• Please use real names.

    Greetings to all who have registered to OPF and those guests taking a look around. Please use real names. Registrations with fictitious names will not be processed. REAL NAMES ONLY will be processed

    Firstname Lastname

    Register

    We are a courteous and supportive community. No need to hide behind an alia. If you have a genuine need for privacy/secrecy then let me know!
  • Welcome to the new site. Here's a thread about the update where you can post your feedback, ask questions or spot those nasty bugs!

White balance color correction

Doug Kerr

Well-known member
It is interesting to reflect on the primary motive for engaging in "white balance color correction". It does not, for example, relate to correcting an "error" in the camera's recording of the color of a scene when operating under a certain kind of illumination.

The color of the light from each "spot" an illuminated reflective scene is determined by spectrum of that light, which in turn is determined by the the product of the reflective spectrum of the surface at the :spot" and the spectrum of the illumination.

So for a "white" piece of paper, the color of the light from it is different when it is illuminated by noon daylight and by the light of the fading sun.

Nevertheless, for quite a range of illuminance chromaticity, the human perceptual system will see that as a sheet of "white" paper. How can this be?

Simplistically, the human perceptual system "determines" the chromaticity of the illumination by considering the chromaticity of the light received from other objects in our view, whose reflective colors we "know" from experience. In then, in effect, "divides out" that chromaticity from the chromaticity of the light received from the paper and gets the "reflective color" of the object, which in this case is "white". That is, we see the paper as "white".

Now consider that we photograph the scene. The light received from the paper, illuminated by the fading sun, is reddish-yellowish, and (if we don't take any special steps) will be recorded as such in the image.

Now, a viewer, sitting at her computer in a noon sunlit room, looks at the picture. Her perceptual system is set up to "divide out" what it takes to be the chromaticity of the illumination in the room, determined from the chromaticity of the light from everything in sight, including the image on the screen, including our piece of white paper. But of course that is not the "proper" operation, since the general illumination in the viewing room is quite different from the illumination under which the scene was shot,

The result is that her perceptual system determines that the reflective color of this object (the white paper) is a bit on the reddish-yellowish side. That is, it looks to her as if this is a slightly reddish-yellowish piece of paper.

Now suppose we don't want that result. We want the viewer to perceive the paper as "white", just as she would if she looked at it in the same situation we had for our photograph.

So when we process the taken image, we do "white balance color correction". That means we shift all the colors to what they would have been if the scene were shot under illumination of the same chromaticity as what will be the "surround" illumination of the viewer's location. But how do we know what that is? Well, we don't.

So we do the color shift predicated on the fact that the chromaticity of the surrounding illumination at the viewer's location will be the same as the chromaticity of the white point of the color space used for our "delivered image".

Now will this always "work"? Of course not. If the viewer is seeing this image on her laptop computer while lying next to the fireplace fire, the paper will look bluish. But that's just how it is.

Now Asher wisely counsels us to not do this mindlessly. Suppose, for example, that we want the viewer to be very aware of the "romantic", sunset-lit nature of the scene. (Perhaps it is to be used as part of a ad for flavored personal lubricants.So we want the viewer to perceive the color of the white paper as reddish-yellowish. So then we do not want to perform the "white balance color correction" step in the classical fashion.

Why do I not say, "we do not want to perform the 'white balance color correction' step at all." Well, there is really no such thing. Our image processing software must work under some assumption of the chromaticity of the illumination on the scene. The way we in effect tell it "don't do this thing at all" is to set the chromaticity to be assumed for the illumination of the shot to the white point of our presumed output color space.

How do we know that that is? Well, the CCT (coordinated color temperature, set with the "Temperature" slider in ACR) for the major standard color spaces is "well known". For example, for sRGB, it is 6500 K.

But what about the other coordinate of the chromaticity, the Planckian offset (in ACR, this is set with the "Tint" slider). Well, just now, I don't know what that is for the white point of the sRGB color space. I'll figger that out a little later.

Best regards,

Doug
 

Asher Kelman

OPF Owner/Editor-in-Chief
I know it's a lot of work, but some diagrams or pictures would make this great article to something beyond fabulous!

Asher
 

Doug Kerr

Well-known member
Well, it seems that for D65 (and I believe the white point for sRGB is supposed to correspond to D65) the Planckian offset is about +0.003 uv units.

In think that in Adobe things, that would be +9 "tint" units.

For D65, the CCT (to state it to four significant figures) is 6504 K.

Best regards,

Doug
 
Top