Pages: [1]   Go Down

Author Topic: Shooting at f/22, f/32 and f/45 with small pixels?  (Read 4799 times)

dreed

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 1715
Shooting at f/22, f/32 and f/45 with small pixels?
« on: September 02, 2011, 12:25:58 pm »

When I look at many classic landscape photographs, I see notations such as "f/32" or "f/45" for film based prints. Down to almost pin-hole size openings, depth of field is huge and detail sharp.

If I were to try creating a similar photograph with a digital camera, I'm faced with two problems..

1) The first is that most lenses that I'm likely to put on a DSLR have their peek f-stop (in terms of sharpness) somewhere between 5.6 and 10.

2) The second is that the "circle of confusion" from diffraction and pixel sizes.

How should these be worked around?
Just ignore both?
For (2), do we need to wait until pixels are much smaller than the circle, so much so that they can actually define the circle rather than be confused by it?
Logged

Tony Beach

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 452
    • http://imageevent.com/tonybeach/twelveimages
Re: Shooting at f/22, f/32 and f/45 with small pixels?
« Reply #1 on: September 02, 2011, 12:41:04 pm »

Different formats.  The challenges of working with a smaller format are different than what you have with a larger format, and you need enlarge the smaller format more.  One work around is to stitch multiple frames, thus making the smaller format into a larger format, then you end up enlarging less and can get more printable detail out of a larger CoC.  To some degree though, changing formats is a zero sum game, enlarging less and stopping down more as you increase focal length and reduce DOF and you end up treading water or only making marginal gains.  I would note too that f/64 on a 4x5 format is something like f/8 on 135 format, so the trick to getting more DOF is to use tilt lenses (or camera movements if you have them) in a clever way.
« Last Edit: September 02, 2011, 12:43:04 pm by Tony Beach »
Logged

ErikKaffehr

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 11311
    • Echophoto
Re: Shooting at f/22, f/32 and f/45 with small pixels?
« Reply #2 on: September 02, 2011, 02:20:47 pm »

Hi,

The problems were always quite similar. LF and MF format lenses were never imune tro diffraction! But if we only have a small format we need to make much more optimal use of it.

I would suggest that stopping an MF lens down to f/22 will yield a sharpness comparable to f/11 on "full frame". To utilize MF fully we need to stop down to f/8 or F/11. Large format lenses are often not as well corrected as smaller format lenses, so they can be stopped down more because they are less sharp, so there is less to loose by stopping down.

Now, LF lenses cower a larger area, so total sharpness may be better anyway. But, would an LF lens be ideally corrected it would also perform best at f/5.6 - f/8. This is indeed true of HR digitars from Rodenstock.

Best regards
Erik

When I look at many classic landscape photographs, I see notations such as "f/32" or "f/45" for film based prints. Down to almost pin-hole size openings, depth of field is huge and detail sharp.

If I were to try creating a similar photograph with a digital camera, I'm faced with two problems..

1) The first is that most lenses that I'm likely to put on a DSLR have their peek f-stop (in terms of sharpness) somewhere between 5.6 and 10.

2) The second is that the "circle of confusion" from diffraction and pixel sizes.

How should these be worked around?
Just ignore both?
For (2), do we need to wait until pixels are much smaller than the circle, so much so that they can actually define the circle rather than be confused by it?
Logged
Erik Kaffehr
 

250swb

  • Full Member
  • ***
  • Offline Offline
  • Posts: 220
Re: Shooting at f/22, f/32 and f/45 with small pixels?
« Reply #3 on: September 03, 2011, 04:25:59 am »

When I look at many classic landscape photographs, I see notations such as "f/32" or "f/45" for film based prints. Down to almost pin-hole size openings, depth of field is huge and detail sharp.

If I were to try creating a similar photograph with a digital camera, I'm faced with two problems..

1) The first is that most lenses that I'm likely to put on a DSLR have their peek f-stop (in terms of sharpness) somewhere between 5.6 and 10.

Classic landscape photographs were made with large format cameras. It should be born in mind that the notations you read of f/32 or f/64 were with large format lenses. Why is this important?

Well as you have alluded, a DSLR lens is optimum at say f/5.6 but you start to get diffraction beyond this. Diffraction is when the light rays bounce off the edges of the shutter blades and cause degradation of the image. It gets worse the more the lens is stopped down. So how do large format lenses survive to f/64? Well its because the opening is much bigger. I don't have a lens thats goes to f/64 to hand, but the one I have here goes to f/45, and the aperture is considerabley bigger at f/45 than a 35mm format lens at f/22. So you need to take these notations in context, there are not equivalent in 35mm photography when diffraction is concerned.

