater viewing distance: I look at pictures at various distances, going closer, retreating further away... I do the same with paintings, sculptures...

The importance of caring for detail also varies with the style of the pictures you take or want to show.

The objective impact of diffraction can be measured and compared to other limitations of lenses. But this is not the only issue. The importance of limiting diffraction is also a matter of practice, depending on the style of pictures you like to take, how you look at pictures. A photographer can consider a marginal benefit as important to him. If limiting diffraction seems to be a matter of opinions, it is only because the presuppositions are never made explicit - or because the experts take their presuppositions as universal. The importance of limiting diffractions is not only a scientific matter, but a matter of jugement, as is very well shown in the citations proposed by Glenn.


Why are there so many people around who absolutely want to be the (only) ones who are right?

Reto

[Colin Southern]

I don't share your viewing scheme. I don't look at big enlargments from a greater viewing distance: I look at pictures at various distances, going closer, retreating further away... I do the same with paintings, sculptures...

And if that works for you then that's great.

The point I'm trying to make though is that I believe that there are far far far far too many folks giving far far far far too much weight to issues like diffraction / pixel counts / ISO noise etc, and far far far far too little attention to things like lighting & composition. As a result, they usually end up with very mediocre (at best) images because they're putting all of their efforts into "saving the pixels" and in the process "ruin the image".

So we see things like:

- Images with camera shake because they didn't want to use a higher ISO because of the "noise"

- Images with insufficient DoF because they didn't want to stop-down the aperture because of the "diffraction"

- Images with poor lighting because the photographer spent thousands on a camera with a ridiculously high pixel count (thinking that that was important for image quality), and then can't afford a decent flash unit.

I'm afraid that when I read "reviews" that start analysing camera luminance noise & diffraction and lens sharpness in 400% crops of test targets pretty much the ONLY thing that comes to my mind is "these folks need to get out more - stop worrying about these things - and go shoot some images in the real world". Personally, I'd far rather just enjoy the photo of the F18 flying under the bridge than to fret / worry about the amount of chroma noise in the shadows of the rivet holes!

[Glenn NK]

I not only found Colin's last post useful, I got a chuckle from: "chroma noise in the shadows of the rivet holes".

How often do we miss the forest because the trees got in the way?

A suggestion for landscape shooters: read the third paragraph of post number 17 that I attribute to Tim Fitzharris - a superb landscape photographer. He knows better than to go "bonkers" over resolution, dof, etc.

G

[Colin Southern]

I not only found Colin's last post useful, I got a chuckle from: "chroma noise in the shadows of the rivet holes".

How often do we miss the forest because the trees got in the way?

A suggestion for landscape shooters: read the third paragraph of post number 17 that I attribute to Tim Fitzharris - a superb landscape photographer. He knows better than to go "bonkers" over resolution, dof, etc.

G

In my mind, the likes of diffraction / pixel counts / SNRs are nothing more than a distraction. If folks would stop worrying about them and start focusing (pun intended!) on lighting / composition / preparation / processing they'd probably improve their photography ten fold.

I think the title of this thread should be "The Myth that Diffraction Matters in real-world Photography".

[Markvetnz]

Its great to see a modicum of common sense in this debate. Obviously diffraction exists and can affect images. It is easy for a physicist ot mathematician to prove this fact. But like a few others, I don't think it should make an ounce of difference to Joe Public trying to take a good image. In my opinion the world is made up 98%ers who are quite happy to look at an image and accept it for what it is. Then there are the 2%ers (and there is nothing wrong with them or their opinions) who tend to be more obsessive and fastidious about fine detail that the other lot don't even know exists.

I also happen to know somebody who is a 0.0002%er and I can tell you it drives me and every one else nuts.

[Polar01]

One of the most imteresting things of note in all this discussion is this burning question. Where is Tresise in all this talk, did she/he or he/she take a hike or what. Me thinks they are hiding under the chorma noise in the shadow of those rivet holes of an F18.

