Display Makers To Use Quantum Dots For Efficiency and Color Depth

ArmageddonLord writes with this news from the IEEE Spectrum, reporting on display industry gathering Display Week: "Liquid crystal displays dominate today's big, bright world of color TVs. But they're inefficient and don't produce the vibrant, richly hued images of organic light-emitting diode (OLED) screens, which are expensive to make in large sizes. Now, a handful of start-up companies aim to improve the LCD by adding quantum dots, the light-emitting semiconductor nanocrystals that shine pure colors when excited by electric current or light. When integrated into the back of LCD panels, the quantum dots promise to cut power consumption in half while generating 50 percent more colors. Quantum-dot developer Nanosys says an LCD film it developed with 3M is now being tested, and a 17-inch notebook incorporating the technology should be on shelves by year's end."

Comments

50% more colors!

[AlienIntelligence]

Enjoy your phone in new psychedelic colors!

-AI

Re:50% more colors!

[gsgriffin]

Great! I always thought my phone was missing about colors.

Re:50% more colors!

[Charliemopps]

I must be getting old... I fucking hate pre-teen sudo-insults. Especially poorly thought out ones like yours.

Re:50% more colors!

[Jeremi]

I fucking hate pre-teen sudo-insults.

A: you're a loser!
B: I know you are, but what am I?
A: sudo you're a loser!
A: Oh God, you're right! I've wasted my life! <sob>

Re:50% more colors!

[I_am_Jack]

Wouldn't that be
% sudo -u loser ./insults.sh


(forgive my lack of command line skills)

Static images

[AlphaWolf_HK]

Any word on burn-in, permanent image persistence, or uneven aging? That's my main concern with OLED and Plasma.

LCD can get image persistence if it shows the same image for very long periods of time (e.g. 24 hours) but on most displays it is only temporary.

I'd be interested to hear if quantum dot might have any of these issues.

Re:Static images

[moosehooey]

Since this is BEHIND the LCD, the light passes through it first. It will degrade evenly, and not be affected by the image displayed on the LCD.

Yeah, I only like my colors 100% "pure"

[jpapon]

the light-emitting semiconductor nanocrystals that shine pure colors

What the hell is a pure color? Something that matches the frequency response of our cones? Fully saturated colors?

Re:Yeah, I only like my colors 100% "pure"

[Anonymous Coward]

A pure color is light with a narrow spectral bandwidth. It doesn't matter which color, just that there is ONLY that color.

Re:Yeah, I only like my colors 100% "pure"

[jpapon]

What good is light with a narrow spectral bandwidth?? The point of a TV is to make images life-like. Light sources in real life have wide bandwidth, and objects generally reflect relatively large swaths of frequency. It would be a nightmare to produce images using lots of pixels with 1 nm bandwidth... it's much better to just choose 3 or 4 primaries and mix them... but mixing works just fine with wide bandwidth primaries.

Re:Yeah, I only like my colors 100% "pure"

[jpapon]

Color != Frequency. They always put that color chart under frequency, but it's rather misleading. Colors are the response of our three types of cones to a particular spectrum of light.

Re:Yeah, I only like my colors 100% "pure"

[jpapon]

Oh, and even if what you were saying was true, it wouldn't really change the resolution at all. That's not how sampling works. If your display is 1024*768, you have that many pixels. Making it so each pixel can show any color wouldn't really increase the resolution. Your ability to resolve spatial changes in color is lower than in intensity. So adding "color spatial resolution" is not equivalent to adding "intensity spatial resolution" - this is why many encoding schemes use more bits for intensity than color information - it's more efficient.

Re:Yeah, I only like my colors 100% "pure"

[Dunbal]

the actual frequency of the color.

I guess you think the whole bit about primary colors is just made up stuff huh? So when you split up white light with a prism you get an almost infinite range of colors only 1nm apart instead of 7 very definite colors (Red Orange Yellow Green Blue Indigo and Violet).

Re:Yeah, I only like my colors 100% "pure"

[zzyzyx]

What's the frequency of magenta ?

