Hacker News new | past | comments | ask | show | jobs | submit login
Rooms can be as bright as the outdoors (benkuhn.net)
740 points by luu on Nov 28, 2019 | hide | past | favorite | 306 comments

(Michael from f.lux): this kind of thing is necessary for most buildings, because otherwise people don't get a "day" signal when they spend time indoors. Most of us spend >90% of our time inside.

However, we shouldn't mess with light levels like these (or even somewhat lower ones) without some limits, and the main one is some automatic timing to turn things down after it gets too late. Otherwise you will find yourself staying at work later and later each day, and it's no accident.

Our clocks don't work that well with more than about a 14-15 hour photoperiod ("day"), so spending 18 hours under 5000 lux will screw up your overall rhythms, making your internal day longer than 24 hours and reducing your sleep. This is considerably worse if you're a natural night owl - it will make you later, faster.

The myth of electric lighting is that we can live in a permanent summer and turn the wintertime into summer, but there are important seasonal questions that we should think about too. For instance, cancer progresses much faster in summer than in winter [edit: cancer progresses much faster in constant light: winter vs. summer is complicated], and the winter schedule even engages a different part of the central clock in your brain than the summer schedule does. There are some important things going on here, which we've mostly tried to remove from our environments.

Knowing how the body works and having the capability to change it really do need to evolve in parallel - I think there is way more capability than knowledge right now.

Is there any solid peer-reviewed, published papers on this stuff? While it sounds intuitively correct and attractive, I've struggled to find anything rigorous that shows light colour and intensity in the evening has an effect of sleep quality - its all anecdotal.

The general field of study is called chronobiology. Doing a search for "chronobiology light sleep" gives all sorts of results. One example:

> Environmental light synchronizes the primary mammalian biological clock in the suprachiasmatic nuclei, as well as many peripheral clocks in tissues and cells, to the solar 24-hour day. Light is the strongest synchronizing agent (zeitgeber) for the circadian system, and therefore keeps most biological and psychological rhythms internally synchronized, which is important for optimum function. Circadian sleep-wake disruptions and chronic circadian misalignment, as often observed in psychiatric and neurodegenerative illness, can be treated with light therapy. The beneficial effect on circadian synchronization, sleep quality, mood, and cognitive performance depends on timing, intensity, and spectral composition of light exposure. Tailoring and optimizing indoor lighting conditions may be an approach to improve wellbeing, alertness, and cognitive performance and, in the long term, producing health benefits.

* https://www.ncbi.nlm.nih.gov/pmc/articles/PMC3553574/

There are 61 foot notes in the above paper, so that should help you get starting going down this rabbit hole. :)

Second this, and I wonder how people who live near the equator respond to seasons and light vs. people closer to the Arctic circle.

Very much this.

I live in Norway. While I'm below the Arctic circle, summer nights are bright. I don't need outdoor lighting to read as the sun goes juuuust below the horizon. This time of year, days are short and the light intensity never really gets all that bright. I don't have issues with any of it - and I'm an immigrant.

There are many genetic differences, even within a given ethnicity, concerning sleep. What may impair sleep for one individual may have no effect for another. This variation may have been beneficial from an evolutionary standpoint, because variable sleep schedules meant that someone in a given community could always respond to threats.

Have you considered manufacturing f.lux for home lighting, such as a wall-mounted controller for Hue or Ikea bulbs that changes color spectrum depending on time of day? Or possibly bulbs with f.lux build in, so no external controller is required?

Yes, we were looking at this quite a while ago. Lighting is a really hard business (especially residential) so we did not make the leap, and mainly have tried so far to work with people who are in that business.

We thought back then that some form of "smart lighting" would be universal. This mostly hasn't happened, so the truth is that the opportunity is different than we had guessed. Today most "smart" replacement lamps are $20+, similar to several years ago, even though normal lamps are $1.

We do support a lot of integrations with f.lux on Windows, so people can use f.lux with their Hue/LIFX/yeelight/etc., but it is somewhere <2% of our users right now.

Out of curiosity, what makes lighting such a hard business?

I have given the issue some thought as well. f.lux integration with Hue is great, but unfortunately not a solution for me, as I'd have to keep my windows pc running all day.

There are apps, such as iConnectHue, which are able to install a preset on the Hue Bridge that dims the lights or changes the color temperature without needing any computer or phone turned on when the timer fires.

Fixture design is fashion and existing products are not expensive.

The fact that bulbs are 1. small and 2. run on mains electricity is a big complicating factor. By far the easiest way to build this would be with dual-colour LED strips. But that's not how most people want to light their homes.

Electrical engineer here.

What is hard about these?

For a few dollars I can convert 120v to whatever I want.

Sell a different version to Europe

I think it's the cost. You can make a cheap driver off 120V that looks like shit and has terrible thermals.

A solid driver with good filtering and good thermal characteristics is expensive. So you're left with a $15 LED bulb that sits on a shelf next to $5 bulbs, selling to customers who see nothing other than two LED bulbs where one is grossly overpriced for no reason they can comprehend

Sell on their website? They get traffic

As a software developer, the fact that I even have to build custom electrical hardware is a significant barrier.

The fact that this hardware runs at 120v (or 240v in my case) makes it the barrier higher as there's now a chance that making mistakes might kill me.

Additionally, the high power levels mean that adequate cooling is needed, and as a non-expert it's hard for me to know whether I've done that correctly.

Thanks for the info! Just learned about bias lighting from your blog post.

Any chance we will see a non-root flux for Android?

While possible (there are tons of "blue filters" on the play store), the effect will always be considerably worse that with root access, since apps can only display overlays, not change the screen's white balance.

Btw, if you don't insist on using f.lux and have root access, take a look at cf.lumen: It allows much more precise configuration (I have it remove almost all blue in sleep mode).

I have some Osram Lightify bulbs along with Home Assistant to do this.

During the day time all my bulbs are at 6000K and 100% brightness, at 6pm the bulbs in my bedroom/office dip to around 4500K and 50% brightness, and at 8pm to 2700K and minimum brightness (we have a toddler).

In my kitchen/living room during the day it's the same, at 8pm to 4500K at 50% brightness and at 11:30pm to 4500K at minimum brightness.

I considered having the schedule run on when the sun sets, but in my location that is 10pm in the summer and 4pm in the winter so it doesn't really work at the extremes (I have the same complaint with tooks like f.lux as well :D).


I've been looking for a hardware solution to this problem so long.

I would kill to have like a programmable dimmer knob that can progressively adjust both temperature and power set to a fixed temperature and power at start (ie: 3750k at 100%) and finish (ie: 2500k at 10%) positions then it auto-calculates the right spot in-between.

Does anyone know if I could do this on a Pi?

I did it with a Pi, HomeAssistant, smart bulbs, and some Python, but it was not as good or as reliable as I'd like.

I wanted the lights to know what temperature and brightness to turn on to based on a schedule, but there wasn't a way for them to pull the desired state upon power-on; instead I'd have to detect power on and push a state change to the desired color temp. This meant they'd turn on at whatever last state they had on power-off, and a few seconds later (or frequently much slower) adjust to scheduled color temp.

This may have worked better had I used smart switches to turn lights on/off rather than wall switches that cut power to the bulbs, but I also had issues with the smart hub losing connection to bulbs occasionally, HA not reliably seeing power-on events for the bulbs, and HA losing connection to the hub. (Bulbs/hub were Ikea tradfri; perhaps other manufacturers' products are more reliable).

This was a couple years ago; I haven't looked into it much since - maybe it would be easier to do now.

Phillips makes a bulb that does this with those temperatures almost on the mark. I put them in every fixture in my home and got lutron casetta smart dimmable switches. Drops to 50% at sunset and 20% at bedtime.

Hey Michael, thanks for commenting! (And thanks for making f.lux which I've found incredibly useful for years!)

I strongly agree that it's important to change light levels throughout the day; it's very obvious that bulbs this bright suppress my sleepiness. I should probably add this to the post's FAQ so that people don't do anything irresponsible.

I do think that some people might get a mistaken impression of how strong the evidence is when you claim that "cancer progresses much faster in constant light" without citing or caveating that this is based on one study, in rats, where the treatment shone a light on them for a full 24 hours.

> Our clocks don't work that well with more than about a 14-15 hour photoperiod ("day")

Regarding only natural light, what would be the ideal latitude region for psychology year-round?

At 30°N and 30°S days swing between 10 hours to 14 hours through a year. Would it be ideal to live in between 30°N and 30°S, or would you expand/shrink the range?

ps. Any resources on geography-psychology relationships would be appreciated.

SAD and cancer rates go up somewhat with latitude, definitely by 35-40N. But it is complicated because it doesn't affect everyone the same way. e.g., most of Europe is living above 40N.

In some cases, availability of natural light (due to tree cover, weather, and position in timezone) is as important as latitude.

I would guess that some people are simply better at living at more extreme latitudes or with a weaker circadian signal. Several researchers have tried to use particular genetic markers to explain this, but I don't know a specific one that explains all of it.

> cancer progresses much faster in summer than in winter

This sounds fascinating. Do you have a source?

Btw, I love f.lux and have been using it since probably like 2010, possibly earlier. Its basically the first thing I install on any computer.

These are pretty bold claims. I'm not sure if a single communication from 15 years ago is enough to to build such a relationship. Please note, that any claims about "cancer" in general are hard to make, because they have such variety of phenotypes.

I made a note about Blask's 2005 work below, which is the main work I was thinking about. We can say that constant light is bad for cancer.

But as I'm refreshing my memory, winter can cut both ways. For some, it makes more darkness and more melatonin (which is thought to be protective of cancer). For others, winter results in not seeing enough light, and so melatonin amplitude goes down. And so this is associated with increased cancer.

My immediate thought is that is very closely related to sleep.

Hey Michael, first off I love f.lux and use it every day. I have a question relating to your comment: is there any reason not to have an extremely high lumen smart light fixture which progressively dims throughout the day and syncs with the solar cycle in the user’s area? That seems like the holy grail, and is simply a combination of Philip’s Hue/LIFX tech and the bulb in the OP.

Yes, I think that's right - we have 20,000 lumens in our office and it is pretty cool, except you can't see your screen when it's up all the way. I still prefer daylighting, because it changes the angle of the light and everything else through the day.

A couple things: I don't think you necessarily need more light in the morning if you have a well-defined (artificial or real) sunset, and if you've reduced light at night to a lower level.

We add light to the morning so that we can add as much light as we want at night - the two cancel out. But I think it would be better to really try to design the evening to look different than the workday. Lights should change, and they mostly don't.

There are for sure some people with long internal clocks (DSPD/N24) who need very custom lighting schedules to have a normal schedule, but I think if we make the overall signal stronger (more day to night contrast), a lot of the differences between night owls and early birds would become smaller.

