6 ms·

Near the end of the article: "Sharp Quattron’s fourth primary color is yellow, and there is nothing for it to do because yellow is already reproduced with mixt

by teilo 16y ago

Near the end of the article: "Sharp Quattron’s fourth primary color is yellow, and there is nothing for it to do because yellow is already reproduced with mixtures of the red and green primaries, he said." This 20-year display "expert" has never heard of hi-fidelity color? Inexcusable. It is accurate to say that yellow is not a primary color of light, so it is true enough that no one should refer to the yellow pixels as a "fourth" primary color. However, it is completely false to say that "there is nothing for it [the yellow pixels] to do". The yellow pixels are used to extend the gamut of the device to cover portions of the visible spectrum that the primary LEDs, in combination, cannot. We do this all the time with Hi-Fi printers. We add RGB or Orange, Green, and Violet to the traditional CMYK primaries to expand the gamut of the device far beyond what traditional process color can accomplish.

evgen 16y ago
> The yellow pixels are used to extend the gamut of the device to cover portions of the visible spectrum that the primary LEDs, in combination, cannot. What color in the visible spectrum cannot be expressed as a combination of red, green, or blue? Octarine perhaps? Adding additional colors only helps if your pixels are large and putting a red and a blue dot next to each other when aiming for purple ends up just looking like two dots of different colors instead of a combined dot.
- teilo 16y ago
  Your sarcasm does not become you. Neither does your ignorance. If you actually worked in the field of color management, like I do, you would know that, as with subtractive pigments, so with LEDs: There is no such thing as a pure primary in the real world. It is impossible to create an LED or a color of ink that is mathematically perfect pure Red, Green, or Blue. Because of this it is impossible for an RGB-only display to display all colors that are visible to the human eye. (Let's exclude UV-florescence from the equation for the sake of simplicity). RGB displays use color correction curves and various profiling tricks to correct for the difference between the pure primaries and the actual light being output by the LEDs. Yellow LEDs can extend the color range of an RGB display into areas of the gamut that the display could not otherwise "hit". The same could be said for (presumably theoretical) Cyan or Magenta LEDs.
  - ugh 16y ago
    How does the technology used to capture the video factor into that? I would assume that with capturing only RGB you already limited the color space – the information is lost – and you cannot extend it, no matter what you try. I know that, when printing, additional colors can help because the RGB space doesn’t map exactly to the CMYK space. There are RGB colors you just cannot get with CMYK. Are the RGB(capture) and RGB(display) color spaces so different that additional colors can help? I would assume they would have to be, wouldn’t they?
    - teilo 16y ago
      The capture technology actually has little to do with it. First of all, yes, only so much can be captured. But no one broadcasts, in raw form, exactly what they capture. They filter and transform it in all sorts of ways, adding, replacing, and shifting color. Also, consider how much content is digitally generated, where any arbitrary color can be specified, by the numbers. There is no capture device. HDMI 1.4 specifies color spaces that, at present, no display can fully reproduce. As time goes on, there will be more hi-fi content that can target such displays when they are invented. Also, it is common to print in the Adobe98RGB colorspace. Hi-fi printers can do it. There is a lot of content in Adobe98. There are, however, very few displays that can display it.
    - dedward 16y ago
      You cannot get back information - no, although you can artificially guess and make something that looks better in some situations. But just because you recorded in one RGB space doesn't mean the particular display you are using can reproduce that exact RGB space - take any dozen RGB flat-panel displays and see if their color gamuts match up exactly. They probably don't - so while I'm ignorant of whatever feature the yellow is trying to achieve here - whether it's just fluff or actually filling out a deficiency in the output of their LCD gamut - either way, their display with the added yellow pixels can produce colours that their previous displays could not.
