Am I misunderstanding the title there or is it simply made to be catchy but wrong?
The actual beginning of the article says this:
"Researchers have shown why intense, pure red colors in nature are mainly produced by pigments, instead of the structural color that produces bright blue and green hues."
So the research has to do with the origin of different colors in animals, not "why blue and green are nature's brightest colors". A quick look at a Scarlet Ibis or a Poison Frog should make it clear that opposing red to blue in a battle for "nature's brightest colors" would be an irrelevant endeavour anyway, and that isn't what the researchers have been doing in the first place.
Yes agreed, they say "structural color" (film interference) only works for blues and greens, and you need pigments for oranges and reds. Title should be changed.
Pigments are inorganic. For example, iron oxide, titanium oxide, burnt umber, and yellow ochre. So the statement is more or less technically sound and somewhat melodramatic, but not terribly informative. It is either trivial if the reader knows what pigments are or confusing if they didn’t take it literally.
Digital photography still uses my father's filter (https://en.wikipedia.org/wiki/Bayer_filter). How did it survive? He made an inevitable choice, this is bit like asking why the number two is so prevalent. Frame the question right (as he did, preparing twenty years for what looks like ten minutes of work), and this grid is the unique answer:
"Checkerboard half the squares green, turn your head, and checkerboard what's left red and blue."
What's most striking here is the prevalence of green squares. With a honeycomb pattern, one could equally distribute RGB cells. However, this would be less efficient in hardware and software. John von Neumann considered base 3 computing, but settled on binary, for a similar efficiency advantage.
From what my Dad understood of the human eye, he decided that green was the best proxy for black and white detail, so he favored green. Digital photography evolved in harmony with the human eye, just as eyes evolved in harmony with the objects of our vision. None of us take in FM radio with our eyes, and few animals see red. Some speculate that our corner of the mammal world sees red to spot ripe fruit, nature's pigment playground.
Very interesting! I remember learning about the Bayer filter in a computational photography class I took. I didn't realize the choice of green was meant to reflect a bias about human perception. It kind of reminds me of luminous flux, a metric adjusted to reflect the varying sensitivity of the human eye to different wavelengths of light.
"Turn your head" means turn your head 45 degrees, and look at the half of the squares you haven't colored green. You see a board of uncolored diamonds. Recursively checkerboard that. Like Huffman coding, green gets one bit, red and blue get two.
In math we can barely understand each other. It is often easier to reverse-engineer code than to read someone else's code. Go look at the picture on Wikipedia, ask yourself how you'd describe it to someone in an elevator, then read my description again?
Equally distributing RGB cells would be less efficient. Not equally distributing RGB cells would be more efficient. We're saying the same thing.
The copper works out easier, not using a honeycomb. In the 1980's I imagined that by now we'd have compilers that could generate a custom circuit for every programming task, and computers would be whiteboards capable of directly simulating any circuit. This would be the primary mode for all computation. On one hand, programmable gate arrays approximate this, but they're only used when one needs that flexibility, as they're not efficient enough to replace general purpose CPUs. Someone on a plane set me straight. Again, one could give an entire course explaining the answer, but the gist is the same: it's the copper. Fewer traces.
My Dad already had in mind post-processing algorithms, and computers in the 1970's were quite primitive. Ask someone to program on a honeycomb at a stressful tech interview, and they won't get the job. Plain arrays are easier, more efficient. Now computers are faster, but the processing has moved into the camera.
Our fixation on rectangles is a cultural conceit. If photographs were hexagons, we might prefer honeycomb sensors. Silicon wafers are round; I'm surprised that the largest telescope camera sensors aren't hexagons; they're piecing together photographs, not putting them up on a wall for display. Yet they piece together square sensors, use a square grid. Huh.
I coded up a Bayer sampler just now based on the Wikipedia page you linked, that was fun. One simple mapping I rediscovered for determining the relevant RGB/BGR channel of a given (x, y) pixel was, in C, (x & 1) + (y & 1).
“The researchers modeled the optical response and color appearance of nanostructures, as found in the natural world. They found that saturated, matt structural colors cannot be recreated in the red region of the visible spectrum, which might explain the absence of these hues in natural systems.”
Sounds like circular reasoning to me: “Red colors cannot be recreated in nature because (according to our modeling) nano structures found in nature cannot produce them”. What do I overlook?
The actual beginning of the article says this:
"Researchers have shown why intense, pure red colors in nature are mainly produced by pigments, instead of the structural color that produces bright blue and green hues."
So the research has to do with the origin of different colors in animals, not "why blue and green are nature's brightest colors". A quick look at a Scarlet Ibis or a Poison Frog should make it clear that opposing red to blue in a battle for "nature's brightest colors" would be an irrelevant endeavour anyway, and that isn't what the researchers have been doing in the first place.