Pages: 1 2 [3] 4 5 ... 8   Go Down

Author Topic: The sky is blue - or should be...  (Read 5775 times)

digitaldog

  • Sr. Member
  • ****
  • Online Online
  • Posts: 14119
    • http://www.digitaldog.net/
Re: The sky is blue - or should be...
« Reply #40 on: May 02, 2018, 04:27:06 PM »

Sorry, Doug, my mistake.
You made so many! Just today.

Quote
It's clear Andrew hasn't edited very many daylight images seeing he isn't a photographer any longer and is just a technology writer.
Magenta is a nasty hue especially when rendering in ACR using Knoll's color engine which provides a very pure hue of this magenta that if applied in the right amount creates perfect renditions of blue skies that I take visual notes on when I photograph my local park. Daylight changes hue in this town and there are a myriad of beautiful blue sky color I want to capture and faithfully render.
Rubbish and more rubbish. How can you edit an invisible hue in ACR or anywhere else. Such comments are compelling reasons not to take you at all seriously again. Par for the course for someone who isn't a professional photographer or a technically correct writer. That’s where we differ Tim!
« Last Edit: May 02, 2018, 04:49:09 PM by andrewrodney »
Logged
Andrew Rodney
Author “Color Management for Photographers"

Doug Gray

  • Sr. Member
  • ****
  • Online Online
  • Posts: 1346
Re: The sky is blue - or should be...
« Reply #41 on: May 02, 2018, 04:28:10 PM »

I'm glad you brought that up because it is an invisible hue when when mixed with a bluish looking daylight image especially with a magenta hued blue sky. How much yellow should be seen in green foliage lit by a noon day sun on a clear day? How warm does the sun make natural objects look no matter whether they are cool or warm toned? To get a convincing daylight appearance it helps to take note how much magenta is in sky blue.

For example stare at the second image I posted (not the flotube WB top version) and then look at a neutral gray area on your display. What hue has it turned into? On my display it makes ACR's neutral R=G=B surround take on a muddy greenish yellow tint. Not too pronounced but very subtle. But note you can't really see magenta in the green foliage or warm concrete. Sample the skin tones and the Lab reading favor a orangish yellow hue and have tint slider in ACR set to +30.

How much a color appears to change when adding/subtracting other RGB colors depends on the color. The ones that affect luminance the most, such as yellow, make a bigger difference because adding yellow increases luminance a lot. The other end of the scale is blue. You can add quite a lot of blue without changing luminance so the effect adding blue to, say a strong yellow is quite low, It mostly results in lower saturation.  Magenta has bigger effect on luminance than blue but is still relatively small. So adding a bit of magenta to greens mostly desaturates the greens. OTOH, adding greens to a magenta color, or yellow to a blue color, will produce marked increases in luminance.

But that's in RGB colorspaces. In L*a*b*, you can hold luminance constant and add/subtract colors on opposite sides of the a*b* center (neutral by definition). You will see very different results doing that.

Logged

Alexey.Danilchenko

  • Full Member
  • ***
  • Offline Offline
  • Posts: 241
    • Spectron
Re: The sky is blue - or should be...
« Reply #42 on: May 02, 2018, 04:31:16 PM »

As far as I understand it he is looking after the reason not how to turn the sky blue.

Reason - rubbish profile. What comes with ACR (profiles) is for most part of it not good so it may be beneficial to build and tune your own profile.
Logged

Stephen Ray

  • Full Member
  • ***
  • Offline Offline
  • Posts: 136
Re: The sky is blue - or should be...
« Reply #43 on: May 02, 2018, 04:35:37 PM »

What is your point with the look you've given those screenshots?

My point was to offer examples of how far off base the originals are from what typical imaging software would render the scenes as using common aim points of black, neutrals, white, and determining a white balance.

Are you saying "Auto" mode favors overly orangy yellow warm daylight? Or that they have a screwed up version of what D50 is suppose to look like?
I think in your Reply 34 you mention D50 as "way over in the warm region." Do you not?

I have never encountered any printer whose "Auto" mode rendered images that way. Or are you just being funny?

You have encountered those printers but you just don't realize it.