Re:Yeah, I only like my colors 100% "pure"

[Anonymous Coward]

If you're going to call somebody out for being wrong, you might want to actually do some research. Those 1024x768 pixels are made up of basically triple that in terms of red, green and blue sites that emit the actual light. If you replace those with ones that can handle the entire gamut you would need a third of them and you lose the overhead from having to have individual shutters on each one.

Re:Yeah, I only like my colors 100% "pure"

[Kjella]

Well, yes and no the chart is technically not wrong if you have a single frequency light source like a laser. The trouble is that most real world objects emit a spectrum of light. This chart shows the cone response relative to frequency so the cone's response is an integral over the spectrum*sensitivity. The problem is that in all commonly current display technologies (CRT, LCD, LED, OLED, 3-chip DLP) you only have a fixed number of frequencies to work with. For example say you have red (600nm), green (540nm) and blue (440nm). Well, it turns out you can't actually produce all combinations with just three wavelengths as real world objects do with infinite wavelengths.

The reason for this if you look at the response chart is that the curves overlap, you can't simply decompose them into three components you can set individually. Any wavelength you send to stimulate the M cones also stimulate the S or L cones. And our vision is particularly good at picking up on those differences, it's a two-stage process like illustrated here. Even if the mix in the SML cones is mostly right the Cg and Cb cells are extremely good at picking up on differences in the relative mix. Ideally you'd like more wavelengths or white light + a color wheel like used in single chip DLP, but it's not that easy and you need a signal with the extended information like xvYCC.

Re:Yeah, I only like my colors 100% "pure"

[diamondmagic]

Fully saturated, i.e. a single wavelength, as a laser produces.

The color filters of LCD panels let through a narrow but not-single-wavelength bandwidth of color. This restricts the color gamut you can reproduce. As explained in TFA.

Re:Yeah, I only like my colors 100% "pure"

[Anonymous Coward]

What good is light with a narrow spectral bandwidth?? The point of a TV is to make images life-like. Light sources in real life have wide bandwidth, and objects generally reflect relatively large swaths of frequency. It would be a nightmare to produce images using lots of pixels with 1 nm bandwidth... it's much better to just choose 3 or 4 primaries and mix them... but mixing works just fine with wide bandwidth primaries.

Good god, the ignorance up in this thread.

One word: Gamut.

Mixing works much better with tight primaries. sRGB cannot correctly depict the selective yellow headlights of an old French car, the ubiquitous green LEDs of early '90s electronics, the GaN blue LEDs Shuji Nakamura cursed us with since, nor an LPS streetlight. Not what I'd call "just fine".

Re:Yeah, I only like my colors 100% "pure"

[wonkey_monkey]

Which would then give you enough space to triple the resolution (which may be what the GGP was driving at) or, if not, would increase sharpness in any case. As demonstrated by the effectiveness of Cleartype, RGB subsampling does have an impact on perceived resolution, and RGB subpixel techniques can be applied to colour images as well as text.

Re:Yeah, I only like my colors 100% "pure"

[thegarbz]

Single frequency peak. That is a pure colour. When you look at a typical incandescent light it is a broadband signal spread across the visible range and well into infrared (hence their inefficient at lighting a room despite being very efficient way of converting electrical energy into photons). For an LCD displaying pure red the peaks actually look rather fat around the red with minor peaks in the green and blue range as well as the backlight bleeding through the display. These imperfections is what makes the primary colours not pure and is also the reason the LCDs can't display black.

A fully saturated colour on your display is not a fully saturated colour. For comparison take a green LED and try and generate the same colour visually in photoshop. It can't be done on an LCD display. Really high end displays do come close in the red and blue areas though.

Pure colours are those emitted by single peak sources. Lasers and diodes are are good example as the energy of the photons emitted is related to the bandgap in the semiconductor and thus is quite well controlled and of a single frequency.

To visualise this on a graphic take a look at this: https://upload.wikimedia.org/wikipedia/commons/thumb/6/60/Cie_Chart_with_sRGB_gamut_by_spigget.png/537px-Cie_Chart_with_sRGB_gamut_by_spigget.png The CIE diagram displays the visual range of colours the human eye can perceive. It's stretched to represent our enhanced sensitivity to greens. Points along the outside edge of the diagram represent pure single frequency colours. The point with the temperatures in Kelvin represent black body sources like the sun which are broadband. Finally the sRGB triangle is formed by the three primary colours which match 99.9% of the LCDs on the market. As you can see our standard displays cover less than half of the visible range of colours.

