How is lens resolution effected by aperture?

[brett_kosmider]

Without getting too deep into MTF or double-line resolution stuff, generally, what

aperture offers the sharpest resolving power on any given medium format lens? Or

more to the point, which apertures should be avoided if the sharpest image is

desired?

 

I have always heard that when shooting 35mm you should stop down from f22 since

it doesn't provide the sharpest image, and some have even told me that the sharpest

aperture on any given 35mm lens is in the f8-11-16 range.

 

Is this the case in medium format? I have since dumped my 35mm gear for a Mamiya

RB ProS. Right now I only have a 90mm C lens but I'd like to get a 50mm C as well as

a 180mm KL at some point. I will primarily be using the 50mm lens and I have heard

that it isn't the shrapest - should I avoid shooting that lens at f32?

 

Thanks

[jbq]

I'd avoid shooting at anything higher than f/16 unless you know the lens to be really poor.

 

At f/19, diffraction robs you of 50% of MTF at 40lpmm - this is huge.

 

Some excellent 35mm lenses "peak" as early as f/4. Some "less excellent" ones, need to be

stopped down all the way to f/11 or even f/16, and at such apertures diffraction is a

definite issue.

[john_banister1]

I've read a rule of thumb that most lenses perform best at about two stops down from wide open. In practice, other considerations such as the depth of field I want for the image, or the aperture corresponding to the necessary shutter speed to stop motion caused by wind that just came up often take precedence.

[q.g._de_bakker]

Brett,<br><br>Maximum obtainable resolution approximately halves (!) every two stops a lens is stopped down. So in a lens free from aberrations, best resolution is produced wide open, and stopping down will always reduce resolution. Always.<br><br>However, lenses that are free from aberrations are rather rare, and while stopping down will still reduce resolution, it will often also reduce the effects of aberrations on resolution. The net result will be first a (slight) increase in resolution and then a steady decline with each stop the lens is closed.<br><br>"Common wisdom" says that the best balance between reduction of aberrations (= increase in resolution) and diffraction (= decrease in resolution) is achieved with the lens stopped down about two stops. This however is by no means always so, and depends on the design of the particular lens.<br><br>And yes, stop a lens down too far, and maximum resolution even drops down below the "level of acceptable unsharpness", a.k.a. depth of field. Depth of field will be enormous, but nothingin the picture will be really sharp. So do indeed avoid stopping down more than absolutely necessary.

[david_holland1]

Light bends when it passes an edge--a phenomenon known as difraction. As the aperture get smaller, the circumference of the aperture relative to its area gets smaller, hence the effect of difraction on the image increases.

 

Difraction decreases resolution, so as the percentage to area ratio increases, the effect increases. That is, the smaller the aperture in absolute terms, the greater the difraction effect.

 

Aperture is expressed as a function of focal length--namely the focal length divided by the diameter. That is, at f/4, the aperture of a 50mm lens is 12.5mm and at f/8 it is 6.25mm.

 

Therefore the theoretical difraction effect (holding constant the other variables as described by Q.G.) is the same for a 50mm lens on 35mm (a "normal" angle of view) and 6x7 (a super wide angle) if the aperture is the same.

 

When angle of view is the same--that is, a 50mm in 35mm as compared to a 100mm or so on 6x7, the aperture doubles in diameter (f/4 is 12.5 for the first and 25mm for the second). Therefore, the difraction effect is less for the 6x7 than for 35mm for the same f/ number, although as you know the same relative amount of light gets to the film when the f/number is constant.

 

And a good thing, too, because you're going to need the extra depth of field in medium format.

 

While the above posts are most useful in understanding how an individual lens will perform, in general you should expect to use an f/ stop or two smaller (but really the same absolute diameter of aperture) with medium format compared to 35mm.

 

Maybe that's why medium format lenses have smaller minimum apertures than 35mm, and 35mm smaller than "designated" digital lenses.

