About Diffraction

For anybody who can help a young lad...This question is about something I have never quite been able to grasp on my own. I understand how to properly select the optimal F-stop according to Mr. Luong, but I still don't know how to quantify the affects of diffraction. Here is an example. Let's say I photograph a scene at F22 and again at F64. I understand that the one shot at F64 is going to have a greater depth of focus that will be softer overall...even at the critical focus point, but is this going to be visible to the human eye at the original 8x10 size. If not, at what enlargement factor will the F64 shot be noticeably softer at the critical focus point than the F22 shot. Basically, what I am asking is if I shoot the same scene at F22 and F64, and I inspect just the critical focus point (not the depth of field limits), at what enlargement size will I notice a difference in sharpness when viewing the resultant prints at the normal minimal viewing distance of 25cm. By the way, I am shooting with an 8x10 and a Nikon 300mm 5.6 lens. Thank you all.

The way to be really sure about this stuff is to learn how to do the calculations yourself. However, most DOF tables are based on a 10x magnification. You need to figure out what stop is the diffraction limit for your set up. Beyond this stop (say, f/32 for a particular lens on 4x5), stopping down actually reduces overall sharpness.

By far the best way of making judgements about this stuff is to shoot some tests.

There are a number of people who can give you better answers regarding diffraction limits than I, but let me complicate your question, Chad. Assuming you're talking about images made out in the field with your 8x10, there are several other practical factors that may be more important than diffraction. That would include things that move your camera (wind, and surface vibration from nearby roads, etc.) and things that move elements, like trees, within the subject (wind mostly). As such, there is benefit to considering all of these aspects to arrive at the optimal shutter speed and aperture based on the circumstances under which you're shooting each individual negative.

With the shallow depth of field you get with most 8X10 lenses, you really need to stop down if you want the whole scene in focus. It is far better to stop down a bit too much and get everything in focus than to worry about diffraction, not stop down enough, and have parts of the scene be out of focus. Shooting at f22-45 will be the norm. Although most lf lenses have a "sweet spot" (where the lens is sharpest and diffraction is minimal) at f16-22, and stopping down past it will result in slight loss of sharpness due to diffraction, this will hardly ever be visible in real world situations at f32-64 unless you are making huge enlargements. Certainly at the original 8X10 size no difference will be visible to the unaided human eye. I have a shot of the Utah canyonlands made on 4X5 film with a 150 mm lens at f45 blown up to 20X24 and it looks extremely sharp to me (certainly as sharp as prints made from transparencies shot at f16-22). I wish I had stopped down all the way to f64, though, as a small area of the photograph is slightly out of focus!

It has been my exerience that there are many more important issues that lead to fuzzy results than small apertures. My major problem is not looking at all regions of the GG with my loupe. In short not focusing or getting the movements correct (same thing). I've trained myself to look at all four corners of the gg. Not locking down the movements causes the same effect. Next problem is camera movement due to wind or shutter or floor movement. Unless you are making bill board size prints and have solved the first two problems you are never going see diffraction effects. I did some tests once to see if I could see diffraction effects. I set up some USAF targets and shot them from F16 to F64 on TMX100 and developed in HC110 dil B for 7 min. I then looked at them with a 50X (combined objective and eyepiece) microscope. Sure enough you could see the slight fuzzyness. It is kind of like a very slight veil over everything. So compared to problems 1 and 2 above I would just use F64 when in doubt or you don't have to kill movement with short shutter speed or need some light on the film. Or maybe you only want one thing in focus. It is ART after all....

K

Chad, some good responses.... of course, testing is best, but sometimes it takes a little bit of simple math, to figure out how much film to waste :-)

Your question concerns only the "plane of sharp focus". I will address this, but you really also need to be concerned about the near and far limits at f64...but anyway, if you look at the diffraction tables, or just use the simple formual, 1500/fstop = max. lp/mm lens can deliver.

1500/22 = 68 lp/mm

1500/64 = 23 lp/mm

Now, keep in mind, this is only the "aperture diffraction" limit, but there are other forces at play here, such as gg / film alignment, shake, other forms of diffraction, MTF curve of the lens, (i.e. how it performs at that given f stop at your focus distance) type of film used, etc. So this does not become overly technical, lets assume you loose another 20% for the above factors, which brings your lens resolution to 54 and 19 lp/mm. Now, if you are using Provia 100F film at the tested contrast ratio, it will resolve up to approx. 55 lp/mm, so you run these two values through 1/R formula, which is the net effect of what the two can deliver together...

