PDA

View Full Version : What is the ideal aperture?



Lightbender
7-May-2016, 13:42
At what aperture do you get the maximum depth of field without sacrificing detail to diffraction?

mdarnton
7-May-2016, 14:36
Someone the other day specified 6mm more or less. I like the idea, so I hope it's right!

Jac@stafford.net
7-May-2016, 14:37
At what aperture do you get the maximum depth of field without sacrificing detail to diffraction?

Generally (which means most, not all) LF photographers to not worry about diffraction. The degree of enlargement and viewing distance are the significant factors. IOW, worry not.

It is something that miniature and MF photographers find difficult to grasp at first.
.

Jac@stafford.net
7-May-2016, 14:40
Someone the other day specified 6mm more or less. I like the idea, so I hope it's right!

OMG, that might have been from my post decades ago. Specific aperture diameter seems to have a congruence with format size, enlargement and viewing distance. Six mm is, therefore, not a universal metric.
.

mdarnton
7-May-2016, 14:59
Maybe I caught it in some of my web searching. I think it was, in fact, you. I just looked on one of my Nikon lenses that I know precisely where it starts to go haywire, and 6mm it was! It doesn't sound too bad for longer lenses, either. I'm not so sure about my 15mm Heliar for 35mm, which may be smaller than that to start with. :-)

Mark Sawyer
7-May-2016, 15:19
At what aperture do you get the maximum depth of field without sacrificing detail to diffraction?

It depends on, (among other things), the degree of enlargement. Edward Weston used a pinhole aperture (and all day exposures) to get maximum depth of field in some of his larger-than-life-size still lifes of vegetables and seashells. Any larger aperture would have given out-of-focus areas, and because he contact printed, the detail is as fine as the eye can see.

Ken Lee
7-May-2016, 15:24
At what aperture do you get the maximum depth of field without sacrificing detail to diffraction?

Forum member Emmanuel Bigler (a professor of Optics) has written about this subject here several times. Here's a sample:


"I have done a compilation of the best f-number recommended by view camera lens manufacturers and we can summarize by the following rule of thumb for classical standard lenses :
N_best = f(in millimeter)/(8 millimeters)...for top-notch modern 6/4 lenses of the last generation and razor-sharp expensive "digital" view camera lenses the rule of thumb would be closer to f(in millimeters)/(11 millimeters)...The consequence is that a good standard 50mm lens designed for 35 mm photography should not be stopped down beyond f/5.6 or f/8, whereas a 150mm lens designed for 4x5 can be stopped down to f/16 to f/22, a 210 mm for 5x7 : f/22 to f/32, a 300 for 8x10 : f/32 to f/45. Assumed that the enlarginf factor is reduced in proportion, of course, for the same final print size."



See Diffraction large format vs smaller (http://www.largeformatphotography.info/forum/showthread.php?53062-Diffraction-large-format-vs-smaller&p=504559&viewfull=1#post504559) for the discussion.

With a good loupe we can point the camera at a detailed subject, inspecting the ground glass as we stop-down the lens. Resolution will reach a peak, after which it visibly declines. That's an empirical rather than theoretical approach.

It's one thing to know the aperture beyond which resolution declines: it's another thing to decide where we want to draw the line for a given image. Are we making contact prints or enlargements ? Can we apply digital sharpening ? From what distance will the image be viewed ? ... etc. Ultimately, we get to decide.

ic-racer
7-May-2016, 15:39
http://www.largeformatphotography.info/fstop.html

mdarnton
7-May-2016, 16:13
Forum member Emmanuel Bigler (a professor of Optics) has written about this subject here several times. To summarize his recommendation, we can divide focal length by 11 to get the number.

[INDENT][I]"...for top-notch modern 6/4 lenses of the last generation and razor-sharp expensive "digital" view camera lenses the rule of thumb would be closer to f(in millimeters)/(11 millimeters)...The consequence is that a good standard 50mm lens designed for 35 mm photography should not be stopped down beyond f/5.6 or f/8, whereas a 150mm lens designed for 4x5 can be stopped down to f/16 to f/22, a 210 mm for 5x7 : f/22 to f/32, a 300 for 8x10 : f/32 to f/45. Assumed that the enlarginf factor is reduced in proportion, of course, for the . . .

In short, using his numbers, around 7mm. Slightly higher standards than Jac, maybe. :-)
I suspect he should have used a wider range of examples than just a bunch of "normal" lenses. :-)

Jac@stafford.net
7-May-2016, 16:18
In short, using his numbers, around 7mm. Slightly higher standards than Jac, maybe. :-)
I suspect he should have used a wider range of examples than just a bunch of "normal" lenses. :-)

See the reference to Emmanuel Bigler, and big thanks to mdarton.
.

