Determining optimum aperture?

EASmithV · Aug 10, 2009

How do I do it? I want to get the optimum aperature at infinity of a fujinon-w 300mm f5.6, a nikkor-s 50mm f1.4, and an ektar 152mm f4.5

Dan Henderson · Aug 10, 2009

my understanding is that most lenses are optimum at 2 to 3 stops down from wide open. Several years ago I did a test of photographing a highly detailed subject at all apertures of my favorite lens, the 150 mm lens of my Hasselblad kit. Wide open (f/4) was markedly worse, improving as I stopped down until f/22, when sharpness began to degrade again. So, at least for me with this lens, f/8, f/11, and f/16 give great sharpness and I try very hard to use one of those apertures. Above and below that things are not as good. Wide open and fully closed are inferior.

BetterSense · Aug 11, 2009

I always just use the smallest one. Since you get more depth of field, overall sharpness is improved greatly over larger apertures.

keithwms · Aug 11, 2009

It's not an easy question to answer because lens resolution certainly isn't constant across the frame- not even close. So stopping down affects the resolution in different ways across the frame... smaller apertures tend to lift edge resolution up to its maximum (basically, by making use of the best-corrected, centralmost portion of the lens), while also bringing center resolution down via diffraction softening. This effect is seen quite nicely in the lens test applets at dpreview (disclaimer: it's a site catering to digital gear... but they do an exceptional job in presenting their lens test data and it makes my point in a tidy visual way). Compare the center and corner performance at f/8, f/11 and f/16 in the link.

With view cameras, the matter is further complicated because we use tilts and shifts, so that the central portion of the lens doesn't necessarily correspond to the central portion of the film. Furthermore, we sometimes stop way down in LF just to maximize the coverage. Actually, that is usually my biggest concern when using ultrawide lenses in LF... actual resolution figures are hardly a consideration when you're worrying about image circle!

Anyway, best resolution with LF lenses ... on average, across the frame ... is usually had at f/11 or so. What do you do with this information? Well, what I often do is focus wide open, do my movements, refocus, get a sense of the DOF required by the scene, and stop down only as just much as I need... or maybe a stop further. In other words I simply let the scene determine how I stop down, while also keeping f/11 in the back of my head.

If I feel like I want the most consistent level of resolution across the frame (e.g. for some macro subjects or for reproduction) then I will tend to stop down further.... going to f/45 or f/64 is fine if that's what you're after, and of course the other major issue is how much you will enlarge the negative. If you are dealing with an LF negative that will be contact printed or barely enlarged then you can stop down with virtual impunity. It's a very different story with 35mm, of course, for which the enlargement factors are typically much larger.

But I think the bottom line is to let artistic vision guide your aperture decisions, not a chart

2F/2F · Aug 11, 2009

I like this explanation: http://www.kenrockwell.com/tech/focus.htm. In short, it is different for each image. It depends on the image depth. Once you know the image depth, you throw it into this formula: f/optimum = square root of (375 x image depth in mm).

brofkand · Aug 11, 2009

BetterSense said:
I always just use the smallest one. Since you get more depth of field, overall sharpness is improved greatly over larger apertures.

That's not necessarily true. You'll actually lose sharpness at higher apertures due to diffraction.

You DO get more DOF, but overall sharpness is not higher.

The way to get the most DOF is to stop down to say f/16 or f/11, and use your distance scale to put the markers between infinity and the lowest number. I believe it's called hyperfocal distance.

ic-racer · Aug 11, 2009

For most pictoral photograpy, optiumum aperture depends on just two things, (1)your acceptable circle of confusion size, and (2) the focal spead between the near and far points that need to be sharp in the resultant picture. There is a lot of math involed but a simple chart will show you the aperture based on focal spread. http://www.largeformatphotography.info/fstop.html (Same thing that 2F/2F mentioned)

If you are taking pictures of flat objects, then you will be using larger apertures (to minimize diffraction) and lens abberations that limit edge sharpness will come into play. To determine the optimum aperture under these conditions you need to look at the MTF (modular transfer function) chart or a lens resolution chart for your lens. http://www.schneiderkreuznach.com/foto_e/an_su_classic/pdf/AN_SU_68_90_R44907_2AE.pdf

BetterSense · Aug 11, 2009

You DO get more DOF, but overall sharpness is not higher.

