What does the AMA know about LEDs and health? There's nothing in this article about why they made these recommendations. And the article compares LED lighting to 'before electricity', instead of to for instance existing sodium lamps?
It is an opportune moment to make the recommendation, since folks are converting to LED so rapidly.
Actually they specifically say why: glare and impact on circadian rhythms. Both of these are very well studied issues with light concentrated in the blue end of the spectrum (as opposed to sodium lamps that have very little blue emissions).
BTW, they also failed to mention increased light pollution due to scattering.
There was a statement saying that too much blue light can harm the retina and some mentions of sleep-disrupting effects, but I tend to agree that the reasons why should have been given higher billing in the article as they're rather important.
Title is misleading; it's not LED streetlights, it's the color temperature of LED streetlights. While the article raises a few good points, color temperature isn't new; it's not like LEDs have some inherent issue that impacts health.
As long as a sizeable majority of LED streetlight in current usage have the blue color temperature (currently, as far as I know all, or almost all of them do), that's an acceptable approximation.
In any case, the original article's title mentions white LED streetlights -- so I guess making that clear in the title here might help.
True, but there's a real problem with people and places using lights with far too high a temperature, and even when shopping for home LED lights it's surprisingly difficult to find low temperature lights that aren't low on lumens output.
Pretty much anything over 3000K looks horribly blue at night, and if it wasn't such a massive pain to find them I'd prefer <2000K.
I have a bigger issue with fluorescent lights: They have a discontinuous spectrum[0] that can mess with your perception of colour as well as a noticable flicker that gets worse as the tubes age.
LIFX makes [IMO] pretty great LED bulbs with controllable temperature. They don't require a central hub like Phillips Hue, and you can find them on sale right now since the newer model is smaller and fits more fixtures. http://www.lifx.com/products/white-800?variant=933783339
Still 20x+ the cost of a normal bulb, and the tech probably won't work in 5 years, but it works well enough as a dumb bulb once the temperature is set that even when they inevitably go out of business or get acquired and shut down by Google, it should still be the temperature you want.
Although changing the colour temperature gets you somewhere its also the extreme spectral spikes of led as a technology which is notably worse than other kinds of lighting
A filter to block out a bit of the excess blue light? This would reduce the effectiveness of the light output though, unless you amped up the output to raise all of the colors and then selectively suppressed the blue to help balance it out. Then this would require extra energy, which defeats the purpose of using LEDs to save energy.
You can just use a different color of LED. Honestly people in Seattle, where these LED lamps were first piloted, have been raising these concerns for at least 4 years. This is the product of some combination of bad design, a lack of user-centered thinking at the utilities, and less than ideal incentives.
For example, daylight led's were touted as making it easier to identify cars involved in criminal activity because you can see the color of the vehicle more easily. But how many times has this actually made the difference in solving a crime?
One of the many hats that I wear at my day job is optics designer. An issue is that the white LED is a particular thing. It's a blue LED coated with a phosphor that absorbs some of the blue light, re-emits it at a broad range of longer wavelengths, while also passing some of the remaining blue light through.
The blue is special -- 473 nanometers -- due to the solid state physics of the semiconductor.
There is indeed an ongoing search for white LED designs that produce less glaring blue without sacrificing overall efficiency, but counterintuitively it's a hard design problem. People also want warmer light in their houses, so the problem is not limited to street lighting.
The safety regulations have a special place in hell for "retinal blue" light. As it turns out every wavelength from UV into the IR has a set of weighting functions for estimating the overall hazard exposure. It's possibly a rare example of a truly science based regulation.
I live in Seattle and I really like the new lights, they're way better than the dingy orange HPS lights. It's easier for my eyes to adjust when walking out of a bright apartment, and I can see more details. Anecdotally, I've also seen fewer vagrant types in the newly lit areas.
Also a Seattle resident. Who's complaining? Neither the Times, PI, nor Real Change have mentioned this recently. I've been enjoying the brighter streets at night.