Steve

ErikKaffehr

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 11311
    • Echophoto
Re: Shooting at f/22, f/32 and f/45 with small pixels?
« Reply #4 on: September 05, 2011, 12:57:35 am »

Hi,

The physical size of the aperture does not matter for diffraction. F-stop does! The reason for this is outside of the scope of discussion, I think. The reason that large format lenses work well with small apertures are:

- They are simply not corrected for large apertures. You need to stop down to f/11 f/16 for optimum sharpness
- The film image is much larger and therefore needs more enlargement

So LF lenses are affected by diffraction exactly as any other lens, but the playing ground is different.

There are high resolution LF lenses made for digital, and they peak around f/5.6, but they also have small image circles.

Best regards
Erik

Ps. The size of the aperture decides diffraction but the focal length decides how it is projected. As it happens f-stop takes all that into account! F-stop is a very nice invention, replacing lots of algebra!

Classic landscape photographs were made with large format cameras. It should be born in mind that the notations you read of f/32 or f/64 were with large format lenses. Why is this important?

Well as you have alluded, a DSLR lens is optimum at say f/5.6 but you start to get diffraction beyond this. Diffraction is when the light rays bounce off the edges of the shutter blades and cause degradation of the image. It gets worse the more the lens is stopped down. So how do large format lenses survive to f/64? Well its because the opening is much bigger. I don't have a lens thats goes to f/64 to hand, but the one I have here goes to f/45, and the aperture is considerabley bigger at f/45 than a 35mm format lens at f/22. So you need to take these notations in context, there are not equivalent in 35mm photography when diffraction is concerned.

Steve

Logged
Erik Kaffehr
 

ErikKaffehr

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 11311
    • Echophoto
Re: Shooting at f/22, f/32 and f/45 with small pixels?
« Reply #5 on: September 05, 2011, 01:12:58 am »

Hi!

The circle of confusion tells what is accepted unsharpness. Diffraction cases a different shape of unsharpness. The difference is that defocus gives a uniformly lit disc (ideally) while diffraction gives a cone. Diffraction can be "deconvolved" better than defocus.

The figures here illustrate the effects of diffraction and defocus:

http://echophoto.dnsalias.net/ekr/index.php/photoarticles/49-dof-in-digital-pictures?start=1

Red is diffraction and green is defocus.

DSLRs and latest MFDBs resolve around 90 lp/mm while "normal" films only resolve around 50 lp/mm, but LF photography was seldom limited by film resolution. So digital is much more demanding, therefore diffraction is more noticeable.

Check this very accurate article by Lars Kjellberg
http://www.qpcard.se/BizPart.aspx?tabId=76

Images at end indicate that T-MAX 100 on 135 at f/5.6 equals Tri-X on 4x5" at f/22! The article is worth a careful study. I include a screendump comparing 135 f/5.6 TMAX-100 and 4x5" f/22 Tri-X.


Best regards
Erik


When I look at many classic landscape photographs, I see notations such as "f/32" or "f/45" for film based prints. Down to almost pin-hole size openings, depth of field is huge and detail sharp.

If I were to try creating a similar photograph with a digital camera, I'm faced with two problems..

1) The first is that most lenses that I'm likely to put on a DSLR have their peek f-stop (in terms of sharpness) somewhere between 5.6 and 10.

2) The second is that the "circle of confusion" from diffraction and pixel sizes.

How should these be worked around?
Just ignore both?
For (2), do we need to wait until pixels are much smaller than the circle, so much so that they can actually define the circle rather than be confused by it?
« Last Edit: September 05, 2011, 01:17:15 am by ErikKaffehr »
Logged
Erik Kaffehr
 

250swb

  • Full Member
  • ***
  • Offline Offline
  • Posts: 220
Re: Shooting at f/22, f/32 and f/45 with small pixels?
« Reply #6 on: September 05, 2011, 03:17:55 am »

Hi,

The physical size of the aperture does not matter for diffraction. F-stop does! The reason for this is outside of the scope of discussion, I think. The reason that large format lenses work well with small apertures are:

- They are simply not corrected for large apertures. You need to stop down to f/11 f/16 for optimum sharpness
- The film image is much larger and therefore needs more enlargement


The size of the hole matters because pro rata as it gets smaller the amount of interference becomes greater. Given the aperture blades of a large format lens are the same thickness as those of a DSLR lens if the hole's were the same size (whatever the respective apertures read) the diffraction would be the similar. But that could mean the large format lens is at f/64, and the DSLR lens is at f/22. Then the next pro rata aspect comes into play in that the diffraction's affect on the image is greater for the DSLR than the large format negative due to the sensor/film size.

If a large format lens is just about breaking into a sweat and affecting image quality at f/64 then the DSLR lens will be breaking into a similar sweat around f/8, beyond which there is nothing to gain and everything to loose, for an average lens. Some prime lenses are so good that an image made at f/22 will be no worse than a zoom lens at its optimum of f/5.6.