I was making the point that although absolute focus in the focal plane itself may be optimal at some moderate aperture, the greater DOF of smaller apertures can result in a sharper final images because more of the image will be in focus. That's my philosophy anyway...F/22 and be there.

2F/2F · Aug 11, 2009

BetterSense said:
I was making the point that although absolute focus in the focal plane itself may be optimal at some moderate aperture, the greater DOF of smaller apertures can result in a sharper final images because more of the image will be in focus. That's my philosophy anyway...F/22 and be there.

You get more depth of field (meaning more on either side of the plane of critical focus that appears just as acceptably sharp as the plane itself), but the plane of critical focus is not as sharp as it could be.

Practically speaking, f/22 is not a problem, except perhaps in super wide lenses. However, f/45, f/64, etc. are not worth the extra depth of field they bring when using a view camera, unless your camera does not allow any tilt or swing, and you are making tiny prints (meaning contact prints, basically). You can just use your optimum aperture, and tilt and/or swing to get what you want sharp.

If you like the effects of diffraction, which you very well may, by all means, use them...but at least understand that they are there.

Q.G. · Aug 12, 2009

I think the inclusion of DOF as a thing that bears considering is not very helpful.
DOF is acceptable unsharpness, and is hardly a measure for optimum lens performance.

What also should not be forgotten is that diffraction always reduces what a lens can do, even at larger apertures.
Many lenses are limited by residual aberrations wide open, and the beneficial effect of stopping down will be greater than the limiting effect of diffraction. When you find the point that both aberrations and diffraction are equally bad, you have found the optimum aperture.

You have to run tests to find it. Each lens design is different, and graphs only show so much. So test.

Now if you think it worth, practically speaking, to stop a lens down to its optimum aperture, you should also not want to stop it down any further.
So forget about DOF.

Stopping down to f/22 (or whatever the smallest aperture is) is a way that is guaranteed to not get the best out of your lens, ever.
The effect of diffraction is quite large: possible resolution is halved (!) about every two stops the lens is stopped down.
DOF calculators are all wrong. One big fault they have is that they do not take this overall image degradation that stopping down causes into account, which makes the difference between 'sharp' and not sharp much smaller, hence DOF much larger, than the formulae suggest.
So you do get huge DOF. But never "sharper final images". Never that what you paid for when you bought the lens, never what it is capable of.

f/22, and miss out.

Ian Grant · Aug 12, 2009

Q.G. said:
I

Stopping down to f/22 (or whatever the smallest aperture is) is a way that is guaranteed to not get the best out of your lens, ever.
The effect of diffraction is quite large: possible resolution is halved (!) about every two stops the lens is stopped down.
DOF calculators are all wrong. One big fault they have is that they do not take this overall image degradation that stopping down causes into account, which makes the difference between 'sharp' and not sharp much smaller, hence DOF much larger, than the formulae suggest.
So you do get huge DOF. But never "sharper final images". Never that what you paid for when you bought the lens, never what it is capable of.

f/22, and miss out.

Tessar type lenses (135mm and over) only reach optimum performance at f22, that's intrinsic to the design, a modern Xenar 150mm f5.6 has a marked aperture scal to f64 and is still razor sharp at f45.

I'm not sure where you drag that myth from, both Schneider and Rodenstock manufacture their LF lenses to be used at f22/32 and with longer lenses f45/64.

Ian

Ed Sukach · Aug 12, 2009

Q.G. said:
...
What also should not be forgotten is that diffraction always reduces what a lens can do, even at larger apertures.

**Interesting** - A possible source to contradict what I've been taught in Optical Theory - and proven by many moons of commercial Optical Quality Control work. Can you direct me to your source ? Hopefully it is NOT "everybody knows".

Many lenses are limited by residual aberrations wide open, and the beneficial effect of stopping down will be greater than the limiting effect of diffraction. When you find the point that both aberrations and diffraction are equally bad, you have found the optimum aperture.

Uh ... you've lost me here. Same source?

You have to run tests to find it. Each lens design is different, and graphs only show so much. So test.