> "In addition to its impact on drivers, blue-rich LED streetlights operate at a wavelength that most adversely suppresses melatonin during night."
I wonder if this ends up saving thousands of lives by decreasing fatigue in drivers. The blue light argument seems like absolute hogwash; we're exposed to orders of magnitude brighter blue light every day. It's called daylight, and the streetlights dont't even emit UV.
> Yes, but daylight only occurs during the day, hence the name.
Only because for more than a century the overwhelming majority of domestic light bulbs aped the colour temperature of candles, so that nightlight unconsciously became associated with yellowness.
There's nothing inherently 'right' about warm yellow light, it's entirely subjective based upon conditioning. For example people in hot, arid climates describe blue-white light as relaxing because it is associated with the cooler early morning. To them, yellow light is the scorching midday Sun. Having lived exclusively with 'daylight' domestic lighting for nearly a decade I agree with them.
The solution to disruptive street lighting is not to fiddle with colour temperatures but to turn off the lighting when it's not required. A residential street doesn't need to be illuminated at 03:00 if there are no cars or pedestrians, but managing that would require motion sensors. Easier just to let the bulb burn away through the night.
And for eons before that, the light we had at night was fire. That's long enough for us to evolve to be sensitive to blue light as a signal of daytime while having decreased curcadian sensitivity to light at lower color temperatures.
If you don't get tired at night, you don't sleep well, and sleep deprivation is much worse than being tired.
If you think the problem with light at night is hogwash you need to read it again. Bright light, especially blue light makes your body react like it's daytime. Do you really think people have forgotten about the sun? You only hurt your own understanding by being so quick to dismiss others.
> Bright light, especially blue light makes your body react like it's daytime.
You're preaching to the choir. I'm well aware of the melatonin-reducing effects of blue light, but the streetlights are a drop in the bucket compared to phones, televisions, and tablets. The one time I actually want to have my melatonin levels suppressed is when it's late and I'm operating heavy machinery.
Even when it screws up those levels for hours after you get out of the car?
We could mandate that people driving at night must drink at least one cup of coffee per half hour, and it would make them be more awake in the short term, but it would be a terrible idea for similar reasons.
Please quote a study verifying this with at least a few hundred people, if not a few thousand. The reason we think it's hogwash is because supporting studies involve only single-digit groups. The study I performed, involving 3 males and 2 females between 18-34, found blue light does not impact ability to sleep.
Not to diminish your study, but a simple "blue light circadian sleep" search in google scholar suggests a pretty mature understanding of the phenomenon, both at biochemical and phisiological level. My little study shows that after covering with black electrical tape each and every blue and white status led around the house everyone involved sleeps better.
Perhaps not the most neutral source for information. I found their summaries of paywalled articles to be a bit... misleading. They'll sell you Swarovski crystals and Himalayan salt lamps though.
The main issue doctors are alarmed about is how the increased blue light affects circadian rhythm and not the fact that we're exposed to it. Our bodies interpret blue light as a signal that it's daylight and we should be active. All sorts of biological processes are tied to it. Significant blue light at night screws the entire system up.
Focused blue light can improve alertness and other functions. That's why light therapy has seen some significant successes in treating seasonal affective disorder, some forms of depression, and other disorders. But there's no free lunch involved here; you're more or less manipulating your circadian rhythm, and eventually, those periods of increased alertness are effectively offset. With the light therapy and SAD example, that's exactly what you want to do: you're effectively telling your body that it's daytime, you need to be active instead of sluggish, and winter is just trying to screw with you. It should help eliminate oversleeping (common with SAD), but that doesn't mean you suddenly need less sleep. The same thing would apply here as well.
Surely then we could use blue-white LEDs on highways an in commercial areas where most of the driving time is spent and warmer colors in residential areas that are just the last minute or two of the drive and are where people are trying to sleep?
That seems like the right answer to me. I still think the streetlights make very little difference compared to the time that the average person spends looking at a screen, however. In my experience the only time I'm looking at streetlights are when I'm in a car, or walking to one. On the other hand I probably average three hours a night looking at a screen.