 

Consider that, for example, at the 9mm shortest focal length of the zoom on an Olympus E-10 at its minimum aperture of f/11, the absolute aperture diameter is less than 1mm, while about the same angle of view in 6x7 (65mm?)is 6mm. A 1mm aperture in the latter would be approximately f/64.

 

All this is not meant to take anything away from the excellent explanation by Q. G. on the relatinship between resolution and difraction, only to indicate that for any given f/stop and angle of view, as format increases theoretical difraction decreases.

[christopher perez]

As you by now realize, your Mamiya RB surpasses your old 35mm for image quality in large part due to the image area the 6cm x 7cm provides.

 

On to the neurotic fine points:

 

In the practical world, the finest grained film resolves around 120 lpmm to 140 lpmm. f/11 diffraction limits at 123 lpmm and f/8 diffraction limits at 174 lpmm. Regardless of the format (35mm, 120, or 4x5) if the lens can resolve this then f/8 and f/11 are your best/sharpest apertures.

 

In 120 format work, the best lenses will record the films resolution limits at all apertures through to f/11. I tested this with my cameras using a USAF Resolution Test Chart. My c.1956 Rolleiflex Planars and Xenotars (f/2.8 and f/3.5) do very very well. My Mamiya RZ optics are consistantly fine. My Mamiya 7 80mm L is simply the best.

 

For your Mamiya RB optics, photograph at f/8 and f/11. Avoid, if you can, using f/32 as diffraction limits resolution to 42 lpmm. Particularly if you make large prints using traditional photographic methods. If your final image size is somewhat small or if you perform digital manipulation on a scanned image after the fact, then just about any aperture would be OK to use.

[brett_kosmider]

Christopher,

 

Thanks for the response. That was the practical info I needed.

 

Brett

[david_henderson]

Can anyone say how reducing resolution from diffraction as you

close down trades off against reduced resolution from

inadequate dof as you open up? This is a dilemma I face

hundreds of times in a year without having much of a clue about

how best to resolve it. I generally settle for achieving the right

dof because I can see it TTL whereas I can't observe the effects

of diffraction before the event. That doesn't make it right though.

 

So is it best for overall sharpness to favour dof , or would I get

better results overall by favouring wider apertures and taking

more of a risk on dof?

[jbq]

David: Ken Rockwell claims http://www.kenrockwell.com/tech/focus.htm that the optimal sharpness for a given image depth is achieved when the diffraction spot is the same size as the circle of confusion. This has the advantage of not involving any arbitrary value in finding what's an "acceptable" circle of confusion.

 

I very slightly disagree with his methodology:

 

-In a perfect world, illumination inside a perfect diffraction spot isn't as even as illumination inside a perfect circle of confusion - I would choose "same 50% linear MTF" as the rule, instead of "same size".

 

-In the real-world, apertures aren't perfect circles, and diffraction through such an aperture results in a larger spot.

 

-In the real-world, some lenses benefit from being stopped down fairly far - the size of the aberrations also gets smaller with smaller apertures.

 

Since those factors counterbalance each other and since some are very hard to quantify, I agree with his numerical results.

[brett_kosmider]

David raises a good point, and I thought of this as well - photographers like Adams

who strove for incredibly deep DOF and even coined a group called f64 (I'm

paraphrasing my brain, not 100% sure if I'm correct, but something to the effect)

where DOF was obviously paramount to anything else.

 

Lanscape photographers in particular (and I am one of them) put a premium on deep

DOF and strive to have every element of a scene in focus. So how to strike a balance

between the sharpest image and the desired DOF - shoot everything at f22 (on a f32

lens)? Or is it a more critical management of DOF, where, say, my nearest subject is

10 feet away and at f16 I can achieve a DOF of 10 - infinity?

 

Though I must admit, I rarely shoot that way and I often have subjects as close as 2-3

feet all the way back to infinity, so maybe I answered my own question - maybe DOF

takes priority in my photography over overall image sharpness. I'm still glad to be

reaping the benefits of 6x7 over 35mm.

