Here’s a highlight recovery challenge.
I’m curious to see what you guys can manage with these intense reds…
Note that the raw file actually has no clipping, so it’s up to your raw processor to make the best use of all the valid data.
License is CC BY-SA
IMG_1360_QP3MHiz.CR2 (20.4 MB)
Here’s Filmulator’s take on it:
I can’t seem to figure out how you would avoid getting excess orange out of RawTherapee, like this:
Thanks for the challenge! darktable 2.7; filmic was the main tool used.
Thanks for sharing. This was a real challenge and I hope you like the outcome 
Oh those reds, tempting and difficult to capture… But a lovely picture nevertheless!
My take in RT 5.6, making use of the excellent Color Regions tool to make sure the oranges are reduced.
@CarVac, here is a result close what you had showed in your initial photo.
IMG_1360_QP3MHiz…jpg.out.pp3 (12.6 KB)
Remark: BTW, if you notice, the name of file has two extra `dots’. RawTherapee 5.6 consistently crashed if I load the attached pp3 file and tried to save without those two dots! I request some of you to download this pp3 file and see if you also encounter this issue. (I am on Win 10 64bit system)
It’s a pity that I could not compare to the original. I have the strong feeling that my monitor can not display colors that come close to reality. Because @CarVac claims that the tones were more red than orange I tried to go into that direction.
IMG_1360_QP3MHiz.CR2.xmp (5.9 KB)Turning off the look table mitigates it. However, I examined the file using various input profiles, including ones from Adobe and some I custom-made specially for this image, and the orange is always there, so eliminating it completely seems to be wrong - it must be part of the scene.
IMG_1360_QP3MHiz morgan passport.pp3 (11.6 KB)
Canon EOS 60D Digital ColorChecker SG.dcp (1.1 MB)
IMG_1360_QP3MHiz morgan sg.jpg.out.pp3 (11.7 KB)
Flowers!! I’m still looking for a steam locomotive fix, but flowers in the meantime…
With rawproc, here’s what I did:
That was fun!
Thanks for sharing!
I wasn’t able to handle the reds, so I got rid of them
I understand that the raw file actually has oranges and yellows, but the flowers were actually entirely red in real life, not even the trace of orange that Filmulator delivered.
I think this is why people complain about digital reds.
Yes, I think I got a better red at the expense of the rich greens in the background.
So, what would this be about, that the red response (and blue also, for that matter) of most cameras where I’ve seen the spectral response curves are “less-definitive” than the expanse of green response? Not a very precise sentence, but looking for a bit of discussion…
If parts of the flower that appears red to human eye, are decoded as orange or yellow, does’nt this mean that the green sites of the sensor are activated?
And thus the sensor green response should be wrong in such cases?
Since no filter is ideal, it could be that a very strong red signal can produce a non-zero input to the green sensors.
I’m just speculating here, extrapolating from my experience with DSP.
Typically the raw green photosites are activated to some extent by red light, but the color profile matrix is supposed to correct for this.
But unless you designed the color profile with tulips in mind it might have errors for extreme colors.
Red by eye or instrument?
Yes, direct sunlight only.
By eye.
Indeed, the transitions between cooler and warmer reds in the petals look very strange, but they are definitely part of the captured values. I have tried both the standard Adobe matrix profile and few DCP profiles for the Canon 60D, with the same kind of result.
Here are two attempts, converted to sRGB from ACEScg with PhotoFlows’s gamut mapping function.
Accurate luminance, no saturation recovery:
With partial saturation recovery:
