The sky is blue - or should be...

[digitaldog]

Yes, I wasn't going for accuracy.

I'd like to see what you'd get trying, because at least the one 'invisible' hue (Magenta patch) is pretty darn close to what it should be IF you adjust brightness of the image. We could analyze the other 23 patches of course. It would probably poke more holes in Tim's comments about your captures; at least the one I examined and he of course didn't!

[Hening Bettermann]

@ TonyW #44
Can it be that simple!
One click on 'Auto Levels' under the Exposure tab in RawTherapee, and the image is basically OK. The sky is pale, which is probably correct, but pale BLUE! And its saturation increases after I discovered  'Raw White Points' under the Raw tab and set it from the default 1.0 to the minimum value which is 0.1. Now I can play a little with contrast without desaturating the sky too much. (If then I re-set AutoLevels, the image becomes VERY dark, so that's obviously why the default is there).

The AutoLevels set Black to +1956, Contrast to +25 and Saturation to +30. No change in Lightness. Normally, I don't like 'Auto'. But the good thing with this one is that I can see what it does and modify if needed. I forgot to note the change in Exposure compensation.

So it looks like my camera and monitor profiles are OK - relief!

Many thanks to all who chimed in!
---
Uff not quite that simple... But this is about RT

Now after I saved the Raw White Point of 0.1 as part of a Processing Profile preset, then choose that, then AutoLevels, it sets Exposure Compensation to +5.34, Black to 1954, Contrast to +24. No change in Saturation.
If I choose the preset with Raw White Point 1.0, then click AutoLevels,  it sets Exposure Compensation to 2.27, Black to 2045, Contrast to 27.  No change in Saturation.

Why does RT change its mind that way?
The 0.1 profile was initially (by mistake) saved under the image file name and re-named in the Mac Finder. Likewise, the 1.0 profile was re-named to include the 1.0 in the name. This could not possibly have any influence on the processing parameters??

[Stephen Ray]

@ TonyW #44Can it be that simple!

It can also be as simple as post #19. Care to post your results when you're happy with them?

[Stephen Ray]

I'd like to see what you'd get trying,

The next time I experiment, which should be very soon, I will include something to reflect an actual specular highlight and a shaded black velvet swatch to reference an actual white point and black point along with the Macbeth.

[Czornyj]

I agree that photography can be an interpretive artform but I've gotten pretty darn close to reproducing daylight scenes in my locale as it appears to me despite the limitations of my old DSLR. In addition I've found photos posted of daylight scenes over in Photo.net's No Words forum from non-professional photographers come really close, better than my camera.

I have a background in photo realistic painting and have extended experience mixing colors made from various pigments and dyes. I know what light does to natural memory colors. I do believe the limitation for not being able to get very close to representing any natural color captured with a digital camera shooting Raw is mainly the photographer's fault.

I have background in oil painting, and never found paintigs nor photography realistic.

[Hening Bettermann]

@Stephen Ray #82

>It can also be as simple as post #19.

But #19 was simple and wrong - this is simple and right. As simple as that ;-)

>Care to post your results when you're happy with them?

Of course. I just was delayed by that RT problem.
After RT AutoLevels, I raised Contrast to 36, Saturation to 30. Nothing else.
Whole image and 2 crops - still undecided. The image idea was not the idyllic scene, but the shades of green and the "furry" look of the willow trees (the yellow ones).

[Hening Bettermann]

Some after-thoughts
and an attempt to make me a rhyme.

So this whole thing was not about profiles nor color memory, it was about saturation.
How come that the sky was desaturated to that degree, since all channels were far from clipping?
The image was underexposed, which made gross lightening in the raw converter necessary. This is done in Lab space, which desaturates colors as lightness increases.
Correct?

[digitaldog]

So this whole thing was not about profiles nor color memory, it was about saturation.

