Color tint red and yellow

[Joanna (Joanna Carter)]

@gregor

The real problems begin when you don’t have a calibrated monitor, as default white balance of most monitors out of the box is 6500K, therefore people without calibrated monitors should set White Balance in their camera or RAW converter to 6500K to get white displayed correctly on their monitor

And the problem with this approach is that, when images taken in this way are shown on a monitor other than the photographer’s (like when being shared on an internet forum), they will appear “wrong”

No, the camera must always remain the SST (single source of truth)

In the days of having to choose your colour temperature when you bought the roll of film, if you didn’t want weird coloured transparencies, you went out and bought a few colour correction filters (80, 81, 82 and 85 series). My argument would be that, if you are remotely interested in colour fidelity, nowadays, instead of filters, you need to calibrate, or have calibrated, your monitor; the cost is likely to be similar to the purchase of a couple of filters.

[Joanna (Joanna Carter)]

@gregor

Won’t they just look wrong on calibrated monitors?

I guess. But, if you want to print, you’re never going to get what you see on paper to match what you prepared on screen.

Or maybe it’s that most people don’t bother ever printing?

[OXiDant (Peter)]

Same here.
To all:

But one thing i don’t understand yet.
the “luminationleds” are aiming towards a whitepoint at 6500k right?
The sun/daylight is a total different light source so why setting my camera at 6500k?
Then i tel my camera the “whitecolor” is 6500k wile it could be 4800 or 7200.
The only thing i could think of is:
i need to change WB of a 1:1 showed WB by my feel of the way i like to change the colors.
(what i suppose do when i watch a image on my 6500k screen anyway)
I try to “see” the advantage of a fixed “5500k” daylight.

@Joanna

The problem with auto white balance, like auto exposure, is that it always tries to “average” everything. So a “warm” sunset will be cooled down and a “cool” light will be warmed up.

that’s why for ooc-jpeg there are “scenery settings” so you can avoid that. But say i use those and shoot “sunset” would my “as shot” in the raw also be different then when shot in AWB? (never thought of that, i used scenery for thing as panaorama and funny shots as candlestars to not influence WBnumbers for my rawfiles.)

something funny about 6500k setting:

See this video
i did a round on my processors but you can’t see the FRV WB (as shot is 4688k/20) daylight is 4788k/20)
weird or not? i can’t select 6500K!
and second observation PL’s WBtool is quite rudimentair in choices.
I notished the “AsShot” differences in numbers which i find rather strainge because those numbers should be stored in the rawfile as numbers not as “well somewhere around here you can start”

Edit: a quick google search:
(have to read this slowly: Color by Kelvin: A better approach to white balance

[Jul]

A few questions:

I’m finding this thread really interesting. Some years ago I used to set my Canon camera to cloudy when it was cloudy or less than brilliant blue sky. Re-processing some 10+ year old raws now, I find they look dreadfully warm without correcting the WB.

A few years ago I just started shooting everything on auto WB.

I only have a Dell laptop, so don’t think I can alter much in the way of screen settings myself. However, I do calibrate the screen regularly with a Spyder.

So, given the above, would I get the best results by applying the preset which applies only the optical/lens corrections, and then applying the PL4 “daylight” white balance to get to the best starting point for further colour adjustments? (I’ll follow Joanna’s advice and in future set “Daylight” as my camera’s default WB.)

I must say that I find the colour options in PL4 a bit confusing. I can choose “Camera Body” and then find my Canon model in the drop-down list. Alternatively, I can choose from a host of alternatives. However, my older Fuji X100 doesn’t appear. What should I use for this? Is it all just a case of “suck it and see”, or is it more “scientific” than that?

[platypus (Canon & Mac (Intel and M1))]

@Jul

I can choose “Camera Body” and then find my Canon model in the drop-down list.

…that’s a setting in the color rendering tool?

[Jul]

Thank you Gregor.

I didn’t realise that I can set a numerical value for WB kelvin in-camera. I’ll look into that in the manual. I’ll set the Colour Rendering as you suggest. As far as my laptop screen is concerned, although I don’t have the ability to set kelvin and brightness myself, I assume that the Spyder software does this as part of the calibration process? Certainly the screen dims somewhat when the calibrated profile is applied a short time after the laptop is started up.

[Jul]

I hadn’t realised that I can set a WB numerical value in the camera menu. I’ll do it now, having found the relevant bit in the camera manual. Would I get an identical result setting say 5,500 in camera and setting 5,500 using the WB slider in PL4? I’m assuming I would, and that setting WB in camera is just more convenient and time-saving for pp.
(Apologies if I’m a bit dim!)

[OXiDant (Peter)]

@gregor

A color temperature of 6500 K is standard for ordinary PC use and for the sRGB standard.

