4 ms·

The FF0000 part of the wider gamut is pretty useless per se, but widening the gamut also gives you a lot of nice non-eye-bleeding colors in other areas. RGB gam

The FF0000 part of the wider gamut is pretty useless per se, but widening the gamut also gives you a lot of nice non-eye-bleeding colors in other areas. RGB gamuts are very limited in light red, dark yellow, light and dark teal, etc. colors. The main downside is that as gamuts get expanded via narrower-spectrum primaries, there is an increasing risk of inter-observer variability (I also wonder if they may cause more eye strain, have worse effects on sleep cycles, ...). A better solution than the current trend of wide gamuts made by 3 primaries with very narrow spectra would be displays with 6+ primaries, but that takes a much more dramatic redesign of the last few stages of the graphics pipeline, a more dramatic hardware redesign, and potentially different content to really show off (content that will moreover be really hard to effectively show on older displays). I’m still hoping it happens within the next couple decades, but I wouldn’t hold my breath. If you know the shape of common RGB gamuts, you’ll start to notice how much all of our RGB-displayed visual media is constrained and nudged toward particular choices. There are e.g. color choices made in some paintings that can’t be effectively used in films. There are plenty of physical scenes that can’t be effectively reproduced. And on the flip side, areas of the RGB gamut where there is more available contrast get used a lot, even when they don’t quite fit the mood.

modeless 5y ago
I've never seen a wide gamut demo that was impressive to me. High frame rate yes, high contrast/brightness for HDR yes, those are definitely worthwhile. But wide gamut just doesn't seem to make a big difference to justify the implementation cost. I'd love to see some content that would justify wide gamut for me. Any suggestions?
- bloak 5y ago
  I think it would only be useful if the colours on the screen have to match colours off the screen. If all you're paying attention to is what's on the screen then it probably doesn't matter much exactly what those colours are because it's all relative, your brain compensates for it, and it's not as if the cells in your eye have a consistent response to colour in any case: it all depends on what you've been looking at recently. If you've ever been in a television shop, back when they still had shops, you could find two screens next to each other showing the same television channel but with radically different colours, when you compare them side-by-side, but if you watch just one of the screens - it doesn't matter which - you get used to the colours of that screen and everything looks fine.
- jacobolus 5y ago
  I sometimes make hue a meaningful attribute in data visualizations, but try to keep chroma roughly consistent for various hues, so that varying chroma does not cause visual artifacts unrelated to the data. RGB gamuts are very limited in blue–green colors (when you additively mix purplish blue “B” and yellowish green “G”, it’s hard to make a bright blue–green), and as a result my whole data visualization ends up with relatively lower chroma than I would like. With an additional few primaries (say, adding a greenish blue primary and a bluish green primary), the chroma-width of the gamut would be much more consistent across various hues, which would let these visualizations be something like twice as intense. It would make a huge difference. You probably haven’t seen any “wide gamut” demos with fundamentally different primaries involved; display manufacturers have been expanding their gamuts by making the three RGB primaries have narrower and narrower spectral power distributions, rather than by adding additional primaries. For a while Sharp made a “4-primary” display, but the additional “yellow” primary was more or less created by an additive mixture of existing “R” and “G” spectra, so didn’t really make as much practical difference as one would hope. https://en.wikipedia.org/wiki/Quattron https://en.wikipedia.org/wiki/Quattron You can get a display with like 8 or 10 primaries if you run a color science research lab and can spend tens of thousands of dollars on it (or maybe more? dunno).
  - modeless 5y ago
    I'm sure if I saw a 10 primary display sitting next to an RGB one for reference, I could tell a difference. But if it was just sitting there by itself, in regular non-ideal viewing conditions, I doubt I would think that it is significantly better than normal displays. Whereas for high frame rate and high brightness/contrast displays I can easily see a big improvement under normal non-ideal viewing conditions without needing a reference.
    - jacobolus 5y ago
