CamCyclopedia
Contents
- Color and light
1. How is light visible?
2. Types of light wavelengths
3. Color temperature
4. Visual perception
2. AWB/CCM in a digital camera
1. Types of AWB algorithm
2. Basic principle of how AWB works
3. CCM
3. Introducing AWB of Galaxy Camera
4. Reference
Color and light
How is the light visible?
How do humans perceive colors? To put it simply, light from a source like the sun illuminates an object, the human eyes are subjected to the light reflected off that object’s surface and the brain takes over to process that image.
The three elements required for this process, i.e. 'light source', 'object', and 'observer/human' are the tri-elements of color.
Types of light wavelength
Wavelengths vary according to their length as shown in the image below [1]. They can be divided into radio waves, infrared, visible light, ultraviolet rays, gamma rays, etc. ‘Visible spectrum’ perceived by human eyes has a short wavelength range of 390~780nm.
The ‘Visible spectrum’ is distinguished by its unique wavelengths of color as shown in the image below [2].
Color temperature is calculated in terms of numbers by using the absolute temperature of a particular light source. It is essentially the temperature of the black body that resembles the color of the light source.
A black body or blackbody is an idealized physical body that absorbs all incident electromagnetic radiation, regardless of frequency or angle of incidence and the color temperature transforms as shown in the image below [3]
Visual perception of humans affects how we perceive colors.
There are multiple effects but the below phenomena affect color perception a lot.
- Color adaptation
Color adaptation lets you experience a singular color on an object under varying lighting conditions (color constancy). If you look at a white paper underneath an incandescent bulb emanating yellow light, the paper feels yellow but it appears white again after going through the color adaptation - The Purkinje effect
Red hue looks darker and blue is brighter in the dark compared to when it is bright - Area effect of color
The intensity and clarity of the same color feel different depending on the area size
AWB/CCM in digital camera
What is AWB (Auto White Balance)?
The dictionary lists AWB as the expression of achromatic colors (white, grey, black) as it is. Broadly, AWB helps form the same object color in the camera irrespective of the light source.
AWB helps apply mechanics similar to a human’s color adaptation in their visual perception system (color constancy) to the camera system.
WB in the camera system adjusts the color of the object exposed to every single color temperature to match that of sunlight (5500K).
Adjusting WB by R, G, B standards present in the camera system means adjusting the R, G, B ratio of the parts that have to look achromatic to human eyes to 1:1:1.
AWB algorithm is broadly divided into chromatic and achromatic algorithms.
- Achromatic algorithm
Gray world: Assumes that the average of all RGB pixels in the captured image is the same
MaxRGB: The part reflecting the source of light reflects the color of it as it is making it assume that the color is white
White Search: Statistically analyses and calculates how each pixel in the segmented areas of the image is distributed in a chromaticity diagram - Chromatic color algorithm
Specific color AWB: This can speculate and calculate the white if it already knows the specific color of an object
Gamut mapping: Gamut mapping calculates the max color area of all wavelengths emitted by a particular light source and the probability of each color radiating out of one beforehand
Basic principle of how AWB works
Once AWB finds the area of achromatic objects in an image, it calculates and adjusts the RGB ratio of that area to 1:1:1. In the image on the left below, you can see how the camera image sensor detects a raw green-hued image overall, but AWB finds the white portion at the middle and processes the image to the one on the right below.
In case there is no achromatic object present in the image, it picks up what looks like everyday items in the picture (sky, grass, etc.), and applies the RGB ratio calculated ahead making the picture resemble the colors in real life.
Your average photo has various objects in different colors thrown into the mix. So, the weighted value for each object is applied and calculated for the RGB gain of each area to figure out the final RGB gain. Camera manufacturers put their know-how to work to run the algorithm. The Galaxy camera system too employs a combination of such algorithms to apply the final AWB calculation. Galaxy is trying to come up with more accurate coloring using AI-AWB deep learning in its newer models.
What is CCM (Color Correction Matrix)?
An AWB-processed image can render achromatic colors as it is but tends to botch chromatic colors in it.
This is because AWB has a short color correction range given the properties of RGB image sensors. The display requires an additional 3x3 matrix color correction adjusted to the displayable color areas. This is called a CCM (Color Correction Matrix).
Let’s understand this using the hexagonal figure below. The black and white vertices, i.e. the AWB reference points align, but the remaining points cannot be represented given their narrow area. You can think of CCM as the process of stretching these to align with the vertices. The final colors on the image are rich and vibrant.
The images below are the before and after versions of the CCM process.
Introducing AWB of Galaxy Camera
The Galaxy camera employs the ‘Auto White Balance’ feature by default, and the MWB (Manual White Balance) feature in Pro Mode elevates usability. The temperature of white can be adjusted in the color wavelength range between 2300K and 10000K. Keep tweaking the color temperature bar for lush images.
In the picture below, white balance is applied at a different color temperature in the same setup to bring out a different tone in each picture.
■ Related Content
Color (Displaying a wide color gamut) - Samsung Members
What is Color Science? - Samsung Members
CamCyclopedia Index - Samsung Members
You can also access CamCyclopedia anytime by going to Community -> Category (app) -> CamCyclopedia -> “CamCyclopedia Index”.
Reference
[1] Electromagnetic spectrum: https://commons.wikimedia.org/wiki/File:EM_Spectrum_Properties_edit_ko.svg (mintz0223. CC BY-SA 3.0)
[2] Visible light: https://commons.wikimedia.org/wiki/File:Linear_visible_spectrum.svg (Gringer. CC BY-SA 3.0)
[3] Color temperature : https://commons.wikimedia.org/wiki/File:Black_body_visible_spectrum.gif (Dariusz Kowalczyk. CC BY-SA 3.0)
2 Comments