1/(1/r1 + 1/r2...)

Where r1 = lens max. lpmm delivered rez., r2 = max resolving capability of film at given contrast ratio.

If you had other variables in the optical train, you could add them also, such as filters = r3)

So, at f22 = 27 lp/mm at plane of sharp focus

at f64 = 14 lp/mm at plane of sharp focus.

Now, pick the lp/mm the human eye can resolve...a good eye can discern up to 5 lp/mm (not neccearily resolve) Average human resolution is closer to the 2 - 3 lp/mm. Using 5 lp/mm, you can enlarge the f64 shot 14/5 or about 3x before the difference will become obvious vs. 4x or 5x enlargement. And at f22 you can enlarge at 27/5 = 5.5x. This seems pretty decent, right? I would say these numbers are fairly accurate for you example, but in the photographic world, pretty accurate is +/- 50%. Just keep in mind, unless you're shooting a brick wall, or at infinity, it's the near / far subjects that can destroy the image, not at the plane of sharp focus. This assumes you can not use lens tilt to perfectly align the plane of sharp focus with the scene. Make sense?

For 8x10 contact prints, I wouldn't worry about diffraction. I shot a 480mm Ronar at f/256 just to see what it did, and the 8x10 contact print looks quite sharp to the naked eye. Insufficient depth of field is very noticeable, and objectionable in certain veins of photography, (not so in others...)

Like Chad, I've also wondered about this very topic off and on over the last number of years. However, in practice, I've just gone with what Brian and Kirk have stated.

I shoot close-ups (indoors) using a 4x5 monorail and 240 f5.6 Nikkor and, usually, depth-of-field is quite a challenge. So, shooting at f32 or f45 isn't all that unusual! And, although I don't like to... I've even gone f64 on the rare occasion.

If I'm not enlarging to large sizes (16x20 or 20x24 being the largest I usually go)... I'd rather have the depth of field and use the smaller f-stops required to get as much of my subject matter in focus as possible.

.

wg,

Great explanation...

Any more applicable comments for folks using 4x5? :)

Cheers

.

" However, most DOF tables are based on a 10x magnification"

10x?

Linhof tables as well as the Rodenstock DOF Calculator/Scheimpflug Calculateor are for far less magnification.

Did you mean 1:10 perhaps?

Thank you all for your thoughtful answers. They have been helpful. I understand all the other factors in getting a sharp negative and I am not ignoring the importance of the depth of field limits, but my question's intent was to single out when sharpness loss due to diffraction becomes apparent at different aperture sizes. I used the critical focus point in my example because that would be the most telling reference point from which one could compare diffraction effects from various apertures at different enlargement sizes. I agree that doing my own tests would be beneficial, but I first wanted to understand the principle, and have a good idea of where to start since testing equals time and money. Thanks again to all of you and a special thanks to wg for his insightful response. Thanks.

Chad

Same analysis for 4 x 5?

Indeed, as far as diffraction is concerned, the degree of enlargement before seeing a difference doesn't depend on the format. So wg's remarks would still apply for 4 x 5. The only way the format enters into the discussion is in the degree of enlargement necessary to produce a given size print. Assuming the same size print, you need to enlarge about half as much with 8 x 10 as with 4 x 5, so you are able to stop down further before the effect of diffraction would be apparent.

Let me quibble about a couple of other points.

First the reciprocal rule is not the only way to combine resolutions. Many people prefer to use the root mean square method: square the reciprocals, sum them, take the square root and then take the reciprocal of that. In reality, all these methods must be considered rough rules of thumb. None of them has any justification from optical theory. The only way to do it accurately is to multiply MTF functions for the different components of the optical system, but that is much too hard to do routinely. Jacobson has an example of such an analysis in his Lens Tutorial, which can be found at the photo.net website. Hence, you should only use such rules as a rough guide. In critical situations, you have to test and see what happens.