They are using these underground to help simulate the true daylight experience. Basically a smart led version of a traditional roof sunlight: https://www.youtube.com/watch?v=aJ4TJ4-kkDw

Wow. It’s hard to believe that’s real.

"Yes, I think that's right - we have 20,000 lumens in our office and it is pretty cool, except you can't see your screen when it's up all the way."

Can you elaborate ? 20k lumens doesn't seem like that much ...

I see residential interior decorating guidelines at about 20 lumens per square foot for non-intense areas like living rooms[1], etc., which implies a 1000 sf office (smallish) conservatively lit with 20k lumens.

I just put 11k lumens of background lighting in a 400sf living room and I am worried it's a bit under-lit, given 12 foot ceilings.

I think that perhaps you are talking about something different than I think you are ?

[1] As opposed to kitchens where I see 50-80 lumens per square foot recommended...

My office is not that big, so yes it is quite intense there. But, I should say just that system is capable of delivering >500 vertical melanopic lux (which is a metric based on holding the meter vertically rather than pointing "up"), and this feels remarkably bright.

> cancer progresses much faster in summer than in winter

Can you point me to a reference for this? This pattern is surprising and I would like to learn more about this.

I think the seasonality data may be mixed, with more diagnoses in the fall and spring in human subjects studies.

This particular claim was a hypothesis from Richard Stevens, with some experimental work that shows rapid cancer growth under constant light (in human tumors implanted in mice) by Blask (2002, 2005): http://www.nel.edu/userfiles/articlesnew/NEL230802A03.pdf

The main mechanism in the Blask work is thought to be suppression of melatonin.

The rat study you cite here is certainly far from sufficient evidence to make such bold claims or implicated that indoor lighting can promote cancer development in humans! I think you should edit your statement in the top comment further to make it clear that you are not referring to any solid human data here at all.

Thank you for f.lux, I enjoy coming home early, being with my family then working later in the night. Having f.lux installed on my computer has improved my quality of sleep tremendously

However to tie it to the article : do you know of the variability between individuals of lighting effect on morale ? I know a lot of people getting depressed when winter comes, but it has 0 effect on me

I love f.lux and I've used it for years, however I have to bring up the one feature I dislike...the "alarm clock" type feature which supposedly could be turned off by an option, but in reality there was no option to turn it off. I was able to turn it off on one computer but not the other, both windows OS!

Are there any lamp brands you can recommend that allow me to set the bulb temperature, but that don't require an internet connected hub and app on my phone? So RF or something.

I believe IKEA's smart bulbs are local only, not internet connected: https://www.ikea.com/us/en/cat/smart-lighting-36812/

(I've been meaning to try them out but haven't done so yet, so I can't vouch for the quality in detail. Their stuff is usually solidly mid-range and very good value for money, though.)

Thanks for the tip. I am a big IKEA fan, their value for money is, as you say, good in general.

Do people in e.g. the arctic have these problems during the summer, when the sun never sets?

A lot of them have very strict routines about going inside and using super-dark curtains at particular times. I've gotten to email some people above 60N and it is just what they do.

In the old Aschoff "bunker" studies where people allowed to keep the lights on as long as they wished, they concluded the human clock was 25.1 hours. We know this isn't true if you provide a proper light-dark cycle, it's much closer to 24. But I treat this old work as a cautionary tale - we can be 25.1 hour animals if we leave the lights on all the time.

I grew up at 66N (Northern Iceland), for what it's worth blackout curtains are rare to nonexistent there in my experience. Sometimes bedroom curtains don't even cover the whole window, and during parts of the year it's daylight outside 24/7.

When I've traveled there with foreigners who grew up at more southern latitudes with year-round "normal" day/night cycles I've sometimes had to tape black trash bags to the bedroom windows for their benefit.

What do the locals do? We simply get used to sleeping in daylight. One way of doing that is getting used to sleeping on your back with one of your arms across across your face and over your eyes (such that your elbow is between your eyes/on your forehead and the back of your palm touches or is near the opposite shoulder).

I still sleep like that out of habit at more southern latitudes out of habit, and a quick poll among some fellow Icelandic expats where I live when this came up the other day revealed that they do too.

In Spain we have lots of problems and room for improvement in many issues, but one thing I can be proud of is that the vast majority of homes have blinds that close completely and get zero light in. I often wonder why it isn't like this elsewhere, and especially in northern places where it would be especially convenient (I get that it's possible to adapt, but wouldn't it be nicer to be able to sleep in any position?)

Of course, the problem with this is that I'm totally spoiled and when I travel abroad (1), I can almost never sleep well. Hotel rooms in most countries have blackout curtains but if you're used to Spanish blinds, even those let a lot of light in from the edges. And don't get me started about LEDs...

(1) PS: Now that I come to think of it, I don't sleep well when I travel within Spain either... curiously, people's homes have blinds but hotels typically don't.

I'm assuming you're talking about the exterior metal curtains common to Spanish houses (persiana con cajor exterior [1]).

Those are only incidentally to block out light, they're mainly for thermal management. You put them down in the middle of the day to keep hot air and direct sunlight out.

A system like that is completely unsuitable in places that get snow. If you installed it on a house in Iceland it would at best freeze solid, and more likely explode as snow slush made it into the joints and froze solid. You don't even need to go outside of Spain to see that, go up to Monachil in Granada and see that no house in the area uses that system.

Of course you could install blinds like that behind double-glazed glass, but at that point you can just buy normal blackout curtains. Since you don't need it for thermal management (any amount of external heat making into the house being a luxury) there's no point in installing it on every window in the house.

So then for practical purposes you're left with only installing blackout curtains in the bedroom, but if you live that far north you're going to need to just get used to sleeping in the light anyway. What are you going to do if you go on a camping trip where you sleep in a tent and forget your blackout eye mask, just not sleep?

1. https://www.google.com/search?q=persiana+con+cajor+exterior

As far as I know, they are for both purposes. In fact, most of them can be lowered enough to provide thermal insulation but still let light through (they have small horizontal holes for that purpose), but if they are lowered more, the holes close and they block light. So it's definitely a planned feature, not a side effect.

In most of Spain they are installed outside of the window glass, but in the north they are installed inside, so they are protected from cold and wind. As in this picture, for example (just found at random): https://www.inmoatuel.com/piso-exterior-de-90m2-reformado-a-...

But they still seal the window hole perfectly so they block light much better than blackout curtains. I suppose the problem with Monachil is that it's a cold place, but within Andalucía which is generally warm, so they don't have a tradition in the area of installing blinds inside as is the case, for example, in Galicia, where you can definitely see them even at high altitudes where snow is common.

When I go on a camping trip... indeed I just don't sleep well, I'm afraid :)

The picture you linked to is taken in A Coruña whose mean temperature is around 10 degrees higher[1] still than relatively mild[2] climates near the arctic circle.

I didn't mean to suggest they weren't also used to let light in. I've spent a lot of time in houses that have these installed, letting light throughout the day while keeping the house relatively cool is a constant balancing act. I meant that without the need for the thermal management you'd go for other sorts of systems, such as internal blinds.

I.e. I'd expect that in places like A Coruña you'd use these for thermal management in the summer with the outer windows open, whereas in Iceland there's maybe 1-3 days of the year where the outside temperature is anywhere near the inside temperature.

1. https://en.wikipedia.org/wiki/A_Coru%C3%B1a#Climate

2. https://en.wikipedia.org/wiki/Akureyri#Climate

I live at 59.3 and dark curtains are not too common, most house I saw have some kind of shades or fluffy curtains, both don't block light

> they concluded the human clock was 25.1 hours

You keep quoting questionable or out dated research, the 25 hours figure seems to be wrong [0]

> Early research into circadian rhythms suggested that most people preferred a day closer to 25 hours when isolated from external stimuli like daylight and timekeeping. However, this research was faulty because it failed to shield the participants from artificial light

[0] https://en.wikipedia.org/wiki/Circadian_rhythm#Humans

Yes this is exactly what I said - it is only a cautionary tale for what we could do wrong, not basic biology.

I usually eschew labels, preferring simply "sleep weirdo", but I fit the pattern of "Non-24 Sleep-Wake Disorder", with my sleep and waking times drifting progressively later at semi-regular intervals. The result is that I'm often awake through the night. (This week I'm going to bed around 9:00 a.m.)

Even with that, I've noticed some pretty big productivity swings based on the season. Interestingly, this hit me very hard in moving from where I'm from (southern Texas) to Michigan (for college) and then to Berlin (for the last 14 years). Right now, in Berlin, which is approximately as far north as Edmonton, Alberta, Canada, North America's northernmost large city, the sun sets before 4:00 p.m.

I usually try to spend a month of winter closer to the equator to keep my productivity levels and spirits up.

Between the seasonal adjustment (where the sunset time swings by a full 7 hours in Berlin) and my sleep weirdo-ness, I've also come to love me some artificial-sun-level lighting.

I tend towards 400-500 watt halogen bulbs. I have an up-firing light of approximately that wattage in every room of my home.

A question for other folks that compensate sunlight with artificial lighting:

Do LEDs really work for you? I'm a wanna-be hippie, and I'd love to use energy efficient bulbs and have tried every generation of them, but the light just doesn't do it for me. I keep going back to halogen as the sweet-spot between black-body radiation and energy efficiency. I can immediately spot the difference between an LED or CFL and incandescent bulb. I've had some success in mixing them in about 50/50 ratios. They've already been banned for sale in the EU, but I have a stockpile that will last me a decade in a pinch. Does anyone else struggle with the light quality from modern lighting?

Have you tried buying higher CRI LEDs? Halogens have very good color rendering compared to typical LEDs (100 vs ~80 for typical LEDs), but you can buy specialty bulbs with closer-to-natural spectra. I haven't dug much into off the shelf options, but one of my coworkers did and got excited about these lights: https://store.yujiintl.com/collections/high-cri-led-lights/p...

Hmm, that's sounds interesting and is at least an avenue to chase down. Sadly, next to my stockpile of halogen bulbs, I also basically have a lifetime supply of LEDs and CFLs that didn't pass muster. Maybe someday I'll have a really big garage that needs lighting. ;-)

The specific bulbs you linked to have a crazy high lumen rating, which appeals to me despite their high price.

I have noticed that name brand (and pricier) LEDs tend to do better for me. Some of the ones from e.g. Phillips and Osram put out a different quality of light (though it's still not up there with incandescents for me).

There's also this company, but I haven't pulled the trigger myself because of the cost:


Their standard socket bulb prices aren't bad (about $15 if you get a 6-pack), and they do produce bulbs for the slightly different European sockets, but shipping to Europe is about $40 (plus I'd then have to deal with picking them up at the customs office, which is always a multi-hour ordeal). But assuming I haven't found something decent before I'm next in the US in March, I'll pick some up then.

Thanks, that's interesting -- I'll have to buy one of their bulbs to compare to my other LED bulbs.

