Finally I feel at home with darktable’s filmic module and can achieve “any” result. But with power comes responsability… I wonder whether this rendition of this very high contrast photograph isn’t over-the-top.
Too me it looks good, definitely better than the OOC jpeg. But I’m not sure whether squeezing 14 EV of DR like this is actually such a good idea. Perhaps I should allow more shadows creep under the mountaineer?
I’m very much interested to hear the opinion of other people.
Very good point about lens distortion, thanks! Given that there are no straight lines in this image, the stereographic projection is perfect.
Up to now I’ve been using the raw developer bundled with my cameras (Olympus Viewer) which does not support anything else than rectilinear, so the idea of using a different projection did not even come to my mind.
Since many MFT lenses require correction, I enabled the lens correction module by default, but I see now that distortion correction for my ultra wide zoom (Olympus 9-18) is actually counterproductive. Naturally, the lens is not perfectly rectilinear, so correcting it to be perfectly rectilinear actually degrades the image in this case.
Hi @alpinist,
this is my version. I hope it’s not way over the top for you, but I like it
I tried to make the image a bit “clearer”, especially the mountains in the distant.
I also colored the rock in the foreground a bit orange as well as the jacket. This splits the image almost diagonal in half with blue and orange
(@Timur thanks for the lens correction projection tip, it looks better this way)
I had a look at your edit history. I see that you use a mask to apply changes locally to the mountains in the right bottom part. (Masks is something I have yet to learn properly.)
Can I ask you some questions?
What part of your edits is a default that you apply to all images? Demosaic “two times” or highlight reconstruction “reconstruct color”? Or do you go through these settings separately for every image?
I guess you increased “adjust autoset parameters” in wavelet denoise because the image features regions that have been brightened heavily. How did you tweak the value? Frankly, I do not see much difference between when I change back your value of 3.52 to 1.
Why do you use two additional exposure instances, both (it seems) with the same mask (the mountains)?
I see that “local contrast” is applied to the whole image, and “local contrast 2” to the masked mountains, but what’s the purpose of “local contrast 1”?
Hi @alpinist,
since in go back and forth in every image i had to look what i did in the end
I think i only apply the lens correction module by default. I use the other modules for each image according to mood I want to achieve.
With the highlight reconstruction I test which mode looks best on the image I currently work on. But I discovered that the “reconstruct color” is very good on reconstructing clouds on most images.
The Demosaic “two times” is a preset added to the module. ( I didn’t put much thought in this )
Yes you are right the “adjust autoset parameters” is not necessary. When i exported the image the first time i had heavy noise on the image because i used two “Contrast equalizers” and no denoise which sharpened the noise heavily and in played around with the parameter. But for the end result it is not necessary
They use the same mask but different gray values for the parameter masks. So that the first only affects the brighter areas and the second only the darker areas. Then i darkened the dark and brighten the bright areas to get more contrast on the mountain. I’m not sure if this is the recommended method because I feel like this should be achievable with the local contrast module?
The “local contrast 1” uses the blend mode: multiply reverse that I picked up on this thread Clarity in darktable
)
