Free idea: native HSB cameras and monitors

What's wrong with RGB?

All existing digital cameras and computer monitors use a format called RGB. In this format, the color of each pixel is represented by three numbers: one for red, one for green, and one for blue. Both the color and the brightness of the pixel are determined by the combination of these three numbers, and each of these numbers has a fixed minimum and maximum. For example, the brightest white is 100% red, 100% blue and 100% green. A very bright purple would be 100% red, 100% blue and 0% green.

This way of storing color is the root of a problem that photographers assume is an inevitable fact of life. When an image is overexposed, color and saturation information are permanently lost. This is because when an RGB camera measures a red, green, or blue level higher than 100%, it has to store it as 100%. So, for example, a bright green pixel the camera sees as 100% red, 200% green, and 100% blue would be stored as 100% red, 100% green, and 100% blue, which is pure white - not green at all. The relative level of green compared to red and blue has been lost. Because of this same problem, no matter how good your image editing software is, adjusting the brightness of an RGB image inevitably destroys color information. If you have a mostly green pixel at 50% red, 100% green, and 50% blue, and you double the brightness, you will be left with a pure white pixel - 100% of each color.

A related problem is the potential relative brightness of colors displayed on a monitor. On an RGB monitor, "pure" colors simply can't be as bright as mixed ones because brightness and color are not independently stored in the RGB format. In other words, a 100% red, 0% green, and 0% blue pixel must be less bright than a 100% red, 100% green, and 100% blue pixel. The format simply isn't capable of expressing a pure primary color that is as bright as the brightest mixed colors or pure white. Monitors can use algorithms to correct for this by making all primary colors brighter, but these corrections are mere guesswork.

The alternative

Cameras and monitors should use HSB, not RGB. HSB also uses three numbers for each pixel: hue (or color), saturation (or washed-outness), and brightness. The HSB color space is, of course, not new. But for some reason nobody makes cameras or monitors that use HSB as their native format. This doesn't seem to be because of hardware limitations - hardware sensors for HSB should be simpler than RGB sensors, because they correspond more directly to physical properties of light (hue is the mean frequency, saturation is the spread of the frequency distribution, and the brightness is its amplitude). Instead, it seems that people are so used to thinking in RGB terms that they assume that the limitations of RGB are fundamental limits of photography, and miss the possibilities of using other formats.

The possibilities

In HSB, because brightness has its own dedicated number instead of being derived from three other numbers, brightness can be adjusted without damaging color or saturation, and any color/saturation combination can be equally bright or dark as any other. It would still be possible to overexpose with an HSB camera - a pixel measured with a 200% brightness would still have to be stored as 100%. But with HSB this ceiling effect on brightness has no effect on hue or saturation.

With HSB, brightness could also be represented using a logarithm or some other non-linear function to allow extremely bright values to be differentiated from each other. With a traditional RGB monitor or printer, the advantage of this is to allow photographers to change their mind on exposure after the picture is taken, without a loss of quality. A native HSB monitor could depict objects that are literally blindingly bright and in any color, emitting a pattern of light that much more closely corresponds to the light that entered the camera when the picture was taken.