So, after my simple button-click exercise on my end, IMO the OP's sky was never as blue as he believes and other's renderings are as underexposed as the first. Look at the unusual histograms. Also, Tim's processing of his photo leaves a lot to be desired. But, just my opinion.


Logged

TonyW

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 472
Re: The sky is blue - or should be...
« Reply #44 on: May 02, 2018, 04:45:40 PM »

Uff what a discussion I triggered! Thanks to all who chimed in. I'll present the result of my own fiddling before I study your posts more closely.
There is no way I can achieve a satisfactory result in RawTherapee. I get my best result with an adjustment layer 'Color Balance' in PhotoLine. However, the histogram becomes jaggy.
Hening,
Is there any reason why your first image has an embedded monitor profile?

As to RawTherapee I can see no reason why a good result cannot be achieved quite simply.  In this case using RT the attached adjusted file only had Auto and +10 saturation (and I normally do not use/like Auto.

Are you using the current version of RT - this might make a difference?

Wait a few seconds for image to change

Logged

Tim Lookingbill

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 2426
Re: The sky is blue - or should be...
« Reply #45 on: May 02, 2018, 04:46:07 PM »


I think in your Reply 34 you mention D50 as "way over in the warm region." Do you not?

Yes, I did say that but D50 in reference to actual daylight doesn't have that much of highly saturated version of orangy yellow as shown in your renderings. 

Also, Tim's processing of his photo leaves a lot to be desired. But, just my opinion.

Again, another unhelpful comment. Care to be specific on what you find undesirable about the look of the image I posted I specifically chose to represent an accurate rendition (not "Pleasing" rendition) of a daylight scene taken in my local park.

Can you see magenta in the foliage shadows or highlights? In the skin tones? The Lab readings indicate not much magenta is in those skin tones but you can really see it in the blue sky and clouds. ACR tint slider was set at +30. That's a lot of magenta.

Come on Stephen! Be helpful. No one's interested in your opinions.
Logged

digitaldog

  • Sr. Member
  • ****
  • Online Online
  • Posts: 14119
    • http://www.digitaldog.net/
Re: The sky is blue - or should be...
« Reply #46 on: May 02, 2018, 04:51:35 PM »

Come on Stephen! Be helpful. No one's interested in your opinions.
Again speaking for others (everyone) illustrating the pot calling the kettle black.
Logged
Andrew Rodney
Author “Color Management for Photographers"

Tim Lookingbill

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 2426
Re: The sky is blue - or should be...
« Reply #47 on: May 02, 2018, 04:52:26 PM »

Hening,
Is there any reason why your first image has an embedded monitor profile?

As to RawTherapee I can see no reason why a good result cannot be achieved quite simply.  In this case using RT the attached adjusted file only had Auto and +10 saturation (and I normally do not use/like Auto.

Are you using the current version of RT - this might make a difference?

Wait a few seconds for image to change

Hening still isn't explaining why the blue sky in the Raw Therapee screengrab looks a muddy yellow as he stated in his OP. Tony, did he message you and tell you he still see this no matter what Raw settings he uses? He's not answering what is causing the muddy yellow blue sky.
Logged

Tim Lookingbill

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 2426
Re: The sky is blue - or should be...
« Reply #48 on: May 02, 2018, 05:06:23 PM »

Here's the skin tone lab reading and ACR white balance slider settings. You can now see on the close up the effects of lots of magenta in the form of an "optical brightener/purplish iridescent " effect on the woman's black hair but her skin does not look pink, purple or reddish. This is why I said magenta is the invisible hue in actual daylight with the sun high in the sky. Not the same for the golden hour.

How much magenta is too much is subjective of course.
Logged

digitaldog

  • Sr. Member
  • ****
  • Online Online
  • Posts: 14119
    • http://www.digitaldog.net/
Re: The sky is blue - or should be...
« Reply #49 on: May 02, 2018, 05:13:30 PM »

Here's the skin tone lab reading and ACR white balance slider settings.
No, it's not. You still have no idea what the Digital Color Meter is showing and how!

It's not measuring anything. It is taking two or three bits of information:

1. The color that an app is actually outputting to a pixel. i.e. an RGB level.
2. The colorspace that the app says should be used for that pixel for ColorSync to correctly display it (defaults to sRGB if the app doesn't specify).
3. The ICC profile associated with the display.

