Sharpening techniques (lizard subject)

Left without sharpening, right with:

Hello everyone,

@heckflosse
Wow!
Can you share your pp3 to take a look at your settings to achieve such a great result.

@paperdigits
Thanks a lot.
Didn’t know this technique.
I am an old-fashioned photographer and I don’t know very well the software part of this trade :slight_smile:
I am just checking its documentation right now:
https://rawpedia.rawtherapee.com/Sharpening

P.s:
one problem with this image is the very end of tail of the lizard which is out of focus.
The animal was still alive and I didn’t take multiples shot (to focus stacking them later on…) :slight_smile:

Sure:

LIZARD_SHARPENING-TEST.NEF.pp3 (11.9 KB)

Hello @heckflosse

First off, thanks a lot for your help!

Just diffed your previous pp3 file.
Only a question…

Why did you opt for the Method=rcdvng4?
I always work with the Amaze method…
Taking a look at rawpedia [1] it looks like your choice might be better off for images with round edges. Is that so?

[1] Demosaicing - RawPedia

1 Like

Well, that’s just my preset. I almost always use one of the dual demosaicers and rcd is faster than amaze (though still vng4 being the bottleneck here).

I prefer Amaze for e.g. architecture (with long straight lines) and rcd for nature (more random lines)

@Silvio_Grosso

Being also a D700 shooter, some hints:

  1. RL deconvolution usually works very well for D700 raw files
  2. if you go past diffraction limit, you can raise the RL Radius even more without getting artifacts. e.g. try 0.85 for your F9 shots or even larger values for F11
  3. Dual demosaicers help to avoid sharpening artifacts in flat (oof) regions

Edit: see also My sharpening workflow for base ISO mages with the new features in RT (aka extreme pixel peeping)

2 Likes

@Silvio_Grosso I’m also a plant pathologist, so if you can share photos of diseased leaves, maybe as a Play Raw, I’ll be on it!

Hello @heckflosse

Thanks a lot again inded for your suggestions!
I do appreciate your work on RawTherapee.
The only commercial software I am still using is Affinity photo for focus stacking tasks (tried Hugin and Enfuse but the duo is a bit too “cumbersome” for me…)

Being also a D700 shooter, some hints…

Yep. I took thousands of macro shoots with this reflex.
Never had any problem so far in 10 years of everyday very intense use.

At work we are saving money for purchasing a Nikon D850 in the near future. Luckilly, its price has dropped a bit lately. Later on, we hope to buy a new macro lens: so far I have been using a macro lens Nikkor 60 mm - f 2.8, which has served me extremely well too.

I am a bit worried about the size of the Raw files produced by the Nikon D850 and I do hope RawTherapee is going to handle them gracefully (together with a beefy computer of course…).
Luckilly, I alwasy shoot my plants diseases subjects with a tripod otherwise I would be also in trouble with such a big sensor as regards the sharpness side:-)

If you don’t need AF, go for a Apo-Rodagon lens. They almost have no CA. Have a look at this one I took with my D700 using an Apo-Rodagon 4.0/105 mm lens:
http://rawtherapee.com/shared/test_images/yellow_poppy.nef

I shoot a D850, and with a decent computer, you’ll be fine. If you get a new computer for it, look at the AMD Threadripper… All those cores!

Hello @heckflosse,

If you don’t need AF, go for a Apo-Rodagon lens.

Thanks. Never used the AF so far with my macro plants subjects :slight_smile:
As a plant pathologist, I mostly take pictures of fungi and bacteria samples.
When are savings are enough, I suppose I am going to opt for a macro lens in the 50-60 mm range. It is not necessary to have it internally stabilized because I always shoot with a tripod in our laboratory.

Hello @sguyader

I’m also a plant pathologist, so if you can share photos of diseased leaves, maybe as a Play Raw, I’ll be on it!

It would be fun indeed but I do fear we are going to “scare” the photographers of this forum with our pictures of very diseased leaves :slight_smile:

Joking aside, here is a picture of a bacterium, on Pepper leaves, quite common in Northern Italy, that is Xanthomonas vesicatoria (leaves + gyca and csga petri dishes with the bacterium at 2 days of growth) :

Here is the NEF raw file in case you are interested:

Thanks! I’m more mycologist and virologist than a bacteriologist.
What is the white stuff in the right petri dish? Paper?

How do you cope with changes in environmental light in your photo setup? Do you fix aperture/shutter speed/ISO?

I am a scientist myself and work with images mainly in high-resolution microscopy.
If I don’t photograph professionally, I do so only from the “artistic” aspect, the accuracy of the reproduction plays hardly any role for me.
The possibilities of post-processing in digital photography make it very difficult for me to find the boundary between necessary improvement and unwanted change of “reality”.
Nevertheless, I would like to show a technique that I sometimes use to “sharpen” and make fine details more visible.
However, this use of tone mapping is not really suitable for scientific photography.

Only one color cuvre in darktable plus (right) and minus (left) tone mapping.


LIZARD_SHARPENING-TEST.NEF.xmp (2.4 KB)

@Silvio_Grosso It depends on what the objective is for sharpening and what you expect from it. E.g., you have noted that

Are you intending to sharpen the unfocused areas so that they are more alike the more focused ones?

Taking @heckflosse processing as a base (really impressive in itself) , maybe you can further improve the look of your subject with wavelets:



LIZARD_SHARPENING-dc wv.png.out.pp3 (11.2 KB)

And you can always control the final sharpening with the global wavelets strength (if it looks oversharpened).

About the end of the tail, I’m afraid there’s no way to recover really unfocused areas, sorry.

2 Likes

Hello @afre

Are you intending to sharpen the unfocused areas so that they are more alike the more focused ones?

Nope, my note about the tail not being completely in focus was just a casual remark.
Actually, I took many pictures of this lizard and one of them it appears most suited to make some experiment about sharpening.

Here it is (jpeg very compressed to post here):

Here is the NEF file (Nikon D700) in case you are interested:

BTW, let me ask you a question since you are extremely knowledgeable about G’MIC :slight_smile:
Can you suggest some unusual G’MIC filter in this domain to try in order to sharpen these pictures of mine?
At present, there are tons of them for this particular purpose, with little documentation, shipped with G’MIC and I am a bit lost…

@Silvio_Grosso, Here is a bright version. Tried to make the paper white and the lizard less dark. Are the colours right? Cropped to make it photographically interesting (for me, of course).
LIZARD_SHARPENING-TEST.jpg.out.pp3 (11.9 KB)

Hello everyone,

@shreedhar

Are the colours right?

Yep. You nailed it!
The white paper and the coin on the picture are both helpful to “white balance” this image :slight_smile:

@Thomas_Do

If I don’t photograph professionally, I do so only from the “artistic” aspect, the accuracy of the reproduction plays hardly any role for me.

As for myself, I take pictures mostly for work, as a plant pathologist.
I consider this both a bliss and a curse.
A bliss because I always strive to take the most accurate picture out of my samples (low ISO, no strong shadows, no diffraction etc). By doing so, in post production I limit myself to some little cropping and a bit of sharpening. Most of all, I never change the colours of my samples in the pictures (they must be accurate as regards the subject displayed).
A curse because I have never learned properly how to work with RawTherapee since my pictures have usually little noise to correct, no strong shadows to decrease and so on. As a consequence, I am not forced to learn this software and I am still a beginner as regards its tools… :slight_smile:

For your use-case I would also consider a Pentax-K1. You are working under well controlled conditions where you really can get an advantage by using pixelshift.

Edit: I mean the old K1, not the Mark II

1 Like