... And the way to "test"? ... Optical Bench? MTF --- ?

Now if you think it worth, practically speaking, to stop a lens down to its optimum aperture, you should also not want to stop it down any further.
So forget about DOF.

I would not forget. Can be useful information.

Stopping down to f/22 (or whatever the smallest aperture is) is a way that i guaranteed to not get the best out of your lens, ever.

Of course. All the lens manufacturers make *SURE* that it will be easy to *destroy* the quality of their lenses by providing easy access to "'small' apertures. They are intent on providing built-in "booby traps" for those who don't listen to "everybody knows".

But ... seriously ... every lens is designed around "acceptable parameters" - OPTIMAL performance may well be at the "middle aperture; "Fast" lenses tend to be "shifted" towards the larger apertures; Copy lenses may ... usually don't... offer adjustable apertures.
Believe me. every reputable lens manufacturer KNOWS that their clients *WILL* at one time or other, use their lenses at the smallest (highest numerical) aperture and they ARE designed to "go" there (I don't know about Holga).

The effect of diffraction is quite large: possible resolution is halved (!) about every two stops the lens is stopped down.

Man!!! Information about THAT progression, "halving"!!??, is something I've got to see in writing.

I hate to tell you this, but diffraction limiting is NOT directly linked to f/stops.
This is not a simple area of optical design ... you might try googling "diffraction".

DOF calculators are all wrong....

No, they are not "wrong".

One big fault they have is that they do not take this overall image degradation that stopping down causes into account, which makes the difference between 'sharp' and not sharp much smaller, hence DOF much larger, than the formulae suggest.
So you do get huge DOF. But never "sharper final images". Never that what you paid for when you bought the lens, never what it is capable of.
f/22, and miss out.

Uh ... can you give me the definition of "Circle of Confusion"?

keithwms · Aug 12, 2009

Oh my

For anyone feeling buried in jargon, semantics and fine technical arguments, look at the link to the lens test applet that I posted above, it can teach you a lot!

RPippin · Aug 12, 2009

This has been an eye opening thread for me. I understand for the first time the true meaning of the "circle of confusion". Next for me is to apply this to a better understanding of the Zone System, BTZS, and "everybody knows, meter for the shadows and develop for the highlights".

Q.G. · Aug 12, 2009

Ed Sukach said:
**Interesting** - A possible source to contradict what I've been taught in Optical Theory - and proven by many moons of commercial Optical Quality Control work. Can you direct me to your source ? Hopefully it is NOT "everybody knows". [...]

You lost me completely.

You have been taught Optical Theory, have worked many moons in Optical Quality Control work, and do not know this?!

Now why don't you teach me how it really is?
Go on!

Ed Sukach said:
... And the way to "test"? ... Optical Bench? MTF --- ?

Take pictures.

Ed Sukach said:
Of course. All the lens manufacturers make *SURE* that it will be easy to *destroy* the quality of their lenses by providing easy access to "'small' apertures. They are intent on providing built-in "booby traps" for those who don't listen to "everybody knows".

Lens manufacturers?

Today’s high quality color films do reach resolutions in the region of 140 line pairs per millimeter with Kodak Ektar 25 leading the field at 200! The full resolution potential of these films cannot be utilized with existing depth-of-field concepts nor f-settings of f/11 and beyond.

Carl Zeiss, Camera Lens News, no. 2, Fall 1997.

Ed Sukach said:
But ... seriously ...

Seriously, what have you been smoking!?

Ed Sukach said:
Man!!! Information about THAT progression, "halving"!!??, is something I've got to see in writing.

Go back to your books then. I bet you haven't touched them since "many moons" ago.

But, conveniently, from the same source, theoretical limits:

f-no. resolution (line pairs per millimeter)
45 - 35
32 - 50
22 - 70
16 - 100
11 - 140
8 - 200
5.6 - 280
4 - 400
2.8 - 560

Ed Sukach said:
I hate to tell you this, but diffraction limiting is NOT directly linked to f/stops.

You definitely need to return to your books!

I was going to suggest that you look up the formula that describes the relation between f-stop and resolution when posting Zeiss' table when i saw what you wrote here.
I now strongly urge you to do so!