Pretty much yes, not that some profiles might produce a more desirable rendering along with differing settings and possibly a different converter could have provided a more pleasing rendering. But no, this idea of memory color being accurate or matching the scene is pretty bogus at least in terms of a match. We saw this below with the two green plant examples (#75) where we are supposed to believe (and we should not) that when sitting in front of a computer display with a gamut and dynamic range that alone is vastly different from what we can see, this user actually accurately recalls what he saw minutes, hours or days prior and is making a match. The key word may be memory, what one thinks he recalls but that in no way means what is produced was a match. Plus one is output referred!*
* http://www.color.org/ICC_white_paper_20_Digital_photography_color_management_basics.pdf

How come that the sky was desaturated to that degree, since all channels were far from clipping?

Exposure such there are one or more channels clipping and saturation are not mutually exclusive per se.

The image was underexposed, which made gross lightening in the raw converter necessary. This is done in Lab space, which desaturates colors as lightness increases.

What was 'done' in Lab space and why is that necessarily pertinent?
Rendering the image is an important part of photography and has little to do with color accuracy and matching:
http://www.lumita.com/site_media/work/whitepapers/files/pscs3_rendering_image.pdf

[Stephen Ray]

@Stephen Ray #82

>It can also be as simple as post #19.

But #19 was simple and wrong - this is simple and right. As simple as that ;-)

>Care to post your results when you're happy with them?

Of course. I just was delayed by that RT problem.
After RT AutoLevels, I raised Contrast to 36, Saturation to 30. Nothing else.
Whole image and 2 crops - still undecided. The image idea was not the idyllic scene, but the shades of green and the "furry" look of the willow trees (the yellow ones).

Please keep trying Hening. Your attempts are still much too dark. If you were familiar with Levels you could easily see so. You could also use the control to lighten the scene to something more normal and then you would be surprised as to how close it appears to my original post #19. When I do the same to your attempt(s), I am not surprised by how grossly cyan your new sky becomes because you've tried to add so much color that was never there. It's very easy for me to determine the original blue sky was only very slightly blue, in fact almost neutral.

Good luck in your processing. It takes time.

EDITED TO MENTION:  To be sure, this is your art and your interpretation. Forward ahead!

[Tim Lookingbill]

Some after-thoughts
and an attempt to make me a rhyme.

So this whole thing was not about profiles nor color memory, it was about saturation.
How come that the sky was desaturated to that degree, since all channels were far from clipping?
The image was underexposed, which made gross lightening in the raw converter necessary. This is done in Lab space, which desaturates colors as lightness increases.
Correct?

There is a lot more wrong with your image than what you've outlined. The entire image is not white balanced for the time of day it was shot which appears to be around mid day. All three attempts of what you posted still have a funky cyan/thalo green cast. I guess my point about magenta neutralizing such types of casts wasn't working for you.

This is my attempt editing the first of the three you posted in ACR 6.7. I had to apply a luminance bump curve so the tops of the foliage appear to be lit directly by the sun. As it is it looks overcast with the funky white balance. Severe HSL adjusts did the rest with a magenta tint slider of +30.

This whole thread has clearly illustrated Raw Therapee is not an app I want to edit my Raw files in if Hening is having this much trouble white balancing it. Sorry, I had to say it but I got more done working on the jpg in ACR than the time it takes to read this entire thread.

[digitaldog]

There is a lot more wrong with your image than what you've outlined.

Not as there is as much wrong with your text here. Hening's image is under exposed; that's it. Had you even examined the raw Histogram (knew how?), you'd see that. Accept it, that's another doubtful error of yours.

Hopefully Hening and other's can see such facts and continue to ignore your misinformed posts here. You've gone on record in just this series of posts, multiple times ignoring the facts provided to you, other's are allowed the same tactic in ignoring you.
There is a lot more wrong with your text than what you've outlined. But other's have done a good job outlining it for your readers even if you ignore the corrections.

[Hening Bettermann]

@ Andrew #87
> What was 'done' in Lab space and why is that necessarily pertinent?

Well I think the lightening. It is my understanding (and experience), that increasing lightness in Lab will desaturate colors. If this is true, it would be pertinent to the case, no?

[Hening Bettermann]

@Tim #89
Sorry Tim, but I find your edition has too much magenta in the sky and too much yellow in the foliage. Yes my tone curve needs more work, but my focus so far were WB/colors, and I find them in the ball park now.