Looks like the lowest cost might be the 100W COB LED ... It would need a bunch of supporting equipment but it might be possible to do it for around $2 per watt.

Edit: I guess the normal bulbs and the led tubes could also come in around $2/watt and they can take mains power so that might be an easier but physically larger route.

I kept seeing the following problem with several different LED lights we were installing until I found one that was better: Take your phone and take a video of the lights when on (normal framerate should be enough, try slow-motion for further investing). Most lights I got from the local hardware store were flickering quite visibly. And I'm pretty sure it's a design insufficiency of the adaptor electronics attached to it, not the LEDs themselves. If I were to guess I'd say most of them probably go the simple path of cutting off the negative half of the AC cycle and feeding the rest directly, without a decent low-pass filter to the LEDs.

What worked reasonably well for me, was getting a normal DC power supply and a few LED strips (with fairly dense and higher power LEDs) and installed that in a room. That light feels pretty satisfying to me.

All that being said, I don't like anything even remotely close to daylight level ambience when I am working on a computer screen; the custom lighting I set up is for other living areas.

> I fit the pattern of "Non-24 Sleep-Wake Disorder", with my sleep and waking times drifting progressively later at semi-regular intervals.

This is a disorder now? I thought that's just how humans worked. It's how I work, anyway - if nothing else (like work, or children waking up at 5:30am) forces me to get up at a set time, I'll slowly cycle through until I'm going to bed at sun-up and waking up at 3pm.

I also thought it was normal.

Once for a couple of weeks I worked and slept as I felt like it. I ended up doing ~30h days. It was the best I’ve ever felt while being quite productive.

Unfortunately it’s not an option to do that long term.

It's great as long as you have no need to deal synchronously with any other humans.

https://en.wikipedia.org/wiki/Circadian_rhythm_sleep_disorde... that affects about 3% of the population.

The most cited paper on this is "Stability, Precision, and Near-24-Hour Period of the Human Circadian Pacemaker" (Czeisler, et al., 1999), which found that their 24 subjects between ages 13 and 65 had a mean circadian rhythm of 24 hours 10 minutes with a standard deviation of 8 minutes. Earlier studies were affected by artificial light lengthening cycles. Non-24 sleep-wake disorder in sighted people is not researched well enough to have accurate rates: there are only about a hundred cases in the existing studies and the disorder is not well known and probably underdiagnosed.

I think there have been studies where people in sleep labs without any natural light, clocks or other cues mostly tended toward ~25-hour days. The "disorder" part is when you can't help but do this, despite having access to all those cues and despite all efforts to the contrary.

No this isn’t typical. Are you particulary good at coping with jetlag?

Depends which way I'm flying. :) Flying westwards is fine, I can absorb the extra hours without much noticing them. Flying eastwards is awful.

I identify with the parent, and I've always been fine with "jet lag" to the point where I often don't even notice it or it only lasts until I go to bed the first night.

(IE; Sweden->Los Angeles, the inverse is harder but not as hard as most people seem to have it, it knocks some colleagues out for an entire week)

A week? Yeah I'm scragged for a day if I'm losing hours on a flight (although this is exacerbated by the fact that if I'm flying east I'm usually going from my 'starts around 9ish' desk job to 'be on site by 6am' minesite commissioning, so I'm effectively waking up at 2am when I'm used to waking up at 8:30am) but after that I'm fine.

Out of curiosity: Assume the answer is yes?

I have this disorder as well, it started when I was 11 or 12. I've tried every trick that sleep doctors recommend (literally, everything) but nothing works. The only way I was able to maintain a 24-hour sleep cycle though school was to drink coffee every morning and take sleep medication regularly (from the age of 12). That was enough to make my sleep cycle 24 hours, but it didn't improve my sleep quality, so I spent most of school in a significantly sleep deprived state (only getting 3-5 hours of sleep a night). That's a pretty miserable state to live in for an extended period of time.

When I started living on my own and I was free to choose my own schedule, I decided to stop "forcing" myself onto a schedule that my body didn't agree with and instead let my sleep schedule naturally align with my circadian rhythms. It took about a month to get adjusted, but I remember waking up one morning and feeling like I had got my first full night's sleep in my entire life - it was an amazing feeling!

Since then I've been living on a ~ 24h 50m sleep cycle and I sleep like a baby (8 - 9 hrs) every night. Interestingly enough, I've discovered that I'm actually a morning person, when I used to think I was a night owl. I'm very grateful that I have the freedom to choose my own schedule (I'm a remote-only contractor), I dread ever having to go back to a 24h sleep cycle.

I haven't noticed much change in how I feel with seasons (I live in southern Ontario, Canada), but there's a big difference in my energy level depending on how much light I get in the morning (my mornings, not Earth mornings). If I don't get a good amount of bright light in the morning I feel sluggish and tired all day, regardless of how much sleep I got. If the sun's up when I wake up, I go for a run or walk every morning; otherwise I have about five 10,000 lux LED lights around my apartment that I turn on all morning (I have two right next to my desk in my peripheral vision, just like the person in the article). The LEDs are pretty weak, but they're enough if they're a few feet away from me. I start turning off all the lights and closing all the blinds in my apartment about 4 hours before bed or else it'll keep me up.

I would be very interested in reading more research about the cause of Non-24 Hour Sleep-Wake Disorder in sighted people; as far as I'm aware there's very little research on the matter. I didn't even know it existed until last year, I expect some more research and awareness could be very helpful to people in the same boat.

Are you familiar with Piotr Wozniak's writings on sleep? I think you might find them interesting https://supermemo.guru/wiki/Science_of_sleep. DSPS is pretty common and fairly natural considering how we've diverged in habits from our ancestors.

What does it mean to be on a "24h 50m" sleep cycle? Does your sleep time progressively become later until you're nocturnal, then continue looping round until it returns to a "normal" night's sleep?

Exactly. 24h 50m is the average I calculated after tracking my sleep for over a year. On average, the time that I wake up and go to sleep gets later by 50 minutes each day relative to a 24 hour day.

I'm currently waking up around 2am local time. In two weeks, I'll be waking up around noon.

If I lived on Mars, where the day length is 24h 37m, I would feel right at home (aside from the lack of breathable air).

Could you please briefly explain the way you cam calculate your cycle based on average time sleeping? I have to get up using alarms at times that are not related to how my body is feeling about waking up, so I don't see how tracking that would help.

I tend to start sleep later and later until I reach some dynamic equilibrium between sleep deprivation due to fixed wake up times and the tendency to delay sleep onset.

I almost never use an alarm anymore - there's no need to, my body knows exactly when to wake up because of my circadian rhythms (that's the way humans were supposed to work after all - we didn't evolve to use alarm clocks). Of course, the time my body "knows" to wake up is almost an hour later than it should, and that's the whole problem.

The way I calculated the length of my circadian rhythm cycle wasn't by tracking the duration of my sleep, it was by tracking the exact time I naturally woke up every day. I then calculated the difference in the time I woke up each day (the daily "drift" in my sleep cycle). The median of those "drift" values turned out to be around 50 minutes, which means my sleep cycle is 24h 50m on average.

It might seem like not using an alarm clock would bias those measurements, but that's actually not the case. Even assuming the fact that I don't use an alarm clock makes me sleep in (it doesn't, but for the sake of argument), that would only effect the average time that I wake up every day, not the average drift in my sleep cycle. There are a number of other biases that would have to be accounted for if this was an actual study, but for my purposes that's accurate enough.

In reality, everybody has a different natural circadian rhythm cycle. If that cycle is close enough to 24h, then you'll have a "normal" sleep cycle. If it's long enough to cause insomnia / delayed sleep onset but your body can still "compensate" (that "compensation" mechanism is called entrainment in chronobiology), then that's called Delayed Sleep Phase Syndrome (I'm not a doctor, but that sounds like what you're describing). If your natural circadian rhythm cycle is so long that your body isn't capable of "compensating" to a 24h cycle, then that's called Non-24-Hour Sleep-Wake Disorder. Delayed Sleep Phase Syndrome is far more common than Non-24.

As for this point:

> I have to get up using alarms at times that are not related to how my body is feeling about waking up, so I don't see how tracking that would help.

I'm afraid tracking it didn't really help me at all - I simply tracked it because I was curious. My "solution" was simply to let my body do it's thing, but that's not an option for most people who have fixed work schedules or other commitments every day.

Seeing this long later, but thank you for the detailed reply.

I absolutely hate LEDs. I don't know why, but I just hate them. Especially the white ones.

The yellower are similar to a mid summer day and I can cope with them, but there's always this feeling of something being off. I guess it's connected to us humans being used to non-sun light being (similar to) a fire - mainly consisting of yellow/orange tones.

Also, I don't exactly remember from where I read this, but when a (town) changed the street lights to LED's, a lot of people started having sleep problems, and it supposedly was connected to the blue light emmiting properties of (lower quality?) LEDs.

> I absolutely hate LEDs. I don't know why, but...

I'm really interested in the why. Most people don't seem to care. My in-laws have CFLs in their living room that just feel horrible to me, but they're none the wiser. I'm very curious as to what it is in the light spectrum that seems to matter to some (small) class of people, myself, and it sounds like you, included.

The word for "fire-like" light, used in my original comment, is "black-body radiation". Stars, fire and incandescent bulbs put out a similarly shaped spectrum.

A few years ago there were scientists working on a black-body (i.e. incandescent) light bulb that reflected infrared emission back to heat the filament and re-emit as visible light making for a theoretically efficient incandescent bulb. I check every year or so to see if the research has been commercialized, but have always been disappointed:


Edit: Another point on the blueness of LEDs is that they seem to amplify macular degeneration, which has made my grandmother mostly blind, and I know based on gene sequencing that I have a more than 50% chance of developing. While most of my rejection of LEDs is based on them not looking nice to me, I am also worried about hurrying the onset of blindness in old age for persons like myself who carry the genes for macular degeneration:


Well, fire does not shine by black-body radiation. Lamps try to emulate a black-body because that's what color standards dictate and labs test against, but fire has a very bad color resolution. (By the way, daylight isn't also like black-body radiation, but it's much closer than fire.)

Personally, flickering makes me ill. I am used enough to 120Hz to survive it, but any other frequency is bad. Also, the 6400K LEDs look way too blue, much bluer than the Sun. That may be because I have a relatively rare kind of color blindness.

huh that's really interesting about fire's emission spectrum!

regarding 6400k - even assuming the sun's spectrum matched the black body spectrum of its surface temperature, its surface temperature is closer to 5770k. but even taking that into account, its spectrum doesn't quite match 5770k in space - and the atmosphere changes it even further. https://en.wikipedia.org/wiki/File:Solar_spectrum_en.svg

I really don't know how much of that change us humans can perceive, but my thinking is that 5000k is probably closer to the center, at least. it might be more of a saturated color, though, since the spectrum is more pointed vs the very flat but spikey spectrum in that plot.