Dan Fromm
7-May-2016, 17:03
In short, using his numbers, around 7mm. Slightly higher standards than Jac, maybe. :-)
I suspect he should have used a wider range of examples than just a bunch of "normal" lenses. :-)

Michael, Emmanuel wrote:


for top-notch modern 6/4 lenses of the last generation and razor-sharp expensive "digital" view camera lenses the rule of thumb would be closer to f(in millimeters)/(11 millimeters)

11 millimeters for these lenses, not 7. And note that the f/number that his rule of thumb yields increases with focal length.

As a practical matter, the only lenses whose apertures are scaled in mm instead of f/numbers are high performance macro lenses for, usually, use above 1:1. It is true that around the turn of the 20th century Zeiss made lenses with apertures scaled in mm. The idea was quickly dropped for general purpose lenses. I'm not sure why, suspect that apertures scaled in f/numbers are easier to think about.

The OP's question is one of those questions for which the best answer is "it depends." There's no one right answer. The answer depends on size of subject, size of negative, size of final print, image quality desired in the final print, ... This doesn't stop people from asking these questions and this one in particular. It also doesn't stop people from giving answers that are incomplete, if not wrong, without all of the conditions spelled out.

It also stops people who ask such questions from thinking about the tradeoffs between sharpness in the plane of best focus and depth of field given magnification from negative to final print. There's no getting around them, one always has to ask what the final print is going to be.

BetterSense
7-May-2016, 17:44
I wish lenses were marked in both mm and f-stops. I often set up portraits with both a LF camera and a smaller camera. I adjust to get the DOF I want on the ground glass; if I could set the same aperture (in mm) on the smaller camera I would know the images would match exactly instead of having to do rough math.

mdarnton
7-May-2016, 17:56
Michael, Emmanuel wrote:


11 millimeters for these lenses, not 7.

No, you're misreading the equation, I'm pretty sure. It's FL in mm, divided by 11 (result in mm) I believe. If it was just 11mm there'd be no reason to specify FL or division. Anyway, if it was 11mm, that wouldn't match any resolution chart on the face of the earth for anything, I don't think.

BetterSense
7-May-2016, 18:39
No, you're misreading the equation, I'm pretty sure. It's FL in mm, divided by 11 (result in mm) I believe.

I made the opposite conclusion. In fact I think you missed the point. Which was that optimum aperture (mm) is constant across formats because diffraction and DOF both depend on only magnification and aperture diameter (and sharpness criteria of course). But 6mm matches my experience better than 10mm but I don't have the best lenses either.


If it was just 11mm there'd be no reason to specify FL or division.

The reason for specifying FL and formula is simply that lenses are (unfortunately) not marked in mm so the photographer needs a way to convert to mm into F-stop so he can set the lens.

Jim Jones
7-May-2016, 19:36
Someone the other day specified 6mm more or less. I like the idea, so I hope it's right!

To avoid most of the math, 6mm is a reasonable figure for the lenses in use decades ago if the image is viewed in correct perspective. The correct distance for viewing the image in correct perspective is the focal length X the degree of enlargement. Wide angle prints are usually viewed at too great a distance for correct perspective. The resulting perspective distortion can be used creatively.

Dan Fromm
7-May-2016, 20:00
No, you're misreading the equation, I'm pretty sure. It's FL in mm, divided by 11 (result in mm) I believe. If it was just 11mm there'd be no reason to specify FL or division. Anyway, if it was 11mm, that wouldn't match any resolution chart on the face of the earth for anything, I don't think.

I quoted it exactly. FL in mm divided by 11 mm. Go back and read the post that quotes Emmanuel.

His equation translates an 11 mm aperture into f/number given aperture. Most lenses' apertures are calibrated in f/numbers, not mm, as I've already pointed out.

Resolution charts have nothing to do with it. Why did you bring them in?

djdister
7-May-2016, 20:11
The Rodenstock brochure "Lenses for Professional Photography" provides this general table of "optimum working apertures" relative to film size in a section comparing depth of field versus diffraction.

150560

LabRat
7-May-2016, 22:51
The fast/easy/practical real-world way to check lenses for this (without film or bench tests) would be to use a DSLR adapter for the film plane of your LF camera... (Buy or make out of a piece of wood that fits in your holder area, with an matching lens mount on an old extension tube mounted on the wood...) You would shoot a test target with this set-up, and shoot a frame at all f-stop settings... You can then blow them up on your monitor, and you should see the point where the diffraction kicks in...

