Hacker News Reader

Comment by mtalantikite | Hacker News Reader

Yeah I agree. NYC went full on LED a few years back and I hate it. I have a lamppost right outside my apartment window and previously the light didn’t really bother me, now it’s this horrible cold LED that’s super bright flooding into my living room.

I get that they’re cheaper to operate, but they throw shadows differently and have more blue wavelengths in them (in nyc they’re 4000 kelvin last I looked). It personally feels much less magical to me and 3-4 years in I still haven’t really gotten used to it.

View on HN

Similar feelings in Chicago.

It also makes me feel less safe at night. I get that the orange glow makes everything look muddy, color-wise, but it doesn't kill your night vision to the same extent as the bright white LEDs do. Sure, things directly under the streetlight are lit up like it's daytime, but I can't see for @$#% in between the streetlights or behind the bushes or whatever.

I'm curious why it seems that no city has tried installing amber LEDs in an effort to try and realize the efficiency gains without quite so many of the downsides.

>I'm curious why it seems that no city has tried installing amber LEDs in an effort to try and realize the efficiency gains without quite so many of the downsides.

Flagstaff, Arizona did. There are some drawbacks, mostly that they simply don't get as bright.

https://arstechnica.com/science/2019/10/how-flagstaff-arizon...

Aha. And that article also explains why. In terms of lumens per watt, it appears that the amber LEDs are actually about 40% less efficient than the low pressure sodium lights they would be replacing.

True. That said, LED lighting has a directionality advantage as well. So it turns out you need less of them to get the same basic coverage as the sodium lamps. Here’s an account of how Tucson, with its many observatories, managed to keep within budget using standard, high-efficiency blue LED fixtures:

“If you absolutely must use white LEDs, you could do what Tucson has done,” Hall said. “They... switched out their whole high-pressure sodium system to 3,000 degree white but reduced their lumen budget for street lighting from 480 million to, like, 170 million [lumens] or something. And you need to do that. For every white LED lumen, you're increasing your skyglow by a factor of about three, but they cut the lumen budget by about a factor of three. So overall, they managed to wash out the skyglow because they’ve got a lot of observatories down there.”

> high-efficiency blue LED fixtures

The extra blue is probably screwing up the circadian rhythms of humans and animals:

* https://www.health.harvard.edu/staying-healthy/blue-light-ha...

* https://www.npr.org/sections/thetwo-way/2016/06/21/482936520...

Lumens per watt is the curse of bureaucratic/industrial mandated lighting. Problem is people hate lighting with low CRI, people also hate dim blue light. But you can't put a number on hate like you can lumens per watt.

Efficiency regulations for residential fluorescent lighting delayed adoption by probably 25 years. Because regulations mandated lamps with terrible CRI.

My opinion is that the whiter LEDs are a welcome change from their orange predecessors. I'm normally very nostalgic, but I make an exception for orange street lights.

This also seems like a choice. There is no reason LED's cannot create warmer light, and there is probably a solution to scatter the rays and not give that "clinical" kind of light which LEDs are known for.

I have an LED desk lamp with a temperature setting, and when it's on the warmer settings it really does produce a pleasant light.

LED lights do in fact have the ability to do different color temps, some fixtures have tunable color temp as well. In my home, I use lamps that output 2200-2700K depending on the dimmer setting. Lower color temps output less lumens per watt, and the goal of exterior lighting is generally to improve visibility, hence the 4K color temp.

Here’s a cutsheet for a Lithonia wall pack fixture showing the different options, there’s also a chart showing what the lumens per watt is based on the color temp: https://img.acuitybrands.com/public-assets/catalog/1008038/w...

The color a LED emits is given by the choice of materials (specifically, the band-gap, with E=hν, ν being light frequency https://en.wikipedia.org/wiki/Band_gap#Light_emitting_diodes... )

"Tunable" LEDs are just multiple LEDs, the goal being to change the emission peak. Peak wavelength can be mapped onto a black body temperature trough Wien's law (λpeak =2.8e-3/T) https://en.wikipedia.org/wiki/Wien%27s_displacement_law

I don't really know how the mapping is done, though, as there are different ways to map it since the LEDs won't have the same emission curve as a black body.

The new NYC lampposts also spread wide light cones, which is good for road safety but means they intrude through house windows. Astronomers have also complained about the lights worsening the "seeing", which is already bad.