For those of us who have LED streetlights right outside our bedroom windows, it makes an enormous difference relative to screens, most of which these days you can at least put into an after-sunset mode even if you don't just voluntarily turn them off. The streetlights are on all night and they don't have an off switch.
I would argue that people with those crazy LED headlights are even worse. Since I often work late, I get the pleasure of being temporarily blinded as a white dwarf passes by at 45mph regularly.
I literally LOL @ "white dwarf". I absolutely agree, but, I have to admit... I have those lights :/
My last car was smashed by an idiot who didn't know how to back out of a parking spot, so rather than waiting for weeks while the body shop replaced the whole right side my car, I just traded it in for a new lease; a 2016 of same model. The new model has those stupid lights and I absolutely hate them. They're even unnatural and obnoxious to the person driving the car, not just the poor souls they're blinding.
Unfortunately, most cars will likely have them within the next few years.
And don't even get me started on the LED eye melters that cyclists are running these days. Some are about as bright as car headlights, but the spread and aim are inferior - especially since the folks that run those invariably seem to point them up from the horizontal. I keep thinking about mounting a 8 D cell maglight on a gimbal on my handlebars to give them a photon in-kind response.
I think part of it is a reaction to the poor bicycle infrastructure in many cities. Bright lights have stopped cars from hitting me at night a couple of times. While I'm simpathetic to other people (I try not to point them in people's eyes) there aren't that many other options.
I live in Portland, OR, e have pretty good bike infrastructure. If the goal is to be seen, there are indeed other options. LED glow or area lights are probably more effective than the eye melters pointing towards the sky and much less hostile.
If you see the biker then I think the lights did a good job. I suppose you could make an argument that it could blind people and they hit someone/something but I never heard of that happing - I only hear bikers getting hit because they were not seen.
i'm a cyclist and i sometimes have to stop and get off my bike and wait for an oncoming cyclist to pass me (eg, on a multipath). it's not necessarily the brightness, it's that it's not constructed to be visible (and/or illuminate the road) without also blinding oncoming traffic. fortunately i haven't ever hit someone or something while blinded by an oncoming cyclist, but that's in spite of the crap bike light, not because it did a good job.
Yes, me too. I'm a daily bike commuter and am quite sympathetic to the being seen aspect. This goes way beyond that in some cases. Case in point: a good chunk of my ride goes through the Portland Springwater corridor. This is a dedicated bike/pedestrian path. Much of it is unlit at night. I've seen these riders a half mile ago (~2 minutes) and by the time we're getting close enough to matter their lights are blinding. It's like a driver who refuses to dim their high beams for oncoming traffic.
Worse, if a cyclist's frontlight is blinding other cyclists coming at him it means they're failing both the others AND themselves, as: The oncoming ones can't see shit and are more likely to make a mistake and the rider themselves isn't illuminating the ground ahead of them so they can't actually see where they're going.
As a cyclist commuter who was hit by a clearly negligent car driver on Sunday in broad daylight, I fully support any measure to make bicyclists more noticed on the streets. In the absence of fully grade separated bike lanes, bicycles can only do so much to protect themselves against 4,000lbs of metal.
More often than not, they're not mounted correctly. When it comes to headlights, too many cyclists (most?) do not think of their bike as a vehicle. The headlight should be a solid light and the light beam should not be parallel to the ground. the primary purpose of your headlight is there to help you see where you're going. If it's not, you don't really ride in the dark. As with any other vehicle, being seen is a secondary reason for lights.
I've been commuting by bike, year-round, for 17 years. While my commute home may require lights, my morning commute requires them. (It's at 3:00 a.m. and some stretches have no lights.) If my headlight was parallel to the ground, it would not let me see where I am going. Oh, it will show a person or vehicle, but it won't let me see the road. It's aimed at a spot X feet in front of me, enabling me to see debris on the road, as well as anything huge.
