[lithedreamer]

In what practical situations would you ever want to go to really narrow apertures? Even on the DFA 100 macro, when trying to eke out maximum DOF for a macro shot, my perception is that it's not worth going beyond f16. (Of course focus stacking would be a better way to achieve greater DOF in macro shots, but let's assume that's not practical.)

On the K-1, you're working with shallower depth of field than APS-C, so I could see myself going to f/22 for some shots.

[UncleVanya]

I remember reading somewhere that f16 is the smallest aperture you should use with DSLR's.
Using smaller apertures would cause softness etc. issues due to light diffraction.
I guess this holds true also for K3?

The f stop where diffraction occurs is debated a bit. Some of this is perceptual and some is that different size sensors have earlier onset of diffraction. However for APSC I most often hear f/8 or f/11 is the max f/stop before diffraction. For Full frame f/11 or f/16 seems more common. Micro 4/3's is typically f/4 or f/5.6.

[Cthulhugan]

The f stop where diffraction occurs is debated a bit. Some of this is perceptual and some is that different size sensors have earlier onset of diffraction. However for APSC I most often hear f/8 or f/11 is the max f/stop before diffraction. For Full frame f/11 or f/16 seems more common. Micro 4/3's is typically f/4 or f/5.6.

It is pixel size that maters, not sensor size. Full-frame and medium format sensors tend to have larger pixels than APS-C and m4/3, however the K-1 and K-5 pixels are about the same size and so should show similar degradation due to diffraction at any given aperture. The K-3 has smaller pixels, and so should so more degradation at the same aperture. The Q's pixels are even smaller. Don't stop down a Q to f/32 ...

[clackers]

There is a calculator at:

Diffraction Limited Photography: Pixel Size, Aperture and Airy Disks

If you go along with its assumptions, a K-3 displays diffraction at f16 that the K-1 doesn't.

A point-and-shoot is starting to get diffraction as early as f5.6, according to that model.

In any case, if your macro or landscape pic needs the aperture closed down heavily, you do it, don't you?

[lithedreamer]

There is a calculator at:

Diffraction Limited Photography: Pixel Size, Aperture and Airy Disks

If you go along with its assumptions, a K-3 displays diffraction at f16 that the K-1 doesn't.

A point-and-shoot is starting to get diffraction as early as f5.6, according to that model.

In any case, if your macro or landscape pic needs the aperture closed down heavily, you do it, don't you?

Nice find. The calculator indicates that the K-1 shouldn't have a problem with diffraction until f/32.

[UncleVanya]

It is pixel size that maters, not sensor size. Full-frame and medium format sensors tend to have larger pixels than APS-C and m4/3, however the K-1 and K-5 pixels are about the same size and so should show similar degradation due to diffraction at any given aperture. The K-3 has smaller pixels, and so should so more degradation at the same aperture. The Q's pixels are even smaller. Don't stop down a Q to f/32 ...

That is correct. I oversimplified.

[Digitalis]

The calculator indicates that the K-1 shouldn't have a problem with diffraction until f/32

wow, just wow.

[BrianR]

wow, just wow.

At what f-stop do you think diffraction gets noticeable for an 8x10" print from a 35mm sensor, given that the viewer is 25cm away coupled with "standard" eyesight (these are the basic assumptions of the calculator, no room to enter # of beers consumed by the viewer)?

[Pete82]

Great information here, thanks!

[Na Horuk]

At what f-stop do you think diffraction gets noticeable for an 8x10" print from a 35mm sensor, given that the viewer is 25cm away coupled with "standard" eyesight (these are the basic assumptions of the calculator, no room to enter # of beers consumed by the viewer)?

Why not just a do a test for yourself? Thing is, the lens has its effect. And so does the scene in front of you - some photos require immense resolution, others not at all. And finally, the Post processing. Some programs use very good sharpening algorithms and can mask some diffraction. It comes down to a compromise of DoF and sharpness. Oh, and printing itself can affect the Image quality. Some print shops will take a bit of time and give you great results. Others will just throw the photo through an automated process, which can reduce overall IQ.

I can't give you an absolute answer for K-1 prints, but i avoid going above f9 with pretty much all my lenses. I use DFA 100mm at more than f9 sometimes, because in macro you need DoF, and the lens is sharper than average lenses

[BrianR]

Why not just a do a test for yourself?

Send me a K-1 and I'd be happy to. I was making a point - that calculator makes some default assumptions about print size and viewing conditions that shouldn't be ignored. While you'll take a hit from diffraction at f/32 on a 24"x36" print viewed with a magnifying glass, diffraction is the least of your problems when viewing the same image on Facebook using your smartphone. It's kind of important to have some context (and also to stay away from Facebook if possible).


And yes, 'test it yourself' is my mantra too. I know what I find objectionable and am pretty good at coming up with the compromise that works for me, my output and my eyes. Not that it's happened here, but I cringe when I see people dispensing hard and fast rules about what f/stops everyone should avoid - experiment for yourself, print, make notes, do better next time...

[Clou]

Nice find. The calculator indicates that the K-1 shouldn't have a problem with diffraction until f/32.

Did you use CoC or pixel size for calculating this?

[lithedreamer]

Did you use CoC or pixel size for calculating this?

Good question. When I check the box that says 'Set circles of confusion based on pixels' and throw in 36 megapixels, the calculator suggests diffraction starts at f/11.

[Digitalis]

# of beers consumed by the viewer

American beer is weaksauce compared to what we get here in oz.

Thing is, the lens has its effect.

Finally someone gets it, lens resolution is very low at f/32 - and this is true for all lenses. Due to the effects of diffraction - no matter how much resolution the sensor has it's only going to record perhaps 20% of the detail that is there due to lens diffraction, resolution figures only get worse for the already interpolated red/blue channels. Hence the need for focus stacking for macrophotography and landscape photography.

Will it be visible on an 8X10 print? possibly. Scoff if you must, but if both images are un-sharpened there will be a visible difference even at 8X10 print size.

[BrianR]

Will it be visible on an 8X10 print? possibly. Scoff if you must, but if both images are un-sharpened there will be a visible difference even at 8X10 print size.

At f/32, I don't think anyone would disagree with this, that calculator sure doesn't.

I can't speak to 35mm, but I do know aps-c and 8x10 prints pretty well, and what I've found is in agreement with the calculator, which puts f/16 just on the cusp of 'diffraction limited' for this sensor and print size. Out of my dfa100mm macro, I do notice a drop in sharpness from f/16 to f/22 at 8x10, and I'd only use f/22 if I have a good reason. F/11 to f/16 also has a drop if I went looking for it, but not so much that I spent time worrying about it.

This was especially true when I was using a 6mp sensor. It just couldn't capture enough detail that the loss in lens resolution from f/11 to f/16 wasn't as big a deal as the loss of DoF from f/16 to f/11 was. This also seems to agree with the calculator when you use pixel pitch instead of print size as your criteria.

Since moving to a k5iis, I've found my same guidelines work fine for 8x10 prints, but when it comes to larger stuff (or more commonly pixel peeping), I've knocked my guidelines back about a stop and a half for general use. That said, f/5.6 to f/8 and focus stacking is a beautiful thing when it's applicable.