Steve K

Emmanuel BIGLER
8-May-2016, 01:57
The Rodenstock brochure "Lenses for Professional Photography" provides this general table of "optimum working apertures" relative to film size in a section comparing depth of field versus diffraction.
150560

Thanks fr D.J. Dister to remember us this interesting table form Rodenstock; and thanks to everybody for quoting my little compilation / calculation that had no ambition to be universal, I only wanted to show how the best F-number scales when you change format from 35 mm photography to large format photography, using a standard focal length for each format.

In fact the demonstration was aimed at people insisting on using sub-centimeter-size sensors, with sub-centimeter-size focal lengths, and claiming that they achieve better image quality than medium or even large format film photography. It is well-known that when you insist on enlarging your digital images taken with half-frame or full-frame [24x36] silicon sensors, you have to worry about diffraction with f-numers as low as 5.6.
This was already well-known to film photographers!
The only difference with what had been known with film for decades, is that with a digital image you have the ability enlarge it beyond reasonable magnifications without ever seeing any silver or dye grain.
So, no surprise if you reach another limit, which is the optical diffraction limit, at least when you have enough pixels in your sensor.
And with sub-centimeter sensor sizes and focal lengths, forget using even using F/5.6 if you want to "avoid" diffraction; or better said: if you want that diffraction effects are hidden beyond other effects, since diffraction is always there, simply you may, or may not, be able to see its effects.

For my compilation of best F-numbers, had taken the same recommended F-numbers as Rodenstock, but made the comparison only for standard lenses across various formats, i.e. from a classical 50 mm in 35 mm photography (24x36 mm format as we say in France), 80 mm for 6x6 [cm], 100 mm for 6x9 [cm], 150-180 for 4x5", 210-240 for 5x7" and 300-360 mm for 8x10".

I have translated into English the resulting graph showing how nicely the best f-number for standard lenses fits with a rule of thumb F (in mm) / (8 mm) for older lenses or F(in mm) / (11 mm) for last generation of digital lenses.
And the consequence of this data fit is that the best aperture has a constant diameter for all formats, namely 8 mm or 11 mm.
But it could be 6 mm or 7 mm, the important is how the value scales with format.
Simply, a constant aperture diameter of 6 mm, in another words a straight line fit of F(in mm) / (6 mm) does not fit well with the series of recommended F-numbers for standards lenses in MF and LF photography.
150569

However a puzzling issue is that Rodenstock recommends the same F-number for all focal lengths covering a given format, which is something actually questionable, I have no idea whether this recommendation is reasonable or not; hence in order to keep the argumentation simple, I had limited my analysis to the family of standard lenses only.
(sotto voce : if Rodenstock gurus state this, I do not argue and I immediately obey like a good soldier : "Jawohl, Herr Optik Guru!")

Doremus Scudder
8-May-2016, 02:27
http://www.largeformatphotography.info/fstop.html

Exactly!

Frankly, I'm surprised that it took so long for someone to link to this.

As far as "ideal" aperture goes: If you're copying flat originals, then you want to find that sweet spot exactly between lens aberrations and diffraction. This varies a bit with lens design and quality, but is fairly fixed as Emmanuel Bigler outlines above.

In the real world, practically all our subjects have depth and require us to choose an aperture that yields acceptable sharpness in the final print. "Optimum" is the operative word here and it all depends on the subject, our required depth-of-field, the degree of enlargement and the sensitivity of the viewer's eye at whatever minimum viewing distance we find acceptable.

I don't really care if my chosen plane of focus has a few more lines per millimeter resolution if the rest of the subject is disturbingly out-of-focus. Nor do I mind a bit of degradation (i.e., less-than-possible resolution in the negative) due to diffraction as long as the things I want sharp in the print are and the degradation isn't visible in the final print. I choose my aperture accordingly.

Best,

Doremus

Lachlan 717
8-May-2016, 04:22
I honestly don't recall this being as big an issue prior to DSLRs and people like Ken Rockwell needing to pad out their blogs with controversial commentary...

mdarnton
8-May-2016, 05:40
Dan, the statement you quote makes no grammatical or mathematic sense. Since EB refrains from explaining himself, I guess we will have to live without explanation.

ic-racer
8-May-2016, 06:17
I wish lenses were marked in both mm and f-stops. I often set up portraits with both a LF camera and a smaller camera. I adjust to get the DOF I want on the ground glass; if I could set the same aperture (in mm) on the smaller camera I would know the images would match exactly instead of having to do rough math.

