7

I'm facing often the same kind of trouble when taking landscape pictures...

I end up with a picture that could be great but that is totally not sharp !

Example :

enter image description here

I set up the maximal focal distance I could, hence f/22 here... But nothing seems sharp...

More information : Canon Objectif EF-S 18-55 mm f/3,5-5,6 IS II
F/22
1/60 sec
ISO-100
focal = 50 mm

enter image description here

Here, nor the trees neither the mountains seem sharp and neat...

More information : Canon Objectif EF-S 18-55 mm f/3,5-5,6 IS II
F/22
1/15 sec
ISO-100
focal = 18 mm

What is the trouble here ?

Michael C
  • 175,039
  • 10
  • 209
  • 561
Andy M
  • 642
  • 7
  • 21

5 Answers5

17

There are two issues at work that are causing your results to be a little soft:

  • Diffraction Since you are using an EF-S lens it is safe to assume you are using a Canon APS-C camera. Most of the recent models have pixels pitches that cause diffraction to begin at around f/6.8-6.9. This is the point at which the affects of diffraction begin when viewed at the pixel level. As apertures are narrowed beyond the Diffraction Limited Aperture (DLA) the results get more and more noticeable at normal viewing sizes. The best way to avoid this is to shoot at around f/8 or wider and at f/6.3 or wider if possible.
  • Slow Shutter Speed Rare is the photographer who can handhold at 1/15 second and get blur free results at the pixel level. You may get useable results for viewing at smaller sizes such as 8x10 or so, but nowhere near the equivalent viewing size of looking at part of an image at 100% on your monitor. If you have an HD (1920x1080 pixels) monitor that measures 23" diagonally you are viewing images at 96 ppi. That means an 18MP image viewed at 100% is being magnified at the equivalent of 54x36 inches!

By reducing the aperture from f/22 to somewhere around f/5.6-8.0 you will be able to increase your shutter speed to around 1/250 second which is much more forgiving when hand holding a shot using a focal length in the 18-50mm range. For the sharpest results you should always use a tripod. And even when you do all of that correctly, the EF-S 18-55mm f/3.5-5.6 IS II is a decent lens but it is by no means the sharpest available at those focal lengths and apertures.

Bear in mind that most of the amazing landscape photos you see at sites such as 500px and flickr have extensive post processing applied that tends to sharpen the results compared to how the image first looks straight out of the camera. In some cases advanced techniques such as focus stacking and highly detailed lens correction are being applied.

Michael C
  • 175,039
  • 10
  • 209
  • 561
6

You should not use the smallest aperture available, you should use an aperture that allows you to focus at an hyperfocal distance that puts both the foreground and infinity in focus.

If you use an aperture that is too small, diffraction will make the image soft. You will get a great DOF, but most of that will be beyond what you need to get your subject in focus.

Small apertures can mean soft images – why?

The ideal DOF would stretch from the object in your foreground to infinity.

There is a formula on the Wikipedia page that I linked to, and applying that to 18mm and f22 gives you:

H = 18² / (22 * 0.03) = 490.9 mm

That means that you have a DOF that stretches from H/2 = 0.24 m to infinity. That is clearly more than you need.

If we try it with f5.6 instead:

H = 18² / (5.6 * 0.03) = 1928,6 mm

That gives you a DOF that stretches from 0.96 m to infinity. That is still more than you need in most cases.

If we try that with the 50mm lens:

H = 50² / (5.6 * 0.03) = 14881 mm

That gives you a DOF that stretches from 7.4 m to infinity, which should be enough for most cases.

So, you have been trying apertures that are way too small for your needs. You can aim for an aperture that gives you the optimum DOF, or perhaps half a stop or a full stop smaller just to be on the safe side. That should still keep you far from the smallest aperture, and also give you a lot better exposure times.

