Jump to content




The above adverts really do help Fotozones. Please click on them if they are relevant to you. Not seeing them? Just exclude Fotozones from your ad blocker. Thanks!


Photo
- - - - -

diffraction limitations


  • Please log in to reply
16 replies to this topic

#1 STS

STS

    Advanced Member

  • Life Member
  • 136 posts
  • LocationManchester CT
  • Edit my pics?:Ask Me

Posted 25 June 2013 - 05:06

I am very new to the world of digital photography; I gave up film around 2 years ago, and I hope to receive a quick education from this group.

 

One of the loves of my life is a Nikon F3 that now sits waiting in the closet, combined with its 10 page manual (put in film shot picture).

 

Around two years ago I picked-up a Nikon D7000, combined with its 100 plus page manual,( photographic  heroin), then a D800(photographic crack) came to visit, another 100 plus page manual, combined with Tom Hogan’s book.

 

A few discovery’s, I now know how to clean a sensor, tune a lens and now for the icing on the cake:

 

Nikon sells lens with all of these handy F-Stops, but due to diffraction I only need the ones from let’s say from f2.8 to F8 or F11 if I feel risky (my 75-150 E with an f32 is history).

 

I read rumors of the 56 m-pixel F4x, cool that almost completely removes the need for F-stops (I hope they can add a phone to it).

 

Assuming current sensor size and fabrication (ignoring the Foveon sensor which changes the pixel density equation significantly), at what point will you stop buying pixel density, verses increases in color depth, dynamic range and/or low light performance?

 

Is there software to reduce the effects of diffraction?

 

Have you ever selected a camera based on depth of field?

 

Does changing the from FX to DX mode on the D800 change the pixel density(point at which diffraction starts).

 



#2 pluton

pluton

    Advanced Member

  • Life Member
  • 700 posts
  • LocationSanta Monica, CA USA
  • Edit my pics?:Ask Me

Posted 25 June 2013 - 05:55

Personally, I could stop buying pixel density right now at 36MP.  Others may have different needs or wants.

The biggest print I can make at home is 17x22", and 12MP was a little weak at that size.  36MP gives me a comfortable margin for error.

The software to reduce the effects of diffraction softening is called sharpening.

As nfoto has elegantly pointed out on these pages in the past, depth of field is an illusion.


Keith B.

#3 EL_guest

EL_guest

    Advanced Member

  • Life Member
  • 6,446 posts
  • Location_
  • Edit my pics?:Ask Me

Posted 25 June 2013 - 07:18

Diffraction starts to kick in already at f/5.6 for most DSLRs...

 

Changing mode from FX to DX mode doesn't change anything in this regard, pixel density is fixed...



#4 nfoto

nfoto

    Fierce Bear of the North

  • Administrators
  • 16,649 posts
  • LocationOslo, Norway
  • Edit my pics?:No

Posted 25 June 2013 - 08:23

While the above is true, few lenses are diffraction limited so stopping them down tends to improve image quality nonetheless. Furthermore, what is important is how image sharpness is perceived and in this context local contrast is crucial. Diffraction plays a lesser role than one might think unless the lens is stopped down to say f/22 or f/32, at which f-numbers few lenses can avoid a deteriorated image quality. However, with the proper subject and lighting of it, even f/22 can give very good results on a DSLR. I regularly shoot my Medical-Nikkor 120 mm f/4 at f/16-f/22 and results are great.


Bjørn

#5 crowecg

crowecg

    Advanced Member

  • Life Member
  • 825 posts

Posted 25 June 2013 - 11:19

f/16 on a D7000  

 

7899708552_2e58ed4b8c_b.jpg
 
This is a handheld shot that would just not be possible below f/8 due to lack of DOF and this bug was way too lively for focus stacking.
 
I think I see degradation beyond f/22 but I don't have a sufficiently powerful flash to get enough light in the image to be able to eliminate motion blur from the equation. 


#6 Andrea B.

Andrea B.

    Lost in the Desert

  • Administrators
  • 11,096 posts
  • LocationNuovo Jersey
  • Edit my pics?:Yes

Posted 25 June 2013 - 13:00

Is there software to reduce the effects of diffraction?

 

Software can't reduce diffraction because it is caused by the physics of light and so forth.