Ed Sukach said:
This is not a simple area of optical design ... you might try googling "diffraction".

Let's make a deal: i'll do that, and you return to school.

Ed Sukach said:
No, they are not "wrong".

They are.
In many ways.

Ed Sukach said:
Uh ... can you give me the definition of "Circle of Confusion"?

I can. Several
One is about the thing your head seems to be spinning in right now.

P.S.
Been to Google, and nothing there i didn't know.

BetterSense · Aug 12, 2009

f-no. resolution (line pairs per millimeter)
45 - 35
32 - 50
22 - 70
16 - 100
11 - 140
8 - 200
5.6 - 280
4 - 400
2.8 - 560

That gives resolution in lp/mm. But does it consider the large effect format size has on acceptable sharpness?

I'm thinking a 4x5 film is 3.33 times longer than a 35mm frame so stopping down to f/32 (50lp/mm*3.33=170) should be about as damaging to resolution as f8 is on 35mm at the same sized enlargement.

Q.G. · Aug 12, 2009

It illustrates the relation between aperture and theoretically achievable resolution.
Which is 'hard', i.e. the way it is, without anything we (including Ed ) can do about it.

What resolution is still acceptable, and when, is another matter.
It depends mostly on what we find acceptable.

Criteria for CoC size traditionally are different for different formats, for exactly the reason you are thinking about.

keithwms · Aug 12, 2009

BetterSense said:
That gives resolution in lp/mm. But does it consider the large effect format size has on acceptable sharpness?

