Why RGB?

The rods gets the level of ilumination and they are the principal responsibles of the peripheral vision and the night vision in black and white. However, they don't give color information. For this task the cones are the responsibles, of these we have three types, the type L (red), the type M (green) and the type S (blue). Each one of these photoreceptors capture the light that oscillates in a range of frequencies different, with more sensibility in the colors yellow, greenish yellow and violet, this doesn't mean that only they can detect those color, but that inside the range of color that they detect, these are the colors that we see with more intensity. Next you can see a graphic with the sensibility of each type of cone for each color, that was obtained doing visual test to human subjects:

In the 20s, the CIE (International Commission on Illumination) choose the colors red, green and blue as respresentative of the ranges of color that can see each cone due to that they are in frequencies that only its asociated cone can detect and because they were colors easily reproducible with the technology of the epoch.

But, how can be formed the rest of the colors from the red, the green and the blue? Each cone alone is blind to the color, but combining its perception with the rest of the cones emerge all the colors that we can see. To understand how this happen I'm going to explain the additive color mixture:

When we are children in the school we learn to mix colors in a paper, using an antique and incorrect color model, the RYB (red, yellow and blue). This way of combine the colors is subtractive, is that to say, we begin with a white paper, in which each time that we paint a color, we substract to the light that arrives all the color except the one that we have painted. In this kind of mixtures, if we mix all the primary colors we get the black. However, the RGB color model is additive, because we part from a black screen and we add colors with light of the same color. In this case if we join all the primary colors we get the white. In this model perhaps the mixtures seem less intuitives because we don't learn them in the college, but is relatively easy learn them, because the basic mixtures of the RGB color model give as result the primary colors of the substractive color model CMYK used in the printers, althought isn't so intuitive as other color models, as the HSL, that is specially thought for that the mixtures are easy to do for the humans. The mixtures of the RGB are: with red and green we get yellow, with green and blue we get cyan, and finally, with blue and red we get magenta.

Think now in the nature of the visible light. This electromagnetic wave/particle can be composed by an infinity quantity of different frequencies, each one of them with different intensities (bigger number of photons). If we could see all the combinations of frequencies and intensities that can form the visible light, we could see infinite or practically infinite colors however, only we can see a little part because our eyes do that different colors in the light be only one on our brains. These colors are called metamers. The way the colors of the reality are assigned to the colors in the brain depends enormously of the type of primary colors that we can capture, because if we have four types of photoreceptors of color, as the birds, we would see colors that we can't even imagine. In the same way, the dogs, that only have two types of cones, they don't can imagine how we see.

Then, if for a light that have different intensities of all the color of the rainbow, we can get a light that produce to us the same sensation of color (metamers) using only three colors, will be a lot of more easy do a screen that can reproduce artificially that color, because we can use that light more simple that only use three primary colors. These is the reason because the RGB model is important. In the following image we can see an example of metamerism, where have been calculated the perceived color for two lights with different components. This calculation is do it applying the filter of each cone (see previous image) to the representation in frequencies of the light, through an operation called convolution:

Nowadays, any monitor can show more colors of which we can differentiate inside the RGB model, however, there are colors that can't be showed and we can see, due principally to that our three primary colors don't correspond exactly with the ones of the RGB model. Due to that, was created a sRGB color model that, with variations in the primary colors, expands the spectrum of represented colors, for its use in professional monitors, but even so, this extended model, still can't represent all the colors that we can see. To finalize here you have the chromatic diagram of RGB (triangle) to compare it with the colors that we can see (figure with form of heel).