0
\$\begingroup\$

I am a beginner and I have been recently trying to do some landscape photography. However what I am not satisfied with is the level of sharpness and muddiness, it feels the photos come out always quite muddy and not very sharp. They look somewhat okayish when zoomed very small like 25-33% however when viewing at full screen (how I like to view any of my photographs at 40-60%) edges start to come out pixelated and overally the photos look unnatural and pixelated.

These photos I took at the rocks by the sea, the clouds and the sky look decent, a little bit noise here and there but nothing terrible. The waves and the sea look quite muddy and grainy, also clearly quite noisy. Rocks look muddy, grainy and unsharp. Blacks are decent, no noise there.

I am not sure what I am doing wrong here. Any tips? I used ISO100 and 1/125s both shot handheld. First with f1.8 and second f4.5. Nikon zf with 20mm f1.8 lens.

enter image description here Original image

enter image description here Original image

\$\endgroup\$
10
  • 1
    \$\begingroup\$ Have you try to close the aperture to f8 (for example). Then you can expect deeper depth of field \$\endgroup\$ Commented Sep 1 at 13:35
  • 1
    \$\begingroup\$ The first one in list (853) was taken with f1.8, second one (755) was taken with f4.5, yes I agree the second one with narrower aperture seems slightly sharper overall. I took multiple shots with slightly different settings yes however I did not take any shots with that narrow apertures like f8. Also I find it a bit hard for me to analyze the sharpness in the field with the camera screen. But thanks for the tip I will try next time with narrower to see if it improves the sharpness. However I might need to bring the tripod next time then. \$\endgroup\$
    – gnok
    Commented Sep 1 at 14:32
  • 1
    \$\begingroup\$ There are very few lenses which are truly sharp at f/1.8 and they cost many thousands of dollars. What lens are you using? \$\endgroup\$
    – Philip Kendall
    Commented Sep 1 at 14:38
  • 1
    \$\begingroup\$ Like most similar question here, you seem to expect results straight out of camera that look like all of those fabulous photos posted on 500px and other sites that have extensive post-processing done to the raw image data. \$\endgroup\$
    – Michael C
    Commented Sep 1 at 21:50
  • 1
    \$\begingroup\$ @scottbb yes I couldn't respond on my mobile when I tried earlier and seem to be able to do so only on the computer. Thanks for the tip, I will try to merge later. \$\endgroup\$
    – gnok
    Commented Sep 3 at 16:36

3 Answers 3

2
\$\begingroup\$

Both images are reasonably sharp.

It may be that your expectations about what you will see straight out of camera may be higher than the reality. Landscape images that have a lot of local contrast, which our brains tend to interpret as being "crisper" or "sharper", usually only show such contrast after significant post processing, especially by using processing actions that increase local contrast.

From an answer to this question: Are award winning landscape images edited?

Pretty much everything on 500px that are at the top of various popularity lists have the everlasting stink edited out of them. Every.Single.One.

Nothing comes out of any camera on the market looking like CGI for a multimillion dollar movie. That's all I see in the "top images" at 500px.

Most of the "Featured Photos & Videos" at Fstoppers also have a lot of editing done to them.

With Flickr, it depends on which communities you're looking at. But many of them there are also highly edited, though not to the point many images at 500px are, and there are also a lot more that are much closer to what one might get from the camera's own raw conversion engine and user selectable options that can be selected before the shot.

Here's an image I shot recently near Page, Arizona. It was taken with a Canon 5D Mark IV and an EF 24-70mm f/2.8 L lens (non-IS) handheld. The lens was zoomed to 43mm, ISO 100, f/8, 1/320.