 

The effects of diffraction are detail loss because of hitting a resolution limit and loss of local contrast around existing detail. So it can look like a blur is setting in. Detail can never be restored or created, but local contrast can be enhanced so that whatever detail is present can look sharper.

 

How you do that is a matter of experimentation and personal choice. So read up on the topics of sharpness and local contrast adjustment. There is lots of material out there on these topics. And there are tons of useful plug-ins and apps for it.

 

Best thing I ever did was to run a series on a fixed subject at all apertures with a chosen camera + lens combo to see for myself when diffraction began to set in and what the effects of it were. Then I ran some experiments with sharpening and local contrast enhancements to see how best to deal with it. A very, very useful exercise which I highly recommend.

 

Don't forget that you also need to control other sources of blur or detail loss by being on tripod and using good lenses.

 

My experience has been that almost always I can go to f/11 or f/16 if I want to or need to and my finished photo (as prepared for printing) is going to be just fine.

 

So basically the message is to learn good processing techniques and don't worry too much about diffraction.

 

Does changing the from FX to DX mode on the D800 change the pixel density(point at which diffraction starts)?

No. Think about it. The sensor "pixels" do not move when you change from FX to DX. You are just cropping in-camera.

 

...at what point will you stop buying pixel density?

There are two considerations: number of pixels and density of those pixels on the sensor. So I'm not sure what you really mean here. For example, consider 24MP on a DX sensor versus 24MP on an FX sensor. Those 24MP are "denser" on the DX sensor in the sense of being more tightly packed to fit on the smaller DX sensor area.


Andrea B.
UltravioletPhotography.com


#7 nfoto

nfoto

    Fierce Bear of the North

  • Administrators
  • 16,649 posts
  • LocationOslo, Norway
  • Edit my pics?:No

Posted 25 June 2013 - 15:02

The law of diminishing return applies to the increas in sheer pixel numbers. To that, add the fact that potential resolution incrase with the square root of the pixel count.

To double the potential resolution of a 6 MPix camera you need 24 Mpix (if the format is equal). Thus, the current samples of 24 MPix cameras will require 96 MPix if twice the resolution is aimed for. At that point we are into empty magnification so further increases are pretty much useless for DX/FX format. Much larger formats would be required for any future increase over that limit.
Bjørn

#8 STS

STS

    Advanced Member

  • Life Member
  • 136 posts
  • LocationManchester CT
  • Edit my pics?:Ask Me

Posted 25 June 2013 - 15:41

First I would like to thank each of you for your thoughtful answers.

 

The questions where triggered by my experimentation this past weekend at f22. I think many people read the image metadata first, see f22, and then “find” diffraction in the image.

 

The software question was one of my own ignorance; at times it is possible to reduce the effects of physics, as with motion blur, this group is the right group in which to ask that question.

 

The question on pixel density, when using a d800 in DX mode, is based on my own experience reading information off a CCD chip. It is possible to do, but damn difficult with the associated matrix of color filters, but Nikon could do it, and I did not know the answer. Once again this is the right group to ask this question.

 

Once again thank you


Edited by tigers@sprintmail.com, 25 June 2013 - 17:07 .


#9 STS

STS

    Advanced Member

  • Life Member
  • 136 posts
  • LocationManchester CT
  • Edit my pics?:Ask Me

Posted 25 June 2013 - 15:47

If the Foveon sensor could be significantly improved, 3 layers of 24 m-pixels per layer would create a 72 m-pixel sensor, with approximately the same attributes as the current 24 m pixel sensors



#10 Scott Murphy

Scott Murphy

    Advanced Member

  • Members
  • PipPipPip
  • 165 posts
  • LocationPensacola, FL
  • Edit my pics?:Ask Me

Posted 25 June 2013 - 16:33

In most modern lenses, diffraction does not really start to noticeably decrease sharpness until you start closing down below f/ 11. In some wideangle Nikkors, f/11 actually seems to be the optimum f/ stop. The results will become more pronounced the shorter the focal length as the actual diameter of the aperture pupil becomes increasingly smaller,


Edited by stm, 25 June 2013 - 16:34 .

“Sometimes I do get to places just when God's ready to have somebody click the shutter.”