 

BK

[christopher perez]

To get that St. Ansel f/64 look of having everything from the foreground to infinity in focus you'll need front and rear tilts/swings. For this you might need a 4x5 view camera. [grin]

[brett_kosmider]

Christopher, yes, this I realize (4x5 is the REAL answer here) but for the rest of us...

What is the answer?

 

There are more than a few notable landscape photographers out there who use

medium format. Notably, James Kay. He shoots a Pentax 67. I wonder what his

reasoning is - sacrifice overall image sharpness for greatest DOF, or balance the two

out?

 

I wonder...

[dan_fromm1]

Brett, you've asked us what YOU should do.

 

The nice people who're responded have told you what THEY would do. The problem with their answers is that none of them knows how you see the world, what your objectives are, or what compromises you're willing to make. YOU know all that, WE don't.

 

The only way you'll find out what balance between sharpness in the plane of best focus (goes away as aperture decreases) and sharpness in depth (increases for a while as aperture decreases, then goes away) suits you is to put the question to your gear.

 

Why don't you do some trials? Try on relatively shallow and very deep subjects. Shoot over a reasonable range of apertures, e.g., f/8 - f/32. Print the negs at a range of sizes. When you're done, you'll know roughly how large you can print and meet your standards (we don't know them) given aperture and subject type. And then you'll be free to get on with making images instead of agonizing over technical constraints.

 

Remember, when you have a question about how to reach your objectives, only you can answer it satisfactorily. Also remember that great images have been made with lenses you think are too fuzzy to use.

 

Cheers,

 

Dan

[brett_kosmider]

Dan, Fuzzy lenses like those on the Holga I love! And points well taken, I'm just

curious as to what others are doing out there - so take it down a notch, man. This is

an exchange of ideas is it not? These are the things we ponder are they not? So don't

be offended by my questions, I'm here for the same reasons you are.

 

BK

[enrico_pocopagni1]

I apologize for this illitterate's question, and maybe someone can indicate me a better place to post, but I used to know that diffraction only arises whenever light passes through a hole or slot wide as or close as wawelenght. Now, how can the even smallest diaphragm hole (several millimeters) trigger diffraction if wawelenght of light is only microns?

 

Thanks to all.

 

Enrico

[jbq]

Enrico - diffraction always happens, but for most aperture sizes it is pretty much invisible to the naked eye. The size of the diffraction spot grows approximately inversely proportionally to the size of the aperture, and proportionally to the distance from the aperture to the image - i.e. it grows inversely proprotionally to the f-stop. At f/16 it is about 5 microns, which results in a visible loss of contrast for details that are about 10 microns in size - which is pretty much invisible, but starts to become visible if you enlarge the picture 10 times and carefully observe the details.

 

If you weren't enlarging, there's be pretty much no problem, and you could stop down an incredible lot (look at just about any slide with the naked eye - it always appears incredibly sharp, even if it's actually not quite in focus when you enlarge it).

[kelly_flanigan1]

One desires a diffraction limited lens. This is the goal; nothing to be feared. Better lenses approach being diffraction limited at faster F stops. For a diffraction limited lens; resolution is higher at the faster openings. <BR><BR>Many MF lenses peak about F5.6 to F11. Typically the central core peaks at a faster fstop; than the edges. The edges require more stopping down; to controll abberations.<BR><BR>This question is like asking what speed a car should be; for peak miles per gallon; without mentioning the specific car. Cars usually have a peak MPG about in the 35 to 50 MPH range; depending on the air drag; and rolling resistance.<BR><BR>

[kelly_flanigan1]

The MF camera has roll film; that wanders more in position; than 35mm film. Thus a single test of a lens will be one data point; maybe better or worse than an average shot. Sometimes lenses in larger cameras are stopped down more; to help quash the film's less uncertain film position. In large precision cameras; aerial cameras; the film is held in a vacuum back; and the repeatablity of focus uis better; and faster fstops can be used.