I measured an OLED display with a spectrometer once. The three peaks were right at the edges of the horseshoe at 460nm, 530nm, and 620nm. Not perfect coverage for the human eye but still amazingly better than what most monitors can do.

Re:Yeah, I only like my colors 100% "pure"

[Anonymous Coward]

Are you sure about this? While you can't create an arbitrary response with just a single frequency, different linear combinations of three single frequencies should be enough to create all possible responses. This is basic linear algebra, and it is equivalent to saying that you only need three linearly independent vectors to span a 3-dimensional space. The lum, Cg, Cb cells only process the *output* of the tricromatic cells, and so do not affect this picture. Of course, if you had tetracromatic vision, one would need four different colors to generate every response, and a normal color screen would be insufficient.

To be a bit more detailed, imagine some arbitrary spectrum a(f). This will create responses S_i = int a(f)*r_i(f) df, where i =1,2,3 indicates the three kinds of cone cells, and r_i(f) is the response of cone cell #i. On the other hand, if we have three fixed frequencies f_i with amplitudes b_i, the response to these will be S'_i = sum_j r_i(f_j)*b_j, where j = 1,2,3 indicates the three fixed frequencies. The latter can be written in matrix form as S' = R B. To achive the same response, we merely demand that S' = S => B = R^{-1} S. R^{-1} exists if the three frequencies were chosen so that they do not have the same response, which is trivial to fulfill. So no matter which response we want to create, we can always find a set of 3 amplitudes which create that response.

Re:Yeah, I only like my colors 100% "pure"

[Kjella]

Are you sure about this? While you can't create an arbitrary response with just a single frequency, different linear combinations of three single frequencies should be enough to create all possible responses. This is basic linear algebra, and it is equivalent to saying that you only need three linearly independent vectors to span a 3-dimensional space.

That is true but your analysis is wrong, that is not the mathematical equivalent because we can only send light with positive intensity. Say you had f1 = [1,0,0], f2 = [0,1,1/2] and f3 = [0,0,1]. Sure with a linear combination of vectors you can express [0,1,0] but only using negative intensity which is impossible.

Re:Yeah, I only like my colors 100% "pure"

[thegarbz]

I don't think that's quite right. If you had three sources of perfectly pure colours positioned at the very peaks of the responses of our rods and cones we could approximate very close to every colour based on the combination of the the three. Our eyes effectively only pick up 3 colours like a camera, monochrome with a response based on the curves above.

As such our interpretation of a pure cyan at say 500nm can be made up of appropriate peaks at 440nm, 530nm, and 590nm as the eye will simply integrate the result of what it sees to come to what we perceive as colour.

In that regard the only colours we really have difficulty approximating are the colours where the spectral response of our rods and cones no longer overlap and are away from the peaks we define. i.e. we can't create violet on a screen as there's no combination of 440, 530, and 590 that could be integrated to come up with the peak at 400nm.

This is best visualised with a CIE diagram. https://en.wikipedia.org/wiki/File:Cie_Chart_with_sRGB_gamut_by_spigget.png As you could see drawing a triangle with points at the very edge of the horseshoe diagram of our visual range would encompass nearly all colours.

I'm wondering if our future is not in RGB displays but rather Red Yellow Green Cyan Blue Violet displays.

Re:Quantum dots?

[Scytheford]

Actually you're dead wrong. Quantum dots are A Thing. Here's how to make them in a basic lab: http://www.youtube.com/watch?v=bNuoYm7Su4o

50% more colours?

[AAWood]

Soooo, any idea what they mean by "50% more colours"? Do these allow the screen to display a wider set of the visible spectrum than LCD screens? Do they allow the same set but at a higher bitrate? Do they simply display the desired colour more precisely? Is this "extra" in the range that consumer GPUs and OSes can display?

Re:50% more colours?