Properly aimed or not, a headlight is only going to let a motorist (or anyone) see you if they are coming towards you. I can see them coming, too. A good reflective vest and secondary lights (helmet, side, etc.) is necessary to let people see you from behind or the side. (The vest is useless if the car is running without its headlights on, which has happened to me a few times in the morning.)
If you were coming towards me, why were you in the wrong lane? If you were drunk or not paying attention, seeing me probably wouldn't make a difference.
If you were coming from behind me, you wouldn't have seen my headlight, anyway. That's why I have a taillight and a reflective vest.
I'm not sure how these are legal, I'm in my twenties with 20/20 and only somewhat sensitive to light but even the normal ones are absolutely blinding to the point where I'm really not sure what I would do (other than have already crashed) without the "focus on the white line to the right" trick.
Most of the offending lights, as far as I know, are badly done retrofit kits that don't have the proper focusing hardware. Owners and cheap shops just drop an LED replacement into a "standard" headlight reflector and get horrible results.
The proper ones, like one sees on BMW, etc., don't blind people.
There's no law or technical consideration that says "LED headlights have to be built 10x brighter". People want them as bright as their electrical system can handle, and they buy them that bright. Blame people. If you like, regulate people; We successfully prevent them from using green lasers as headlights, this should be no different.
Not my area of expertise by any means, but maybe quantum dots could be used to filter only a small sliver of the visible spectrum produced by streetlights. Maybe there could be very-nearly transparent foils that you could overlay over your windows that would block just that light.
And people go install 4,000 and 5,000K LED lights in their kitchens and homes because it feels modern, can you imagine the disruption. I always insist on installing 2,800K at a max with my clients.
I have a couple of 4k LEDs in my kitchen, but also have much warmer LEDs in other places around my house. The reason the kitchen is different is so it feels sterile and clean. The cooler light gives this appearance more than warmer light. I guess the same reason many places have white counter tops.
"Hot" and "cold" have different meanings in different contexts. In the article, they're referring to color temperature, which is about the temperature of the source, and assumes black-body spectrum. So 3000K means a source at 3000 Kelvins. Light from a 5000K source has more blue, and so is cooler. But the color temperature is higher (5000K vs 3000K).
Color temperature of "white" LEDs is due to the amount and type of phosphor to downconvert the blue LED spike to warmer colors. The more phosphor and conversion that goes on, the less total output (no free lunch in physics). So in markets where amount of light (like exterior lighting) is a selling point, vendors sell their cooler temperature LED's for performance.
The article is saying "cool it and dim it" but I think they mean to say "warm it and dim it". They suggest the temperature should not be above 3000 kelvin.
That is because redder colors correspond to a cooler or lower color temperature. Colloquially, we think of blues as cool and reds as warm, but blue light is actually more energetic (or "hot") than red light.
Yes. It's the blue component (large with high color temperature lights) that is thought to most strongly affect human melatonin secretion and sleep. Here's one of many studies that indicate this:
At posting, I didn't have time to provide more references. You'll see I've given three more references below. Sources include Acta Physiologica, SPIE, and Neuroscience Letters. They discuss different effects of illumination color temperature in human systems.
I posted the initial reference as an entry point to follow literature on potential effects of illumination color temperature on human sleep, or perhaps circadian rhythms, through action on a hormone known to be critically involved in sleep. I didn't suggest it was the end of the story. My post was on topic and it contributed to the discussion.
Would be glad to review data you think interesting.
The first article isn't valid as it's paywalled and the abstract doesn't mention any studies; it could be someone's well-educated opinion.
The second article covers how bright light can be without being painful, and doesn't mention sleep or melatonin. Or at least the abstract does, again the article is paywalled, but at least it involves 500 participants.
The third study only covers 12 participants. I'll grant there may be thousands of these studies, each with a handful of subjects. Is there some statistical way to stitch the results together in any meaningful fashion - or would that just result in a low p-value?
I really like gtk-redshift, you can play with the settings in ~/.config/redshift.conf as well I find that a lower gamma even at midday works well and I tweaked even more towards the red than the default.