      Again, it would depend on the content shown. If you watch a 24 frame/second video on a 120 Hz computer display, I’m not sure you’ll notice much difference. If you use an extremely high contrast display to show a low-contrast image, you won’t see any obvious difference. If you use a 4K display to show a 240p video, you won’t see any obvious difference. If you watch a video with monophonic sound on a 7.1 sound system, you won’t hear any obvious difference. Etc. If you look at a photo produced by an RGB digital camera, color graded in some standard RGB space, and then presented on an RGB display vs. a 10-primary display, you aren’t going to see any obvious difference, because the content does not take any special advantage of the medium. In comparison, if you look at a bunch of synthetic images created to highlight the differences / stretch the capabilities of the displays, or e.g. a video produced by a whole hardware–software–author pipeline expecting the more capable display, the differences will be pretty obvious.
mncharity 5y ago
> If you know the shape of common RGB gamuts, you’ll start to notice [...] I'm interested in exploratory questions of the form "If education content about mumble wasn't such a disaster, then... what? What might doing it well look like? How might doing it well then be used to teach other topics better?" Color is widely taught preK-6. And yet first-tier physical-sciences graduate students not infrequently reply to "What color is the Sun?", with something like "it doesn't have a color; it's rainbow color". Suggesting that teaching is very not going well. One might imagine perhaps adapting tablet/phone clip-on spectroscopes down towards K, and developing a spectra-centered content and learning progression for color. So let's hypothesize students with a firm grasp of color. What might they do with it? > [...] you’ll start to notice how much all of our RGB-displayed visual media is constrained and nudged toward particular choices. There are e.g. color choices made in some paintings that can’t be effectively used in films. There are plenty of physical scenes that can’t be effectively reproduced. And on the flip side, areas of the RGB gamut where there is more available contrast get used a lot, even when they don’t quite fit the mood. Any suggestions on resources for exploring these further? Any other favorite stories or examples, useful or fun, where understanding color well gives insight or skill? Any other thoughts? Thanks!
- jacobolus 5y ago
  > "What color is the Sun?" This is a hard question to answer. My answer would be something like “intense pain followed by blindness”. :-) But you could also say that sunlight is “white”, because it excites all of our cone cells and is by far the brightest object in any field of view. You could change the spectral power distribution of the sun quite a bit one way or another and it would remain “white”, because our visual system adapts. > One might imagine perhaps adapting tablet/phone clip-on spectroscopes I’d start by handing out diffraction grating glasses, and looking at various kinds of lamps.
  - mncharity 5y ago
    > This is a hard question to answer. [...] Oh, nice - I'd forgotten how the brief version of the question sends conversations in unexpected directions, and turns up a wide variety of conceptions. Long ago, to narrow the diversity of responses, I switched to asking longer versions, like "A 5-year old asks you: 'I'm told the Sun is a big hot ball! Awesome, I love balls! What color is the ball?' What color is the Sun?" Which has gotten a bit boring - same old failure modes grown familiar. Diverse could be useful again, for thinking about teaching color, and more fun. > “intense pain followed by blindness”. :-) Sooo... one can't tell what color a light is, unless you can look at it directly...? :P > sunlight is “white”, because it excites all of our cone cells So does my red shirt, no? Necessary but not sufficient. I'd love to find some content which tells a K-6 accessible story of spectra being convolved with L/M/S response curves. Related common misconceptions from textbooks include "white means a mix of all colors" or "all colors equally". > remain “white”, because our visual system adapts. IIRC, the range of colors for which you can't tell there's a tint, rather than merely ignoring it, isn't very large. You get used to 5000K incandescent, but can still tell it's "warm" (ie, cooler/redder than white). Hmm, I suppose that's an increasingly less-than-wonderfully-accessible exemplar... maybe tablet/display night mode? > diffraction grating glasses Nifty. I wonder how they might be tweaked to be better for education. Maybe increase clarity, by using a strip of one-axis grating, cutting across a neutral-density filter background - so you still get the bright spectra, but can more easily switch between seeing source and spectra. Maybe bend the grating to spread the spectra out more?? Thanks!

