SCoombs
Well-known member
My daily experience goes a little something like this:
I take a photo. I look at it in the viewfinder/LCD and am relatively happy with it. I come home and load it into LrC, which is set to import with camera settings. I view the photo in the library module where it shows the embedded jpeg preview and think it looks okay. I change to the develop module where now Adobe's own processing comes into play for the first time in this chain of events and I'm suddenly very disappointed in the sharpness of the photo. I load it up in NX studio for comparison and think it looks fine, or at least like something I can work with.
The difference in this one example is not coming through nearly as pronounced here on the forum as it actually is, but you can at least see somewhat of the difference. These are at 100% view. One place where I think you can still see it fairly noticeably in these screenshots as they have posted to the forum is in the "ring" around the edge of the eye which is basically a gradient from red to the yellow of the rest of the eye. In the NX studio version, you can clearly make out the ring and the gradient. In the Lightroom version, it's difficult to make out the color at all - it just looks like a darker version of the yellow - and the gradient is essentially nonexistent. (by the way, there are differences in color between LrC and NX studio, but after comparing hundreds of images like this I am pretty confident the fact that the red color is hard to make out in the LR version is NOT because of color differences, but differences in the sharpening and how the texture is being rendered).
Lightroom:
NX Studio:
I have learned over time how to adjust LrC's rendering to come close to - but not match - Nikon's processing. The first thing to do is usually to reduce the luminance noise reduction. For this image, LrC chooses 46 as the default. Typically I find putting it at 20 is closest to what Nikon gives me. This image is in the "Standard" picture control, and LrC usually matches this by importing with texture +8, clarity +4, sharpening amount at 40, radius at 2.0, and detail at 25. My standard tweaks to get as close as I can to Nikon's processing is to change to sharpening 60, radius 1.0, and detail to 40. It is closer:
This is closer, but does not match Nikon's output. It also yields noise/grain/sharpening artifacts which are more prominent and of a more unpleasant quality than Nikon's own processing. In the example posted here, I haven't adjusted the masking slider to help clean up the background. That does help reduce the quantity of the artifacts in the background, but their quality is still worse than with Nikon's out-of-camera processing, and of course we can see the roughness of all of this on the body of the bird as well.
There's something else pretty striking here. Say that we look at our "matching" version now and notice it still isn't quite what Nikon gives us so we go to 200% in the hopes of getting a better idea of what the difference actually is:
LrC
NX Studio:
Look at the difference here: LrC's is very pixelated, whereas Nikon's is smooth. Look around the edge of the pupil: we get distinct pixels vs. something a lot smoother. Could this just be a matter of Lightroom choosing to show us the image as is, which at 200% is going to be pixels since it's been blown up so large, whereas Nikon is "hiding" the pixels and choosing to smooth them out for the display? To answer this question, I exported the image from each and viewed it at 214% using Windows photo viewer (214 being as close to 200 as it would let me go):
LrC:
NX:
I'm not sure how well it comes through here, but the file exported by LrC displays more pixelization at this size than the version exported by NX Studio, which is somewhat smoother. It's not nearly as pronounced as viewing them at this size in the programs themselves, so it seems that at least part of the difference is smoothing in the screen rendering as done by Nikon, but not all of it.
A few followup questions people might ask. First, how does it compare if I denoise with LrC? As far as the overall look, probably a bit better. As far as the sharpness level, not really much better. However, advantage is that the denoised version can be pushed a bit further in terms of sharpening without being ruined by artifacts. On the left is the denoised version, and on the right is the denoised version with sharpening increased to 100, which is probably oversharpened but you get the point:
Second, how much does this matter practically speaking? Am I pixel peeping here, and none of this would really matter if I were just viewing the image at the full size? To that I'd say two things. One is that if LrC is giving me worse results compared to Nikon's own processing at any level of crop, well, it still is what it is and the difference will impact images even viewed at more normal crops. Number two is that, as a corollary to number one, I have seen the difference at images which are not being cropped.
Third, what about trying another processing solution like DxO PureRAW? The answer to that is that I have tried this option and it can indeed be a solution at times. I found that a lot of the time the results were clearly quite superior, while other times I would say I thought they were worse. It's worth having as something to try with photos like this even if it doesn't always give results I'm going to want to use. However, to go back to the point made above, the bottom line is that Lr is the tool used most frequently and I'm trying to sort this out.
What are your experiences with any of this?