 

- Ansel Adams


#11 Dave Rosser

Dave Rosser

    Advanced Member

  • Life Member
  • 664 posts
  • LocationThornbury, near Bristol

Posted 25 June 2013 - 16:47

You might find this blog an interesting read http://blog.phaseone...e-angle-lenses/

#12 EL_guest

EL_guest

    Advanced Member

  • Life Member
  • 6,446 posts
  • Location_
  • Edit my pics?:Ask Me

Posted 25 June 2013 - 20:31

In most modern lenses, diffraction does not really start to noticeably decrease sharpness until you start closing down below f/ 11. In some wideangle Nikkors, f/11 actually seems to be the optimum f/ stop. The results will become more pronounced the shorter the focal length as the actual diameter of the aperture pupil becomes increasingly smaller,

 

Well, then you have a very loose interpretation of noticeably IMHO.

 

What modern Nikkor wide angle lenses would that be that has f/11 as optimum aperture?



#13 EL_guest

EL_guest

    Advanced Member

  • Life Member
  • 6,446 posts
  • Location_
  • Edit my pics?:Ask Me

Posted 25 June 2013 - 20:40

If the Foveon sensor could be significantly improved, 3 layers of 24 m-pixels per layer would create a 72 m-pixel sensor, with approximately the same attributes as the current 24 m pixel sensors

 

A foveon sensor of 24 mp is 24 mp, it is not 72 mp

A 24 mp RGB sensor is more like a 6 mp + 12 mp + 6 mp camera guessing/calculating a result of 24mp



#14 nfoto

nfoto

    Fierce Bear of the North

  • Administrators
  • 16,649 posts
  • LocationOslo, Norway
  • Edit my pics?:No

Posted 25 June 2013 - 20:43

.. which works because the colours have significant spatial autocorrelation.

 

Even higher MPix counts would improve this even further. (the condition being that pixel smearing is well controlled)


Bjørn

#15 EL_guest

EL_guest

    Advanced Member

  • Life Member
  • 6,446 posts
  • Location_
  • Edit my pics?:Ask Me

Posted 25 June 2013 - 20:44

While the above is true, few lenses are diffraction limited so stopping them down tends to improve image quality nonetheless. Furthermore, what is important is how image sharpness is perceived and in this context local contrast is crucial. Diffraction plays a lesser role than one might think unless the lens is stopped down to say f/22 or f/32, at which f-numbers few lenses can avoid a deteriorated image quality. However, with the proper subject and lighting of it, even f/22 can give very good results on a DSLR. I regularly shoot my Medical-Nikkor 120 mm f/4 at f/16-f/22 and results are great.

 

I agree, I was referring to 'my' use of 'optimum' image quality/sharpness with as little diffraction as possible.

Sure stopping down lenses is allowed for :) as well as opening up fully.



#16 nfoto

nfoto

    Fierce Bear of the North

  • Administrators
  • 16,649 posts
  • LocationOslo, Norway
  • Edit my pics?:No

Posted 25 June 2013 - 20:46

True. Each application has to be evaluated on its own. An optically superior lens will show diffraction effects more clearly as well.


Bjørn

#17 Dave Rosser

Dave Rosser

    Advanced Member

  • Life Member
  • 664 posts
  • LocationThornbury, near Bristol

Posted 25 June 2013 - 22:05

Surely there is an optimum aperture where overall image sharpness in the plane focused on is optimum, typically 2 or 3 stops down from maximum aperture. Any further stopping down can only be to increase depth of field, the blog I referenced in my previous post shows that, for wide angle lenses, diffraction effects rapidly cancel out the sharpening effect of stopping down for objects near the limit of the depth of field.




0 user(s) are reading this topic

0 members, 0 guests, 0 anonymous users



Receive a FREE CAMERA BAG from Think Tank Photo


An appeal to all Fotozones visitors: please help me to keep this site going by starting your gear purchases using any one of the affiliate links shown below:

Amazon.com | Amazon.co.uk | Amazon.de | Adorama.com | thinkTank Photo | DigitalRev.com | OWC | B&H or Donate via PayPal

Starting your shopping here doesn't cost you anything more, but by using the links above (or any others found on the site) you are advising the affiliate that you support this website. This results in a small commission that helps with the running costs. If your preferred outlet isn't among those listed above you can also support the site by making a donation of any amount via PayPal (no PayPal account required). Any donation will be most appreciated.