In an earlier post, I discussed the difference between purple and violet, and explored some of the color limitations of electronic display technology. Phil recently pointed out an article in Wired that discusses the use of adjustable diffraction gratings to produce arbitrary colors. (In practice, the gratings don’t produce the colors; they diffract a white light in such a way that the desired color can be made to pass through a pinhole). In theory, an array of these can be constructed to produce vivid-color televisions and monitors.
There’s something I find a bit suspect about the article, though. I mean, yeah, it’s full of the traditional Wired-style junk science (e.g., using relative voltage to compare power efficiency without taking current into account — plus, it includes a diagram of all the colors monitors can’t display [pause two beats here to let that sink in]), but in terms of color rendering, it says one thing that stands out as really bizarre.
The researchers are quoted as saying they intend to use white LEDs as the light source for this technology.
LEDs are diodes made with materials specifically chosen so that electrons crossing the p-n junction cause a photon to be released. The wavelength of these photons (color of the produced light) depends on the exact materials being used. Note I said “wavelength,” not “wavelengths” — LEDs produce a single color out of the spectrum at a time. (Strictly speaking, they produce a very narrow range of wavelengths, typically about 20 to 30 nm wide, with very steep drop-offs — but this is as close to a pure color as to make no difference for this conversation).
White LEDs can be produced by mixing together two or more carefully chosen single-color LEDs, but this is rarely done. Almost all white LEDs produced today use a blue LED as their base (gallium-nitride based, with a wavelength of ~460 nm); on top of this LED, they layer a phosphorescent substance (cerium-doped yttrium aluminum garnet) which absorbs part of the blue light and emits a yellow light centered around 580 nm.
If you take the light from one of these LEDs and pass it through a prism, you’ll get a very thin, bright line of blue, and a slightly wider beam of orange/yellow/green.
By now, you should see where I’m going with this. If you use a white LED as your color source for a monitor that uses a diffraction grating, the results will be no better than today’s color display technologies, and arguably worse. Not only will you lack the ability to display colors below 460 nm (keeping in mind that s-cones peak at 420 nm: no violet for you!), but you’ll have gaps in the lower green and upper red spectrum as well.
Nonetheless, the adjustable diffraction technology is fascinating, and I hope something like this eventually gets to see the light of day — hopefully using something more wide spectrum than what the article implies for a light source.
Now all we need is a CCD that can record a full-spectrum scene, and we’re good to go.