Bayer Matrix GRGB. Why?

[Mark D Segal]

Maybe the principle is to make the camera perceive colour in a manner similar to how the human eye perceives colour - in which case the configuration makes sense - and it works!

[RichardChang]

As the concept was explained to me, each color of R, G, and B have a contribution to luminance (that's the detail of the picture, rather than the color of the picture).  Green has the most contribution, red has less, and blue has the least contribution of the three.

Human visual perception is based mostly on luminance.  This allows us to hunt and gather in very low light, which typically lacks color.  The human brain "fills in" the color information we need, which is why we perceive normal color in flourescent rooms that film and digital "see" as greenish.

Our retina is composed of rods and cones.  Rods are luminance receptors and cones are chrominance receptors.  We have more rods than cones, because seeing shapes is more important to us than the color of those shapes.

The Bayer pattern is nothing more than a design that is actually patterned after human visual perception.

Richard Chang

[Rob Keijzer]

Hello everybody,

Every book or article about digital capture sooner or later comes up with a description of a sensor.
Also the Bayer Matrix sensor that has a GRGB configuration. the reason for two green-sensitive sites in stead of one is that the human eye is more sensitive to green, than for other colours, say the books and articles.

I appreciate that our eyes are more sensitive to green, but why does my camera's sensor need to boost the greens on top of that?

IMO this way, greens are boosted twice: once in the camera, and another time in me.

Rob

[marcel123]

The sensor is not "boosting" the green - it is just sampling it more often.

The end results is that the the green channel has twice as much information (i.e. resolution) as the red or blue channels.

Thus, when you construct the final image from the matrix, the green channel is sharper. It's not "boosted" in the sense that it is greener, but rather in the sense that it has more detail.

Because of the rectangular nature of the matrix, and the fact that we only use 3 color samples, it was reasonable to simply have an extra sample for one of the 3 colors. Since the eye is most sensitive to green, it makes sense to add resolution to that color channel because it will have the greatest perceptual result.

This is very similar to how color TV (NTSC) works. The luminence (black and white) part of the signal is given a lot of resolution, while the color is given relatively little resolution, yet because the eye is more sensitive to the luminence, it perceptually works.

[Rob Keijzer]

Thanks, all of you for throwing some (green) light on this. I think I understand now.
Particularly what's been said about NTSC made it clear to me.

I guess it's a bit like: what can't be seen needn't be recorded, and what the eye sees best should be recorded in the finest detail.

Rob