You are right that the image is not white balanced to the time of the day, which is indeed around 13h. So the true WB may have been closer to 6000K than 5000. However, if I move the WB in that direction in the raw converter, the image becomes more yellow, which did not comply with my then fresh memory.

Whatever failures my edit still may have, they are most likely not the fault of RawTherapee. RT lets you manipulate more parameters than  I understand (e.g. wavelets). WRT WB, it allows an adjustment not only of the blue-yellow and green-magenta, but also the red-blue balance (not that it helped me in this case though).

[digitaldog]

@ Andrew #87
> What was 'done' in Lab space and why is that necessarily pertinent?

Well I think the lightening. It is my understanding (and experience), that increasing lightness in Lab will desaturate colors. If this is true, it would be pertinent to the case, no?

In true Lab processing, it shouldn't. Lightness is a perceptually scaled component of color, the axis seen in Lab (Lstar) from light to dark. It IS the L in HSL. Such an edit shouldn’t affect color (aStar and bStar) but the appearance 'may' appear as less saturation.

[digitaldog]

You are right that the image is not white balanced to the time of the day, which is indeed around 13h. So the true WB may have been closer to 6000K than 5000. However, if I move the WB in that direction in the raw converter, the image becomes more yellow, which did not comply with my then fresh memory.

Don't get caught up in this rubbish idea that the numbers are meaningful, the color appearance, output referred, as you view it on (hopefully) a calibrated and profile display does. Tim hasn't any idea about this; the WB setting at the scene wasn't measured with the correct device so it's unknown aside from a large guess the camera is providing, the numbers define a very large range of possible colors, even if you did measure the scene with a Spectroradiometer, and even IF the converter's idea of a CCT value was in sync, it's quite possible you, as the image creator wouldn't like it as much as something you pick moving about the sliders. This has absolutely nothing to do with 'accurate' color (colorimetric matching color), and much more to do with rendering the image as you the photographer desires. Tim's numbers/WB rabbit hole is just that, a big hole you don't need to join him inside of. Again to dispose this hole:
http://www.lumita.com/site_media/work/whitepapers/files/pscs3_rendering_image.pdf

[digitaldog]

Hening, this is the kinds of tools necessary to measure the color of light correctly, Tim has none. The camera doesn't provide anything like this data!

[Hening Bettermann]

What strikes me is that you have to choose an illuminant (D50) on beforehand - wasn't this what was to be measured? And even if I had that spectrophotometer and would carry it in the field - how would I translate its reading (the graph) to WB settings in the raw converter?

[digitaldog]

What strikes me is that you have to choose an illuminant (D50) on beforehand - wasn't this what was to be measured? And even if I had that spectrophotometer and would carry it in the field - how would I translate its reading (the graph) to WB settings in the raw converter?

D50 is a specific color, any CCT value is a range. D50 was developed by averaging hundreds of measurements around the world, using differing spectroradiometers too! You don't have to choose a Standard Illuminant (of which D50, D65 are), nor do you have to choose any CCT Kelvin value per se at any specific time other than in the raw converter when you wish to adjust the WB to appear as you desire. So you would have zero need to measure anything in the field unless you wanted actual colorimetric data as I did show last post, for your own information. But above and beyond the processing of the raw data, the numbers are not needed or useful.

You can take the same raw, with the same CCT metadata embedded into two different raw processors and get two different WB appearances! What looks 'right' to you as the image creator? That's the right number. The stuff Tim's talking about in terms of numbers is absolutely off topic and not pertinent (and rubbish) to how you process your images. Wrong data from the wrong measurement device (if I can be so kind), and that wrong data does nothing useful you can't control by altering the sliders or controls in your raw processor until you get to the desired color appearance.

[digitaldog]

@Tim #89
Sorry Tim, but I find your edition has too much magenta in the sky and too much yellow in the foliage.

Proof of wrong concept about editing on CCT WB numbers alone while ignoring how an image appears. Just ignore this silly concept.

[Hening Bettermann]

@Andrew
> In true Lab processing, it shouldn't.
Then *why* is the sky desaturated in an underexposed image? (Still trying to make me a rhyme.)