      - ugh 16y ago
        I have no doubts about its ability to display a wider range of colors. It’s just that after attending a few lectures about data transmission I rather got the feeling that those responsible for designing the compression and coding would rather quit and sail the South Pacific than transmit stuff that’s not actually used. That might just have been a by-product of all the compression and coding I learned about being really old (PAL, NTSC), newer stuff might be a lot more forward looking (as teilo suggested).
- chadaustin 16y ago
  I recommend spending an evening reading about color theory and human perception and CIE 1931 on Wikipedia. http://en.wikipedia.org/wiki/CIE_1931_color_space http://en.wikipedia.org/wiki/CIE_1931_color_space In short, the "cowlick" is the range of human color perception, and no triangle made of linear-combinations of three visible primaries can cover the entire cowlick.
  - FlorinAndrei 16y ago
    Yup. You need a nearly-infinite number of primaries to cover exactly 100% of the human eye's gamut.
- OpieCunningham 16y ago
  I'm not familiar with the details of Samsung's technology. And certainly it is true that display manufacturers create wholesale fabrications of supposed technical improvements in order to sell displays. But theoretically, adding additional dedicated colors to a display would increase the gamut (the field of colors a display can present, as represented by a gamut chart: http://www.pcmonitors.org/wp-content/uploads/2009/12/colour-gamut.jpg http://www.pcmonitors.org/wp-content/uploads/2009/12/colour-... ). In a perfect world, a red, green and blue color combination would be capable of presenting every possible color. This not being a perfect world, the presentation of color is limited, sometimes dramatically across displays and display manufacturers (if this were not the case, all displays would be perfectly equal, excepting white and black levels). Whether Samsung has implemented their yellow in a way that increases gamut is a separate question I cannot answer.
Robin_Message 16y ago
Wasn't the point that yellow was a poor choice of colour to add since there is not much missing from the gamut between the red and green primaries? I'm no expert, but looking at the diagram I'd say a cyan colour would be far more gamut expanding.
- dedward 16y ago
  Your eyes are sensitive to those areas in different amounts - while it looks like there is a lot more space to expand towards cyan - you probably wouldn't notice it as much as the extra R+G pop you're going to get with the added yellow.
- teilo 16y ago
  The edge of the blob on that chart represents the color that can be achieved by mixing the three chosen primaries. If you wanted to expand the curve outward along the top edge of the blob, you need yellow pixels with a more intense color than can be produced by mixing red and green. What this will achieve is not just brighter yellows. The whole curve expands upward, leading to more intense colors along that entire portion of the gamut.
  - Robin_Message 16y ago
    No, the triangle is what can be achieved by mixing the primaries. The blob is what the eye can distinguish between. I think the blob is the http://en.wikipedia.org/wiki/CIELUV_color_space http://en.wikipedia.org/wiki/CIELUV_color_space but I'm not certain. Can I suggest we look at http://en.wikipedia.org/wiki/CIE_1931_color_space#Definition_of_the_CIE_XYZ_color_space http://en.wikipedia.org/wiki/CIE_1931_color_space#Definition... instead? There you can see that three primaries cover a large range of human colour perception. In particular yellow is well covered, but there is a big gap around cyan, hence my earlier comment that a cyan primary would seem more useful. This corresponds to the requirement for a negative red colour to reach between 440nm and 540nm. Note red is the only primary that goes significantly negative. Also since we expect television to be displayed using three RGBish primaries, the triangle is also the only colours that are actually transmitted in any TV signal or recording. Now, it is possible that the small gain in yellow is noticeable, and leads to a nice visual "pop" when viewing the screen, but it really isn't gaining you a lot. I'd be more interested in a cyan primary (think of the skies!) or an HDR screen with a real range of brightnesses.

