The “better efficiency” of privileging the blue light coming directly from the diode is oversold.
At night, what happens is that directly visible blue lamps in your peripheral vision cause a severe adaptation response from your eyes (a kind of “oh hey, it’s daytime” misjudgement), which causes everything that isn’t directly lit to recede into dark shadow as the rods are swamped. You end up using mostly cone vision, where the blue part of the spectrum doesn’t really contribute to brightness response anyway, defeating the original purpose.
Longer wavelengths don’t cause the same kind of adaptation, allowing the rod cells in the eye to remain useful down to much lower levels of overall environmental lighting.
Industry has made up bogus metrics/arguments about “mesopic efficiency” of high-CCT street lamps and car headlamps which are IMO based more on marketing needs than dispassionate scientific investigation.
I’m hoping that someday non-industry-funded color scientists will do some proper research on the glare and adaptation effects of different outdoor lighting sources.
You may have solved an issue with a new tunnel here in the Netherlands. The contractors and operators are puzzled why the accident rate of the new tunnel is so much higher than the old one, what with all the efficient and 'bright as day' white light. But if you're right and that adaptation happens then likely going into the tunnel your nightvision gets 'switched off' long before you still need at (at the tunnel entrance) because you're staring into the well lit tunnel, and at the end of the tunnel your eyes are likely still overloaded and are re-adapting to darkness.
The monchromatic lights they were using before probably did not cause any such adaptation and therefore did not mess up the nightvision of the drivers to the same extent.
They have tried to remedy the problem by dimming the lights overall and gradually upping the intensity as you get further into the tunnel but it is still a problem.
So do I understand you right that would mean that during the night they should use monochromatic lights and during the day white?
Edit: I've sent an email to 'rijkswaterstaat', the federal infrastructure department of the government here with a link to this thread. Let's see if they reply.
What they should probably be using is a lower color temperature, and likely a lower intensity. Light with a lower color temperature has the majority of its energy in the red/yellow end of the spectrum, and that has a lesser impact on night vision.
I see the article says the intend to dim the lights, which is probably the most important factor.
When Caltrans rerouted the Bay Bridge here in SF, they also took the opportunity to clean the central tunnel through Yerba Buena and replace its lights with similar "bright as day" lighting. At night, I notice I unintentionally squint as I enter the now blindingly-white westbound tunnel, and my eyes feel "disoriented" (for lack of a better word) when I come out.
I thought I'd get used to it by now, but I haven't. Maybe similar.
I've taken to putting on sunglasses when I drive through that tunnel. Yes, even at 11pm, which is when I usually find myself going through it. I think I might be more sensitive to light color and brightness than most people.
I could be wrong but it seems to me that the near-unreadability of blue commercial signage at night is probably another aspect of this whole category of concerns.
It depends whether you're myopic or the opposite. Things are blurred for me (myopia) on the red side of the spectrum, but blue is sharp; Other people have sharp reds and blurred blues.
Well, many transport authorities don't care and would rather save money on electricity costs and materials (less LEDs and less powerful drivers) than actually make the road more driveable at night.
They would save even more money on electricity by just cutting the brightness of the lamps. Whatever official body or industry recommendation sets the standards for urban illumination really needs to bring things down.
My street’s new LED lamps could have their intensity chopped by 2/3 without adversely affecting anyone’s vision.
The electorate believes that brighter street lighting leads to less crime, and dimmer street lighting leads to more crime. A municipal politician's career would be eviscerated by the first violent crime on a street they voted to dim.
"Soft on light pollution" doesn't kill political careers.
Well, they might even be right. See, if the electorate believes it then likely so do the muggers (them being part of the electorate). And so the muggers will seek out the streets with the dimmer light.
Even if the science is lousy it might actually work out that way if there are streets with and streets without dimmed lights.
So the only way to control for this is to have the whole city dim the lights on some nights and not on others to compare when there was the most criminal activity.
