The sky is blue - or should be...

[digitaldog]

@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.)

It's one, 'incorrect' rendering interpretation, nothing more. You've seen the same raw file rendered differently with more saturation right?
Oh, and to dismiss Tim's idea and backup my last post about differing raw converters both interpreting the WB differently and producing the rendering differently, here is one .DNG in two different raw processors with defaults: LR and Iridient Developer. Note what the later states for WB!

[Tim Lookingbill]

@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).

If it isn't Raw Therapee's fault then why doesn't that image look like it was shot in daylight even if the blue sky looks right? I've edited 100's of daylight landscape, city scape and my local park and never would I settle for the results demonstrated in this thread including my own effort.

I mean you can speculate all you want but there's more to a landscape than just getting a "correct" looking blue sky. Can you post a different daylight landscape image processed in Raw Therapee that does look right including the blue sky? I can't believe we've all spent so much time and effort just to get the blue sky to look right while the rest of the image looks so wrong!

I still don't know what can be learned about processing color images from this thread.

[Tim Lookingbill]

About that different daylight landscape processed in Raw Therapee, make sure it has greens, browns, blues, yellows or any color variety but don't post a picture of sand dunes which is too simple of an image with very few memory colors.

[digitaldog]

About that different daylight landscape processed in Raw Therapee, make sure it has greens, browns, blues, yellows or any color variety but don't post a picture of sand dunes which is too simple of an image with very few memory colors.

Too simplistic for simpletons is more like it;snapshot you provided below that falls into simplistic camp!

I've edited 100's of daylight landscape, city scape and my local park and never would I settle for the results demonstrated in this thread including my own effort.

That many snapshots?   
Tim is of course, wrong again:

[Hening Bettermann]

@ Tim #101
Hi Tim,
> If it isn't Raw Therapee's fault then why doesn't that image look like it was shot in daylight even if the blue sky looks right?

Well it looks OK to me, may need more work on the TRC, but I find colors in the ball park.

> Can you post a different daylight landscape image processed in Raw Therapee that does look right including the blue sky?

I'll try.
Here is another daylight landscape, shot on May 10th, 2016, at 17:21 h. Opened in Iridient, in RT, and in RT with contrast +20 to match Iridient. No other edits. Same profile, my custom created in DCamProf (Combo LUT). I have to admit that I like the sky in Iridient a little better. 
Difference is visible even in the jpegs. (This is not my final edit!)

> I still don't know what can be learned about processing color images from this thread.

Well what I learned is that RT's AutoLevels saved me an underexposed image, which I was unable to do manually. And that the problem was this underexposure, and the saturation, not the profiles, color memory or white balance.

[digitaldog]

> I still don't know what can be learned about processing color images from this thread.

Well what I learned is that RT's AutoLevels saved me an underexposed image, which I was unable to do manually. And that the problem was this underexposure, and the saturation, not the profiles, color memory or white balance.

He doesn't know what can be learned; true and by brute force of his own desire not to learn.
Inidient Developer is a very good package.
Yes, part of your problem was under exposure and yes, you can 'fix' that if you understand the tools available to do so and the net result is just a bit more unnecessary noise in the shadows due to the under exposure.

[BobDavid]

The sky is often tricky to capture: cameras, lenses, sensors, RAW processors, human sensation and perception, atmospheric conditions, UV index, etc. make for lots of variables. And, we all have slightly different expectations of how the sky looks.

The green or magenta signatures that often shows up in digital photos are problematic; human vision is different from machine vision.
We've learned to expect the sky to be blue (on a clear day, depending on time of year and time of day, it generally falls within a dominate range of cyan) and so that's how we expect to see it. What's particularly interesting is that color perception is rather plastic. The brain white balances differently from how a camera works.

Another consideration is that the sky looks and always will look different on a screen than it will in a print.

So my advice, is to practice. There are lots of ways to alter the color of the sky. One technique is to add a layer, fill it with an appropriate hue, and then lower the layer opacity level to taste. Mask out the non sky areas to avoid contamination.

It's also fairly simple to make a composite by taking the sky from another picture. It's a good idea to build a stock library of skies taken with the same camera at a range of ISOs (to match noise).

Walk into a museum and see how painters render the sky. There's a lot of variation.

[nirpat89]

The lesson is look at the histogram in the back of your camera before taking it to the RaeTherapy or whatever post-processing tool is to be used.  So you can immediately take another shot or two to correct the exposure.  And before starting to think about color management issues, make sure the histogram looks proper. 