IF you knew how to use ACR, you'd know how to set it for Lab read-outs without futzing with the Digital Color Meter you've shown here today and pasted below, with ACR set clearly to RGB. Would you like to learn how to use ACR to show you the Lab values of the actual data within ACR or continue to confuse yourself with the numbers that are divorced from the actual data?
Probably the later..... :-\
Logged
Andrew Rodney
Author “Color Management for Photographers"

digitaldog

  • Sr. Member
  • ****
  • Online Online
  • Posts: 14119
    • http://www.digitaldog.net/
Re: The sky is blue - or should be...
« Reply #50 on: May 02, 2018, 05:18:37 PM »

How much magenta is too much is subjective of course.
Only if it's visible.  :P
Logged
Andrew Rodney
Author “Color Management for Photographers"

Tim Lookingbill

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 2426
Re: The sky is blue - or should be...
« Reply #51 on: May 02, 2018, 05:44:18 PM »

Still ignoring nimrod.
Logged

digitaldog

  • Sr. Member
  • ****
  • Online Online
  • Posts: 14119
    • http://www.digitaldog.net/
Re: The sky is blue - or should be...
« Reply #52 on: May 02, 2018, 05:47:03 PM »

Still ignoring nimrod.
"The highest form of ignorance is when you reject something you don't know anything about". -Wayne Dyer

Modern Nīmrōd, Tiberian Nīmrōḏ, Aramaic: ܢܡܪܘܕ‎, Arabic: النمرود‎ an-Namrūd), a biblical figure described as a king in the land of Shinar (Assyria/Mesopotamia), was, according to the Book of Genesis and Books of Chronicles, the son of Cush, therefore the great-grandson of Noah. The Bible states that he was "a mighty hunter before the Lord [and] .... began to be mighty in the earth".[2]
« Last Edit: May 02, 2018, 05:57:15 PM by andrewrodney »
Logged
Andrew Rodney
Author “Color Management for Photographers"

mouse

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 255
Re: The sky is blue - or should be...
« Reply #53 on: May 02, 2018, 05:49:36 PM »

So to continue if necessary  ;D , raising the ISO would have zero effect on exposure. Yeah, it's under exposed and there lies the beauty of a raw Histogram and RawDigger.

Apologies for sticking my oar into this thread, but this reply has me confused.  It has been my belief (and experience IIRC) that increasing the ISO (while not, strictly considered, an exposure parameter) has a very direct effect on the raw Histogram as displayed in RawDigger. 
Logged

digitaldog

  • Sr. Member
  • ****
  • Online Online
  • Posts: 14119
    • http://www.digitaldog.net/
Re: The sky is blue - or should be...
« Reply #54 on: May 02, 2018, 06:40:35 PM »

Apologies for sticking my oar into this thread, but this reply has me confused.  It has been my belief (and experience IIRC) that increasing the ISO (while not, strictly considered, an exposure parameter) has a very direct effect on the raw Histogram as displayed in RawDigger.
See: https://www.rawdigger.com/howtouse/rawdigger-histograms-overexposure-shapes
Different cameras, even if based on the same sensor, may render extreme highlights around the clipping point differently; and differently, even with different values of clipping points, depending on ISO setting. It is important to recognize the look and calculate the practical clipping point, which is not always the same as the maximum raw value. Here we will try to demonstrate the typical “looks” of the histogram of the clipping zone.
Logged
Andrew Rodney
Author “Color Management for Photographers"

TonyW

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 472
Re: The sky is blue - or should be...
« Reply #55 on: May 02, 2018, 06:43:17 PM »

Hening still isn't explaining why the blue sky in the Raw Therapee screengrab looks a muddy yellow as he stated in his OP. Tony, did he message you and tell you he still see this no matter what Raw settings he uses? He's not answering what is causing the muddy yellow blue sky.
Tim, I am not seeing the muddy yellow sky he is seeing either in Firefox, PS or Raw Therapee from the raw file or the screen grab in the first post (at least from memory-I am away from PC now)

I have had no messages telling me what he is seeing

Selecting areas with the PS colour picker indicated sky hue to be grey/blue.