Which is why some people mention enlargement factor in this context.... Of course, pinhole photography is an interesting albeit extreme case in point.

~~~

Overall, in my opinion, far too much time is spent discussing the high frequency center resolution, to the neglect of the rest of the frame.

Tonal and focus transitions across the frame can be more important than maximum center esolution: you can have tack sharp center resolution and crappy edges, and that effect will be especially noticeable because the viewer of the print is able to see both side by side....

...which brings me to a basic principle of perception that sometimes gets lost in the technicals: we are not generally able to perceive absolute resolution. I.e., your eye probably cannot tell the difference between a print from a 100 and 80 lp/mm neg at typical enlargement sizes, that kind of thing is background noise if the composition is effective. Critical detail that is lacking is usually not missed.... unless it is viewed comparatively.

So, if you have a big falloff in resolution across the frame, that can certainly be noticed. And actually that is seen quite often. An extreme example would be the results from holgas etc. where the effect is so strong that it becomes a defining part of the composition.

The effect of resolution falloff is manifest in the high frequency (critical detail) as well as in the tonal smoothness and perhaps also in the lower frequency characteristics e.g. the amount of swirl and general smoothness of the bokeh.

So... apologies for the coffee-time ramble, but my point is that, in these discussions, too much emphasis is placed on the paradigm of front-to-back sharpness and how aperture selection affects the absolute resolution at frame center. There is a lot more to aperture selection than lp/mm at frame center!

Ian Grant · Aug 12, 2009

Q.G. said:
It illustrates the relation between aperture and theoretically achievable resolution.
Which is 'hard', i.e. the way it is, without anything we (including Ed ) can do about it.

What resolution is still acceptable, and when, is another matter.
It depends mostly on what we find acceptable.

Criteria for CoC size traditionally are different for different formats, for exactly the reason you are thinking about.

Nothing you say falls into line with reality and the fact that LF lenses are designed to be used at f22/32 and have excellent resolution across the field.

Some lens designs don't achieve edge/corner sharpness until f22m andthen how do you explain the outstanding performance of an f5.6 150mm Xenar at f32 & f45, it stops down to f64 ?

Ian

Ed Sukach · Aug 12, 2009

Q.G. said:
You have been taught Optical Theory, have worked many moons in Optical Quality Control work, and do not know this?!

No, I DON'T "know" this!

Now why don't you teach me how it really is?
Go on!

Considering where you are starting from (and your attitude)...

Even though I am under NO obligation to instruct you in anything - you will just have to learn for yourself, I will direct you to:

This from one of the important text books:

What is Light? By A.C.S. van Heel and C.H.F. Velzel,World University Library - Library of Congress Catalog Card Number: 67-24448

Page 97, Section 4, Diffraction:
We can now turn to another facet in the study of light: the diffraction effects, the existence of which has already been mentioned. (See section 2 on the rectilinear propagation of light, and section 18 on the distribution of light at the focal plane of a lens upon reduction of the size of the diaphragm). To explain such phenomena, the great physicist and optician, Augustin Fresnel (1788 - 1827) developed a theory of the propagation of light, with which we begin our discussion.

Until now we have used as a model for the propagation of light, the theory of Huygens, which we define as follows.
One can imagine a wave front to originate out of the previous one by supposing each point in the latter to be a secondary source of spherical waves. The envelope of these spherical waves forms the new wave front..."

**Fascinating**. I will admit to having read this section a number of times, but I finally gained a fairly good "grasp" of what was going on.

Polarisation, section 5, was another matter. I passed the "tests" - but I'm still not completely sure...

Take pictures.

Ah! Obviously an infallable way to test a lens, free from subjective judgement and preceptual bias.... NOT!

[QUOTE Lens manufacturers?

Todays high quality color films do reach resolutions in the region of 140 line pairs per millimeter with Kodak Ektar 25 leading the field at 200! The full resolution potential of these films cannot be utilized with existing depth-of-field concepts nor f-settings of f/11 and beyond.

Carl Zeiss, Camera Lens News, no. 2, Fall 1997. [/QUOTE]

As interesting as that my be ... I doubt that it has anything to do with diffraction. Is the article available on-line ? Or ...?

Seriously, what have you been smoking!?

The last refuge of one insecure in his argument: an "Ad Hominem" attack.

But, conveniently, from the same source, theoretical limits:

The same source? Zeiss? Now I am really interested!

f-no. resolution (line pairs per millimeter)
45 - 35
32 - 50
22 - 70
16 - 100
11 - 140
8 - 200
5.6 - 280
4 - 400
2.8 - 560

Nice table. Anything to do with diffraction?

I was going to suggest that you look up the formula that describes the relation between f-stop and resolution when posting Zeiss' table when i saw what you wrote here.
I now strongly urge you to do so!

I understand some of the factors that influence lens design... What happened to "diffraction"?

I can. Several
One is about the thing your head seems to be spinning in right now.

Well, that really says it all. There is only ONE definition recognized as "Circle of Confusion" - an important criteria in lens design.

Q.G. · Aug 12, 2009

Ian Grant said:
Nothing you say falls into line with reality

That's what you think.
It is reality. Inescapable law of nature.
Even Ed's favourite mr. Fresnel knows this.

Ian Grant said:
and the fact that LF lenses are designed to be used at f22/32 and have excellent resolution across the field.

Some lens designs don't achieve edge/corner sharpness until f22m andthen how do you explain the outstanding performance of an f5.6 150mm Xenar at f32 & f45, it stops down to f64 ?