[Hening Bettermann]

@nirpat89
There was a strong wind, I had to use a short shutter speed, and was reluctant to open the aperture wider than f/8. I forgot to raise ISO - and I also forgot that it wouldn't have helped, as Andrew has pointed out in this thread. The value of the in-camera histo is limited as long as it is only based on the in-camera jpeg, not on the raw.

[digitaldog]

The lesson is look at the histogram in the back of your camera before taking it to the RaeTherapy or whatever post-processing tool is to be used. 

When shooting a JPEG yes. When shooting raw, no. Not unless you've got a raw Histogram on the back of that camera.

[nirpat89]

@nirpat89
There was a strong wind, I had to use a short shutter speed, and was reluctant to open the aperture wider than f/8. I forgot to raise ISO - and I also forgot that it wouldn't have helped, as Andrew has pointed out in this thread. The value of the in-camera histo is limited as long as it is only based on the in-camera jpeg, not on the raw.

Are you using manual mode?  ISO would have helped only in that case.  Yes, in-camera is a jpeg, but difference between an uncompressed raw or tiff file and a jpeg as far as histogram is concerned is not so great is to grossly misconstrue the exposure.  I bet the day was hazy, foggy or smoggy to start with as the dynamic range of the scene is quite narrow, not something you will get on a typical clear, sunny day with plenty of shadows - not in my experience.  Perhaps a polarizing or UV haze filter might have helped.

P.S. I would say the shot was not actually "under-exposed" in the classical sense as the shadows on the left were still not clipped.  Typically folks would add exposure in a scene like this to move the right side of the histogram all the way to the edge (exposing to the right or ETTR) which would help minimizing the noise in the shadows as Andy mentioned.

[Tim Lookingbill]

@ Tim #101
Hi Tim,
> If it isn't Raw Therapee's fault then why doesn't that image look like it was shot in daylight even if the blue sky looks right?

Well it looks OK to me, may need more work on the TRC, but I find colors in the ball park.

> Can you post a different daylight landscape image processed in Raw Therapee that does look right including the blue sky?

I'll try.
Here is another daylight landscape, shot on May 10th, 2016, at 17:21 h. Opened in Iridient, in RT, and in RT with contrast +20 to match Iridient. No other edits. Same profile, my custom created in DCamProf (Combo LUT). I have to admit that I like the sky in Iridient a little better. 
Difference is visible even in the jpegs. (This is not my final edit!)

> I still don't know what can be learned about processing color images from this thread.

Well what I learned is that RT's AutoLevels saved me an underexposed image, which I was unable to do manually. And that the problem was this underexposure, and the saturation, not the profiles, color memory or white balance.

Hening, is there a reason your shots of a mid day scene look so dark? What happens when you brighten the image to look as bright as a noon day should?

I just shot the image below yesterday of my local park. Just another color perception test shot to check for yellows in greens, magenta in blue sky and neutral look of tree trunks. It's pretty darn close.

What's up with your dark daylight shots?

[Stephen Ray]

SNIP<The camera profile is created with DCamProf>SNIP

This is where I stopped reading the OP's very first post. Honestly. As I said in my first post #19, "Had the photographer not interfered."

[digitaldog]

Interfered how, with his own profile compared to what?

[Stephen Ray]

Interfered how, with his own profile compared to what?

Yes, with his own profile, which I believe is inferior to something more normal to what the camera factory settings should provide. 

[Stephen Ray]

Just another image for comparison against the OP's original post image.

Late August, 1:PM, viewing south-west, Canon Rebel using kit-lens, exposure set by camera, Lightroom WB 5500k/+10M, no other LR changes.

The scene is similar with blue sky areas, sunlit treetops, and green lawn tones (center distance.)

[Stephen Ray]

I just shot the image below yesterday of my local park. Just another color perception test shot to check for yellows in greens, magenta in blue sky and neutral look of tree trunks. It's pretty darn close.

Just another image for comparison against your posted image, Tim.

May 22, 2016, 1:PM, viewing West, Canon Rebel using kit-lens, image straight out of the camera, no PS or LR adjustments.

[digitaldog]

Yes, with his own profile, which I believe is inferior to something more normal to what the camera factory settings should provide.

What camera factory setting? It's raw.

[Stephen Ray]

What camera factory setting? It's raw.

Does not the OP's camera profile options show in any of his raw processors?

[digitaldog]

Does not the OP's camera profile options show in any of his raw processors?

Which one? In which product? And even if they did, the same profile would very likely produce differing results in two different converters. Just like CCT Kelvin numbers already provided.