> Well, fire does not shine by black-body radiation

Flame, not entirely; embers, yes.

> I'm really interested in the why. Most people don't seem to care. My in-laws have CFLs in their living room that just feel horrible to me, but they're none the wiser.

My first suspicion would be the spectrum of light emitted, which is mostly reflected in the CRI [1].

CFLs in particular "cheat" by having a couple really strong peaks of very specific wavelengths that to your eyes (or reflecting off a white surface) look like the right color, but when reflecting off anything they have reproduce colors inaccurately -- it's beyond my ability to explain this in detail, but you can see this in some spectrum charts [2].

LEDs vary greatly in this respect, so if you're concerned, you should really be paying close attention to the CRI (and note, as CRI goes up, so does cost).

[1] https://en.wikipedia.org/wiki/Color_rendering_index

[2] https://www.ledsmaster.com/color-rendering-index-cri-versus-...

psa: you can get this thing, which isn't the most precise ever, but lets you see this spectral information about lights for cheap: https://www.amazon.com/EISCO-Premium-Quantitative-Spectrosco... - I got one and WHEW CFLs' band lines are really obvious. Also, I feel kind of tickled that I saw the band gap on LED lights before seeing a description of what it is in the source for the spectrum plots in this article (the source being https://commons.wikimedia.org/wiki/File:White_LED.png ).

I'm a bit miffed it seems so hard to find lights that don't have this problem, though. maybe we can improve it by getting the word out that these cheap little diffraction devices can give you a pretty good approximate reading of the smoothness of the spectrum of a light source. Hmm, I just realized I have yet to take this to home depot...

I'm really curious about those MIT incandescent bulbs. If they worked well and haven't been brought to market, it's possible that contacting the people involved in creating them could have good results in making them happen. Perhaps they could be convinced to prioritize it if a case can be made that it can have a significant positive impact on the world?

I like CFLs but hate most LEDs because they flicker! It's at 120hz, but I can see it and it drives me crazy.

Another thing I hate is LEDs with such a sharp point source them make speckle patterns like a laser.

But I did find some good LEDs and I'm using them now, them seem as good as the CFLs they replaced. Had to return a bunch of bad ones though.

My experience is that daylight-blue (5000K–6500K) LEDs or fluorescents feel unnatural unless they are quite bright; something in the animal part of my brain worries that the sun is broken. Low CRI bulbs also feel off, which makes intuitive sense I suppose.

I've found that high CRI warm white (2700K) bulbs are the way to go for me.

I have these at home, and I like them a lot: https://www.lowes.com/pd/GE-Relax-60-Watt-EQ-A19-Soft-White-... . (I bought them at Home Depot, but it looks like they no longer sell them.)

You need more that 60 watt.

It might be the flicker. Most LED lights I've found have pretty bad flicker: https://youtu.be/eWut5Qx6mkM?t=285

Phones are pretty bad too. Check out the Nexus 5x vs Pixel 3: https://drive.google.com/drive/folders/1f7-ChF4zFiisFdmEZVEz...

I found that Phillips non dimmable bulbs are the cheap and don't flicker.

Same experience here. I film all the led bulbs I purchase in 240fps to see if they flicker at 60Hz. I discovered that Philips bulbs NEVER flicker. Occasionally I stumble upon other cheaper brands that don't flicker, but there is never any guarantee. I might buy 2 batches 6 months apart of bulbs from the same model from the same third party seller on Amazon, and they slightly changed the product (updated packaging and markings on the bulb) and one batch doesn't flicker whereas the other does... So I end up nowadays just buying Philips.

A ton of LEDs flicker at 120hz which makes me crazy. It's like a strobe effect that I only see when moving.

The Edison style are often like that. They use four led elements configured as a full wave bridge rectifier but because the diodes are themelves the load ther is no place to put the lowpass and so you get 120.

Oops I wrote 60Hz but meant 120Hz.

I think it's the lack of heat (AKA infrared light) also reaching your body, which, when receiving bright visible light, is evolutionarily wired to expect heat coming from that same light source (e.g. the sun). Feels off for you to be cold and for it to be super bright.

I'm currently experimenting with 'emulating outdoors' in my office environment for health and productivity, and so far I've started with an 8,000 lux fanless corn LED (bought off Amazon) hanging from the wall and it's brilliant. I immediately feel 'happier' when I turn it on in the morning, every single time, but after a while, I agree, something feels off. I feel 'cold', temperature-wise.

I have read (no source for now, just from memory) that having bright white light indoors (e.g. fluorescent) without the presence of other light spectrum (including possibly natural UV) can actually increase cortisol.

So I think that more than just visible light is needed to complete the picture. Careful, low enough levels of UV and infrared that doesn't point directly at you but more 'above' you like the sun itself, may be the optimum to get some benefits but not cause dangerous skin or eye problems.

Possibly due to the lights being driven from AC power making them pulse. You gotta find ones with big enough caps that they don't flicker, or buy some grow LEDs and a decent DC driver to emulate the sun.

A lot of the LED bulbs out there are poor quality, they have bad CRI and they flicker. I’d be interested to know if you find high CRI LEDs better, especially warmer ones.

I've recently got amber ones, at around 2000k and they are such a pleasure. Many of them look like classic incandescent bulb too, imitating wires in normal bulb.

Unless you source high quality LED luminaires, they look terrible let alone flicker... most of the shit on amazon is just that unfortunately.

The flickering. It’s not nice.

LEDs do not flicker.

Cheap led light systems do

Yeah, I get the impression there's a massive difference in quality between cheap and good LEDs, and it's not always easy to tell which is which. In theory, when done right, LEDs can have a perfect spectrum with no flicker. But many manufacturers cut corners.

"I keep going back to halogen as the sweet-spot between black-body radiation and energy efficiency."

Permies put out an article arguing for incandescent if your goal is light AND warmth: From https://richsoil.com/electric-heat.jsp - "This last one was the most important. A standard incandescent light bulb heats something to the point that it glows white hot. So I used this to heat myself and it doubled as a light source. And, I should point out that in a few months this light bulb will be banned by the US government. It is already banned in many countries. The comedy is that it is being banned to save energy. And yet, I think people can save far more energy by keeping it."

But incandescent light bulbs are a terrible heat source from an efficiency perspective. You're losing an incredible amount from the power plant, to the transmission, and pretty much everything else. Once it arrives, yes, the energy gets converted nearly 100% to heat eventually, mostly in your proximity. But a heat pump or even gas heating is far more efficient if the goal is to get heat.

If you have electric heating anyway, it makes no difference. You’re right that a heat pump would be better, but for many people it isn’t an option.

If you read the article, the author is focusing on the scenario heating a person while sitting at their desk, not the entire room.

In those cases you could argue the methods he prescribes are more efficient, though not practical for times where you want to heat up a whole room or home.

Interesting, reminds me of several articles on this site.


Yeah, love that article. I'm slowly building a large foldaway 'kotatsu' coffee table using a mixture of ikea (ovraryd) and amazon/ebay parts (foldaway hairpin legs, centerpost, kotatsu heater).

yup. like my low flow toilet that i need to flush twice...

Not anymore, technology has improved a lot! Get this one (or variant on it): https://smile.amazon.com/gp/product/B001KAFKRM

I've owned it for 5 years and it's clogged only once (it happened in the first week of ownership) then never again.

And one time (in 5 years) I needed to flush it twice.

That one is bragging about a 3" valve. There are other toilets available now that have 4" valves.

> Do LEDs really work for you?

All of them suck (objectively and subjectively) compared to full spectrum lamps.

> Does anyone else struggle with the light quality from modern lighting?

Besides poor light quality, another issue are dimmable LED lights which are often just PWM'd at a couple hundred Hz. This is a migraine trigger for me.

I'm going through this currently. It doesn't help that my inspiration for work can sometimes make me work through the night, which 'messes up' the following day. I'm my own boss, so can do this without repercussions, but I sometimes think I'd be more productive if I stuck to a more regimented sleep schedule.

I did have a normal-ish sleep schedule when I had my job in finance, but I remember feeling exhausted all the time, especially weekday mornings.

I certainly feel healthier now, just sleeping when I'm tired. On occasion this can mean missing a bunch of daylight. Not sure if that's actually all that healthy though...

It sounds like we're similar. My personal, entirely unscientific self-diagnosis is that I have a well-defined circadian rhythm like most people, but it's easier for me to override by obsessive interest than for most people. Even sans coffee (which I'm pretty careful about as I near 40), if I'm really interested in something, I can work on it in a block of time up to and over 24 hours. Naturally that futzes with my sleep. My "tired-impulse" is relatively weak compared to my "this is exciting" or "people are waiting on this" impulse.

Before I ran a company (about 1/3 of my career), I mostly managed to hold things down on a "normal" schedule, but I was always tired and frequently had performance reviews where the only complaint was my tardiness.

There are some interesting data points: while camping (which I do a lot of -- several weeks per year), I do tend to sync up with the sun, even though in northern Europe that means waking up before 5:00 a.m. There's also limited screen time. (Though I did work from tent on a remote beach in Crete for a month this year.)

I picked up some melatonin tablets in my wife's home country (where it doesn't require a prescription), and will experiment with that in the near future for I-really-need-to-sleep-now times, but as much as my sleep is weird, it's so intertwined with my personality now that I'm hesitant to medicate it away even given the chance.

Try some of these:

[1] https://www.osram.de/ecat/Professional%20LED-Lampen%20mit%20...

[2] https://www.osram.com/ecat/Professional%20LED%20lamps%20with...

They work fine for me, while i'm considering myself sensitive to all sorts of flickering, or "strange" colors. Don't be afraid of the "professional", they aren't much more expensive, 8 to 9€ per piece instead of something like 4 to 6€ for the others. They make a warm light but are too bright to directly look into, even the candlelike things at only 2 Watts. I bought the first ones out of frustrations with broken CFLs maybe three to four years ago (not a single CFL lasted over a year!), their warmup times, noise, and such, as a test. Now the only other light i have is some flexible/movable Halogen spot light, which i rarely use. Anything else is that stuff from the links. It just works for me, while even saving some energy. So far no defects. Instant on. No noise. No flickering for me. They are commonly available. I got mine at the Bauhaus. Test one or two and see for yourself. Though you won't get a single thing with 500 Watts. They max out at 75W equivalent while using 8Watts. So you'd need many of them.

> Berlin, which is approximately as far north as Alberta, Canada, North America's northernmost large city

Presumably you mean the city of Edmonton. Alberta is about twice the size of Germany. ;)

> Do LEDs really work for you?

Mostly, yes. It's easy to buy halogens that all have pretty much the same good quality of light. LEDs you need balance CRI, R9, dimming quality, flickering, etc., but I've found LEDs that work for most of my use cases.