Cheers:

Allan

[Markvetnz]

One of the most imteresting things of note in all this discussion is this burning question. Where is Tresise in all this talk, did she/he or he/she take a hike or what. Me thinks they are hiding under the chorma noise in the shadow of those rivet holes of an F18.

Cheers:

Allan

Me thinks that we are perhaps no longer worthy of having pearls cast before us. Ignorant infidel that we are.

[xpatUSA]

In my mind, the likes of diffraction / pixel counts / SNRs are nothing more than a distraction. If folks would stop worrying about them and start focusing (pun intended!) on lighting / composition / preparation / processing they'd probably improve their photography ten fold.

I think the title of this thread should be:

"The Myth that Diffraction Matters in real-world Photography".

I would possibly agree if I knew what "real-world Photography" is. Would anyone be able to define it?

Is the real world one of sweeping panoramas, gorgeous sunsets, speeding cheetahs, horse shows, impressive architecture, weddings, etc - perchance to the exclusion of all things macro?

Genuinely curious because I see the phrase "real world Photography" quite often while trawling around the fora.

Ted

[McQ]

Its great to see a modicum of common sense in this debate. Obviously diffraction exists and can affect images. It is easy for a physicist ot mathematician to prove this fact. But like a few others, I don't think it should make an ounce of difference to Joe Public trying to take a good image.

Absolutely. I also wanted to add a bit about the original motivation for this article since it was first put online (over 5 years ago!). At that time, the megapixel race was in full-swing, and people were increasingly concerned with whether their camera had 8 megapixels vs. 10 megapixels of resolution. The point of bringing diffraction into the spotlight was that nitpicking over megapixel increases didn't really matter for an increasing range of photos -- even if one were using perfect technique. Other fundamental and unavoidable factors would always limit resolution. For that reason, it needed to delve into the physics in order to really make that case. Understanding all of that certainly isn't necessary for standard photography, but many find it interesting, and want to know the "why" in addition to the "how" with cameras. That's the group it was targeting.

As has been discussed, the vast majority of mainstream photos wouldn't be any different if diffraction were or weren't known. Too often people lose sight of other more important considerations:

1. Subject matter that achieves one's artistic goals (interesting/evocative composition, lighting, etc.)
2. Camera settings to make the most of that goal (focal length, depth of field, etc.)
3. Proper sharpness technique (focusing, tripod use, post-processing, optimal aperture, etc.)

Diffraction only comes into play near the end of the considerations within #3; everything else is almost always more important.

That all being said, diffraction absolutely matters for those cases where one wants to push the limits of detail in a large print -- such as with landscapes, panoramas and intricate architecture. For example, f/22 and above will become visible** in an 8x10 inch print on a DSLR with a crop factor. At f/32, this can even make your DSLR camera the equivalent of roughly 2(!) megapixels. In other words: if you have a photo that needs to be printed large AND you're not in need of more depth of field, then using too high an f/number is ill-advised. That's really all one needs to take away from this.

One last thing to note is what this all means for lens purchases. Once one knows that diffraction is unavoidable and sets in at small apertures, this makes it easier to understand why expensive L-series and other lenses only really improve image quality at large apertures. At f/16+, one gets very similar image quality regardless of the lens price -- at least in the center of the frame.

**according to the standard definition for depth of field, which is a third of 20/20 vision

[Simon Garrett]

I very much agree with the recent comments.

So just remind me, the reason we need 36M pixels is...?

I know the mere 12M pixel in my D300 is now next to worthless, I read it all the time. It's just that I seem to remember the same people saying 12M pixels was really great 5 years ago.

[Colin Southern]

So just remind me, the reason we need 36M pixels is...?

Um, to sell cameras to folks who have been lead to believe that the number of pixels is one of the most important factors in creating great images?

[jcuknz]

Another reason why some need 36Mp is so that they can point and shoot and then crop in editing to the picture they want
Why when I got my 5Mp camera I felt a degree of relaxation from shooting with my 3.3Mp.