Second, do you really want to base your decisions on what the print looks like from 25 cm? Some people do have an irresistible urge to get their noses right up against any print, no matter how large. But, if you think about it, that places a severe upper limit on the size of the print, even if you use an 8 x 10 negative. Normal viewing distance for a print is usually taken to be the diagonal of the print. So an 8 x 10 print might be viewed from about 12 inches and a 16 x 20 print at 24 inches. As you get further away from the print, you can tolerate progressively less resolution, and if an 8 x 10 print looks sharp at 12 inches, a 16 x 20 print, made from the same negative, will look sharp at 24 inches. An 8 x 10 print will not show appreciable diffraction effects even at f/64, and it follows that any larger print, viewed from the appropriate distance will also not show diffraction effects at usable apertures.

A small peeve. There was lots of helpful advice, but I had to go to the sixth response to find one that actually answered the question. This seems a common occurence. Someone asks a sensible question, and people answer by explaining why the question is foolish and why he/she should be doing something else. I just did it myself in the above paragraph. I don't object to that kind of advice, but perhaps one should also make an effort to answer the question first.

Now that the question has been answered, a few tid bits to tidy this up.....

Chad, glad we got you in the ballpark...and as I have mentioned, and Leonard supported me, in the ballpark is at best + / - 50% , more on this later......

Henry, I think Leonard answered your 4x5 question.... All things remain equal, IF, you open up 2 stops for each format, (as its half the fl) which then gets eaten up by the extra 2x enlargement required to get back to the same final size, this is at the plane of sharp focus.

Leonard, as for the reciprocal method vs. root mean square method.... agreed, i have quibbled with both myself..and in the end, they produce relatively similar answers, so not a huge deal, I stick with 1/R as its supported by Fuji, and it's their film I use :-) But regardless, since the purpose is only trying to land in the ballpark, it's not that signficant.

Also, Leonards comments on viewing distance are VERY important consideration often overlooked, in my example, I assumed the closest focus distance, whatever that is for the viewer, 10 - 18" based on how close the person can focus. But as a general rule, you double the viewing distance, you can half the resolution and have the same net effect of apparent sharpness. This is why 35mm film can be used to shoot billboards. The ideal viewing distance is often the image diagonal. But that does not stop someone from walking up close to a 60" print, it's just human nature to inspect even further.... specially in this high tech era where everyone has a mangifying glass on everything, computing speed, storage capacity, Mega Pixels, etc. etc.....

As for the application of this math to photo optics..... when i started in photography, I too was fooled into the application of otpical math to determine outcomes. And although there is some very basic and fundamental uses for the math, which can at least get you in the ballpark, there is rarely a precise means. As Leonard pointed out, having detailed information on the lens MTF data and the film MTF data is the closest you can get to achieve a "non tested" outcome. However, rarely if ever, in the photographic world do these two pieces of information exist complete enough form, to arrive at such a conclusion. And even then, it's not quite clear how to combine the two to get the optimum outcome, although there is several theories on such. However, the MTF theory is the ONLY method that ties together YOUR lens and YOUR film, and although it does not include many of the other factors in the image taking process, it does give you some ceilings to work with.

The application of optics math to generic lenses, well, this is where things fall apart. If you take the standard Depth Of Field equation to find out your focus point, so the near and far acheive a certain resolution....well, now IMO, you have enter voodo optical math. As for the formula, it doesn't know if you are using the sharpest lens made with the sharpest film made of the modern era, or 1800's vintage lenses and film.... it does not know if you are using a lens optimized for close up work, but yet shooting at infinity. It does not know if you are shooting at an f stop the lens was NEVER designed to be used at, as all lenses are designed to reach their optimum performance over a 2 - 4 f stop range. So what do you end up with? At best, you end up with some theoretical value that is often useless, and this is where testing becomes mandatory. The DOF tables or scales used with certain cameras were sometimes created with testing, and sometimes theory, different enlargement factors were considered, different films were used throughout different eras, etc. etc... hence why this has become an ever changing and evolving area of photography, but fortunately for the better.

Although I was quite befuddled without the joy of pluggin a few numbers into a spreadsheet and seeing the outcome of a given photographic situation, I have finally come to accept it... very painful indeed.... but, now I work in a reality based world.... :-)

Leonard, wg...

Thank you... I appreciated your responses. :)

Cheers