I just look in the lens [to view the aperture size for comparison]. Seriously.

ic-racer
8-May-2016, 06:26
However a puzzling issue is that Rodenstock recommends the same F-number for all focal lengths covering a given format

Not a bad rule of thumb for a 'same format' discussion. Long lenses magnify effects of defocus and diffraction on the film. Short lenses minimize both. Therefore, a larger aperture size is needed for long lenses and smaller aperture size is ok for short lenses. Turns out [long focal length] divided by [larger aperture] is the same f number as [short focal length] divided by [smaller aperture].

BetterSense
8-May-2016, 08:07
I just look in the lens [to view the aperture size for comparison]. Seriously.

I do the same now. It works ok.

Bruce Watson
8-May-2016, 08:59
At what aperture do you get the maximum depth of field without sacrificing detail to diffraction?

This question has meaning in a laboratory setting, but I submit it has very little meaning to actual photography. As you gain experience you quickly find that loss of detail due to diffraction is not nearly as important as DOF. If you need the small apertures to give you the level of DOF that the photograph you are making requires, then do it. Diffraction will take care of itself. In my years in photography I've never rejected a single shot due to diffraction limited sharpness, but I've filled trash cans with my DOF mistakes.

Make more photographs. Worry less.

Dan Fromm
8-May-2016, 09:17
Dan, the statement you quote makes no grammatical or mathematic sense. Since EB refrains from explaining himself, I guess we will have to live without explanation.

Michael, re Emmanuel's explaining himself, see post #19 above.

Emmanuel, I can't find the table you posted on Rodenstock's site or in archive.org's copies of that site. Bummer.

Emmanuel BIGLER
8-May-2016, 11:20
Dan, I was simply referring to a chart posted above by D.J. Dister.
Here is a pure text transcription of this chart. I have omitted the list of focal lengths, because the focal length for a given format does not change Rodenstock's recommendation!

Format focal length best F-stop
24x36 18 to 135 mm 5.6
6x6 30 to 250 mm 11
6x7 35 to 280 mm 11 - 16
6x9 45 to 360 mm 16
9x12 - 4x5" 55 to 480 mm 16 - 22
13x18 - 5x7" 90 to 600 mm 22 - 32
18x24 - 8x10" 115 to 900 mm 32 - 45

Dan Fromm
8-May-2016, 12:49
Emmanuel, this discussion is very interesting. It seems too highly abstracted from photography. In particular, it seems peculiarly focused on some measure or other of lens performance on-axis, isn't concerned with image quality off-axis, and isn't concerned with image quality in the final print.

Of course, since I'm a Certified Ignorant Barbarian Insensitve To The Fine PointsŪ, I could easily have missed all that I say I didn't find.

Emmanuel BIGLER
9-May-2016, 09:29
It seems too highly abstracted from photography.

Well ... Dan .. how to say .... you should tell this to the Rodenstock engineers and marketing guys who created this chart ;)
I'm just making a transcription of the chart into a graph ;)

Unfortunately, those who designed the above-mentioned "LF film lenses" are probably retired now: too bad, since we would love to have them here arguing with us!

djdister
9-May-2016, 10:49
It may be helpful to read the explanation about those "optimum working apertures" from the Rodenstock brochure, as seen on this page (http://www.prograf.ru/rodenstock/largeformat_en.html#table2).

Tin Can
9-May-2016, 11:01
LOL

I find this both interesting and amusing.

I will be doing my own 'experiments' and not just follow.

An easy one will be 'improving' meniscus lens with the 'correct' aperture for format.

Tin Can
9-May-2016, 11:09
It may be helpful to read the explanation about those "optimum working apertures" from the Rodenstock brochure, as seen on this page (http://www.prograf.ru/rodenstock/largeformat_en.html#table2).

Yes, the linked whole page is better and more inclusive.

I like the shift limits column in particular which is derived from 'Field angle at recommended working aperture'.

Although here is where Bob says that it is an unauthorized source, implying inaccuracy I believe.

Moot point these days.

Drew Wiley
9-May-2016, 11:15
Forget the math, forget the tables. Everything depends on how YOU want the image to look at a given degree of enlargement. It depends on the subject matter,
the specific lens, what kind of depth of field options happen to work best at the time (and view camera movements are a real advantage in this respect). But since
I personally like the option to make relatively large prints, my own habit of avoiding diffraction isn't much different from those Rodenstock recommendations. Yet
there are many possible exceptions, especially with smaller formats when "selective focus" comes into play.

djdister
9-May-2016, 12:28
I'm sure it sounds somehow unscientific, however the Rodenstock table of "optimum working apertures" seems more like "rule of thumb" advice, based on a need to balance the best depth of field against diffraction. And their advice in the table would seem to suggest that diffraction is less of an issue when larger film formats are used. I find these two statements from that page to be the most interesting, and suggest this is a rule-of-thumb type of advice:

"Under optimum circumstances most large format lenses should be stopped down to f/22; smaller apertures may be permissible with larger formats (see chart). If a great deal of depth of field is required, the aperture can be stopped down by one more number without any substantial loss in the image sharpness."

