F16 the limit?

Messages
6,257
Name
Steve
Edit My Images
No
Is it true that on a crop sensor camera (1.6 /1.5) that with any lens the image quality degrades beyond F16?
 
Last edited:
In Kelby and Bryan Peterson books they seem to advie using real small apertures for landscapes... Beyond f/16... How come they don't take diffraction in to account?
 
Do some tests with your setup and see if you can notice a difference. I haven't found it to be a major issue for landscapes personally.
 
Bryan Peterson doesn't give a 4x. He says f22 is fine :shrug:
 
This was shot at f22 - it's sharp enough for me :)

4871670696_1ebbab983c_z.jpg
 
I guess the positives of having larger depth of field outweigh the negatives introduced by diffraction.
 
you start to see diffraction at around f8 on an APS-C sensor however it depends on the lens/camera how much of an effect this is.

you see more of an effect on a 50D compared to a 350D due to the resolving power of the sensor.
 
I use f22 on a tripod for most open landscapes, if i am near water and want to slow the motion it is a must.
If i am taking a more specific thing like a Church within a landscape i normally do f11 to f16
 
you can always use a pinhole..... f128 is a starting point!

The overall sharpest aperture is generally thought to be 3 stops down from full - so a f2.8 should be at the maximum quality around f8

The shaprness does degrade due to diffraction - but if you need the depth of field that's more important....
 
I've found that on my 24-70, the images are just as sharp at F/22 as they are at F/11. Now, whilst I would, in general, agree that smaller apertures introduce more diffraction, to my knowledge, the latest offerings in the high end Nikon market are optimised to be as sharp as possible at F/22. This goes against what I used to think, however, testing this in Scotland last summer, I could not notice any difference in sharpness what so ever when using both aperture values.
 
my 85 is pretty naff at f/16 (it's min aperture). All down to the individual lenses where their 'sweet spot' is. and isn't.
 
never had any issues shooting past f/16 - qualities basically the same
 
Diffraction is more of a problem for cameras with small sensors (at any given aperture) because in order to produce a print (or online image) of any given size (inches, not pixels) you will need to enlarge the smaller image more, thus magnifying the effects of diffraction, shake, blur, misfocus and poor lens IQ more.

If you shoot with medium or large format you will need to use much longer focal lengths for equivalent framing and will need to stop down even more in order to maintain DOF, which will make diffraction worse at the film/sensor plane, but since you won't need to enlarge so much it is not a problem.

Quite simply, the aperture you need to use to obtain a given DOF scales with sensor size. It follows that if you "need" f/16 or f/22 when shooting with a 35mm camera you would need no more than f/10 to f/14 max on an APS-C body. Stopping down beyond f/14 on a cropped camera makes no more sense than stopping down beyond f/22 on 35mm. You just don't need to do it, and if you do then puppies and kittens will perish. Sorry. Them's the facts.

e.g. if you were to shoot a scene with a 5D2, using a focal length of 32mm, an aperture of f/16 and a subject distance of 5' you would get a DOF (according to DOFMaster) covering from 2.94' to 16.6' and totalling 13.7' altogether.

To shoot with an APS-C body you would only need to use a 20mm focal length for the same framing, and stopping down to just f/10 for the subject, still at 5', you will get a DOF covering from 2.91' to 17.9' and totalling 15' altogether. You get roughly equal DOF for the much larger aperture, because you are shooting with a much shorter focal length. Now that's close enough to be the same for practical purposes. If you chose an aperture 1/3 stop either way from f/10 the mismatch between DOFs would be greater.

So, if the crop factor is 1.6X, your focal length choice will vary by a factor of 1.6X for the same framing and you aperture will vary also by 1.6X for the same DOF. When it comes to producing the final image your enlargement factor for the smaller image will be 1.6X greater. See the pattern?
 
Last edited:
The exact aperture at which diffraction will start to degrade image quality will depend on the size of each pixel (a function of sensor size and MegaPixel count).

When I did calculations for my D80 I seem to recall the max aperture I could use before diffraction would set in was f11

TBH it is only really in Macro where really small apertures e.g. f22 are necessary. Landscapes you can use the Hyperfocal focusing method e.g. on 18mm on a Crop @ f8 if you focus at 2m everything between 2m and infinity will be in focus.
 