Eek, yes, meant Edmonton -- typed too fast. Corrected.

You're probably never going to see this so I'll keep it short. I just want to say that your disorder might be due to a circadian rhythm disruption. Instead of just focusing on light, you should also focus on darkness. Try not to look at any unnatural light after sundown.

If you want to cheat for convenience, they make red glasses that block all blue and green light which can interrupt your melatonin. You want to find red glasses that have barriers on the side to prevent the light from coming through from the sides. They sell 2 different kinds on Amazon. I found the effect to be very profound. You get tired quickly and when you fall asleep, you feel like you're in a coma and wake up refreshed. Unnatural light after sundown really does interrupt your sleep cycle.

Good morning to you Scott, Much the same here, I never could keep a rhythm.

>I tend towards 400-500 watt halogen bulbs.

Are those really 400-500 watt or just give light equivalent to 400-500 watts worth of incandescent light bulbs?

You might want to try a 28-hour day.

Did you mean Edmonton?

> The lamp flickers when other current-hungry appliances turn on

That should not happen. There is likely a problem somewhere in the electrical system. We used to have similar problems in our home, hired an electrician to help track it down... they gave our home a clean bill of health, but called the city... long story short, they found a problem in a transformer a couple blocks away from us. It had been impacting many homes, and they were shocked nobody had call it in as a problem before us. But apparently, everyone just thought, "Oh, flickers are normal when appliances kick on, right?"

Ah, this does seem quite plausible--my local power utility doesn't seem super competent, and my landlord doesn't seem particularly excited about maintaining my house :)

Wait. This isn't normal?!

I have the exact same thing happen here.

Whenever the bathroom heat lamp (well, lamps - 2 x 275W, 240V) turns on, I clearly see the lights in whatever room I'm in dip for maybe 10ms.

This used to happen with the old incandescent lamps from years back, but still happens with the newer CCFL types too. Has never not happened.

Huh. Something to keep in mind, particularly for whenever I'm in an environment with access to competent management/engineering ;) (I just know if I tried bringing it up with the current landlord I'm almost certain I'd be authoritatively told "that's normal"...)

I have the same issue every time the air compressor is turned on.

Is it not? I thought the motors inside of them caused that to happen?

Not according to the electricians who fixed it for my neighborhood.

Yes and no. Motors pull a much higher current when they're starting up than when they're running, so your wall socket voltage is going to dip when you start a motor. That said, it shouldn't dip enough to be particularly noticeable, and if it is then that can point to a problem in your power supply or wiring.

It might not be noticeable, or as noticeable, with a 150W incandescent bulb, which has thermal inertia in the filament. What's going on here is that these are fast LEDs.

On the flip side, mains-voltage LEDs generally have built-in power supplies which should lead to them actually showing less dimming during brief voltage dips.

Some mains-voltage LEDs in fact flicker at twice the mains frequency! This is visible when you turn your vision, or when something moving, like rotating fan blades, is illuminated.

The Phillips flood lights in the track lighting fixture in my kitchen are like this.

They contain some sort of very light-weight power supply to rectify the line voltage and adapt it to the LED, but there is no LC capacity in it to even smooth out the 120 Hz ripple.

There is no room in those bulbs for the electrolytics/inductors that would be required.

If you got enough of a draw from one device in your house that lights flicker, the safety switch should probably trigger.

The corn cob bulbs are very hard on the eyes unless you have some kind of diffuser.

I found a better solution is a 1 to 7 bulb splitter with 7 150W equivalent bulbs for a total of about 15000 lumens.


I'd be curious trying a spotlight (something like this: https://www.amazon.com/Primelux-8-inch-14400-Lumens-Driving/...) that is farther away from me.

The parallel rays of a spotlight would probably look and feel more sun like compared to the radial rays of normal bulbs.

I was thinking about getting this myself, too. I looked at dozens of corn cob light bulbs on eBay and AliExpress and found none with a CRI of 90 Ra or better. In Germany you can buy cheap 13W E27 LED light bulbs at Aldi every January with 95+Ra.

I’ve got 5 of these in my office. https://www.superbrightleds.com/moreinfo/industrial-led-ligh....

They are fantastic. I have several more of the longer fluorescent replacements in my wood shop. More light is magnificently better for energy and productivity in pretty much all environments.

I would still prefer skylights to bulbs, but these are a good solution for dark rooms.

Have you checked the CRI? Its not bad but not great either - 84 as per the specs.

There are a lot of Chinese cheap LED fixtures available on Amazon and no one seems to care about the CRI - it defines how colors appear in your room.

I bought some LEDs on Amazon and uploaded charts showing the wavelength distribution. The LEDs were awful and the charts made it very clear why. Amazon deleted my review and the item currently has 5 stars.

The best lights I have are some warm white LED strips from eBay. I glued them to the top of my workbench (it's a desk under some shelves basically) and it is a joy to work there. I measured them and they are pretty close to daylight in terms of CRI, something like 93 if I recall correctly.

The only thing that makes me unhappy is the current state of LED strip driving. I have a device that takes 120VAC and turns it into 12V for the light strips. It "interprets" the output of a triac dimmer in front of it to PWM the LEDs. Very wasteful and stupid, but there are no constant-current drivers that just let you use something to adjust the current output unless you build your own. (A friend of mine did just that; the lights are amazing.) PWM naturally causes the cheap capacitors in the DC/DC converter to make noise, so I pretty much never dim them.

Did the CRI for the LEDs you bought match that which was listed or did they even have a CRI stated? I'm willing to spend extra for high CRI but so much junk on Amazon outright lies about being UL listed, it wouldn't surprise me if they started lying about CRI once they realize it is a selling point.

Yeah, they claimed a CRI of 91 but it was closer to 70, I think. (Measured with ArgyllCMS and a Colormunki Photo.)

You could tell it was wrong just by looking. It wasn't a measurement that only appeared by looking at it with expensive equipment; you turned the lights on and instantly though "this is completely unusable". I only measured it to see how bad it was.

For those like me who didn't know what CRI is, it stands for color rendering index, and appears to represent how well the colors of things appear when lit by a light source.

[0]: https://en.m.wikipedia.org/wiki/Color_rendering_index

[1]: https://en.m.wikipedia.org/wiki/High-CRI_LED_lighting

Unfortunately it’s relatively limited, there are only a small number of wavelengths which are measured, and you can game it by tuning the peaks of the bulb to match up with those points.

There are better measures, but almost no manufacturers use them. I think just having the output spectrum and eyeballing it will do the trick.

CRI is the real key. A lot of manufacturers have cut back on quality in the war over pricing. Philips had the very best consumer LEDs then started pushing their “SlimStyle” garbage with a CRI of barely 75. Amazon is littered with products advertising CRI “over 90” but in reality are probably barely breaking 80.

What's the effect of different CRI on humans? I'm curious what the difference is between good and bad CRI, and if a person would be able to notice it.

I don’t know about physiologically but perceptually it feels dimmer and drearier. The room feels gloomy although there is certainly sufficient “quantity” of light. It’s almost like everything takes a few milliseconds longer to resolve.

That, of course, is apart from the actual being unable to see certain shades of color and the spectrum being skewed one way or the other.

CRI is useful, but it's not a perfect measurement. You can make an 80 CRI bulb that subjectively looks pretty good for many use cases, or a 90+ CRI bulb that looks bad. (e.g. with poor R9 measurements.)

What kind of fixture are they in?

84 CRI is not very good.

How much UV light is emitted?

It is possible to recycle old LCD monitors into articficial windows that have a very real feel to them. The diffraction lenses for the backlight direct the light in a way that makes it appear sun like. Of course this is only about the direction and not about the spectrum, but it might be possible to find LED strips with a high CRI. Any feedback would be appreciated.

DIY directions on YT: https://www.youtube.com/watch?v=8JrqH2oOTK4

Great video. The light produced is so soft and convincing.

Do you know of any consumer version for sale? It feels like there should be a product offering with how little changes need to be made to convert a tv or laptop screen

Came here to recommend the same thing, and I think it's an even better solution, as it makes the light appear to come from further away and the diffusion prevents bright spots.

> Other than “we’ve been doing it for a while,” there seems to be no reason to expect that being in a 100x dimmer environment all day wouldn’t be awful. Indoor darkness seems to be one of those things that we don’t question only because it’s been that way forever.

An interesting question is: what average light levels existed in the Environment of Evolutionary Adaptedness for humans, given the types of things our ancestors spent their days doing?

Consider: many humans were savannah persistence-hunters; but many humans were also jungle and arboreal-forest gatherers. And either way, we probably liked to find trees to climb and use as protection, when we could (as chimps do.)

So, given that, are human brains calibrated for "sun-baked African plains" daytime brightness, or "under a forest canopy" daytime brightness, or somewhere in between?

It could be different for everyone. My ancestors likely grew up in dense equatorial jungles and caves, not plains. A lot of my family members perform much better in dim light. My optimal seems to be 2-7 PM on a rainy day.

I'm not sure about the genetics of it (none of my family members seem to agree with me), but I also perform computer tasks far better in a dim light. A friend of me, who is also a night-owl, has the theory that when we get exposed to daylight, our cavemen heritage's natural reaction is to feel like going hunting/food gathering. While I don't buy the argument verbatim, I feel more like doing outdoor/physical activities on sunny days.

I've been wondering the same about ‘normal’ levels of activity. Pretty sure most of us are horribly motion-deprived by the standards of hunters or gatherers. But then again, presumably someone also was staying behind at home, and I don't remember any folk wisdom saying housekeepers are all dystrophic.

I mean, I think most of us spent a lot of generations inside caves, didn't we? I think we adapted.

I'm extremely light sensitive, and so have built quite a few DIY solutions in order to alleviate my S-A-D symptoms.

One of my systems may have been a bit over powered and over used, as I developed what a doctor friend later diagnosed as "light induced hypomania", which was quite an experience.

Experiment at your own risk.

Could you provide specs on this system and clarification on how much use you believe was too much?

It was tens of thousands of lumens, directly next to computer for at least 8 hours a day. Beat the depression, drove me mad.

Would you have any pictures? This sounds fascinating (with the caveats you mentioned of course)

That sounds delicious

> One of my systems may have been a bit over powered and over used, as I developed what a doctor friend later diagnosed as "light induced hypomania",

I would note that light-therapy lamps for SAD have clear instructions to not sit in front of them for more than an hour-or-so at a time. Your body really does just go on producing more dopamine as long as you're sitting there.

That sounds a bit crazy. Does that mean you get the effects of Adderall from a light?

> Isn’t using this much power bad for the planet? Coal has a carbon intensity of about 1 kg CO2e / kWh (source), so one coal-powered lamp-day produces 2 kg, approximately one-third to one-half of a cheeseburger (source).

I don't think the author gets a cheeseburger everyday! And even if it is true, it adds on top of the cheeseburger.