Guffa
  • 21,309
  • 3
  • 58
  • 85
  • Thank you very much for the detailed answer. Very interesting ! I'll study Hyperfocal distance for I have heard about it but never tester properly... Thank you ! – Andy M Jan 25 '14 at 14:26
  • Note that all of the formulas above are only valid for standard viewing conditions: an 8x10 viewed at 1 foot by a person with 20/20 vision. Just as with DoF, the more you enlarge a photo the further the hyperfocal distance moves from the camera for the same aperture value. – Michael C May 07 '16 at 11:11
  • @MichaelClark: What do you mean? How could the conditions when the image was taken change by enlarging the photo afterwards? – Guffa May 10 '16 at 08:48
  • Viewing conditions also affect DoF. When you enlarge you also enlarge blur circles and thus need to reduce the size of the CoC in order to get the same percieved DoF. – Michael C May 10 '16 at 08:52
  • For more, please see http://photo.stackexchange.com/a/70016/15871 and http://photo.stackexchange.com/questions/75471/why-i-am-getting-different-values-for-depth-of-field-from-calculators-vs-in-came/75474#75474 – Michael C May 10 '16 at 08:56
  • @MichaelClark: The formulas doesn't give a percieved DoF, they give absolute measurements in the real world. It doesn't matter what you do to the percieved DoF, the absolute measurements doesn't change. – Guffa May 10 '16 at 08:57
  • There's only one distance that is really in focus. All other distances are out of focus to one degree or another. What we call DoF is what appears to be in focus because the amount of blur is so small that we perceive it as a point. There is no "magic barrier" within which everything is equally in focus and beyond which everything is not in focus. – Michael C May 10 '16 at 08:59
  • Almost all DoF charts and calculators assume a viewing size of 8x10" viewed at a distance of 12" by a person with 20/20 vision. That, plus the size of the image sensor (which determines the magnification factor from the virtual image projected on the sensor to the display medium) is what determines the CoC that must be used to calculate DoF. It is why a CoC of about .03mm is used for FF cameras while a CoC of about .02mm is used for APS-C cameras: because the image on the APS-C sensor must be magnified by a factor of 1.5X more than the FF camera to get the same viewing size. – Michael C May 10 '16 at 09:04
  • http://www.cambridgeincolour.com/tutorials/dof-calculator.htm Click the "advanced" button, change the display size and see what happens to the DoF and hyperfocal distance. – Michael C May 10 '16 at 09:18
  • 1
    Your calculations don't seem to take into account the crop factor of the sensor. – shrx Mar 05 '17 at 09:48
1

Use a tripod. Shutter shake is a very frustrating aspect of modern cameras. When I had my old Fujica 801 I could shoot at 1/60th of a second and take a sharp photo. I would not trust myself below 1/250th on my Fujifilm Finepic. Try shooting on a clear day - you will have more light and less problems to solve

MikeW
  • 33,408
  • 10
  • 88
  • 123
Will MCcarter stock photos
  • 11
  • 1
1

Note also that the old rule of thumb that a shutter speed of 1/30 second is fast enough for a sharp picture is not good enough in the digital age. In the analogue era you would not typically zoom a picture to see very small details, so a picture would look sharp and that was it. Today, we take one picture and zoom in expecting to see small details clearly right until the pixels become visible. But if you magnify a picture by factor of 10, then motional unsharpness is also going to be magnified by a factor of ten. So, just like when you are using a zoom lens, you should reduce the shutter time to take this into account.

Count Iblis
  • 3,616
  • 1
  • 13
  • 17
0

Apart from diffraction and slow shutter you also have the problem that you are shooting with a kit lens. Those images look more to me like kit lens material than shaky hand material.

You could try to shoot F5.6-F8 and use tripod with mirror lockup or iso 400 so you get 1/100s but that doesnt change the fact that you are shooting with a cheap lens that they toss at you for free when you buy the camera, because it is worthless.

