tl;dr: minus yellow could be a better tint adjustment than minus green.
I noticed that some incandescent bulbs I have (General Electric "Reveal" series) emit light that is noticeably nicer than normal incan bulbs, and that they have a powdery light-blue coating on the interior surface. Here's an article explaining what it is (Nd2O3) and what its transmission spectrum looks like. (The article does claim at the end that incans have CRI 80-85, which seems dubious.)
This coating differs in function from, say, a minus green filter, in that the coating absorbs over a narrower interval of wavelengths that is centered closer to orange, while the minus green filter absorbs a broader range of wavelengths that is more shifted toward green. A simplified way to think about this is that the coating subtracts yellow, not green. The coating results in light that makes colored objects look nice (the same way a minus green filter would), without making whites seem pink.
Yellows look a bit strange (you can tell it's a red/green mix instead of monochromatic yellow), but nobody ever complains about that, as yellow is a somewhat "redundant" color: in additive color theory, you can make yellow from red and green, but you can't make red from any combination of yellow/green/blue, or green from yellow/red/blue.
Wonder if there's a way to source this coating (or doped glass) as a flashlight lens accessory for tint snobs, as it makes my dedomed 519A look low CRI in comparison!
Thanx. I’d wondered about those Reveal bulbs, and the potential implications of some of that background info on LED applications is interesting to consider.
Thank you so much for the data--this is exactly the type of stuff I'm looking for! Is there a chance your device could also measure R9 (saturated red)? I suspect that the light oversaturates reds so much that the R9 score would be low.
Also, that's the thickest coating I've ever seen on a bulb...I didn't know it was possible to raise it all the way to 4700K! The one I have appears to be just a bit over 3500K to my eyes, probably similar to your left one. The duv was actually much higher than I expected, given how nice and red-saturated the light looks! I guess this is evidence that minus-yellow can make light look nicer without making it pink.
My conjectural explanation is that yellow is the only range of wavelengths that corresponds to a color that been doubly-counted by the human eye: we see the yellow wavelengths as yellow, but we also see red+green as yellow. Suddenly, an unmodified light source now has two copies of yellow in its perceived spectrum, which would make things look, well, yellow!
Deleting yellow wavelengths takes care of this problem and also corrects for the double-counting of orange and yellow-green.
yep
adding more layers of energy use
filters
phosphors
it IS interesting though that converting blue to yellow makes it more visible
due to the eye’s weirdness
the lumens might go up while the actual electricity-to-light factor goes down
.