Any sources to studies done on the effects of light at different Kelvins? I'm really surprised as I would have expected LED light with more blue and looking natural to be preferable to orange Sodium Vapor lights. The sun which is our main light source puts out a wide spectrum. I'm surprised the lower wattage LED street lamps could cause actual health problems. I'm usually in bed during sleeping hours and not really exposed to the new blue LED street lights installed recently in my neighborhood though.
As someone who has always been interested in Astronomy I do worry about light pollution. Having worked at Observatories I do know that only red light is used due to it's low impact on light pollution.
LED spectra are usually tiny, sharp spikes, the home ones are getting slightly better, but the good ones are rather pricy. Incandescent or halogen bulbs feel more natural to me and their spectra look different
Blue indicator LEDs on consumer electronics are a real plague. I had to cover the LEDs on a noise machine with electrical tape because of the harsh blue glow. I had to wonder what on earth the manufacturer was thinking -- this product is marketed to parents of newborns to help their babies sleep! Heartbeats, whalesong, blinding blue LEDs, white noise: one of these does not belong.
88 comments
[ 2.6 ms ] story [ 173 ms ] threadIt is an opportune moment to make the recommendation, since folks are converting to LED so rapidly.
BTW, they also failed to mention increased light pollution due to scattering.
In any case, the original article's title mentions white LED streetlights -- so I guess making that clear in the title here might help.
Pretty much anything over 3000K looks horribly blue at night, and if it wasn't such a massive pain to find them I'd prefer <2000K.
Some of the newer "warm tone" CFL bulbs appears to do better better on these aspects. [0]: http://minerva.union.edu/newmanj/Physics100/Light%20Producti...
http://www.designingwithleds.com/the-new-cree-tw-series-led-...
Still 20x+ the cost of a normal bulb, and the tech probably won't work in 5 years, but it works well enough as a dumb bulb once the temperature is set that even when they inevitably go out of business or get acquired and shut down by Google, it should still be the temperature you want.
For example, daylight led's were touted as making it easier to identify cars involved in criminal activity because you can see the color of the vehicle more easily. But how many times has this actually made the difference in solving a crime?
The blue is special -- 473 nanometers -- due to the solid state physics of the semiconductor.
There is indeed an ongoing search for white LED designs that produce less glaring blue without sacrificing overall efficiency, but counterintuitively it's a hard design problem. People also want warmer light in their houses, so the problem is not limited to street lighting.
The safety regulations have a special place in hell for "retinal blue" light. As it turns out every wavelength from UV into the IR has a set of weighting functions for estimating the overall hazard exposure. It's possibly a rare example of a truly science based regulation.
I live in Seattle and I really like the new lights, they're way better than the dingy orange HPS lights. It's easier for my eyes to adjust when walking out of a bright apartment, and I can see more details. Anecdotally, I've also seen fewer vagrant types in the newly lit areas.
I wonder if this ends up saving thousands of lives by decreasing fatigue in drivers. The blue light argument seems like absolute hogwash; we're exposed to orders of magnitude brighter blue light every day. It's called daylight, and the streetlights dont't even emit UV.
Only because for more than a century the overwhelming majority of domestic light bulbs aped the colour temperature of candles, so that nightlight unconsciously became associated with yellowness.
There's nothing inherently 'right' about warm yellow light, it's entirely subjective based upon conditioning. For example people in hot, arid climates describe blue-white light as relaxing because it is associated with the cooler early morning. To them, yellow light is the scorching midday Sun. Having lived exclusively with 'daylight' domestic lighting for nearly a decade I agree with them.
The solution to disruptive street lighting is not to fiddle with colour temperatures but to turn off the lighting when it's not required. A residential street doesn't need to be illuminated at 03:00 if there are no cars or pedestrians, but managing that would require motion sensors. Easier just to let the bulb burn away through the night.
If you think the problem with light at night is hogwash you need to read it again. Bright light, especially blue light makes your body react like it's daytime. Do you really think people have forgotten about the sun? You only hurt your own understanding by being so quick to dismiss others.