And he starts with incomplete data, as only 40% of the energy in coal is converted to electrical energy. Then, from that 40%, additional 5% are lost in transmission and distribution, so what reaches his flat is around 38% of that, or 2.6 kg of coal per kWh.

So one his coal powered lamp day is already, assuming his other factors are correct, around 5 kg of coal(!) or almost 2 tonnes of coal per year! Which is a huge footprint. Imagine that the same guy would have to just unload that much coal, delivered to his house, once a year, and to feed his lamp every day, would he still consider it "nothing"?

Burning hydrocarbons roughly produces 3 times more CO2 in mass, resulting in 15 kg of CO2 for one of his "lamp days". In volume, it's even more easy to see the size of that all: a kg of CO2 has volume of more than half of cubic meter (0.56 actually), so his lamp day produces more than 8 cubic meters of pure CO2 per day or 3000 cubic meters of pure CO2 per year. Enough to fill 25 European city flats with pure CO2 (otherwise, in atmosphere there's only 0.04% of CO2 -- that's 400 ppm talked about here: https://e360.yale.edu/features/how-the-world-passed-a-carbon... ).

"But the atmosphere is huge" -- yes it is, but still the change of CO2 concentration affects us. There is an experiment that demonstrates the effect with ink:


We don't see CO2 with our eyes, but the molecules of it efficiently block the parts of heat emissions (it is just electromagnetic radiation, just like visible light is) which would cool our part of the atmosphere. So the heat around us increases, just like it's warmer under a blanket (we are practically under always thicker blanket, and we add to it with our actions, just like adding the droplets of ink in that video). Everybody should understand that much.

Every few years some company/research lab will claim they have a technology that’ll simulate sunlight coming through a window, but it still isn’t something one can buy.

The moment we can buy eg a 3ft by 1ft panel that we can just mount on a wall and feels like an actual sunlit window for a few thousands dollars or so, interior design is likely going to get radically different.

(And that’s not to mention the sci-fi dream of a screen that can simulate a window looking out on any landscape, but these are probably way further out)

When we were building out an office that only had a few windows I went looking for, and found one of these. Unfortunately it’s $60k instead of $2k. According to reviewers that saw it, it’s amazing.

Apparently it uses some nanoscale material that produces (mimics?) Rayleigh scattering.


I don't think the trick is the Rayleigh scattering, I think it's the parallel rays. Notice how all the pictures have well defined shadows of parallel rays?

I started to look into this and from a brief readings of the patents it seems that the magic sauce is Rayleigh scattering by shining an LED light onto a sandwich of two transparent panels and a solid layer of silica aerogel. I couldn't find the thickness of the aerogel layer, which will determine if you get more blue or red light out of the sandwich.

It's quite ingenious, the aerogel consists of silica nanoparticles with a size distribution smaller than the wavelength of the light, leading to Rayleigh scattering to dominate. The non-uniformity of the aerogel structure is what gives it the more nuanced lighting, it lets a bit of yellow light, a bit of this and that, making a non-monochromatic source that resembles sunlight very well.

hmm, that explains the color, but it doesn't explain the coherence. it is cool though, aerogel makes a lot of sense to use for this.

I remember seeing a DIY video on how to make a light panel that simulates light coming from very far away using large Fresnel lenses from old TVs


Have a look at https://www.coelux.com/. (Not affiliated, I just like the concept. But AFAIK their 'panels' still are 50+ cm thick and outside your as well as my price range.)

I was hoping someone would mention this. Have you seen the product in person? I’d be very curious whether it lives up to how good it looks on video.

For those unfamiliar, the panel simulates Rayleigh scattering to give the appearance of a blue sky, with a simulated sun and close to incandescent color rendering. Too bad the price tag still seems to be in the tens of thousands though.

What you want is, first, a strong source and second parallel rays (not radial rays like omnidirectional light bulbs). The later is the hard part. Because of conservation of etendue, (https://en.wikipedia.org/wiki/Etendue#Conservation_of_etendu...) it's difficult to have very parallel rays unless the source is small and/or far away. Most spotlights still have relatively wide angles of radiation. I wonder if something like this might work if well placed: https://www.amazon.com/Primelux-8-inch-14400-Lumens-Driving/...

The main limitation office workers face when working in sunlit environments is the reflective glare on most, if not all, computer displays.

I conjecture if/when e-ink displays match conventional LCDs in response time and resolution there will be a movement towards “outdoor offices”.

Imagine being able to work productively in a garden, atrium, sun room, or any other natural outdoor environment. Many of the pitfalls of indoor, fluorescent lighting would be diminished. The Bay Area and its relatively fantastic weather would be fully realized as a perk to tech workers.

> The Bay Area and its relatively fantastic weather would be fully realized as a perk to tech workers.

I built a Linux box with an rpi3 I had lying around and a boox max3 I got. So far working out doors:

+ Insects: you have no idea how many seem to come out of no where just to annoy you.

+ Sun: you need sun screen even on overcast days. I look like an overripe tomato.

+ Rain: when you least expect it one tiny cloud will go past and drop enough rain to soak you and any exposed electronics.

+ Children: loud and obnoxious, seem to be everywhere at any time.

+ Homeless: turns out they really like parks and assume you're one of them if you're there for more than 30 minutes.

There's probably more but those are the things I had to deal with yesterday. Today I just sat on my balcony instead. Sunlight is nice, but there's a reason why people have been building shelters for millennia.

Another limitation is battery technology, but that's changing. I have a 20K mAh USB C power bank and it's made me more mobile. It can almost fully recharge my laptop or recharge my phone 6 times before needing to be recharged.

> I conjecture if/when e-ink displays match conventional LCDs in response time and resolution there will be a movement towards “outdoor offices”.

I think you're right. Working outdoors with a laptop is already better than indoors. It'd be easier with e-ink.

Mitsubishi's synthetic-light "windows" announced last year appear to be pretty spectacular. I want one.



Many offices, mostly in places like Amsterdam, Paris etc. in art-deco building, take pride in the amount of natural light they're pumping in the office. With mirrors and sunroofs (with sails as shade) and a few plants, you can really give a beautiful, open air feeling to your office.

This, feels like crutches.

edit: there is something to be said about starting an hour earlier and finishing an hour earlier if the sunset is early. It may be a good idea to live with the natural rhythm of the earth rather than some clock on the wall.

When it gets dark by 3:30pm and there’s so much cloud cover it never gets past dusk, you’ll appreciate the crutches. Biggest downside to living in seattle atm.

Ugh yeah I wanted to kill myself more and more every year I spent in Seattle. Moved somewhere sunnier and that urge just vanished.

Yup. Half my life is now managing that problem. I’ve gotten better at it but I really do need to move somewhere sunnier, especially for the winter.

A 2’ square sun lamp from the Indoor Sun Shop helped but only so much.

You exaggerate. Seattle looks bright compared to half of Europe.

Compare the sunset times with London, Copenhagen, Oslo.

You had 1½ hours more daylight today than Copenhagen.

Compared to Copenhagen sure. I’m from California - the lack of daylight is not something I’m used to at all.

You're going to be awake for 16 hours on the average day. We get 9 hours of daylight a day in New York in the winter, so some portion of that is going to be spent in darkness no matter how wonderful your windows are.

In the summer, things are much better; 16 hours of daylight. So you really can put yourself on a schedule where the appearance of light wakes you up and its disappearance puts you to sleep.

The second sentence of the article explains that he's trying to figure this out because the sun sets in Boston at 4:15 PM now.

You don't live in Boston, so you don't understand. It's dark by 4-4:15 here

It gets dark in Amsterdam around the same time as well. And Amsterdam overall gets less sunlight as it is much more north than Boston.

> Isn’t this expensive to run? It draws about 2 kWh per day (assuming 8h of usage), which costs about $0.30 at typical rates. This is comparable to one day of a modern fridge or one load of laundry. Power is cheap, folks!

But it will offset your other heat sources.

If you live alone, you can treat it like a room heater and lower the temperature for the rest of your home.

Since the bulb is radiating heat some degree of heat, you may be able to keep the temperature of the room you’re in lower without feeling cooler.

The fan is a bummer though.

2 kWh is kind of nuts. Yes, it's in the same order of magnitude than the fridge and washing machine, but those already tend to be the largest consumers of electricity in a household, unless you're doing cooling or heating with electrity. And a modern fridge suitable for 2 to 3 persons will be more like 0.5 kWh per day and a load of laundry is < 1 kWh.

Three standard 60 W incandescent bulbs will draw just under 2 kWh in 10 hours, and not too long ago it was completely normal to have many more of these on at the same time.

And if you're trying to light your whole house this way you may need three or four of these bulbs. I guess you'd just keep whichever is in the room you're in on though.

While LEDs do produce some waste heat, efficient ones convert 50%+ of the input power to visible light, so it's less than you might think (compare to ~10% for incandescents, iirc). The bulb I have doesn't have a perceptible effect on my room temperature (although I guess it probably does cause the thermostat to turn on a tiny bit less often).

That visible light is still mostly ending up as heat in your home though, other than what spills out of windows.

Can you give an example of an LED bulb installation with 50%+ luminous efficacy? Perhaps you are thinking of the raw lumens/watt coming out of lab-grade LED material without considering other system draws (cooling, power conversion)?

The poor CRI is really a dealbreaker for me. What I've done for my DIY "ghetto" photography lighting is I mixed 4 led strips of 2 temperatures from 2 different manufacturers. Much better, but still poor, to be frank.

I've been looking at this, but it's not cheap:


I'm most pleased with the page's non-Javascript based image comparison slider. Nice one.

It didn’t work for me on mobile, and only in a weirdly rendered way on tablet. I would have liked it if there were just plain image links available.

Doesn't work for me :/

> Instead of having a strong urge to stop working whenever it got dark out

Lots of us have the opposite urge! My theory is that darkness induces a combination of mild-depression and subliminal sort-of-pleasant fear that HELPS us ADD-iacs focus waaaay better especially on tasks like coding and debugging.

PLEASE never use OPs conclusions when designing public spaces, this would kill the productivity of people like me!

> ...that HELPS us ADD-iacs focus waaaay better

That's interesting since I too can only really focus at night or late evening. I knew someone with ADHD and I could see some glimpses of me in their behaviour.

Not sure if it generalizes. And if it does, not sure if it does do all ADDs:

I'm ADHD without the Hyperactive component, subclinical, ADD or (ADHD-PI or Inattentive-ADHD more precisely). Or with the H component purely mental/internal, which kind of sucks because you get the "benefits" of ADHD plus a higher propensity to slide into an unhealthy ultra-sedentary lifestyle.

That too seems familiar. I'm not sure if I am hyper or not I used to be but I think I outgrew it finally but it took decades.