[SuricouRaven]

The whole field of computing is built on three-primary color specification anyway. Either RGB, or HSV, or YUV, or some varient of them. Or CMYK, in which the K is really a fudge-factor used to account for real inks not behaving like mathematically ideal inks. So even if someone built a display of a wider gamut, good luck finding any content to use it. I suspect this is just marketing being allowed to write the press report.

Re:50% more colours?

[Dunbal]

They mean "we need funding and investors, and 50% more sounds great without actually saying anything".

Re:50% more colours?

[Jesus_666]

But if they refer to your average TN display that often can't even display TrueColor without dithering, 50% more colors may essentially mean "it's not the trash your laptop probably uses".

Re:50% more colours?

[Dcnjoe60]

It doesn't matter because 24 bit color produces more color variants than the human brain can actually distinguish. Having more color won't look any different.

Re:50% more colours?

[Dcnjoe60]

The whole field of computing is built on three-primary color specification anyway. Either RGB, or HSV, or YUV, or some varient of them. Or CMYK, in which the K is really a fudge-factor used to account for real inks not behaving like mathematically ideal inks. So even if someone built a display of a wider gamut, good luck finding any content to use it. I suspect this is just marketing being allowed to write the press report.

RGB has nothing to do with computing, but everything to do with the physics of light. Printing uses CMYK also because of the physics of light. The difference is RGB is when light is emmitted and CMYK when it is reflected. That is why blue and yellow paint make green, but blue and yellow light make magenta. With light, mixing colors is additive, with painting/printing, it is subtractive.

Re:50% more colours?

[SuricouRaven]

RGB were the chosen wavelengths only because by mixing them in appropriate ratios it is possible to reproduce the perception of most colors to human vision. If there were any non-humans animals smart enough to judge, they'd tell you that all the colors on television look wrong. Humans see subjective colors, not spectrographs. To represent a color with precision would require storing the entire spectrum, which is impractical.

The mathematics of CMYK say that f you have full use of C, M and Y all absorbing you should get nothing reflecting at all - pure black. But real inks don't work quite like that, they reflect a little light even in regions of the spectrum the user would rather they didn't, so what you'd actually get is a murky blackish-brown. That's why the extra K: The additional black allows for the imperfect nature of ink to be corrected for.

Re:50% more colours?

[MagusSlurpy]

Soooo, any idea what they mean by "50% more colours"?

It means they let someone with a marketing degree write a blurb about technology.

Re:50% more colours?

[Prune]

RGB is a small subset of the visible color gamut. It's the triangle in this graph: http://www.antigrain.com/doc/introduction/cie_1931.jpg

Re:50% more colours?

[Prune]

RGB is not even close to most visible colors. The overall graph here shows the full visible color gamut, and RGB is the small triangle inside it: http://www.antigrain.com/doc/introduction/cie_1931.jpg

Re:50% more colours?

[QQBoss]

I have a buddy who used to teach ophthalmic surgery at Georgetown U. and did research in this area. He also did computer animation as a hobby (one that actually made him good money, to where I think his teaching later became the hobby). Wish I could locate one the papers, but most of his work was done pre-WWW and probably has never been put up. His info showed that most people can resolve 8 bits of red, 9 bits of green, and 8 bits of blue. That extra bit sucks from a memory usage point of view, though, and compressing it out only affects side-by-side comparisons for a huge portion of humanity, and even then not all that meaningfully.

But 24 bits is definitely not more than what the average human brain can distinguish.

The problem with quantum dots...

[outsider007]

You won't know how many pixels are dead until you open the box.

Re:The problem with quantum dots...

[Anonymous Coward]

You won't know how many pixels are dead until you open the box.

I thought all LCD's were like that.

Schrödinger's cat quantum thought experiment, http://en.wikipedia.org/wiki/Schr%C3%B6dinger's_cat

Re:The problem with quantum dots...

[I_am_Jack]

Which raises the question that perhaps the quantum dot monitor will display the correct color only when you're not looking at it.

Re:Quantum dots?

[Scytheford]

Because the energy levels of the electrons are at quantum levels. They transition between these levels and emit light. This is an absolutely correct usage of the word "quantum". You are a foolish troll.

Re:Quantum dots?

[Anonymous Coward]

Perhaps because they're semiconductor particles whose electronic properties are size-determined due to quantum confinement, rather than bulk material properties.

