What do you do when you are caught with a beautiful sunset and your dog decides that he is too cool to look at that spectacular giant explosion? Let’s see what people at pixls.us can do to rescue the backlit image with FOSS software!
DSC02168.ARW (24 MB), released under CC BY-NC-SA DSC02168.ARW.xmp (12.3 KB), note that you need the latest darktable in git because it uses the “color checker lut” module
Tip: Manually select “Samyang 12mm f/2” to correct distortion and TCA if your software of choice supports lensfun.
I’m going to try this for sure (you’ve effectively wrecked my afternoon productivity), but just wanted to say that you did a cracking job on your result.
Auto CA correct on, lowered film area, max drama + overdrive, raised black clipping point (to clear out some veiling flare), lowered white clipping point to raise the brightness after max drama, and raised shadow brightness.
Ok, developed in RawTherapee (DSC02168.ARW.pp3 (9.6 KB)), mostly to recover highlights and boost mid/shadows to some degree. Then GIMP for some color toning, masking, freaky details (masked to the dog mostly), and cropping:
With rawproc, mainly to see if I can open any raw other than NEF. Just gamma to drag the data out of linear, a curve to lift the shadows, and a bit of HSB-based saturation. Resized to 800x534 and a smidge of USM sharpening for posting.
I can never back away from a backlit challenge. Or a dog photo, for that matter! Here’s mine with some basic adjustments in DT and enhancements in GIMP.
Somehow the colors on my screen are a bit whaky at the moment. But I hope it looks decent on other screens too.
The edit is darktable + some sharpening in the gimp.
@Jonas_Wagner, I like this edit alot. It combines a decent structure and sharpness with the washed-out characteristic of the backlit situation. The only thing I would probably change is reducing the yellow/saturation in the blown-out highlights in the sky.
I wonder how you achieved this result, it would be great if you could share some processing details.
@chrisDSC02168.ARW.xmp (23.1 KB) should reveal all the important bits. The only other step was some final sharpening in gimp/gmic (RL deconvolution to a new layer, added a mask and selectively painted it in). I agree on the sky something is off there. Likely because of the graduated filter I applied. I might give it another shot at some point.
@patdavid I like the cinematic approach you went with. As @djotaku said, this could very possibly be a movie still as the camera slowly pans over the grass field into something very… dramatic.
@Jonas_Wagner Nice faded look! I definitely like the second version better as well. The warm color cast also gives the image a nice autumn-ish feel.
@djotaku The main challenge is to realistically recover details from shadows and highlights without producing many artifacts (mainly halo). The latest entry from @RT-Noob actually did a surprisingly good job in this department here!
Hm, it looks worse when I use your xmp with my darktable. What’s wrong here? A version conflict (dt 2.0.1 here)? That’s what it looks like (screenshot, but export does not differ):
Hm, what’s going on here? Is it your screen calibration/output colour profile that makes the difference?
Edit: Maybe a missing input colour profile? Step 48→49 does not change anything.
I’m on 2.0.4 but I doubt that’s where the difference comes from. I also doubt that it’s the display profile. It could be an input color profile thing or that the xmp I uploaded was outdated. Let’s try again with the xmp for the second edit: DSC02168.ARW.xmp (29.8 KB)
@Jonas_Wagner Mine also looks similar to @chris’s screenshot (with your second XMP). I think a input color profile is missing here. From darktable’s stderr:
[colorin] could not find requested profile `a6000_darktable.icc'!