On Thu, Oct 01, 2009 at 09:57:32AM -0700, Brian Fisher wrote: > Your facts are all basically correct, but some things you are missing is > the fact that human vision is both dynamic in terms of it's ability to > perceive ranges of intensity and color and able to support a very wide > contrast ratio (around 1,000,000 to 1) compared to what contrast ratio > monitors are currently pumping out (somewhere between 350:1 to 1000:1) > (btw, our perception is basically decibel based, so it's more like the > best monitors are 1/4th of what we can do rather than 1/1000th) > > What that means is our monitors are pretty much crap when it comes to > maxing out our eyes contrast-wise, and that even though our eyes have > limited color perception, the limits and use of that limited color > perception depend on what we are looking at - or to put another way, which > 7-10 million colors the average person is distinguishing depends largely > on what colors and intensities there are to look at. (btw, this is why > contrast ratio doesn't matter so much for home or movie theatres with > paltry contrast ratios of 1:500 but the environment is blacked out - our > vision readjusts very well to just using the range provided) > > The big thing about all this new monitor stuff is the new high-contrast > displays, as Pierre said "new TV with > retro light using LEDs has higher contrast (up to 2000000:1 wich is about > 126 DB dynamic range)". People who have seen those new monitors have told > me the pictures looked like real life, that it was like looking out a > window, not at a monitor. > > With those new high contrasts though, if you aren't using the full > contrast range for a particular scene, RGB888 can be way too small. So > more range is needed if you wanna do something like go from a cave to the > out of doors - if the cave is half as bright with RGB888, you dropped half > your range, to RGB777, and your eyes can start seeing bands and such > better.
So the big difference, if I inderstand you correctly, is: RGB888 is over 16 million colors MAX, but the average human eye, being dynamic, distinguishes somewhere around 7 to 10 million colors AT A TIME. > RGB is designed to be close to our rods and cones, which for most people > see an R with very good color range, a G with good color range, and a B > with mostly OK color range (relatively speaking). That's why many > restricted bit color schemes with uneven bit allocation put the extra bits > in the R or G but never the B. But there are actually some people (I've > heard they are mostly if not all women) with an extra cone (I think) that > is pretty close to the B, and those people have amazing color perception > much much better than RGB888, and usually work in color and print related > fields cause they can do things like match paint samples and colors at a > level beyond what us normal people can. I have heard of that! Tetrachromatism! Apparently the most common forms of red-colorblindness and green-colorblindness are not that the red or green cones are missing, but that they are present but picking up the wrong wavelengths. So if you are Red-blind that means your cones are GgB Green, Different Green, and Blue. And if you are green-blind then your cones are Red, Different Red, and Blue. RrB Both mutations are carried on the X chromozome, and both are recessive traits. Since women have two copies of the X chromosome, they rarely have any colorblindness, because they almost always have a "good" copy of the Red or Green gene on the other X, but men having just one X are far more likely to have colorblindness. Tetrachromatism (If I recall correctly) happens when a woman gets one different kind of colorblindness on each X. Somehow, sometimes this results in the "Other Red" and "Other Green" cones combining to behave like a population of "Orange" cones, so her vision is effectively Red-Orange-Green-Blue and she can see orange as if it was a primary color, instead of as a secondary color like most of the rest of us do. --- James Paige