You're preaching to the choir. I'm well aware of the melatonin-reducing effects of blue light, but the streetlights are a drop in the bucket compared to phones, televisions, and tablets. The one time I actually want to have my melatonin levels suppressed is when it's late and I'm operating heavy machinery.
We could mandate that people driving at night must drink at least one cup of coffee per half hour, and it would make them be more awake in the short term, but it would be a terrible idea for similar reasons.
Focused blue light can improve alertness and other functions. That's why light therapy has seen some significant successes in treating seasonal affective disorder, some forms of depression, and other disorders. But there's no free lunch involved here; you're more or less manipulating your circadian rhythm, and eventually, those periods of increased alertness are effectively offset. With the light therapy and SAD example, that's exactly what you want to do: you're effectively telling your body that it's daytime, you need to be active instead of sluggish, and winter is just trying to screw with you. It should help eliminate oversleeping (common with SAD), but that doesn't mean you suddenly need less sleep. The same thing would apply here as well.
I'd urge some caution there.
My last car was smashed by an idiot who didn't know how to back out of a parking spot, so rather than waiting for weeks while the body shop replaced the whole right side my car, I just traded it in for a new lease; a 2016 of same model. The new model has those stupid lights and I absolutely hate them. They're even unnatural and obnoxious to the person driving the car, not just the poor souls they're blinding.
Unfortunately, most cars will likely have them within the next few years.
Based on what?
No, absolutely not.
Lights are to help the cyclist SEE.
To be SEEN you need reflectors. [1]
Worse, if a cyclist's frontlight is blinding other cyclists coming at him it means they're failing both the others AND themselves, as: The oncoming ones can't see shit and are more likely to make a mistake and the rider themselves isn't illuminating the ground ahead of them so they can't actually see where they're going.
[1] https://www.youtube.com/watch?v=CfWzeGlaFvI
I've been commuting by bike, year-round, for 17 years. While my commute home may require lights, my morning commute requires them. (It's at 3:00 a.m. and some stretches have no lights.) If my headlight was parallel to the ground, it would not let me see where I am going. Oh, it will show a person or vehicle, but it won't let me see the road. It's aimed at a spot X feet in front of me, enabling me to see debris on the road, as well as anything huge.
Properly aimed or not, a headlight is only going to let a motorist (or anyone) see you if they are coming towards you. I can see them coming, too. A good reflective vest and secondary lights (helmet, side, etc.) is necessary to let people see you from behind or the side. (The vest is useless if the car is running without its headlights on, which has happened to me a few times in the morning.)
Sorry. Didn't see you.
If you were coming from behind me, you wouldn't have seen my headlight, anyway. That's why I have a taillight and a reflective vest.
The proper ones, like one sees on BMW, etc., don't blind people.
You don't want to "cool it" - that means more blue! You want to warm it with less.
http://www.ncbi.nlm.nih.gov/pubmed/8979406
I posted the initial reference as an entry point to follow literature on potential effects of illumination color temperature on human sleep, or perhaps circadian rhythms, through action on a hormone known to be critically involved in sleep. I didn't suggest it was the end of the story. My post was on topic and it contributed to the discussion.
Would be glad to review data you think interesting.
http://onlinelibrary.wiley.com/doi/10.1111/apha.12552/abstra...
http://proceedings.spiedigitallibrary.org/proceeding.aspx?ar...
http://www.sciencedirect.com/science/article/pii/S0304394008...
The second article covers how bright light can be without being painful, and doesn't mention sleep or melatonin. Or at least the abstract does, again the article is paywalled, but at least it involves 500 participants.
The third study only covers 12 participants. I'll grant there may be thousands of these studies, each with a handful of subjects. Is there some statistical way to stitch the results together in any meaningful fashion - or would that just result in a low p-value?
As someone who has always been interested in Astronomy I do worry about light pollution. Having worked at Observatories I do know that only red light is used due to it's low impact on light pollution.