Exactly, I wish they at least had some sort of shade on the back towards the sidewalk that stopped it throwing that much light back on the houses. Do we really need that much light on a one way, residential street?

"Seeing" is actually about how steady the atmosphere is. The old sodium lights emit in a narrow part of the spectrum which can be mitigated against by using light pollution filters. These allow all light through, except sodium frequencies. Sadly, filters are ineffective against modern LED lighting which emits across a lot of the spectrum.

It's actually garbage for road safety because bluer lamps spaced so far apart creates less even lighting, blows out everyone's night vision, causes distracting glare in peripheral vision, and makes it impossible to see into the shadows.

The ideal for road safety would be more uniform lighting from a larger number of closer-spaced lamps, placed at lower height, diffused, shielded from the sides, and colored orange (and optionally uniformly dimmer).

It would also help road safety to ban the new bluer types of car headlamps, which cause much worse glare than previous yellower ones.

> which cause much worse glare than previous yellower ones

The color has almost nothing to do with that. Newer headlights are much brighter, and mostly use projectors now. The increase in glare is mostly due to the projectors, because they emit all their light from small point. You'll notice that newer cars with halogen projectors are almost as bad as one with HIDs.

It would be interesting to see how a headlight would look if it were able to emit the same amount of light on the road but from a rectangle spanning the entire front of the car. Minimum glare, still effective.

> color has almost nothing to do with that

The color makes much of the difference. (What really matters is the total intensity in the shorter-wavelength part of the spectrum, so sure, making the lamps much brighter and less diffused is also bad.)

Shorter wavelengths cause significantly more glare, because the peripheral part of the human retina is packed with rods and S cones. Intense light including shorter wavelengths also causes the eyes to become bright-adapted, making it harder to see in the dark.

The super-bright halogen headlamps emit much more blue light than previous incandescent headlamps, and are included alongside LED headlamps in my complaint here.

Bigger, much dimmer, more diffuse lamps would help, but they should also be less blue. (And people should stop mounting headlamps on their SUVs and pickup trucks at other drivers' eye level and aiming them straight ahead.)

I’ve seen exterior fixtures that have shielding options to prevent light from being cast at houses, but it’s an added cost. We sometimes install them on commercial buildings that abut residential areas to prevent light pollution into neighboring homes.

Isn't 4000k far more "natural"? A full moon is 4000k.

Not at that brightness. Moonlight is extremely dim.

Moonlight for some hours of the month, never exceeding a rather low intensity, a minuscule intensity compared to a single led streetlight.

The night time lighting humans(and our non-human co-inhabitants of earth) have evolved with and known for a very longtime comes from fires of various types, flames with a much warmer color temperature.

Humans have had fire at night on a regular basis for a pretty short time, evolutionarily speaking. And that goes double for what animals might have seen; actual fire is so resource intensive that until a few hundred years ago the vast majority of the night was lit only by the moon and stars.

Also, intensity is orthogonal to color temperature.

Use of fire predates anatomically modern humans by some hundreds of thousands of years.

Consider the spread of lactose tolerance, a recent (no more than 10,000 years and perhaps less) mutation which has spread to 35% of humanity in that time. We have had plenty of time to co-evolve with fire, and doubtless, we have done so.

Well, we certainly didnt co-evolve to be fireproof in all this time, so what adaptation to fire do you suggest we did evolve?

Its probably to do with cooked food, not vision - which is already co-evolved with the Sun, for when we are active/hunting/etc.

For one thing, we find a controlled fire soothing and attractive.

Wild animals very much do not. It's likely that our mucus membranes have also adjusted to the presence of more smoke: this doesn't prevent cancer, which is a post-reproductive problem for the most part, but it does keep smoke inhalation from being immediately debilitating.

Does it mean that fire doesn't fool our circadian rhythm into thinking it's still daytime? Sure, why wouldn't it? We've had a very long time to adjust, and getting appropriate amounts of sleep increases fitness in an obvious way.

We haven't had more than a hundred years of adjusting to bright blue-wavelength light after dark, however. So it could affect our circadian rhythm, and as studies show, it does.

You might be able to get the city to add some metal to shade the light coming into your room. I don't know about the US, but here in Germany it took me one call to the city office and two weeks of waiting.

Reply on news.ycombinator.com