I take a photo. I look at it in the viewfinder/LCD and am relatively happy with it. I come home and load it into LrC, which is set to import with camera settings. I view the photo in the library module where it shows the embedded jpeg preview and think it looks okay. I change to the develop module where now Adobe's own processing comes into play for the first time in this chain of events and I'm suddenly very disappointed in the sharpness of the photo. I load it up in NX studio for comparison and think it looks fine, or at least like something I can work with.
The difference in this one example is not coming through nearly as pronounced here on the forum as it actually is, but you can at least see somewhat of the difference. These are at 100% view. One place where I think you can still see it fairly noticeably in these screenshots as they have posted to the forum is in the "ring" around the edge of the eye which is basically a gradient from red to the yellow of the rest of the eye. In the NX studio version, you can clearly make out the ring and the gradient. In the Lightroom version, it's difficult to make out the color at all - it just looks like a darker version of the yellow - and the gradient is essentially nonexistent. (by the way, there are differences in color between LrC and NX studio, but after comparing hundreds of images like this I am pretty confident the fact that the red color is hard to make out in the LR version is NOT because of color differences, but differences in the sharpening and how the texture is being rendered).
Lightroom:
You can only see EXIF info for this image if you are logged in.
NX Studio:
You can only see EXIF info for this image if you are logged in.
I have learned over time how to adjust LrC's rendering to come close to - but not match - Nikon's processing. The first thing to do is usually to reduce the luminance noise reduction. For this image, LrC chooses 46 as the default. Typically I find putting it at 20 is closest to what Nikon gives me. This image is in the "Standard" picture control, and LrC usually matches this by importing with texture +8, clarity +4, sharpening amount at 40, radius at 2.0, and detail at 25. My standard tweaks to get as close as I can to Nikon's processing is to change to sharpening 60, radius 1.0, and detail to 40. It is closer:
You can only see EXIF info for this image if you are logged in.
This is closer, but does not match Nikon's output. It also yields noise/grain/sharpening artifacts which are more prominent and of a more unpleasant quality than Nikon's own processing. In the example posted here, I haven't adjusted the masking slider to help clean up the background. That does help reduce the quantity of the artifacts in the background, but their quality is still worse than with Nikon's out-of-camera processing, and of course we can see the roughness of all of this on the body of the bird as well.
There's something else pretty striking here. Say that we look at our "matching" version now and notice it still isn't quite what Nikon gives us so we go to 200% in the hopes of getting a better idea of what the difference actually is:
LrC
You can only see EXIF info for this image if you are logged in.
NX Studio:
You can only see EXIF info for this image if you are logged in.
Look at the difference here: LrC's is very pixelated, whereas Nikon's is smooth. Look around the edge of the pupil: we get distinct pixels vs. something a lot smoother. Could this just be a matter of Lightroom choosing to show us the image as is, which at 200% is going to be pixels since it's been blown up so large, whereas Nikon is "hiding" the pixels and choosing to smooth them out for the display? To answer this question, I exported the image from each and viewed it at 214% using Windows photo viewer (214 being as close to 200 as it would let me go):
LrC:
You can only see EXIF info for this image if you are logged in.
NX:
You can only see EXIF info for this image if you are logged in.
I'm not sure how well it comes through here, but the file exported by LrC displays more pixelization at this size than the version exported by NX Studio, which is somewhat smoother. It's not nearly as pronounced as viewing them at this size in the programs themselves, so it seems that at least part of the difference is smoothing in the screen rendering as done by Nikon, but not all of it.
A few followup questions people might ask. First, how does it compare if I denoise with LrC? As far as the overall look, probably a bit better. As far as the sharpness level, not really much better. However, advantage is that the denoised version can be pushed a bit further in terms of sharpening without being ruined by artifacts. On the left is the denoised version, and on the right is the denoised version with sharpening increased to 100, which is probably oversharpened but you get the point:
You can only see EXIF info for this image if you are logged in.
Second, how much does this matter practically speaking? Am I pixel peeping here, and none of this would really matter if I were just viewing the image at the full size? To that I'd say two things. One is that if LrC is giving me worse results compared to Nikon's own processing at any level of crop, well, it still is what it is and the difference will impact images even viewed at more normal crops. Number two is that, as a corollary to number one, I have seen the difference at images which are not being cropped.
Third, what about trying another processing solution like DxO PureRAW? The answer to that is that I have tried this option and it can indeed be a solution at times. I found that a lot of the time the results were clearly quite superior, while other times I would say I thought they were worse. It's worth having as something to try with photos like this even if it doesn't always give results I'm going to want to use. However, to go back to the point made above, the bottom line is that Lr is the tool used most frequently and I'm trying to sort this out.
What are your experiences with any of this?