The ultra-sedentary lifestyle seems very accurate. I make a plan for example study for my CCNA and here I am six month later finally doing it. This is with daily reminders to myself, set alarms, notes, collecting study material etc. not just thinking but some prep.

And when I am in the zone I tend to overdo it often way too much detail. Brevity is not my forté.

Self medicating with caffeine seemed to be working it would help me focus even during the day. But I think it contributed to my fatty liver I never had it before or after my energy drink addiction. Now it's one cup of coffee per day max.

then sharing what helped me a lot: (1) forcing yourself to do physical exercise: running + gym + ~ twice a month mountain hiking works for me, (2) drop ALL refined sugars (caffeine and other stuff can stay, even some alcohol, but NO SUGAR!) [after ~3 months of (1) + (2) your brain will be totally reshaped and waay better performing], (3) probably the most important: build external accountability - like in make sure there are friends/lovers/family/clients/employees/etc. that know about your goals like "study for my CCNA" and will shame you, hold you accountable or even sue you for failing to do what you said you'll do, (4) meditation (mindfulness) - it's important to get it that when you feel the absolute least inclined to meditate and you couldn't stand doing it is when you most actually need it and you'll get the biggest benefit from it, and (5) a general more "stoic" attitude of accepting that "life is pain" and of "learning to love the types of pain that lead to growth and development"... for me mostly the pain of forcing myself to finish stuff and to stay-on-track instead of goind sideways or starting something new again and again.

But I don't want my office to be as bright as the outdoors. I want it to be dark.

I’m the same. It is very interesting how people have radically different experiences with light. I feel the best in winter with cloudy dark days. My wife is completely opposite and so are most people it seems, they feel good in sunlight. I’m afraid of it and the summer heat.

I’ve recently started buying high CRI LEDs and I’ve noticed a difference in how the room looks and feels. The older LEDs while can be bright just aren’t bright in the same way.

Light temp is mentioned. It's worth looking at CRI which gives you an indication of how much of the spectrum is represented. E.g. 18,000 watt HMI lights for film are tungsten (3200k) and have a CRI of 99 which is insanely good. But they're 18000 watts. Many high output low watt LEDs have horrendous gaps in the spectrum.

Unfortunately to test CRI you need a $1500 color meter. I'm lucky enough to have one and access to a range of film lights and it's crazy how much CRI varies and the horrendous gaps in the spectrum for some lights.

Do you really need a single mega-light? Seems to me for a lot of situations having more dispersed lamps would work better and light a room more evenly and allows more fine tuning of light levels.

This is great to see, for I've been wondering lately why being outdoors is more pleasant (in clement weather) than indoors. Is it the unpredictability? the humidity? the noises? the sunlight?

In that same vein, I've wondered how to simulate outdoor conditions inside, especially something approximating more light, in part because I think a family-member is mildly susceptible to seasonal affective disorder. I can't wait to see what the author has to say.

I’ve also been wondering this lately. In particular if I’m feeling slightly ill, I feel much better outdoors. At least in my experience, it isn’t necessarily sunlight as I think it’s also better at night. And it isn’t super “fresh” air or hearing birdsong or the wind rustling through trees (positive things I associate with being outdoors) as I live in a city. I hadn’t thought about humidity. Currently my best guesses are:

- a correlation between being outdoors and stood up and moving

- colder temperature (but I guess in the summer when it’s not stifling it can be preferable to be out so maybe this is wrong)

- some psychological/placebo effect

- less CO2 in the air. I can’t really put any numbers to this hypothesis as I don’t have a meter.

Furniture, flooring and carpeting contribute towards indoor air pollution. For example, glues used in laminate flooring can release formaldehyde into the air for decades at room temperature or warmer.

Obligatory Tom Scott video on indoor CO2 levels:


(or the studies themselves https://www.gwern.net/docs/co2/2015-stafford.pdf and others)

I.e. I think air quality is likely a big factor. When it's windy and I get good airflow through my room it doesn't feel as different from outside. It's probably not only CO2, but also dust, VOCs, funghi, microorganisms, etc.

My guess is more sunlight, less CO2.

> Isn’t 5000K very “cold” light? “Cold” lights (above 2700k, the typical incandescent color temperature) have a bad reputation, but 5000K is actually less cold than the sun (which is about 5500-6500K)

I don't think the author is taking into account that a lot of the blue spectrum from sunlight is scattered in the atmosphere resulting in a much 'warmer' light hitting us.

I'm building my own bright lamp recently. I bought 20 Philips bulbs, 1500 lumens each. I attached sockets for them in 5 by 4 array to a 40 by 40 cm board. I'm gonna attach the board to the ceiling and hang transparent ikea kitchen cabinet lining below the lamp so it spreads light a bit so it's not too bright to look directly at it.

It takes about 250W and gives over 10 lux from 1m.

My current setup are two 100W led reflectors pointed at the white ceiling. Thet give over 200 lux at my desk but are very greenish. It shows up on the camera (with any whitebalance) and is even slightly visible to the naked eye.

New lamp should be significant upgrade in terms of colors and light strength. Total cost is around 50$ because I bought lightbulbs very cheap and used cheapest available sockets, plus few hours of very relaxing work. Reflectors were even cheaper. Just 20$ and hanging them was less than an hour.

I'm in Canada, so the bulb he selected isn't readily available. Anyone found a good one with high CRI?

Also, it's not clear to me how important the watts/lumens are in a single bulb vs. multiple. If I can only find 100W/6000 lumen bulbs, are three of those the same as one higher wattage bulb with 18000 lumens?

I'm a little bit worried about the amount of <450nm blue light emitted by these things. Even if they look warm white, there still might be a very prominent blue peak. Blue light at about 420…430nm has a very prominent retinal phototoxicity, and prolonged exposure can lead to loss of vision.

I have an LED build-out that was inspired by the Lumenator (https://arbital.com/p/lumenators/). Since that post 1600 lumen A19 bulbs have gotten much cheaper and so I run 9 of those in a 3-fixture tree lamp with 4-way spliters.

I also experimented with commercial string lights and ended up hanging one strand with around a dozen bulbs over an ikea pyramid coat rack behind a Japanese style paper screen and that made for a bright and beautiful standing lamp.

I also found that having a timer/dimmer was a great way to wake up in the morning. Much better than any alarm, but if you don't sleep alone you need to get your partner on board.

I highly recommend trying a build for yourself!

Anyone know a way to set up lighting that is (tiers of priority):

Tier 1:

* high-CRI

* high-lumen

* dimmable, and can be set up to turn on slowly in the morning

Tier 2:

* not glary

Tier 3:

* easy to install

* inexpensive

* power efficient

This immediately reminded me of the scene from Silicon Valley where Gwart is working directly in front of an extremely powerful light, which Jared rushes to shut off because it's too bright

This is a pretty big claim:

>The effect was huge: I became dramatically more productive between 3:30pm and whenever I turned off the light. Instead of having a strong urge to stop working whenever it got dark out, I was able to keep working my normal summer schedule, stopping just before dinner. I estimate the lamp bought me between half an hour and two hours a day, depending on how overcast it was.

Does anyone know if people really do become "dramatically more productive"? It should be something that could be tested...

I use a flood-light bulb in my torch floor lamp (pointing up to the ceiling). This fills the space with a lot of light without having a bright spot that cannot be stared at.

There are a bunch of calculations about this in Dercuano, in particular based on indoor-growing setups, but with a chair in them instead of a plant pot. It turns out that fluorescent tubes or especially halogen bulbs are cheaper than LEDs, and at the small scale of a desk and chair, the extra energy is affordable.

One big problem for hackers is that very few screens are readable even in 10klux, let alone 100klux direct sunlight (although that has the disadvantage of burning you).

What about transreflective displays?

Annoyingly I've still not seen one IRL. They go black and white, apparently, but are still 100% legible.

I've been positively impressed with the readability of OLPC XOs in sunlight, which I think are what you're talking about. It's too bad they weren't willing to consider a capitalist rather than communist model for making them available; I think they would have been popular. I think there are also color transflective displays but I've only seen them in photos.

E-ink displays are even better than the XO, and machines with them are much more easily available, but most of them only come in the surveillance-capitalist hood-welded-shut model. They use orders of magnitude less energy than transflective LCDs for some usage patterns, but refresh slowly.

Oh my god this is the most insane description of OLPC I’ve ever seen.

Not “neoliberal private-public partnership bull crap” but somehow “communist”? Lololol.

The only way they would sell you an XO was if you were going to give one to every kid in the country as part of a centrally planned education policy. While this could have been very effective, there's no denying that it's structurally much more Lenin or Castro than Pinochet, Eisenhower, or even Yew. If that's not obvious to you, it says more about your knowledge of political ideas than about my mental health; perhaps you only know “communist” as a pejorative due to US jingoism?

A capitalist approach to the problem would have stopped them from failing as completely as they did but also would have eliminated the possibility of the far-reaching transformation of education they were seeking to impose.

I like natural light, but office ceiling lights hurt my eyes.

I always feel better with bright lights and have considered trying something like this before, however two things always stop me.

1) Concerns about eye damage.

2) Concerns about skin damage.

I was thinking about that. Putting a microsun in your peripheral vision? Can't that mess with your eyes in the long run?

I strongly, strongly prefer indirect light.

I don't really need this, but if I did, the fan would probably be a deal-breaker for me. Instead, I'd spread this out of multiple weaker light sources. That would spread the light out more, spread the heat of course, and probably also make it easier to tune it to exactly the light level you want.

I do understand the desire to have daylight-level lighting in your house. It's a shame it takes so much power.

I'm planning to move to northern Alaska in a few years, so this information is very welcome.

I know I will need a setup against S.A.D. because it's very real up there.

If I get 4 of the 120 watt fanless bulbs, the daily cost of the bulbs and running them 8 hrs/day (assuming 30,000 lifespan) would cost only ~$0.29/day. That's an easy choice against SAD.

i have seen articles about excessive blue light from led bulbs being bad for your long term eyesight

i wonder if this has similar problems

When David Chapman's "You need more lumens" article came out (which is referenced in the original post), I was on the verge of building a similar setup but at some point read similar concerns on light therapy. Not necessarily effects on circadian rhythm, but retinal damage from staring close to the amount of light.

From https://meaningness.com/metablog/sad-light-lumens : "The clinical studies were done with 10,000 lux provided by a bank of fluorescent tubes in a box that directed most of the light forward. Very roughly, if you are a couple feet from a box like that, 1 lumen produces 1 lux." It'd be interesting to see studies that have these lighting setups but don't involve staring towards them.

The setup that I'm happy enough with: a set of LIFX bulbs in normal light fixtures. They seem to put out more lumens (1100 lumens) than other color-changing led bulbs. If you don't want 'smart' lighting but still want the ability to change color temperature, Philips makes "SceneSwitch" lights (800 lumens) that will cycle color temperature based on turning on/off.