Diffraction is more of a problem for cameras with small sensors (at any given aperture) because in order to produce a print (or online image) of any given size (inches, not pixels) you will need to enlarge the smaller image more, thus magnifying the effects of diffraction, shake, blur, misfocus and poor lens IQ more.

If you shoot with medium or large format you will need to use much longer focal lengths for equivalent framing and will need to stop down even more in order to maintain DOF, which will make diffraction worse at the film/sensor plane, but since you won't need to enlarge so much it is not a problem.

Quite simply, the aperture you need to use to obtain a given DOF scales with sensor size. It follows that if you "need" f/16 or f/22 when shooting with a 35mm camera you would need no more than f/10 to f/14 max on an APS-C body. Stopping down beyond f/14 on a cropped camera makes no more sense than stopping down beyond f/22 on 35mm. You just don't need to do it, and if you do then puppies and kittens will perish. Sorry. Them's the facts.

e.g. if you were to shoot a scene with a 5D2, using a focal length of 32mm, an aperture of f/16 and a subject distance of 5' you would get a DOF (according to DOFMaster) covering from 2.94' to 16.6' and totalling 13.7' altogether.

To shoot with an APS-C body you would only need to use a 20mm focal length for the same framing, and stopping down to just f/10 for the subject, still at 5', you will get a DOF covering from 2.91' to 17.9' and totalling 15' altogether. You get roughly equal DOF for the much larger aperture, because you are shooting with a much shorter focal length. Now that's close enough to be the same for practical purposes. If you chose an aperture 1/3 stop either way from f/10 the mismatch between DOFs would be greater.

So, if the crop factor is 1.6X, your focal length choice will vary by a factor of 1.6X for the same framing and you aperture will vary also my 1.6X for the same DOF. When it comes to producing the final image your enlargement factor for the smaller image will be 1.6X greater. See the pattern?


Thanks Tim :thumbs:

Very good explanation... :cool:
 
tdodd says it how I sort of understood it when I was reading up on this, it can bit a bit of a mouthful to swallow though, but yes you take f stops into account on a cropped sensor just like you do focal lengths.

my 85 is pretty naff at f/16 (it's min aperture). All down to the individual lenses where their 'sweet spot' is. and isn't.

I was under the impression that f8 is pretty much the starting block for all lenses sweet spots??
 
Landscapes you can use the Hyperfocal focusing method e.g. on 18mm on a Crop @ f8 if you focus at 2m everything between 2m and infinity will be in focus.

A convenient rule of thumb for choosing an aperture when shooting landscapes is to divide your focal length by 2 to get the aperture you need for a large DOF. This guideline works equally well for any sensor size because the simple equation scales exactly with focal length, which in turn scales with sensor size. e.g....

for a 10mm lens you can shoot landscapes pretty well at f/5.0
for a 12mm lens you can shoot at f/5.6 (close enough)
for a 16mm lens you can shoot at f/8
for a 22mm lens you can shoot at f/11
for a 35mm lens you can shoot at ~f/16

If you want to use longer focal lengths for landscape photography on croppers or 35mm then you are going to struggle to get an extensive DOF, at least close to the camera, without pushing well into diffraction territory.

If you want a large, but not huge, DOF then try dividing focal length by 3. It should certainly get you in the ballpark.

I offer this guideline based on the work of Harold Merklinger and his article - "DOF Revisited".
 
Last edited:
How could bryan peterson have got it so wrong?

Because different people have different opinions on what is acceptable, what you may see as being fine others might see as not.

I've shot many at all sorts of apertures and as long as I get the shot I want I am happy :)
 
Last edited:
How could bryan peterson have got it so wrong?
Assuming you're talking about "Understanding Exposure", it's because Bryan Peterson was using a 35mm film camera and, despite the pretence that the book has been updated for digital, it really hasn't. One massive clue - no mention anywhere of the histogram. When you're supposed to be "understanding exposure" I think the histogram would be a tremendous tool to aid learning, and an integral element of using the camera properly and to its full. "Histogram" is not listed at all in the index. Unbelievable! Is it any surprise that other, more technical, topics such as diffraction effects when using different sensor sizes are not covered, nor the way that choice of focal length and aperture will also vary between "full frame" and "digital"? Of course, we get to read little stories about his wife, and his trips around the world, but not the things that are actually important in operating your camera (with respect to exposure) to the best of its potential.
 