Here's what it would have looked like with the in camera settings at the time it was shot. Auto White Balance. Standard Picture Style with in-camera adjustments to contrast (-1) and saturation (-1) [These don't affect the final image since I can set them to whatever value I want in raw conversion, but they do make the histogram, based on the JPEG preview attached to the raw file and used to display the image on the back of the camera, show a little more accurately what's in the raw file]. It's shot a bit dark because if I'd shot it any brighter, the brightest highlights would have been blown out.

enter image description here

Below it's been converted from raw using global adjustments. White Balance was set at 5800K and a 2.5 Mired shift towards green. It was brightened by almost a full stop in overall brightness (with the highlights then pulled back considerably to prevent them from clipping) and the shadows boosted another one-third stop or so. There were adjustments to contrast (-), highlights (-), shadows (+), and minor adjustments to color using an HSL tool. Lens correction included chromatic aberration, color blur, and distortion corrections.

enter image description here

Here's the same raw file with the exact same adjustments as above with additional tone mapping applied to increase local contrast. Various raw processing applications call this 'Clarity/Vibrance', 'Detail Enhancement', or even "Local Contrast'.

enter image description here

Also understand that, no matter how well it was shot and with what high end equipment it was shot with, nothing will look as good magnified to 100% on a large monitor as it will at a more normal display size. Pixel peeping has raised expectations to ridiculous levels! A 24MP image displayed on a 23" HD (1920x1080) monitor at 100% is the equivalent enlargement of a 60x40 inch print! Viewing a 50MP image at 100% on such a monitor would be like looking at a small section of a 120x80" print! How often do you critically examine a 120x80" print from a distance of only 18-24 inches (the distance between most people's eyes and their computer monitor)?


An interesting aside: All images in the above answer were uncropped and resized to 2160x1440 pixels when exported before uploading them. istatic has resized each one to be the same (decompressed on display) file size. As the contrast is increased in each image, the resolution goes down to maintain the same decompressed file size. From 2160x1440 in the "straight out of camera" photo, to 1868x1245 in the second photo after initial raw conversion settings, to 1736x1157 in the photo with the same initial raw conversion settings plus additional tone mapping to increase local contrast.

\$\endgroup\$
4
  • \$\begingroup\$ "It's shot a bit dark because if I'd shot it any brighter, the brightest highlights would have been blown out." - it's bad for the sake of comparison. If you blew some highlights it would improve image significantly and make it comparable to following images. \$\endgroup\$ Commented Sep 2 at 10:40
  • \$\begingroup\$ I like the unedited photograph. It looks very crisp to me. I agree, perhaps you are onto something on my expectations however I have compared mine not to some award winning photographs but to some sample photographs in reviews etc. and of course to those I took myself with mobile phone. Thanks for the detailed answer. \$\endgroup\$
    – gnok
    Commented Sep 3 at 16:49
  • 1
    \$\begingroup\$ @EuriPinhollow I had absolutely zero intention of using the shot straight out of camera. I knew when I shot it I was going to boost the mids & shadows while keeping the highlights in check. That's what Adams meant when he talked about "previsualization". Blown highlights in the raw file are not recoverable. Perhaps I was a bit too conservative, but I got the end result I wanted with no unacceptable noise levels. \$\endgroup\$
    – Michael C
    Commented Sep 3 at 23:12
  • 1
    \$\begingroup\$ @gnok Phones do a LOT of "extra" processing that dedicated cameras usually do not. \$\endgroup\$
    – Michael C
    Commented Sep 4 at 7:48
0
\$\begingroup\$

I took a look at the exif of the first image and I don't see any obvious issue. 20mm f/1.8, 1/125, 100, VR on...

But I do see one potential issue and that is active D-lighting was enabled. Active D-lighting causes the camera to underexpose in order to preserve highlights, and that affects both raw and jpeg images.

Image noise has more to do with lack of light/information per area; and underexposure certainly doesn't help. The only thing ISO has to do with noise is that it is compensating for a lack of light; it makes the noise visible (may actually reduce noise).

However, ISO does not compensate for a lack of information... e.g. blues (sky/water/etc) are almost always noisy because the typical sensor only has 25% of the photosites sensitive to primarily blue wavelengths; 75% of the sensor is lacking adequate information... and that leads to noise/color noise (usually magenta).

Pixelation is an entirely different/unrelated thing... if you are seeing pixelation it is due to a magnification or monitor setting.

