Re: New(ish) Alternative 10 Bit monitor solutions

At a fundamental level, the print and motion picture worlds are on different planets.
When we aim for a Rec709 rendition, its the result of a lot of history, and the
SMPTE/CCIR recommendations and standards that have been established. These have evolved over time with a specific light emitter in mind -- the excited SMPTE/EBU phosphor in a CRT, and everything has to either be that or emulate it.

Here's an example of the lengths the manufacturers have implemented in an attempt to circumvent the problem: the Sharp "Quattron" LED flatscreens with "4-color technology". What a load of bollocks. The problem they have is that they can't generate SMPTE white with only their RGB LEDs. Its like this. In a simplified Trinitron, there were three stripes of phosphor; one that passed through Red, one Green, and one Blue, and we figured out how much energy to generate so that the RGB melded in proper proportion to give the human visual system a synthesis of nature, or at least NTSC. That was then. There is a "gamut" of possible colors that can be generated within any signal path. In a sampled system, there is a finite number of elements which defines the so-called resolution of a system, not only in terms of its physical aspect, but its color depth, as well. Enter HD. Even at 720P, a simple grey scale sawtooth would require 720 separate grey levels to avoid stairstepping. In 8-bit processing, there are how many? Anybody? Not 255. Not in TV -- its actually 219 levels mapped 0 IRE to 100. So the situation is worse than you thought, not the same as photography at all. 10-bit processing?... that'll be good, right? 1024 greyscale levels? Not so fast. Same deal, 0-100 is mapped 64-940, so actually that number is 876 steps. It works very well for 720P, good enough for 1080P if it isn't displayed too big, works great in interlaced 1080i, in fact more than adequate.
But what about those phosphor thingies? You refer to the resolution trickery going on in DSLRs -- no better example in digital motion picture cameras can be found than in the R3D-cam, a "Bayer-pattern" sensor. Its a single element with an array of RGB-sensitive receptors -- one line of GRGRGRGR... and the next line of BGBGBGBGBG... and so on. You will also find this pattern in some of the older style of pixel boards like the Mitsubishi Diamondvision, in which one "white" pixel was generated by two green phosphor tubes and one each red and blue in a 2x2 array. So that is why Galt and Thorpe of Panavision and Canon really talk about "4K" as being a mathematical fudge, at best. Its really a kind of 2:1:1 system. How does this translate to the Sharp Quattron "4-color" screen? Well, yellow is Red+Green. Its a way of putting that second green emitter in, which they should have, but then things would probably have gotten really complicated. My guess is that they also have other problems like blue being way too bright in the LED world, and they can't get the black performance at the other end without stepping all over that. And it will be the same or similar with all these other graphics displays.

You may have gathered I don't have much confidence in the system and I wish George would just shut up some times, because the audio on that spot is not a masterpiece of dialogue editing either... no offence. Couldn't he get the lines? The phrase inflections are very discontinuous. Or maybe there wasn't enough budget to auto-tune that out. Whatever, but it makes me feel like I'm being deceived.

Dithering is a semi-random (noise-like) introduction of neighboring values intended to confuse the eye into an impression of gamut smoothness. Its like a blur, but in color depth.

BTW, the problem with yellow in the CRT world is that it is generated, of course with R+G or -B, and it usually winds up being interpreted as "greenish" -- color bar yellow displayed on its own is not what most people would pick as a pure hue. And time after time, over the decades, and many systems, I am always being asked to fix the green foliage in a shot (which of course has no vectorscope presence on or near the axis where the "G" target is painted) -- its almost always floating a little bit west-southwest of color bar yellow.

So, in brief, video displays and graphics monitors -- they look the same, do similar things... not interchangeable. While it may be technically feasible to simulate some of their functions back and forth, you will always be looking through yet another filter or interface.

jPo

You mean "Old Ben"? Ben Kenobi?