Hello Gregor,
It suspect that 6500K is chosen for office enviroments.

and i understood your example. one thing i eh keep missing is:
what about the captured light temp as measurement?
As far as i understand sensor=>rawfile conversion:
1 a sensor has no colorreference other then it’s wavelenght restrictions which is done by a physical “filter” placed before the wells so for instance infrared is blocked. (every sensormanufactorer has it’s own “idea” of perfect hue range which may enter as “exposure” on the sensor.
this range can be measured with a scale of Kelvin. (which is much wider then our human visible range.)
2 cameramanufactorers build in a software based Hue filter (human visible range) and shift this around over the wider huerange accoording to the actual sourcelight temp in order to capture as much as possible the humen eye perspective of the scene. Aka it tries to let everything looks as we should see it.)
This is the AWB feature.
3 if we are manual WB then we shift this filter/framerange our selfs.
(if i would use a Philips huecolorbulb and change in a white room with a red and green object his lightsource the AWB would change right in order to try keep the white wall white? (in fact i think i will test this if i have some time. )
4 this measurement of AWB is a number which the Demosiac algorithm of arawconverter is using to create the RGB values which the screen colorprofile algorithm is using to replicate the “color” of capture. (sadly every rawdeveloper seems to interpreted this measurement number different.)

(it is much more complicated because every human eye has it’s own sensitivity of wavelengths and thus everyone sees a color differently and sensors having tolerances and screen too … rabit hole.)
5 WB setting influence “Exposure” understandable because not every wavelenght hitting the sensor has the same luminance strenght and thus a different exposure target.
i can grasp the principles of this.

So you state/say that if i stop worring about what’s captured as White and basicly stop shifting the filter around on the original exposure and just “capture” a hue-range pinned on 6500k as White. i eliminate one variable which can screw my white balance and exposurevalue?
So if i do my philips hue test(scrolling through its hue settings emitting all kinds of source light) and use awb and a fixed Wb in my camera the end result of the rawfile to jpeg are for both the same?

some one send my a colorchecker rawfile:
as shot is around 8074/5sp and 8607/6dxo “fine shade temp is around 7500K” so it was inside the house on daylight i suppose.
Daylight shows it gets “blueisch” i think that’s because of the glas of the window(house) blocking redhue more then bluehue
second thing when i zoom in 100% or more both applications come closer to each other but not “the same”
video

And you state/say that if he was shooting at 6500k i have less variations in both applications? and less trouble to find “white”? and a proper tonecurve aka nice rendering of my image?
And why would i not choose say “natural daylight 5500K” as WB value?
(i recon the developer program wil render the 5500k as 6500k on screen to show white if the light source was 5500k to begin with.)

i definiatly are going to play with such a huelamp to fool the camera’s WB in order to see with my own eye’s what the camera is “thinking”. and do the same in a fixed WB to see if this is no problem in post. (not that i don’t believe you but i need to see it in practise to get the mental picture in my head.)

[OXiDant (Peter)]

That video guy is speaking at my speed! finially my thoughts arn’t run ahead but in sync with his

@gregor

The hue lamp sounds like a very worthwhile experiment and I look forward to the results.

ok i did a setup and shot a custom WB (make this as white) it endded up with 4612/5 as default
then i waited a bit to get darker because the lamps are not that strong.
silkypix shows this:
video
And now something “weird” is going on!
Watch to the RAW WB numbers when i click on the 4612/5 MWB.!
video
1- dxo default thought of wb daylight is 5400k/0
2 even if i have MANUAL WB SET it just does a estimated gues! As Shot isn’t AS SHOT!
this makes me a bit confused about the quality of the raw White Balance tool.

Then i looked at the i shine colored light direct in the lens to mesh with the sensors WB detector flooding it with colored light.
i must say AWB is quite good.

multilightsource in different colors:

i hope the zipfile can be uploaded:
It will be here for a few days

So look for your self.

conclusion @sgospodarenko, could you point someone to this to explain why “As Shot” isn’t as shot? i am confused about this.

edit: and how can i dial in 5600K as manual WB in my camera?

[Joanna (Joanna Carter)]

I’m not sure what you are trying to prove here, especially with the amount of colour noise due to high ISO

Here’s an exercise I did for a demo for our photo club :

First two shots taken with auto WB (the sky was clear blue)

As you can see, the one on the left was taken against the light and the one on the right was taken with the sun behind me.

If you ignore the surroundings, you will see that there is a slight but definite difference between the rendering of the chair fabric; this is due to the difference in blue content of the sky against and with the light.