modeless 5y ago

I've never seen a wide gamut demo that was impressive to me. High frame rate yes, high contrast/brightness for HDR yes, those are definitely worthwhile. But wide gamut just doesn't seem to make a big difference to justify the implementation cost. I'd love to see some content that would justify wide gamut for me. Any suggestions?

bloak 5y ago
I think it would only be useful if the colours on the screen have to match colours off the screen. If all you're paying attention to is what's on the screen then it probably doesn't matter much exactly what those colours are because it's all relative, your brain compensates for it, and it's not as if the cells in your eye have a consistent response to colour in any case: it all depends on what you've been looking at recently. If you've ever been in a television shop, back when they still had shops, you could find two screens next to each other showing the same television channel but with radically different colours, when you compare them side-by-side, but if you watch just one of the screens - it doesn't matter which - you get used to the colours of that screen and everything looks fine.
jacobolus 5y ago
I sometimes make hue a meaningful attribute in data visualizations, but try to keep chroma roughly consistent for various hues, so that varying chroma does not cause visual artifacts unrelated to the data. RGB gamuts are very limited in blue–green colors (when you additively mix purplish blue “B” and yellowish green “G”, it’s hard to make a bright blue–green), and as a result my whole data visualization ends up with relatively lower chroma than I would like. With an additional few primaries (say, adding a greenish blue primary and a bluish green primary), the chroma-width of the gamut would be much more consistent across various hues, which would let these visualizations be something like twice as intense. It would make a huge difference. You probably haven’t seen any “wide gamut” demos with fundamentally different primaries involved; display manufacturers have been expanding their gamuts by making the three RGB primaries have narrower and narrower spectral power distributions, rather than by adding additional primaries. For a while Sharp made a “4-primary” display, but the additional “yellow” primary was more or less created by an additive mixture of existing “R” and “G” spectra, so didn’t really make as much practical difference as one would hope. https://en.wikipedia.org/wiki/Quattron https://en.wikipedia.org/wiki/Quattron You can get a display with like 8 or 10 primaries if you run a color science research lab and can spend tens of thousands of dollars on it (or maybe more? dunno).
- modeless 5y ago
  I'm sure if I saw a 10 primary display sitting next to an RGB one for reference, I could tell a difference. But if it was just sitting there by itself, in regular non-ideal viewing conditions, I doubt I would think that it is significantly better than normal displays. Whereas for high frame rate and high brightness/contrast displays I can easily see a big improvement under normal non-ideal viewing conditions without needing a reference.
  - jacobolus 5y ago
    Again, it would depend on the content shown. If you watch a 24 frame/second video on a 120 Hz computer display, I’m not sure you’ll notice much difference. If you use an extremely high contrast display to show a low-contrast image, you won’t see any obvious difference. If you use a 4K display to show a 240p video, you won’t see any obvious difference. If you watch a video with monophonic sound on a 7.1 sound system, you won’t hear any obvious difference. Etc. If you look at a photo produced by an RGB digital camera, color graded in some standard RGB space, and then presented on an RGB display vs. a 10-primary display, you aren’t going to see any obvious difference, because the content does not take any special advantage of the medium. In comparison, if you look at a bunch of synthetic images created to highlight the differences / stretch the capabilities of the displays, or e.g. a video produced by a whole hardware–software–author pipeline expecting the more capable display, the differences will be pretty obvious.

mncharity 5y ago

> If you know the shape of common RGB gamuts, you’ll start to notice [...] I'm interested in exploratory questions of the form "If education content about mumble wasn't such a disaster, then... what? What might doing it well look like? How might doing it well then be used to teach other topics better?" Color is widely taught preK-6. And yet first-tier physical-sciences graduate students not infrequently reply to "What color is the Sun?", with something like "it doesn't have a color; it's rainbow color". Suggesting that teaching is very not going well. One might imagine perhaps adapting tablet/phone clip-on spectroscopes down towards K, and developing a spectra-centered content and learning progression for color. So let's hypothesize students with a firm grasp of color. What might they do with it? > [...] you’ll start to notice how much all of our RGB-displayed visual media is constrained and nudged toward particular choices. There are e.g. color choices made in some paintings that can’t be effectively used in films. There are plenty of physical scenes that can’t be effectively reproduced. And on the flip side, areas of the RGB gamut where there is more available contrast get used a lot, even when they don’t quite fit the mood. Any suggestions on resources for exploring these further? Any other favorite stories or examples, useful or fun, where understanding color well gives insight or skill? Any other thoughts? Thanks!

jacobolus 5y ago
> "What color is the Sun?" This is a hard question to answer. My answer would be something like “intense pain followed by blindness”. :-) But you could also say that sunlight is “white”, because it excites all of our cone cells and is by far the brightest object in any field of view. You could change the spectral power distribution of the sun quite a bit one way or another and it would remain “white”, because our visual system adapts. > One might imagine perhaps adapting tablet/phone clip-on spectroscopes I’d start by handing out diffraction grating glasses, and looking at various kinds of lamps.
- mncharity 5y ago
  > This is a hard question to answer. [...] Oh, nice - I'd forgotten how the brief version of the question sends conversations in unexpected directions, and turns up a wide variety of conceptions. Long ago, to narrow the diversity of responses, I switched to asking longer versions, like "A 5-year old asks you: 'I'm told the Sun is a big hot ball! Awesome, I love balls! What color is the ball?' What color is the Sun?" Which has gotten a bit boring - same old failure modes grown familiar. Diverse could be useful again, for thinking about teaching color, and more fun. > “intense pain followed by blindness”. :-) Sooo... one can't tell what color a light is, unless you can look at it directly...? :P > sunlight is “white”, because it excites all of our cone cells So does my red shirt, no? Necessary but not sufficient. I'd love to find some content which tells a K-6 accessible story of spectra being convolved with L/M/S response curves. Related common misconceptions from textbooks include "white means a mix of all colors" or "all colors equally". > remain “white”, because our visual system adapts. IIRC, the range of colors for which you can't tell there's a tint, rather than merely ignoring it, isn't very large. You get used to 5000K incandescent, but can still tell it's "warm" (ie, cooler/redder than white). Hmm, I suppose that's an increasingly less-than-wonderfully-accessible exemplar... maybe tablet/display night mode? > diffraction grating glasses Nifty. I wonder how they might be tweaked to be better for education. Maybe increase clarity, by using a strip of one-axis grating, cutting across a neutral-density filter background - so you still get the bright spectra, but can more easily switch between seeing source and spectra. Maybe bend the grating to spread the spectra out more?? Thanks!