by teilo 16y ago

evgen 16y ago

> The yellow pixels are used to extend the gamut of the device to cover portions of the visible spectrum that the primary LEDs, in combination, cannot. What color in the visible spectrum cannot be expressed as a combination of red, green, or blue? Octarine perhaps? Adding additional colors only helps if your pixels are large and putting a red and a blue dot next to each other when aiming for purple ends up just looking like two dots of different colors instead of a combined dot.

teilo 16y ago
Your sarcasm does not become you. Neither does your ignorance. If you actually worked in the field of color management, like I do, you would know that, as with subtractive pigments, so with LEDs: There is no such thing as a pure primary in the real world. It is impossible to create an LED or a color of ink that is mathematically perfect pure Red, Green, or Blue. Because of this it is impossible for an RGB-only display to display all colors that are visible to the human eye. (Let's exclude UV-florescence from the equation for the sake of simplicity). RGB displays use color correction curves and various profiling tricks to correct for the difference between the pure primaries and the actual light being output by the LEDs. Yellow LEDs can extend the color range of an RGB display into areas of the gamut that the display could not otherwise "hit". The same could be said for (presumably theoretical) Cyan or Magenta LEDs.
- ugh 16y ago
  How does the technology used to capture the video factor into that? I would assume that with capturing only RGB you already limited the color space – the information is lost – and you cannot extend it, no matter what you try. I know that, when printing, additional colors can help because the RGB space doesn’t map exactly to the CMYK space. There are RGB colors you just cannot get with CMYK. Are the RGB(capture) and RGB(display) color spaces so different that additional colors can help? I would assume they would have to be, wouldn’t they?
  - teilo 16y ago
    The capture technology actually has little to do with it. First of all, yes, only so much can be captured. But no one broadcasts, in raw form, exactly what they capture. They filter and transform it in all sorts of ways, adding, replacing, and shifting color. Also, consider how much content is digitally generated, where any arbitrary color can be specified, by the numbers. There is no capture device. HDMI 1.4 specifies color spaces that, at present, no display can fully reproduce. As time goes on, there will be more hi-fi content that can target such displays when they are invented. Also, it is common to print in the Adobe98RGB colorspace. Hi-fi printers can do it. There is a lot of content in Adobe98. There are, however, very few displays that can display it.
  - dedward 16y ago
    You cannot get back information - no, although you can artificially guess and make something that looks better in some situations. But just because you recorded in one RGB space doesn't mean the particular display you are using can reproduce that exact RGB space - take any dozen RGB flat-panel displays and see if their color gamuts match up exactly. They probably don't - so while I'm ignorant of whatever feature the yellow is trying to achieve here - whether it's just fluff or actually filling out a deficiency in the output of their LCD gamut - either way, their display with the added yellow pixels can produce colours that their previous displays could not.
    - ugh 16y ago
      I have no doubts about its ability to display a wider range of colors. It’s just that after attending a few lectures about data transmission I rather got the feeling that those responsible for designing the compression and coding would rather quit and sail the South Pacific than transmit stuff that’s not actually used. That might just have been a by-product of all the compression and coding I learned about being really old (PAL, NTSC), newer stuff might be a lot more forward looking (as teilo suggested).
chadaustin 16y ago
I recommend spending an evening reading about color theory and human perception and CIE 1931 on Wikipedia. http://en.wikipedia.org/wiki/CIE_1931_color_space http://en.wikipedia.org/wiki/CIE_1931_color_space In short, the "cowlick" is the range of human color perception, and no triangle made of linear-combinations of three visible primaries can cover the entire cowlick.
- FlorinAndrei 16y ago
  Yup. You need a nearly-infinite number of primaries to cover exactly 100% of the human eye's gamut.
OpieCunningham 16y ago
I'm not familiar with the details of Samsung's technology. And certainly it is true that display manufacturers create wholesale fabrications of supposed technical improvements in order to sell displays. But theoretically, adding additional dedicated colors to a display would increase the gamut (the field of colors a display can present, as represented by a gamut chart: http://www.pcmonitors.org/wp-content/uploads/2009/12/colour-gamut.jpg http://www.pcmonitors.org/wp-content/uploads/2009/12/colour-... ). In a perfect world, a red, green and blue color combination would be capable of presenting every possible color. This not being a perfect world, the presentation of color is limited, sometimes dramatically across displays and display manufacturers (if this were not the case, all displays would be perfectly equal, excepting white and black levels). Whether Samsung has implemented their yellow in a way that increases gamut is a separate question I cannot answer.

Robin_Message 16y ago

Wasn't the point that yellow was a poor choice of colour to add since there is not much missing from the gamut between the red and green primaries? I'm no expert, but looking at the diagram I'd say a cyan colour would be far more gamut expanding.

dedward 16y ago
Your eyes are sensitive to those areas in different amounts - while it looks like there is a lot more space to expand towards cyan - you probably wouldn't notice it as much as the extra R+G pop you're going to get with the added yellow.
teilo 16y ago
The edge of the blob on that chart represents the color that can be achieved by mixing the three chosen primaries. If you wanted to expand the curve outward along the top edge of the blob, you need yellow pixels with a more intense color than can be produced by mixing red and green. What this will achieve is not just brighter yellows. The whole curve expands upward, leading to more intense colors along that entire portion of the gamut.
- Robin_Message 16y ago
  No, the triangle is what can be achieved by mixing the primaries. The blob is what the eye can distinguish between. I think the blob is the http://en.wikipedia.org/wiki/CIELUV_color_space http://en.wikipedia.org/wiki/CIELUV_color_space but I'm not certain. Can I suggest we look at http://en.wikipedia.org/wiki/CIE_1931_color_space#Definition_of_the_CIE_XYZ_color_space http://en.wikipedia.org/wiki/CIE_1931_color_space#Definition... instead? There you can see that three primaries cover a large range of human colour perception. In particular yellow is well covered, but there is a big gap around cyan, hence my earlier comment that a cyan primary would seem more useful. This corresponds to the requirement for a negative red colour to reach between 440nm and 540nm. Note red is the only primary that goes significantly negative. Also since we expect television to be displayed using three RGBish primaries, the triangle is also the only colours that are actually transmitted in any TV signal or recording. Now, it is possible that the small gain in yellow is noticeable, and leads to a nice visual "pop" when viewing the screen, but it really isn't gaining you a lot. I'd be more interested in a cyan primary (think of the skies!) or an HDR screen with a real range of brightnesses.