"These recommended values should only be exceeded in the interest of best possible sharpness when the depth of field is much more important than good sharpness."

rich815
9-May-2016, 12:50
This question has meaning in a laboratory setting, but I submit it has very little meaning to actual photography. As you gain experience you quickly find that loss of detail due to diffraction is not nearly as important as DOF. If you need the small apertures to give you the level of DOF that the photograph you are making requires, then do it. Diffraction will take care of itself. In my years in photography I've never rejected a single shot due to diffraction limited sharpness, but I've filled trash cans with my DOF mistakes.

Make more photographs. Worry less.

Yup. Last time this was discussed I shared how impressed I was with the work of Charlie Waite in this book:

The Making of Landscape Photographs: A Practical Guide to the Art and Techniques

https://www.amazon.com/dp/1855851490

Later I learned he almost always shot his images on a Hasselblad and closed completely down to f/22. But what about diffraction, I thought. Did not seem to matter. Of course it does, but not in practical results anyway.

This starts to get into the useless "which lens is sharpest" discussion.

Will Frostmill
9-May-2016, 15:04
A tip:
An easy way to find out if you are working at the same absolute aperture in mm is to take a picture of both lenses side by side, and then measure the size of the apparent aperture in the image. You only have to do this once, and then you can mark your shutter or keep a notecard that lists the real equivalent apertures. (Since manufacturers lie lie lie about true focal lengths.)

Jim Noel
9-May-2016, 15:35
Every lens, image, and application has its own "ideal aperture". There isn't a one case fits all.

Drew Wiley
9-May-2016, 15:42
Diffraction is an issue with any format. But it's related to how that format is used. If you make a 30x40 print from 8x10 film, that's only a 4X linear enlargement.
So the difference between f/32 and f/45 is going to be subtle, not nonexistent. With Mr Waite's Hassie, you'd end up with an 18X linear enlargement, so basically
it would be a case of wretched versus really wretched. But if it's just a small book image, we're back down to no big deal. This has nothing to do with "sharpest
lens" discussions. All lenses suffer from diffraction at some point. It is just a matter of when diffraction becomes noticeable and distracting. With large format work we need to think about lens coverage and evenness with significant movements, a less than ideally flat film plan, and very often, uneven subject planes. How we use or abuse our options involves a variety of things. A contact printer might find the whole discussion silly; but some of use generally enlarge and like our prints crisp, and I mean optically crisp, not inkjet.

Argentum
9-May-2016, 18:05
It seems to me that all lenses have a degree of residual aberration of various kinds and at various apertures. There is a point, when closing down the aperture, where the diffraction becomes more significant than the other aberrations and just before or at that point is the sweet spot of then lens. i.e. it will give optimal DoF without compromising resolution due to aperture diffraction.

Some lenses are so well corrected for all the aberrations that the lens wide open may be the sweet spot. Other lenses may have a lot of abberations such that you can close down further than you might expect without it showing a loss of resolution in the negative and gain a lot of extra DoF.

Typically however, the sweet spot is usually somewhere in the middle of the lenses aperture range. A 50mm lens @ f5.6 on 35mm format has an aperture of 8.9mm and at f8, an aperture of 6.25mm neither of which takes the actual entrance pupil into consideration but the typical sweet spot will be somewhere between those two values. A 150mm lens @ f22 on 4x5 format will have an aperture of 6.8mm.
So the sweet spot aperture is more about focal length rather than film format but if you take the standard focal length from each film format you will find that they all have a ball park sweet spot of an aperture of around 7mm. But as I said, it really does depend on each specific lens. My suspicion is that with modern lenses the sweet spot will be wider open than it used to be in days of yore but that it is a bit of a sweeping statement as there are always exceptions to the rule.
So remember, diffraction is NOT the only aberration that any lens might have and it may not be the limiting factor to high resolution leaving aside local contrast lighting ratios.

So the moral of the story could be that if you want maximum DoF then use poor quality lenses becasue you can close down further before diffraction kicks in more than the other aberrations and if you want highest possible resolution then use high quality lenses.:D

Perhaps it might be a better idea to do a bit of testing to find the sweet spot for your lenses at different focussing distances if you're overly worried about it.