My gut feeling is that the issue may be a poor monitor profile causing the yellow muddy tones and leading to potential over correction which looks to be the case with a corrected image in reply#35; appears too blue overall to me

In any case there is no reason I can see that suggests he will not get a pleasing rendering from RT (which may be very different to my play!) or as Andrew has shown in ACR.  Assuming of course an accurate monitor profile

Logged

digitaldog

  • Sr. Member
  • ****
  • Online Online
  • Posts: 14119
    • http://www.digitaldog.net/
Re: The sky is blue - or should be...
« Reply #56 on: May 02, 2018, 06:57:15 PM »

I have had no messages telling me what he is seeing
Nobody here but the OP does (and Tim of course), and speaking for everyone who's not sitting in front of his display is OK in this context. We simply have no idea what he's seeing but that doesn't stop someone from suggesting what he's seeing as he uses English words (muddy/yellowish, Magenta that isn't invisible, blue) mean and appear like. Utter speculation. And even the numbers provided by someone is bogus, not that numbers tell us what a color appears like, certainly without color management going full circle to needing to be in front of the OP's display with him.
Logged
Andrew Rodney
Author “Color Management for Photographers"

GWGill

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 464
  • Author of ArgyllCMS & ArgyllPRO ColorMeter
    • ArgyllCMS
Re: The sky is blue - or should be...
« Reply #57 on: May 02, 2018, 08:50:43 PM »

I'm not doubting you think your display set to D55 looks neutral to you, but the OP has stated his display set to D50 makes the blue sky image he posted look muddy yellow which doesn't look that way on my display.
Sounds like your display isn't reproducing in an sRGB like EOT then.
Quote

Does the blue sky sample image posted by the OP look muddy yellow to you on your D55 calibrated display?

Yes it does, because that's the color it is (in sRGB space). It's low saturation :- b* is -9.

And it looks (subjectively) the same on my D65 white point Macbook Pro retina display.

Quote

That's the problem with folks associating a color temp number to some hue of neutral for display calibration. At some point the yellowish/orangish hue of D50 compared to D65 is going to contaminate certain colors such as blue sky because clearly ICC display profile matrices and/or LUTs aren't fixing it calibrating to D50 if blue skies look muddy yellow as the OP indicated.

Umm. what ???
 
Even in a mixed WP adaptation situation (something that is hard to organize - the display tends to dominate your perception if it is large and you are concentrating on it), a WP shift will alter hue perception, but won't alter saturation.


Logged

Hening Bettermann

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 902
    • landshape.net
Re: The sky is blue - or should be...
« Reply #58 on: May 03, 2018, 07:30:28 PM »

I highly appreciate all the concern I got for my problem! Now I'll try to reply to (some of) your posts.
First off, I made 2 mistakes in my initial statement:
1-My camera profile was not shot by me - I used the target shot kindly supplied by Imaging-Resource.com. Unfortunately, I can't retrieve the link to it, nor the raw file on my Mac.

2- my initial basic editing in RawTherapee was Lightness +40, Contrast +80, not the other way round as I posted.

Now some comments

@Andrew
Yes I'm aware of that WB only refers to the blue-yellow balance and neglects the magenta-green balance. But the 5000K camera WB complies with my fresh memory in most cases - why not in this, and why the sky in particular? This is not a scene with a tricky lighting - it's about-noon daylight with a sunny sky, with moving clouds.

@DP, reply#13 on May 01, 2018

Thank you for your effort.

@Tim, reply # 14
Of the 2 versions of your park scene, if viewed one at a time, both look credible to me. If viewed simultaneously, I would prefer the 1st one. The sky looks more credible to me, and I would find that more important than neutral shadows. Why should shadows be neutral?

@nirpat89, reply #18
Your sky looks VERY good to me! I will later have to try if I can duplicate it using your method. I'm not familiar with Levels. - The fore- and middle ground I find too dark. So if I succeed duplicating your sky, I'll apply it using a mask.

@ Stephen Ray, reply #19
In your rendering of my image, the sky looks even worse, and the trees are all too yellow. I'm afraid I will have to go on interfering...

@andrew, reply #20
That sky is blue at least, and it looks to me like it would be good if lightened.

