I'm focused on visual perception.

Each eye has 5 million light sensitive cells called "cones". Cones are responsible for daytime color vision. This is sort of like a 5 megapixel camera.

But each eye is connected to the brain by only 1.2 million optic nerve fibres.

The answer is nature's data compression scheme, called the opponent color process. After the three cone types sense light, fourteen types of ganglion cells combine and reduce the signals to a "dark to light" channel we call luminance. Most of the visual information is encoded this way.

The remaining color information becomes a yellow to blue channel, and a red to green channel. This gives us the four "unique colors" of human vision: red, yellow green, and blue.

Yellow and blue are "opponents", when combined, the color goes away, and become the darkness to lightness channel.

Red and green are opponents, when combined, the color becomes yellow.

Red, Green, and Blue are the minimum number of lights needed to stimulate "a somewhat full range of color."

This is because there are three types of cone cells in the eye (LMS), each covering three slightly different ranges of visible light.

LMS stands for Long, Medium, and Short, meaning the wavelengths of light they are most sensitive to.

But in a funny way, large, medium, small are also descriptive. We have more L cones than any other; they make up the largest group of cones. We have very few S cones, so they make up the smallest group.

The L cone is sensitive to deep reds, through orange, yellow, into green. It is most sensitive to yellow.

The M cone is sensitive to orange, through yellow, green, and into teal and cyans. It is most sensitive to green.

The S cones make up only 2% to 7% of cones, and are loosely scattered in our peripheral vision. They cover from cyans through blues and into violet.

The L and S cones can be combined to sense purples and magentas.

From the description, it should be clear that the L and M cones mostly overlap. it is the small differences between the peak sensitivities, that gives us the range of colors from red through turquoise.

The L and M cones combine together to create luminance, the darkness/lightness channel.

The luminance channel contains all of the fine details, edges, thin lines, etc. The color channels contain very little detail, in fact less than a third of the resolution of the luminance channel.

What we see is divided into a lightness/darkness channel cslled luminance, with colors separated into a red/green group and a blue/yellow group.

The luminance channel and the color channels support different functions and are used for different tasks. Small details are held in luminance. Therefore luminance is most important for reading text, detecting edges, understanding textures.

Colors are processed separately. They lack fine details but are used for object recognition (an orange vs a baseball) and discrimination (ripe orange vs an unripe one).