Re:Quantum dots?

[Dunbal]

Because nano is so 2009..

Sweet little lies...

[Anonymous Coward]

" beautiful displays that would be inexpensive and easy to manufacture."

But expensive to buy for sure. And will only be slightly cheaper when the next superior tech is at the door. Rinse and repeat...

Re:Sweet little lies...

[Jeremi]

But expensive to buy for sure. And will only be slightly cheaper when the next superior tech is at the door. Rinse and repeat...

Well, yes, that's how capitalism works. Someone invents something useful, and then they try to maximize the profit from their labor by selling it for as much as the market will bear. Eventually the price comes down due to competition. You can either pay top dollar for the new hotness now, or wait a while for the price to come down, your choice.

It's a feature. Note that you can buy a $99 LCD display at Walmart today that performs better in all respects than the $9,000 LCD display of the same size you could buy in 1995.

Another problem

[a_claudiu]

They are working fine until you look at the TV.

Re:Quantum dots?

[Anonymous Coward]

So... what energy levels of electrons aren't at quantum levels?

They transition between these levels and emit light.

How is this different than anything that absorbs and releases photons as electrons move up and down these levels. In other words how is this differentiated between everything else that has atoms and electrons. Don't the electrons in neon raise quantum levels when neon is excited. So do I have quantum beer sign? Still seems to be a buzzword.

Re: microdots

[tfigment]

This technology is nothing new. Its been used heavily since the sixties to bring out vivid colors in all manner of displays (its actually even older than traditional color tv displays). Sometimes they refer to the technology as microdots. I'm not sure I need a LSD screen yet or one that uses PCB bus instead of a PCI bus one.

Re:Quantum dots?

[Anonymous Coward]

The term is related to Quantum Well and Quantum Wire. A quantum well is a system where particles (electrons) are confined to move in 2D by two very large potential barriers on either side of the well. It's generally one of the first systems studied in quantum mechanics. Quantum wires are like quantum wells except the potential barriers also exist in a second dimension, so that the particle is confined to move in 1D along the "wire". A quantum dot is a small box which is confined by potential barriers in all directions so that the electron can only exist within the extremely small dot.

Obviously quantum dots are going to be around the nm range so that they can actually confine the particles in any meaningful sense, but the point is the effects that QM predicts for that particular configuration. The size and shape of the dot allows us to precisely tweak the energy levels and wavefunction symmetries involved, something fairly particular to the "nano 3D potential barrier" system.

Wider Gamut, not usually an advantage for TV.

[guidryp]

LCD TVs already easily match Rec. 709 color primaries (similar to sRGB used in standard color destkop monitors).

Since TV signals and Blu Rays are all using this standard, using a non standard wider gamut emitter, just gets you unnatural colors.

If you like artifical, oversaturated hues, great, but if you want natural looking color this does nothing for you.

IIRC, LGs new 55" OLED TV will be corresponding to Rec. 709 color primaries, not the outlandish Neon of OLED smartphones.

For a TV, what you want is properly calibrated Rec. 709 color, not, nonsense about 50% more colors.

Re:Wider Gamut, not usually an advantage for TV.

[Prune]

Uh, Rec. 709 is a small portion of the visible color gamut. It's represented by the triangle in this graph: http://upload.wikimedia.org/wikipedia/commons/8/8f/CIExy1931_sRGB.svg Note the area it covers of the overall visible gamut is maybe 50%.

Re:Wider Gamut, not usually an advantage for TV.

[guidryp]

Uh, Rec. 709 is a small portion of the visible color gamut. .

Uh, So?

Standards exist for a reason. Just about all available Media is produced for Rec. 709/sRGB.

Showing it with wider color primaries will not make it look more real, it will make it look more unnatural.

Wide gamut PC monitors were all the rage 3 or 4 years ago, until people started realizing it made it nearly impossible to get neutral color and the tide turned back to sRGB screens.

Gamut isn't simply a case of more == better. In the vast majority of cases, more == worse.

Re:Wider Gamut, not usually an advantage for TV.