Another way to control for it is to increase or decrease lighting in some locations while leaving others the same and see what changes occur to crime rates. That has been done at least once, in Chicago where low-output lights were replaced with much brighter lights in some alleys but not others, with controls for various demographic factors and prior crime rates.
The result was that crime increased where brighter lighting was used. One plausible explanation is that committing crimes, like most human activities is easier when there's sufficient light to see what you're doing, and using a flashlight is a poor option when you want to avoid calling attention to your illegal behavior.
That's only true for victimless crimes, and the bulk of the increase in non-index crimes in the study was drug-related. The study mentions increased reporting due to visibility as an explanation.
The index crimes increased as well. Those include homicide, assault, burglary and theft, all of which have victims.
I'm not sure why you're getting downvoted (lack of a citation?) but I think you're right. To the extent that most people think about this at all I bet they would say they prefer bright street lights to dimmer ones in their neighborhood.
Thank you. You've confirmed a long-standing suspicion regarding how white LED lamps are actually worse than yellow sodium for streetlights and headlights.
He had a suspicion, but no logical reasoning, until one was offered here on HN. That logical reasoning seemed rational and persuasive so he chooses to believe it.
Your snark is awfully unwarranted and entirely unwelcome.
The effect would have to be stronger than the differences in luminous efficacy to work, which is about 10% between 3000K and 4000K. This is a great theory, but you provide no supporting evidence. The "bogus" metrics you complain about are based around real color research done in 1931, far before people were marketing LEDs: https://en.wikipedia.org/wiki/Luminous_flux
No, I’m talking about stuff like https://scholar.google.com/scholar?q=mesopic+efficiency (add keywords about street lamps and car headlamps if you want more recent and specific results). The relevant literature is mostly from the past few decades, and ongoing.
Luminous efficiency is not the only thing we should care about here. What we want is for the road and objects on it to be clearly visible to someone with dark-adapted vision, and for individual sources of light to be visible but not cause glare or distraction. Obviously a brighter light does a better job of making itself visible and illuminating things directly in its path; the problem is that brighter lights which are intense in the blue part of the spectrum cause more glare and more adaptation than dimmer yellower lights, which makes everything that isn’t being directly lit by them effectively less visible. If objects directly under the sweep of car headlamps are lit to the same extent as during a sunny day, then the shadowy bushes off to the side of the road and the kid running out from behind them are going to be very dark by comparison.
I’m not an expert in this specific topic (I’ve studied much more about color vision under well lit conditions), but several of the papers I skimmed when I started looking into it because I was frustrated about my street’s new LED lamps seemed to be funded by industry, and I’m not convinced that their experiments mirror real-world conditions of nighttime streets, or properly measure the effects of adaptation and glare.
The lighting industry wants to use a measure of luminous power based on “mesopic conditions” (i.e. mix of photopic + scotopic) because if they use the photopic function then blue light doesn’t really contribute much and their lamps don’t look as efficient or bright. My personal experience is that bright sources of blue light in my peripheral vision cause enough adaptation that I’m not convinced that a “mesopic” function as they’re using it is actually appropriate.
I should probably email some real color scientists to ask for clarification someday. If you know of a nice survey paper written by someone with industry-independent bona fides, I’d love to skim it.
At night, what happens is that directly visible blue lamps in your peripheral vision cause a severe adaptation response from your eyes (a kind of “oh hey, it’s daytime” misjudgement), which causes everything that isn’t directly lit to recede into dark shadow as the rods are swamped. You end up using mostly cone vision, where the blue part of the spectrum doesn’t really contribute to brightness response anyway, defeating the original purpose.
Longer wavelengths don’t cause the same kind of adaptation, allowing the rod cells in the eye to remain useful down to much lower levels of overall environmental lighting.
Industry has made up bogus metrics/arguments about “mesopic efficiency” of high-CCT street lamps and car headlamps which are IMO based more on marketing needs than dispassionate scientific investigation.
I’m hoping that someday non-industry-funded color scientists will do some proper research on the glare and adaptation effects of different outdoor lighting sources.