Well, there is excellent and there is excellent.

Q.G. · Aug 12, 2009

Ed Sukach said:
No, I DON'T "know" this!

Well, there we are!

Based on your bragging, you should though.

Ian Grant said:
Considering where you are starting from (and your attitude)...

That's perhaps because we cannot all be as learned as you, right?

Ian Grant said:
Even though I am under NO obligation to instruct you in anything - you will just have to learn for yourself, I will direct you to:

This from one of the important text books:

What is Light? By A.C.S. van Heel and C.H.F. Velzel,World University Library - Library of Congress Catalog Card Number: 67-24448

Page 97, Section 4, Diffraction:
We can now turn to another facet in the study of light: the diffraction effects, the existence of which has already been mentioned. (See section 2 on the rectilinear propagation of light, and section 18 on the distribution of light at the focal plane of a lens upon reduction of the size of the diaphragm). To explain such phenomena, the great physicist and optician, Augustin Fresnel (1788 - 1827) developed a theory of the propagation of light, with which we begin our discussion.

Until now we have used as a model for the propagation of light, the theory of Huygens, which we define as follows.
One can imagine a wave front to originate out of the previous one by supposing each point in the latter to be a secondary source of spherical waves. The envelope of these spherical waves forms the new wave front..."

**Fascinating**. I will admit to having read this section a number of times, but I finally gained a fairly good "grasp" of what was going on.

Nice quote.
But it says absolutely nothing about this matter.
So not well chosen.

Ian Grant said:
Polarisation, section 5, was another matter. I passed the "tests" - but I'm still not completely sure...

Who is?

Ian Grant said:
Ah! Obviously an infallable way to test a lens, free from subjective judgement and preceptual bias.... NOT!

Who said anything about an infallible test?

But you do know (i assume, but tell me if i am wrong) that photography is a visual medium, and that what you see is what you get?
Good!

Ian Grant said:
As interesting as that my be ... I doubt that it has anything to do with diffraction. Is the article available on-line ? Or ...?

Yes.
It has everything to do with diffraction (you really need to get those books out again)
And yes, it is available online.

Ian Grant said:
The last refuge of one insecure in his argument: an "Ad Hominem" attack.

That's why you started by pointing out your personal qualifications?

I don't doubt your personal qualifications. So not even an ad hominem.
But given that your argument is based on authority, it is not possible to dismiss your mistake as such without touching your proffered authority.

Ian Grant said:
The same source? Zeiss? Now I am really interested!

You know it is optics 1.0.1.
So you really should know this, and not find this interesting, but old hat.

Ian Grant said:
Nice table. Anything to do with diffraction?

It is about nothing else but diffraction.

Ian Grant said:
I understand some of the factors that influence lens design... What happened to "diffraction"?

What are you asking?

Ian Grant said:
Well, that really says it all. There is only ONE definition recognized as "Circle of Confusion" - an important criteria in lens design.

Only one? You think so?
I think 'the place where your head is (was?) spinning' qualifies as a perfectly good definition.

Now just stop digging.

richard ide · Aug 12, 2009

Out for lunch and lost his way back!

Ed Sukach · Aug 12, 2009

Q.G. said:
I]Todays high quality color films do reach resolutions in the region of 140 line pairs per millimeter with Kodak Ektar 25 leading the field at 200! The full resolution potential of these films cannot be utilized with existing depth-of-field concepts nor f-settings of f/11 and beyond.[/I]
Carl Zeiss, Camera Lens News, no. 2, Fall 1997.

Click to expand...

I've finally reached the Zeiss website, and D'Ld the .pdf file. I see where you are coming from, but I will submit that you are extrapolating to reach an unwarranted conclusion. Here, Zeiss is commenting on theoretical limits of resolution, tied in some manner, to f/stops. I would really be interested in their methods of determining these approximate limits, but no information is provided.
According to Zeiss, a resolution of 45 l/mm (one way to look at resolution) is reaced at f/11. The normally accepted design criteria for a 35mm frame is 1/60th mm, so there will be some degradation at an aperture smaller than f/11. Notice that I wrote "some" ... the image will not abruptly disappear into a ball of "fuzz".

One thing to remember: the "practical limit" of any lens system is not entirely and wholly dependent on "diffraction"... it would be, IF the remaining parameters were all perfect... and the chances of that happening are not great.

I will ask one favor: Direct me to web site where I can find Zeiss's formula for determining "diffraction limitation". Until I can check that out, I will not continue to discuss their assumptions.

It is interesting that you consider "Diffraction" as one of the simple. basic phenomena. I think it is one or two notches above "basic".

BTW... I've been considering the "tone" and syntax of your posts ... in some way, vaguely familiar. Would you happen to be posting from St. Louis, Missouri?

Seriously, what have you been smoking!?

Click to expand...

Ad hominem, again. You do know ad hominem means, don't you?

Go back to your books then. I bet you haven't touched them since "many moons" ago.

Click to expand...

I hate to break the news, but, that time span is about 15 minutes.

I was going to suggest that you look up the formula that describes the relation between f-stop and resolution when posting Zeiss' table when i saw what you wrote here.
I now strongly urge you to do so!

Click to expand...

Sounds good to me --- WHERE??

I can. Several
One is about the thing your head seems to be spinning in right now.

Click to expand...

Hilarious. Indicative of your sharp wit, I suppose.

richard ide · Aug 12, 2009

Ed,

I thought it could be HIM also.

Determining optimum aperture?

Privacy & Transparency

Privacy & Transparency