Last edited:
Tim no i was talking about the video linked above. He specifically talks about using f/22 on a crop camera. Anyway I guess the trick would be to simply take multiple shots at different apertures and see for myself. It's not like it is going to cost me anything in film I guess :)
 
OK. I have not watched the video. From what I've read in "Understanding Exposure" I have no interest in hearing any more from him.

Back to diffraction - The effect is real, and if you pixel peep you will certainly see it. However, whether or not it is a problem in the real world will depend on how much you enlarge your shot. If all you will do is to make a little 800x533 web image (without first cropping) then the diffraction probably won't be (easily) visible, if at all. If you want a large print then you may well see a problem emerging. For educational purposes, don't stop at f/22 either. If your lens goes to f/32 then why not really see how bad things become? Truly the diffraction effect is significant when you stop down that far.

When shooting on APS-C I rarely stop down beyond f/8, as there is seldom much to be gained, and plenty to be lost. I guess f/11 is my prefered maximum, and f/16 my upper limit. On 35mm I prefer f/16 as a maximum and f/22 as an upper limit. For APS-H I tend to stick closer to my APS-C preferences, but then I don't generally use my crop body cameras for scenes that require a lot of DOF. More usually I am looking to shoot with shallow DOF or for maximum IQ, and that means rarely going above f/8 for either crop format.

Here you go. Here's a series of shots at f/32, f/22, f/16, f/11, f/8, f/5.6 from my 7D and 70-200, focused using Live AF and monted on a tripod with mirror lockup and timer release. They were shot raw and cropped to 100% and converted in DPP with Standard picture style and sharpening at 3.

f/32 :
20100811_120557_0713_DPP.JPG


f/22 :
20100811_120553_0712_DPP.JPG


f/16 :
20100811_120549_0711_DPP.JPG


f/11 :
20100811_120545_0710_DPP.JPG


f/8 :
20100811_120541_0709_DPP.JPG


f/5.6 :
20100811_120537_0708_DPP.JPG


I would show f/4 and f/2.8, but the six image limit means I can't include them, and to be honest it is hardly necessary. You can see that even at f/16 there is visible image softening, and beyond that things just get worse and worse.

Apart from the visual clues to the problem, if we look at the file sizes of the crops, which I think you'll agree look identical, there is a clear trend, with softer files having smaller file sizes....

f/32 = 229KB
f/22 = 251KB
f/16 = 274KB
f/11 = 290KB
f/8 = 295KB
f/5.6 = 294KB
f/4 = 280KB
f/2.8 = 278KB

So while it's a little risky to take those figures as hard absolutes, there is undoubtedly a pattern and the lens may well be sharper wide open than at f/16. When you also consider that the gap between f/16 and f/2.8 is 5 stops, that is a lot of wiggle room to play with on shutter speed and ISO, both of which could radically improve your IQ. So, by all means stop down as far as you need to, but don't go further than necessary. You may end up doing more harm than good.
 
Last edited:
Tim no i was talking about the video linked above. He specifically talks about using f/22 on a crop camera. Anyway I guess the trick would be to simply take multiple shots at different apertures and see for myself. It's not like it is going to cost me anything in film I guess :)

That's saved me typing! Yup, to see what your kit's capable of and what YOU'RE happy with, do your own testing.
 
Tim,

Excellent explanation there :thumbs:, very clear and illustrative. I have to say, your edit of the post of you vs. Bryan Peterson is a shame, I enjoyed reading it last night, and love to hear non-conventional points of view.

I've always preached sticking to f/8 ~ f/11, for the exact reason you mention, and still don't understand why do many of the better-known names always suggest to go up to as small an aperture as possible :shrug:. I can't see their logic :thinking:

Again, thank you for your clear input on the matter.
 
Sorry for the edit. I thought I should stay a bit more on topic rather than allow my usual rant about that book continue, especially as I wasn't sure that the book was in fact the point of concern.
 
Back
Top