[rhadorn]

That all being said, diffraction absolutely matters for those cases where one wants to push the limits of detail in a large print -- such as with landscapes, panoramas and intricate architecture. For example, f/22 and above will become visible** in an 8x10 inch print on a DSLR with a crop factor. At f/32, this can even make your DSLR camera the equivalent of roughly 2(!) megapixels. In other words: if you have a photo that needs to be printed large AND you're not in need of more depth of field, then using too high an f/number is ill-advised. That's really all one needs to take away from this.

One last thing to note is what this all means for lens purchases. Once one knows that diffraction is unavoidable and sets in at small apertures, this makes it easier to understand why expensive L-series and other lenses only really improve image quality at large apertures. At f/16+, one gets very similar image quality regardless of the lens price -- at least in the center of the frame.

Thanks Sean for making your statement once more in a very clear manner.

In a thread dealing with one single technical factor, it is pretty out of topic to argue that there other factors. It is not very helpful to present photographers as dealing with unimportant things like pixel counts, noise or diffraction (of course, 'other photographers', non oneself). If all these matters were meanlingless, Sean could bin a good half of the entries in the CiC tutorial. Photographic technique is the search of the best compromise between contradicting factors: it does not help to point to underachievements as the result of a concern with a single technical issue.

To achieve the best, the photographer has to master technical matters, even where their effect is marginal. But at the end, he has to make choices, his choices. This in turn does not mean that technical matters are the only aspects in photography - which could be the topic of one or several other threads.

Reto

[arith]

Another reason why some need 36Mp is so that they can point and shoot and then crop in editing to the picture they want
Why when I got my 5Mp camera I felt a degree of relaxation from shooting with my 3.3Mp.

When I upgraded from 6 to 15MP I thought great now I can relax and crop away. But instead I learned to use all the frame and all the pixels, composing and choosing exposure to get as much information as possible.

I can calculate dof but rarely go outside my diffraction limit, but have done a couple of times for instance getting something a metre away sharp falling off DoF over 50 metres away.

Then mostly I blur it using Gaussian blur or Orton; but I have a choice and I have an A1 pin sharp print, and numerous A3+ ones including a couple of really good ones. Most of the extra fine resolution is achieved using expensive software but I have files full of useful info to start with.

AS for CA; my Tamron lens supposed to suffer with over 1px at 17mm and all ap[ertures, but it isn't that bad. However, by stepping back after reading about magnification from 2 to 4 times distance, I can get the same DoF with a different perspective, larger background, but with negligible CA at 35mm.

Still if I have a high contrast aberation, I can just desaturate the edge and I can't see it in a print. TA

[Glenn NK]

Um, to sell cameras to folks who have been lead to believe that the number of pixels is one of the most important factors in creating great images?

And some seemingly very good photographers are falling for it too. What was that famous quote supposedly attributed to PT Barnum?

Every once in a while, one must get re-grounded with reality, because the step from the sublime to the ridiculous is often not noticed.

This thread seems to have steered us back to reality.

Glenn

[Colin Southern]

Another reason why some need 36Mp is so that they can point and shoot and then crop in editing to the picture they want

Although ...

- If the image is only for internet display - and at a typical (and generous) 1200 x 800px - then we're already able to throw away around about 95% of the captured data with a 21MP camera.

- Although I do this myself (and to a not inconsiderable amount) (although in my own defence it's more to do with changing the aspect ratio), one might present the argument that using a higher MP count so that one can be "lazier" with compositions probably isn't "best practice".

[Colin Southern]

To achieve the best, the photographer has to master technical matters, even where their effect is marginal.

This is certainly a driving force for some. Personally, I feel it's a personality trait that walks a fine line between "perfectionism" and "Excessive/Compulsive Disorder" - and I don't say that whilst "looking down my nose" at others, as I very much walk that line myself every day.

Having just made that confession, I might add that I have to fight hard not to cross that line; it's often been said that photographers never really finish an image - they just abandon it at some point (and I can relate to that). What I've had to learn is that - in the real world - nothing is ever perfect and that THAT'S OK. If an image has a bit of diffraction because it was shot at F22 when F16 would have done the job - and it's not as perfect as it could have been (but still looks great) then THAT'S OK.