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!
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.
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!
No, the light reflected by your red shirt is not like the light from the sun, because it is nowhere near the brightest object in view. But you are right that if you made a lamp with the same spectral power distribution as light reflected by your red shirt but several orders of magnitude brighter, it would still probably be outside the range considered “white”, even after adaptation.
> how they might be tweaked to be better for education.
I think the current disposable cardboard ones which cost like 50¢ each (maybe less?) when bought in bulk are pretty good for education.
If you walk around at night with diffraction grating glasses looking at different kinds of lamps, you can learn a lot about what kind of light they produce.
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
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).
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.
If you want a somewhat relevant test, you can try comparing images produced by a photo printer with images printed in old newspapers. Having a larger gamut makes the former much more vibrant.
Current displays produce very intense orangish-red, purplish-blue, yellowish-green, and reddish-purple colors. So the differences with a hypothetical >5-primary display are going to be most noticeable in the blue-green range, and oranges and yellows.
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.
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.