When discussing white balance, we need to start with Color Temperature (CCT). As the name suggests, color temperature refers to the "temperature of color". However, it differs from the cool and warm colors we perceive in daily life. It describes the color emitted by an absolute black body (a theoretical concept proposed by physicists) when continuously heated starting from absolute zero (-273°C). For example, when an iron block is heated in a steelmaking furnace, its color changes progressively: black → dark red → bright red → golden yellow, before the iron melts. Since the melting point of iron is 1535°C, we cannot observe the colors produced at even higher temperatures. In contrast, the blue flame seen during welding radiates from an extremely high temperature. These color variations stem from different proportions of light wavelengths. As shown in the figure below: at low color temperatures (reddish tones), light with longer wavelengths dominates, and the opposite is true for high color temperatures. This causes white objects to appear bluish under high-color-temperature light, and yellowish under low-color-temperature light.

To the human eye, pure white always satisfies **R = G = B**. What white balance achieves is normalizing the image under different color temperature conditions. Specifically, it multiplies the RGB responses from the image sensor by respective gain coefficients to ensure "true white appears as white": R' = R × R_Gain G' = G × G_Gain B' = B × B_Gain This adjustment makes **R' = G' = B'**. In short, the core purpose of white balance is to restore an object’s inherent, original color, no matter what kind of light source illuminates the scene.