\$\endgroup\$
9
  • \$\begingroup\$ "Active D-lighting causes the camera to underexpose in order to preserve highlights" - that statement requires some explanation to not be misleading. Nikon's D-lighting does not change ISO range which would be the case if it was the way you are describing it. \$\endgroup\$ Commented Sep 1 at 17:23
  • \$\begingroup\$ @EuriPinhollow, Active D-lighting affects the way the camera meters the scene which changes the exposure; and the camera will change whatever setting it is in control of in order to accomplish that. For raw files it is similar to using highlight weighted metering. I just tested my Z9 and it did change the ISO when set to auto with the camera in manual mode. \$\endgroup\$ Commented Sep 1 at 20:03
  • \$\begingroup\$ "because the typical sensor only has 25% of the photosites sensitive to primarily blue wavelengths; 75% of the sensor is lacking adequate information..." This is incorrect. As any look at the sensitivity curves of any sensor used in ILCs shows, there are no hard pass filters in the CFAs of such cameras. Some "red' gets through the blue filter, a LOT of red gets through the green filter. A LOT of green gets through the "red" filter (which is actually yellow-orange) and blue filter, and some blue gets through the green filters. \$\endgroup\$
    – Michael C
    Commented Sep 1 at 22:01
  • \$\begingroup\$ @MichaelC, I said "primarily blue"... I said nothing about "only blue" or hard pass. Yes, there is overlap (specifics vary by camera/sensor)... just as there is probably some "red" wavelengths in the light from a blue sky. And it is that cross contamination (might not be the best term) that causes the resulting noise/color noise... if the sky/light was purely blue as we perceive it, and the CFA was hard pass, then the demosaicing would have no issue with such situations. \$\endgroup\$ Commented Sep 1 at 23:45
  • \$\begingroup\$ That "Cross contamination" is what allows our brains to create the perception of color. The vast majority of the colors we perceive can not be stimulated with a single wavelength of light. Our cameras mimic that. \$\endgroup\$
    – Michael C
    Commented Sep 2 at 0:45
0
\$\begingroup\$

F1.8 matters a lot for cameras with large sensors. For a deep scene like yours you will see significant difference between F1.8 and F4.5 (but F8 might look the same at this focal length).

Noise is almost always present. The strength of noise depends on a lot of factors:

  1. smoothness of objects. Noise will always be more apparent on objects with no details
  2. image settings: there's obviously noise reduction but all other settings will have effect on noise visibility: increasing saturation, contrast and sharpness through image settings will increase noise
  3. colour of objects. Camera sensors are generally biased towards green frequencies (both sensitivity and number of green pixels), so saturated blue and red objects are in most disadvantage and even more so if sensor requires strong colour conversion (colour matrix)

As for your two examples I am almost sure that you've increased sharpness setting too far for DSC0755 (it's clearly different value than what DSC0853 has) which will make image look bad when there are rough textures such as uneven waves. You can confirm it yourself by looking at detailed image information either in the camera or in Nikon software. By the way waves are clearly different between your two examples what also contributes to the difference. Waves on DSC0755 clearly have more small details because of different wind.

These photos I took at the rocks by the sea, the clouds and the sky look decent, a little bit noise here and there but nothing terrible. The waves and the sea look quite muddy and grainy, also clearly quite noisy. Rocks look muddy, grainy and unsharp. Blacks are decent, no noise there.

DSC0853 looks very normal to me besides smaller DoF. Can't any problems there. That's exactly like I would imagine waves and dry and wet rocks to look. Also when you are not viewing image at full resolution it's important to take into account what image viewer you are using. Scaling full image to your display is not a standardized operation and will differ between viewing programs.

\$\endgroup\$
2
  • \$\begingroup\$ Yes I believe it had different sharpening setting now that I have analyzed the photograph information in the Nikon studio. I did not realize it at first. I usually do not like the adjustments and like to use only the flat settings so the photos look as natural as possible. "Also when you are not viewing image at full resolution it's important to take into account what image viewer you are using. Scaling full image to your display is not a standardized operation and will differ between viewing programs." This is also good information to know. Thank you! \$\endgroup\$
    – gnok
    Commented Sep 3 at 16:27
  • 1
    \$\begingroup\$ @gnok Unless you've decided it's against your style (everyone's a little different) it's probably best to think of every digital image from a camera like it's a starting scan of a negative: it doesn't become become the 'real' image until you 'print' it with at least minimal settings to find the exposure range you want for the image; using something like levels or curves. It's just the raw data you measured, you've still got the rest of the work to do to make it a photo. \$\endgroup\$
    – davolfman
    Commented Sep 3 at 21:46

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

Not the answer you're looking for? Browse other questions tagged or ask your own question.