PL tells me that the temperature for the one on the left was 5253°K (tint of 1) and the one on the right was 5083°K (tint of -6)

I have a colour meter that I use for my large format film photography work, where you need to use filters to make sure the WB is right for the reversal film. In this experiment, I measured the light coming from behind me with the colour meter and dialled that in to the camera.

Now we have two shots taken with manual colour temperature: the one on the left at 8580°K and the one on the right at 5150°K.

It might look like we have a colour difference but, if you crop out the centre of both shots, you get this :

As you can see, it is not easy to tell which was taken against and which with the light.

Although I don’t carry my colour meter around with me for digital work, I certainly wouldn’t let my camera tell me what it thought the temperature should be. And, not only does it tell me the temperature, it also tells what tint it thinks is right, which can be far more disturbing when you are faced with an image with no discernable shade of neutral grey/white for the pipette tool.

Thus my practice of knowing precisely the temperature and tint in the camera, which gives me a reliable starting point for anything I might need to do in PL.

I entitled a talk I gave to our photo club “It’s the camera that lies - or, at least, it doesn’t know the truth”

The only thing I find curious is that, despite manually setting the WB to 5600°K / 0, PL always shows me 5510°K / -3. And I have read others who say that different processing tools can often give slightly different results like that. Nonetheless, most of the time, I am happy with the results that PL gives me, only occasionally tweaking the balance to where I intended it to be

[OXiDant (Peter)]

@Joanna

I’m not sure what you are trying to prove here, especially with the amount of colour noise due to high ISO

nothing, i just try to understand what’s happening in DxOPL in the RAW WB tool.
(i didn’t had strong lights and colorchanging capabilities so it was this or nothing)

@Joanna

Now we have two shots taken with manual colour temperature: the one on the left at 8580°K and the one on the right at 5150°K.

this you do with a white card to set a mwb? or type the numbers in the camera?
(i can’t i think in my DMCG80 typing in numbers only shoot a white card to set mwb.)

That said arn’t you finding it strange that silkypix is neatly show a stable as shot in mwb and Dxopl is mucking about on the mwb rawfile’s? (see the two video’s which shows the different approach.)
then the hole purpose of setting a whitecrad to set wb on a scene is out the window.

[Jul]

Thank you Gregor for your explanation

[Joanna (Joanna Carter)]

@OXiDant

(i can’t i think in my DMCG80 typing in numbers only shoot a white card to set mwb.)

I just looked up the manual for your camera and it seems to show that you can set the WB manually (page 111?) Look for this symbol in the WB menu

[Jul]

Hi Joanna,

I’m now going to set WB in camera using the k value as you suggest in this thread, but I’m a bit unsure about handling the green-magenta tint slider in PL. Is there any way of doing this other than on a sort of “trial-and-error, what seems to look OK” basis?

[Joanna (Joanna Carter)]

Just set it to 0. Like the temperature, you can play around if things don’t look right, but that is likely to be for very few shots.

[OXiDant (Peter)]

This is totally new to me
i always assumed Rawfile’s doesn’t be effected by WB so why bother to set something. As shot is a suggestion by my camera and i choose which i like best.
(i think i will shoot a day in this modes see what happens if i do.)

It’s behind the four "mwb presets by shooting in the scroll list. (out of sight unless you scroll to the end.)
ok thus i need to set this to 6500K? or DxO’s default like 5200K?
or just 5500K (sunlight/daylight).

And to resume if i use this how does this work properly?
i set my camera on manual 6500K and just use it as i would in AWB mode. just “point and shoot” in A- mode or P-mode or any of the psam mode’s?
and just post choose in rawdeveloper’s RAW WhiteBalance tool my preferred Kelvin and tone?
By the way the “correction” tab is the colortone number in the Raw WBtool? diagram like

does set this number?

funny how a small question can be grown in this.

[Jul]

Thank you Joanna

[Joanna (Joanna Carter)]

@OXiDant

ok thus i need to set this to 6500K? or DxO’s default like 5200K?
or just 5500K (sunlight/daylight).

You’re better setting it to 5600°K. That’s “average” daylight

I usually shoot in A mode because that allows me to choose depth of field by setting the aperture

[George]

@gregor

According to DxO staff @Lucas and @Benoit PL4 first applies optical corrections, including Demosaicing, before it applies WB, exposure compensation, color rendering, tone curve etc.

I learned different. WB and exposure compensation are done on the raw data during the demosaicing proces. Both aren’t available when opening a jpg. Or behave different.

George

[OXiDant (Peter)]

I think we need to dig in some other threads.
But before demosiacing it’s doing (deep)prime, CA, and optical correction?
After demosiacing all the rest is added for preview.
A lineair DNG can have a full sensor colorspace by export so i assume WB is after demosiacing.