[Prune]

Gamut is about being able to represent all colors that are perceptible in the real world. You're making an argument that because of immature technology, we should handicap our displays. It's the dumbest thing I've read in a long time on this site. The difficulties with neutral color are based on 1) poor calibration, and, more importantly, 2) insufficient quantization -- due to the extended gamut you need around 10 to 12-bit quantization _per channel_ to have sufficient precision. Considering most LCDs and OLEDs can't even do 8-bit and have to dither, the technology has a ways to go. Your argument basically reduces to this: we must use lower gamut since all implementations are shitty, whereas it should be--we should fix the implementation. You might as well say we shouldn't build faster airplanes because they'll fall apart, rather than that we should refine the technology in question.

Re:Wider Gamut, not usually an advantage for TV.

[guidryp]

No I am simply arguing for standards, instead of ill defined "wider" color gamut, that is simply, bigger number is better nonsense.

Unless you have media AND players, AND displays all working in lockstep, you get worse results, not better.

You need standards body to meet, create a new wider gamut standard and build new product at all stages to meet it.

Going it alone is pointless spec whoring.

Why do we need this?

[Dcnjoe60]

Why do we need this? The power savings is a plus, but the human brain can only "see" and distinquish an estimated 10 million colors ( http://hypertextbook.com/facts/2006/JenniferLeong.shtml ) and current display technologiy produces 16.7M colors (24-bit True Color). Having a display show 24M colors (50% increase) won't look any different since current technology already exceeds our ability to percieve the differences.

Re:Why do we need this?

[Nyder]

Why do we need this? The power savings is a plus, but the human brain can only "see" and distinquish an estimated 10 million colors ( http://hypertextbook.com/facts/2006/JenniferLeong.shtml ) and current display technologiy produces 16.7M colors (24-bit True Color). Having a display show 24M colors (50% increase) won't look any different since current technology already exceeds our ability to percieve the differences.

You answered your own question. It's worth it for the Power Savings, IMO, the fact that it shows colors possibly better then we can see them is just the bonus.

Re:Why do we need this?

[foniksonik]

Apparently from all the other posts, the 16.7M colors we can get now do not overlap 100% with the 10M colors we can see. I believe this is called the Gamut range of colors being produced vs the Gamut we can see.

Supposedly these light emitters can create a Gamut of light frequencies (colors) that overlaps more, thus can produce more colors (that we can see).

Re:Why do we need this?

[gstrickler]

Because the gamut of 24-bit RGB doesn't cover the entire range of visible colors and intensities. While we can only distinguish ~ 8M colors, we can distinguish a huge range of intensities. 24-bit displays cover 16M colors AND intensities, so in this case, 16M is not > 8M because they're counting different things.

While current displays are adequate for most purposes, they do not display all of the colors we can see, nor all the intensities we can see. Typical displays only cover 45%-75% of the AdobeRGB (1998) color-space, which itself is a subset of the visible gamut. Some (more expensive) displays cover a greater percentage of the visible range, but none cover the entire range.

Re:Why do we need this?

[Prune]

It doesn't even overlap 50%. Full RGB color space is the triangle in this graph; note the area it covers of all visible colors: http://upload.wikimedia.org/wikipedia/commons/8/8f/CIExy1931_sRGB.svg

Re:Why do we need this?

[Prune]

This is one of the dumbest comments I've read on slashdot. You're confusing quantization with extent. The article is very obviously talking about covering a larger part of the visible color gamut. RGB is represented by the triangle in this graph: http://upload.wikimedia.org/wikipedia/commons/8/8f/CIExy1931_sRGB.svg You'll note it doesn't even cover 50% of visible colors. Most TVs and displays can't even reproduce the full RGB space. The 24-bit/16.7M merely refers to the number of colors and affects how smooth gradients are, and has nothing to do with the range of colors that can be reproduced.
For fuck's sake, I didn't expect this level of stupidity from someone with a sub-1M user ID!

Re:Why do we need this?

[Dcnjoe60]

This is one of the dumbest comments I've read on slashdot. You're confusing quantization with extent. The article is very obviously talking about covering a larger part of the visible color gamut. RGB is represented by the triangle in this graph: http://upload.wikimedia.org/wikipedia/commons/8/8f/CIExy1931_sRGB.svg You'll note it doesn't even cover 50% of visible colors. Most TVs and displays can't even reproduce the full RGB space. The 24-bit/16.7M merely refers to the number of colors and affects how smooth gradients are, and has nothing to do with the range of colors that can be reproduced.