Steve

Jack Varney

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 413
    • http://
Re: Shooting at f/22, f/32 and f/45 with small pixels?
« Reply #7 on: September 05, 2011, 08:01:53 pm »

The size of a lens opening does affect diffraction. The reason is because diffraction is caused by the diffraction of the light at the edges of the aperture blades and the image is made over the area of the opening. Since the edge is linear in relation to opening and area is is exponential, in smaller openings diffraction is appreciably larger relative to image. Therefore. large format lenses show less diffraction when stopped down to apertures not suitable to MF and smaller formats. Certainly the larger format contributes, also.
Logged
Jack Varney

ErikKaffehr

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 11311
    • Echophoto
Re: Shooting at f/22, f/32 and f/45 with small pixels?
« Reply #8 on: September 05, 2011, 09:34:12 pm »

Yes,

But you also need to take the focal distance into account because the diffraction peaks are at different angles. So a longer focal length will project a larger diffraction ring. Using F-stop takes this into account.

Best regards
Erik



The size of a lens opening does affect diffraction. The reason is because diffraction is caused by the diffraction of the light at the edges of the aperture blades and the image is made over the area of the opening. Since the edge is linear in relation to opening and area is is exponential, in smaller openings diffraction is appreciably larger relative to image. Therefore. large format lenses show less diffraction when stopped down to apertures not suitable to MF and smaller formats. Certainly the larger format contributes, also.
Logged
Erik Kaffehr
 

Bart_van_der_Wolf

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 8913
Re: Shooting at f/22, f/32 and f/45 with small pixels?
« Reply #9 on: September 06, 2011, 03:29:10 am »

But you also need to take the focal distance into account because the diffraction peaks are at different angles. So a longer focal length will project a larger diffraction ring. Using F-stop takes this into account.

Correct, the f/# is a fraction, in which the focal length is the numerator. Thus the angular resolution is constant, and diffraction is only influenced by the F-number (and wavelength). IOW f/22 is f/22 regardless of actual focal length, and the amount of diffraction (diameter of the first minimum of the Airy disk from the center) is the same (2.44 x N x Lambda, where N is the f/# and Lambda is wavelength).

Cheers,
Bart
Logged
== If you do what you did, you'll get what you got. ==

Jack Varney

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 413
    • http://
Re: Shooting at f/22, f/32 and f/45 with small pixels?
« Reply #10 on: September 06, 2011, 11:40:48 am »

"Certainly the larger format contributes, also."

My error, it is the larger format and a presumed larger circle of confusion for larger formats, since airy disk radius is given by-

r=1.22λ F/a.
Logged
Jack Varney

ErikKaffehr

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 11311
    • Echophoto
Re: Shooting at f/22, f/32 and f/45 with small pixels?
« Reply #11 on: September 06, 2011, 12:52:05 pm »

Hi,

Keep in mind that with a DSLR you can use a larger fstop for similar depth of field, so f/32 may correlate to f/11 (depending on LF-size and DSLR format).

The issue is a bit complex, as it also depends on your definition of what is acceptable sharpness.

Best regards
Erik


When I look at many classic landscape photographs, I see notations such as "f/32" or "f/45" for film based prints. Down to almost pin-hole size openings, depth of field is huge and detail sharp.

If I were to try creating a similar photograph with a digital camera, I'm faced with two problems..

1) The first is that most lenses that I'm likely to put on a DSLR have their peek f-stop (in terms of sharpness) somewhere between 5.6 and 10.

2) The second is that the "circle of confusion" from diffraction and pixel sizes.

How should these be worked around?
Just ignore both?
For (2), do we need to wait until pixels are much smaller than the circle, so much so that they can actually define the circle rather than be confused by it?
Logged
Erik Kaffehr
 

EricV

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 270
Re: Shooting at f/22, f/32 and f/45 with small pixels?
« Reply #12 on: September 06, 2011, 01:17:25 pm »

But you also need to take the focal distance into account because the diffraction peaks are at different angles. So a longer focal length will project a larger diffraction ring. Using F-stop takes this into account.

A longer focal length enlarges not only the diffraction ring, but also everything else in the photograph, so less enlargement is needed when printing or displaying the image.  The size of the diffraction disc on the sensor does depend on F/stop only, but the size of the diffraction disc on a final print depends on the diameter of the lens opening only.  