@sebbe, reply #21
> As far as I understand it he is looking after the reason not how to turn the sky blue.
Yes, mainly. Of course I want to cure this image also.
Yes the lens was a 135 mm Zuiko, which has a greenish tint - not a yellow one.
> I would add some points on tint towards magenta and also a few kelvins towards yellow. See below.
No, on the contrary...! I think the sky needs more blue/cyan, not yellow, and not at all magenta...Unlike Tim in the next post, I find the foliage all too yellow in your rendering. I find my own image quite OK in THAT regard. That is the problem - the SKY is off - why?

@Doug Gray reply #28
> Most likely the problem isn't the ICC display profiles but partial adaptation when using monitors with different whitepoints.
This has to be considered. But now that I have gotten all these replies I can compare them to my initial rendering on the same monitor, and the problem is the same. Some have provided better renderings of the sky, but at the expense of the foliage, so they would have to be applied separately to the sky.

@ Doug Gray reply #31
> If that happened to me I would measure the Lab reading from the monitor and compare it to samples from the image in the same area.
I tried it using the Digital Color Meter, but I'm not familiar with Lab, so I don't know what to make of the results:
image sky left side: L*82.26, a* -2.96, b* -9.70
screen area: L*100 a*0.00 b*0.00

@Tim, reply #39
> Hening, are you seeing how cyanish green that entire image looks
Yes I do, but I find the sky more to my memory, so I would have to apply that rendering to the sky only.

@Alexey.Danilchenko, reply #42
> Reason - rubbish profile. What comes with ACR (profiles) is for most part of it not good so it may be beneficial to build and tune your own profile.
Alexey, I don't use Adobe profiles. As I stated initially, I built my own with DCamProf, in this case using a target shot supplied by Imaging-Resource.com.
Yes I also suspect the profile - I just don't know how to improve it.

@Stephen Ray reply #43
> IMO the OP's sky was never as blue as he believes
That is certainly a possible bias to be considered. I am aware of that the sky is "blue" near the zenith, but not so much near the horizon, where the blue fades out. But on a day like this, it does not become muddy or yellowish.

@TonyW, reply#44
Uff! That rendering seems perfect to me! Theoretically unsatisfactory that 'Auto' does it. I'll try to duplicate it and try to find out what 'Auto' does.

> Is there any reason why your first image has an embedded monitor profile?
I would think it is because it's a screen shot?

The version of RT I used was 5.3-602.

@ Tim, reply #47
> Hening still isn't explaining why the blue sky in the Raw Therapee screengrab looks a muddy yellow as he stated in his OP. Tony, did he message you and tell you he still see this no matter what Raw settings he uses? He's not answering what is causing the muddy yellow blue sky.

I don't understand. If I knew why that sky looks the way it does, I would not need to bother you guys with my question - ?

@ TonyW, reply #55
> My gut feeling is that the issue may be a poor monitor profile
The monitor profile is created with the Eizo Color Navigator and an i1 Display, to the target as mentioned.

So next stop will be trying to duplicate Tony's rendering in RT - tomorrow!
Good night - and thanks a lot!




digitaldog

  • Sr. Member
  • ****
  • Online Online
  • Posts: 14119
    • http://www.digitaldog.net/
Re: The sky is blue - or should be...
« Reply #59 on: May 03, 2018, 07:41:32 PM »

@Andrew
Yes I'm aware of that WB only refers to the blue-yellow balance and neglects the magenta-green balance. But the 5000K camera WB complies with my fresh memory in most cases - why not in this, and why the sky in particular? This is not a scene with a tricky lighting - it's about-noon daylight with a sunny sky, with moving clouds.
My point was that the numbers are kind of meaningless when you get them from a camera which isn't the correct tool for measuring and reporting a correlated color temperature in Kelvin. That's a range of possible colors for one. IOW, that your camera 'states' CCT 5000K doesn't mean that's what the illuminant is. If you had say a Spectrophotometer and the right software, the reported values would be more valid but in no way does that mean your display is producing that color value, or that the two MUST match if you could get two such measurements.
See:
http://digitaldog.net/files/22Thecolorofwhite.pdf

Quote
That sky is blue at least, and it looks to me like it would be good if lightened.
Easy to do and one of the problems editing someone else's image. The data (raw) is there, and you should be able to render a blue sky as you desire.
Logged
Andrew Rodney
Author “Color Management for Photographers"
Pages: 1 2 [3] 4 5 ... 8   Go Up