Michael Nielsen
  • 10,644
  • 35
  • 48
  • I wouldn't exactly call them worthless. Today's kit lenses are better than most of those used by the great photographers of the first half of the 20th century. Why do you think the technique of an unsharpen mask was developed in the 1930s? – Michael C Jan 26 '14 at 08:31
  • The review you linked to is for a Canon kit lens several generations ago. The EF-S 18-55mm f/3.5-5.6 IS and IS II version were a new design with significant improvements over that lens, which was known to have wide variations in quality from one lens to the next. – Michael C Jan 26 '14 at 08:35
  • OK, KenRockWell states IS II is not better than IS and digital picture has mouseover comparisons between nonIS II vs IS and IS vs F2.8 USM. While the reviews do say the IS is "acceptable" in the wide end (which is where OP is using it, while to me a std lens has to perform best at the long end if there is such a diff.), when I look at those results, they moved from unacceptable (nonIS2) to slightly less unacceptable (IS1) to acceptable (F2.8USM). – Michael Nielsen Jan 26 '14 at 10:15
  • and even though you find them good, it doesnt change the fact that his pictures has that see-through plastic look that kit lenses give you , that you also see in the digital picture images from those lenses. http://www.the-digital-picture.com/Reviews/Canon-EF-S-18-55mm-f-3.5-5.6-IS-Lens-Review.aspx – Michael Nielsen Jan 26 '14 at 10:17
  • how good old lenses were are irrelevant. it is relative to current tech and it is how much of an upgrade do you get over your compact and high end cellphone are you getting that matters to the consumer. And a kit lens on your nice DSLR investment just makes you disappointed. I know I was. and I see ppl come here disappointed, asking how to improve it with the kit lens. And unsharp mask do not bring back the details lost and covered up by the noise floor, and it looks ugly compared to the true signal. – Michael Nielsen Jan 26 '14 at 10:31
  • Then your link is also irrelevent since it is not about anyone's current kit lens! When unsharp masks were developed in the 1930s there was no such thing as noise floor. Grain, yes. Noise floor, no. And anyone should know if you plan to use any kind of advanced post processing method such as strong sharpening or tone mapping to use base ISO so that the noise isn't enhanced in post. – Michael C Jan 27 '14 at 00:16
  • Saying a cheap lens isn't as good as the best lenses is not the same thing as saying it is worthless. The original post illustrates that there are improvements in technique that the shooter can make using this lens. For every newbie that gets discouraged by the lack of quality from their consumer grade gear (mostly because of their lack of good technique), there are dozens who get discouraged and quit because they are told they can't produce anything of value without spending thousands of dollars on the best cameras and lenses. – Michael C Jan 27 '14 at 00:20
  • The advice to find the limitations of your current gear before you spend large amounts on upgrades is applicable here. Anyone who expects perfect results shooting at f/22 using a camera with a DLA of f/6.8 and 1/15 second handheld obviously has a ways to go before they've reached the limits of even the most basic gear with all of its imperfections. – Michael C Jan 27 '14 at 00:23
  • Then there are those who spend a lot on gear upgrades and get discouraged and quit when they don't see much improvement in their results because they are still using the same bad technique they were using with the cheaper gear. – Michael C Jan 27 '14 at 00:27
  • Im not saying that it is worthless because L lenses are better. I'm saying it is worthless because it is not good enough to make ppl feel good about their DSLR purchase, esp. if they upgraded from a decent compact (350$ and up). – Michael Nielsen Jan 27 '14 at 08:33
  • If that is the case then no one should upgrade from a decent compact to an entry level DSLR. Because for most shots the DSLR is a lot harder to learn to use to get the same results you can get from a good compact that decides everything for you and masks bad technique behind a smaller sensor and much deeper DoF. – Michael C Jan 27 '14 at 09:29
  • There is something very interesting in this conversation. I'm indeed a Newbie that has lot and lot to learn and the answer of Michael Clark is very interesting in that case. However, as you practice, having a better lens can show you more significally the difference between the bad use and the good use of your gear! I mean, setting everything badly will give you bad results no matter what gear you have. But if you have good gear, when you use correct settings, the difference between bad and good will be more obvious! – Andy M Feb 18 '14 at 12:07
  • if you cant see the improvement in usage and settings, how can you ever learn? any neural network need feedback to get trained. – Michael Nielsen Feb 18 '14 at 15:02