If I take a picture of a tree with a 50mm lens at f/8, diffraction will blur everything on the sensor by around 10um.  Let's say the tree is so far away that a single leaf has a size of 1mm on the sensor.  So diffraction blurs the image by around 1/100 the size of a leaf.  If I take the same picture with a 200mm lens at f/32 (perhaps switching to a larger format sensor, to maintain the same field of view, to capture the entire tree from the same location), diffraction blur will be 40um, but the leaf will be 4mm, and diffraction will still be 1/100 of a leaf.  When I make prints from both images, with the tree and leaves printed the same size, diffraction blur will be the same size on both prints.
« Last Edit: September 06, 2011, 01:24:43 pm by EricV »
Logged

joneil

  • Full Member
  • ***
  • Offline Offline
  • Posts: 163
  • This is what beer does to you....
Re: Shooting at f/22, f/32 and f/45 with small pixels?
« Reply #13 on: September 06, 2011, 05:32:49 pm »

When I look at many classic landscape photographs, I see notations such as "f/32" or "f/45" for film based prints. Down to almost pin-hole size openings, depth of field is huge and detail sharp.


-snip-

   There are two other issues that come up in large format that you will not always see in a digital SLR.   The first one is size of image circle and movements.  For example, a lens may just barely cover a sheet of 4x5 film without movements, but stop it down to say F22, and the image circle increases a wee bit, so you are able to use more movements.  This is true for some, but not always all large format lenses.

    Secondly - and i don't know why - on 4x5 film, the 210mm lens for years was the "standard".  Especially in some photography schools.    For example, my local community college photography course, the large format section they always sent out students with 210mm lenses.   On 4x5, a 150mm lens is roughly the same as  50mm lens on 35mm film or a full frame DSLR.  So 210mm is kinda like a 85mm lens - a short telephoto with shallow depth of field.   So to get a full depth of field you are pretty much forced in to F22 or more.  Especially if you are into tilting and/or shifting a lens, stopping down is often the way to go. 

  I guess that's where that old saying in large format come from :   "F64 and just be there".  :)

   On a DSLR, especially wide angle lenses, even when I want good depth of field, personally I seldom find the need to go past F8.  I usually make that my concern first - what I want the final image to look like, and I worry about lens diffraction, etc, last.  For example, one of my personal favourite photographs I shoot was of a creek on the family farm in infra-red film, 35mm HIE.  That old stuff form Kodak had grain the size of small pebbles.  :)

   Enlarged to 11x14" print, man it was anything but full of fine detail.  But the coarse look combined with the infra-red gave it a really dreamy, surrealistic look that I personally liked.

   So sometimes it's more important to get how you want the image to look, and worry about the effects of diffraction, pixel size, etc, secondly.
Good luck

joe

Logged

torger

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 3267
Re: Shooting at f/22, f/32 and f/45 with small pixels?
« Reply #14 on: September 07, 2011, 10:49:37 am »

The reason one could use f/64 on large format film is that the film size is so large that despite large diffraction you still get high resolution.

However, for "maximum DoF" images format does not matter in theory (assuming perfect lenses etc), since with a larger format you need longer focal length for the same angle of view which must be compensated with smaller aperture. With a smaller format you indeed need larger aperture to minimize diffraction, but can compensate with shorter focal length to get the same DoF.

The reason larger formats have "higher resolution" for maximum DoF images is because lenses for larger formats can be made to resolve more over the total film/sensor area. While a smaller format lens is typically sharper (due to smaller image circle), the format is more small than the lens is sharper so you in total get less resolution in the image. In the film era I guess resolution advantages of the larger formats was more due to film grain than lens sharpness, but I don't know. In digital we can make pixels really small, but film grain will be same size regardless of format.

Roughly speaking, the current sharpest digital lenses for 645 medium format seems to have a resolving power of 80-100 megapixels, 35mm full frame about 30-35 megapixels and APS-C about 12 - 14 megapixels. In other words, the latest APS-C sensors (16 - 24 megapixels) are maxing out the lenses, while 35mm full-frame and medium format could use some more (although IQ180 is up there).

The new IQ180 80 megapixel medium format sensor has sensel/pixel size which would yield about 32 megapixel 35mm fullframe. The same aperture that would give diffraction problems on 32 megapixel 35FF will thus give problems with IQ180. So you'd probably want to use f/8 or so if you want to get full use of those 80 megapixels.
« Last Edit: September 07, 2011, 11:00:10 am by torger »
Logged

EricV

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 270
Re: Shooting at f/22, f/32 and f/45 with small pixels?
« Reply #15 on: September 07, 2011, 12:54:09 pm »

When I look at many classic landscape photographs, I see notations such as "f/32" or "f/45" for film based prints. Down to almost pin-hole size openings, depth of field is huge and detail sharp.
These are *not* pinhole size lens openings.  On a 50mm lens, for 35mm format, an aperture of f/5.6 provides a lens opening of 9mm.  On a 400mm lens, which provides a comparable field of view for 8x10" format, an aperture of f/45 provides the same lens opening of 9mm.  When images of the same scene are taken using those lenses at those apertures on both formats and printed to the same size, diffraction effects will be similar (see previous post).
Logged
Pages: [1]   Go Up