For fuck's sake, I didn't expect this level of stupidity from someone with a sub-1M user ID!

Has nothing to do with how much the TV or screen can reproduce. It has everything to do with how well the brain can discriminate the various wavelengths. So while it is theoretically true that the technique may produce more colors, whatever that means exactly, if the human brain cannot discriminate between them, what good does it do?

This is not an issue of physics, but of biology, but then maybe I'm just to much of a dumb fuck to know what I'm talking about.

Re:Why do we need this?

[Dcnjoe60]

Because the gamut of 24-bit RGB doesn't cover the entire range of visible colors and intensities. While we can only distinguish ~ 8M colors, we can distinguish a huge range of intensities. 24-bit displays cover 16M colors AND intensities, so in this case, 16M is not > 8M because they're counting different things.

While current displays are adequate for most purposes, they do not display all of the colors we can see, nor all the intensities we can see. Typical displays only cover 45%-75% of the AdobeRGB (1998) color-space, which itself is a subset of the visible gamut. Some (more expensive) displays cover a greater percentage of the visible range, but none cover the entire range.

As stated in another post, the color problem you are referencing is one of physics -- producing the various wavelengths. What we see, however, is one of biology and the human brain cannot differentiate between similar wavelengths. Therefore, including all of them does not mean that we will see the image any better. Intensity is an issue, but the summary is talking about color, not intensity, although they are related.

The limiting factor in all of this is not going to be the production of the visible wavelengths. It is going to be the limitations of the human brain.

Re:Why do we need this?

[gstrickler]

Yes, the human eye and the brain are going to be the limits. And given the range of intensities (e.g. contrast) one can see at any given time, and the ability to discern continuous color gradients, it appears that we'll need somewhere between 24 and 36 bits driving displays with contrast ~5000:1, using at least 3 color narrow band color sources centered on the frequencies to which the eye cones respond, and capable of delivering more than 1000 lux (the brightness of an overcast day) at the viewer's position to have displays that can adequately reproduce the color range we can see. The intensity range we can see is much larger than that, but we can't see all of it at once, so as long as the contrast is sufficient, we don't need full sunlight type brightness.

Re:Why do we need this?

[Prune]

It _is_ an issue of biology. And that's exactly what the larger encompassing graph represents: the perceptual color space for humans. Humans can see all colors in that; RGB can only represent the colors in the triangle, and most monitors are a subset of the triangle. This has nothing to do with physics so I'm not sure why you brought physics into the discussion. Next time I recommend counting to 10 before letting an itchy Submit-clicking finger take action. It gives you time to save later embarrassment.

Re:Quantum dots?

[Pinky's Brain]

In a gas discharge lamp the effect is actually elemental, not dependent on quantum confinement within a particle. You have an elemental discharge sign.

MOAR Pixels!

[Culture20]

I don't care about colors or power savings. Get me better DPI or just more pixels overall.

Re:MOAR Pixels!

[Prune]

Troll much? RGB doesn't even cover 50% of colors visible to the human eye. It's represented by the triangle here: http://upload.wikimedia.org/wikipedia/commons/8/8f/CIExy1931_sRGB.svg The larger superset is the full CIE XYZ color space visible to the human eye.

Re:MOAR Pixels!

[Culture20]

That's nice. I personally have no need for more colors in the currently limited screen space. I need to have a bigger view, even if it drops to 16bit color.

Re:MOAR Pixels!

[Prune]

It's not the number of colors but the color gamut. You seem to lack reading comprehension. The issue is not quantization (bit depth) but the saturation that can be achieved. One is completely unrelated to the other.

Re:MOAR Pixels!

[Culture20]

It's not the number of colors but the color gamut. You seem to lack reading comprehension. The issue is not quantization (bit depth) but the saturation that can be achieved. One is completely unrelated to the other.

And you seem to lack comprehension of simple concepts. I said I want pixels. Not color. Hell, give me monochrome, but give me 19200x12000.