I use a bank of blue LEDs aimed up at a white ceiling for the winter so I spent some time looking into the question. I didn't find anything very definite or concerning.

Note that there is a whole snake oil thing going on with blue light these days. Some of what we see about this stuff might be coming out of that particular industry.

Here's a good selection of research papers https://www.bioshumanlight.com/research/ . It seems like there might be some adverse effect due to the peak at the lower end of the blue spectrum. I'm waiting for the BIOS team to do a consumer product.

You can choose the colour temperature you want. Go for 2500K if you prefer less blue light.

The issue is that all white leds are made by shining a blue led at a yellow phosphor. Rhe blue and yellow balance to whatever temp you pick, but it is still a lot of blue.

It can't be much worse than being under the blue sky.

The blue sky's spectrum doesn't have a giant spike like this:


Yes it does! A giant spike in the blue region is literally what makes the sky blue in the first place. If there weren't such a spike the sky would be white. When you look up at the sky on a typical clear day, you're getting more blue light than almost any one of these indoor setups.

I see your point, I was thinking of daylight in general. But even the blue sky is not quite as spiky:


If you look at 460-470 nm where the spike is in the LED spectrum, it's much lower in the blue sky spectrum.

I think the concern is something about the relative amounts of blue light. Not sure what exactly but something like, the human perception of brightness and therefore the self protection of the eye is calibrated for natural light, so the pupil contraction, looking away, etc, is not done correctly in with that unnatural distribution. Anyway I don't know if that's a real effect, but that's what people (should) mean when they're talking about "too much blue".

Blue sky peaks between 500-550nm, this is 100nm more than LEDs, and much more diffuse.

this comment makes no sense. even assuming these things are equal, no one is just out there staring at the sky for hours on end for the small percentage of time the sky is actually blue.

So when you're at the beach, hiking, or driving on the highway, what are you looking at instead? The blue sky is going to be most of your field of view unless you just stare at the ground.

> But as of recently, it’s totally practical to fill your entire house with light that’s as bright as full daylight.

If you actually get your house as bright as sunlight, that means you have ~1kW / m2 of light in your house, which is impractical no matter how cheap and efficient LEDs get.

It's 445W/m^2. More than half of these 1kW is infrared.

I have a one-person office with glass walls in a coworking space. I've been looking for a lighting option that could possibly go above or below my monitor. I like the bulbs in this article, but I would prefer something directional so that I don't disturb my neighbors. Any suggestions?

If a 120-degree cone is good enough for you, try searching for "high bay lights"? E.g.: https://www.amazon.com/s?k=high+bay+lights

If you're willing to pay more for much better color rendering than mine, my coworker got excited about these: https://store.yujiintl.com/collections/high-cri-led-lights/p...

Note that it's 1/3 the lumens of my corn bulb, but the smaller angle means the lux may be about the same (or you might need more--I haven't tried these).

I'm not quite following - which direction do you want the light to shine?

I've got bias light setups at both my home and office with bulbs behind my monitor shining at the wall.

Actual outdoors light levels are much higher than 600 lux. More like 50,000-100,000 lux.


I built my own lamp for that and it cost me roughly the same. It's best to use metal-halide lamps as their spectrum is very close to the Sun's. It's important to get one with a uv filter though.

> stop working whenever it got dark out

Also, this might be a good urge - just stop working and stop trying to use light to squeeze every ounce of productivity from yourself until you burn out!

Anyone knows a decent and inexpensive way to measure CRI?

It seems like a simple enough problem with a prism and a monochrome CCD sensor, yet the cheapest device I can find is over a thousand dollars.

I wonder if it would be possible to write an app that could do it using the camera in a smart phone? The idea is that you would take some photos of some reference objects of known color under sunlight, and then take photos of those same objects under the lighting you are testing, and then figure out from the differences in the photos what the CRI must be.

The reference objects could be well known, easy to obtain household items, like a Pepsi can, a plastic bottle of Tide detergent, or box of Arm & Hammer baking soda. If the app included a reasonably good sized database of such items, there would be a good chance an average user would have a suitable set of color references already on hand.

I'm afraid that phone camera sensors are filtered to R/G/B, so a strong monochromatic orange, for example, may end up under-represented or even absent in the captured image, and will likely be hard to distinguish from an overlap between red and green.

Maybe modern image processing techniques can figure it out, but this seems like a huge problem to me.

> Why doesn’t everybody do this?!

Because I am more productive when it is dark or overcast. I get nothing done when it's bright.

I struggle to focus during the day and become more productive when it is dark.

I put a bulb like this only in my office. Has helped tremendously although you notice the difference in light color between it and the other lights in the house right away.

Any recommendations for spotlight table lamp in the office ? Looking for something when studying or working with room lights off and have only table lamp on for focus.

Kind of depends if you want something with integrated LED, something where you can screw in your own bulb, controls for dimming, what type of bulb, etc.

Personally I like the Ikea Forsa.

Any suggestion for smart light bulbs that work without the cloud? Don‘t want to have the server shut down in five years when the bulb still has 15 years in it.

Basically anything that can be flashed with open firmeware, such as Tasmota (https://github.com/arendst/Tasmota/wiki/Prerequisites) plus a local instance of Homeassistant. It needs a bit of work and sometimes soldering skills, but some devices can be flashed without soldering and in general it’s not too hard to do. :-)

Pretty much any zigbee bulb (Philips Hue, Osram Lightify, Ikea Tradfri, and tons of chinese brands) can be controlled with a cheap CC2231 USB dongle (~20€ including programmer for flashing) and zigbee2mqtt. I use homeassistant for a nice UI and Node-RED for automation.

this is a cool idea, but, the "is this bad for the planet" question seems like it's saying it isn't, but the comparison seems to indicate that actually it's quite bad for the planet, like very bad as far as lifestyle choices go? not to say that alone makes it a bad idea to do, but that a sane society would make buildings that capture and use more sunlight from the actual sky.

Rooms can even be brighter than the outdoors you know.

A colleague points out that the COB rather than SMD based lights have a better spectrum.

why though?

> isn't this expensive to run

comparing lighting to the consumption of a fridge, which is one of the most energy intensive appliances in an average house, is kinda ridiculous. Its like saying your lawn mower energy use is just fine because it consumes the same as your SUV.

Tl;dr: 2kwh per day for a light is a LOT. A standard led light would consume about 5% of that.

Three incandescent light bulbs that would have been standard 10 years ago would consume as much energy while putting out far less light. It's not a crazy amount of energy, standard LED lights just use so little energy that it's hardly worth measuring.

He states in monetary terms what it costs (30c a day), he's not obfuscating it.

That it costs a lot more than a regular light is neither here nor there (and I think anyone reading would have already guessed that).

text: fluorescent vs. an incandescent the graph below: fluorescent vs. _LED_

which is it? ;)

please do a follow-up on the affiliate revenue from this article

I can think of one more potential downside: light polution is bad enough already. In the author's case, 8x higher lux values. Insect population in Germany has dropped by about 75% since the 70's. A recent The Guardian article names light polution a key factor.

In principle, it's possible to shield your windows with thick curtains, but I know many people won't bother, some will openly boast them, and it may even lead to more of them being hung outside when people see how great they are.

I've kind of intuitively gravitated towards this ever since energy efficient bulbs started becoming popular. Instead of buying light bulbs to save energy per se, I've been buying the highest lumen efficient light bulbs in a given form factor and color temperature. I prefer 3000K light indoors, and daylight for working lights. My desk lamps have 100W equivalent daylight "high CRI" light bulbs. My house lights are basically the highest lumen lights available in each form factor be it candelabra, "regular" or "fake luminescent". The light fixtures that do not have light bulbs per se are the highest possible lumen rating as well. I figure I'm spending less money anyway, might as well have more light. Unfortunately R30 flood lights aren't generally available in power rating equivalents beyond 75W incandescent. If some marketing person from e.g. Osram, Cree or Philips is reading this, could you guys make 150W equivalent R30 flood light? Charge pretty penny for it, too. I'll pay. I want punishing light levels in my kitchen/bar area.

> For instance, cancer progresses much faster in summer than in winter [edit: cancer progresses much faster in constant light: winter vs. summer is complicated]

I don't know you, you make a very successful app that people really swear by. I have no qualms about people using f.lux, if you want to take your $4,600 laptop and $1,400 phone, on which people spent decades perfecting color reproduction, and you're paying $13/mo for Netflix to watch shows with multi-million dollar budgets and dedicated colorists, and you start watching Orange is the New Black and you make it yellow, that's your prerogative. So you don't need my opinions, so don't downvote just because you like f.lux and hate opinions.

Why did you have to say anything about cancer at all?

What's your personal line for, "I'm not sure" versus "This is actionable evidence for change in behavior?"

There is no clinically controlled evidence showing that interventional blue light, the kind that would be emitted from an LED, causes significant change in sleep quality with measures like minutes spent sleeping or an insomnia scoring system. I found this by searching "blue light" and "sleep" in clinicaltrials.gov. There are 4 or 5 studies on the matter, all negative results. The best way to interpret this evidence is that blue light will not make you sleep less or experience insomnia more.

Indeed, if you consulted a doctor, you might learn that waking up too early is the most common sleep disturbance in adults. This interferes with their daily lives, by causing them to feel drowsy in the middle of the day. That's why blue light is typically investigated as an intervention: to get people to fall asleep and wake up later. It just doesn't turn out to work for that. Its absence is not causing people to fall asleep earlier either.

You can't post in the flamewar style like this to HN, and we ban accounts that do it. Would you please review https://news.ycombinator.com/newsguidelines.html and stick to the rules when posting here? Note this one: Have curious conversation; don't cross-examine.

We detached this subthread from https://news.ycombinator.com/item?id=21661852 and marked it off-topic.

Someone might not hate opinions, just not appreciate the condescending tone in yours. You can still rely on the color reproduction of your expensive device when it matters.

I am a colorist, working on a very expensive Eizo and I still have f.lux on my machine. Obviously I added exceptions for all applications that are color critical and switch it of for an hour while watching a movie late.

For me it is just a natural sign that I shouldn’t go on forever on the computer. A way to bring the outside night/day cycle onto my machine, you could say.

> if you want to take your $4,600 laptop and $1,400 phone, on which people spent decades perfecting color reproduction, and you're paying $13/mo for Netflix to watch shows with multi-million dollar budgets and dedicated colorists, and you start watching Orange is the New Black and you make it yellow,

You're projecting way too much my man. What about the people with 70$ monitors using flux while coding at night to avoid fucking up their sleep cycle ?

> The best way to interpret this evidence is that blue light will not make you sleep less or experience insomnia more.

Type "blue light circadian rhythm study" on google and you basically only find studies supporting that blue light has a negative impact. I got a pair of anti blue light glasses recently and I'm 100% confident they indeed do reduce eye strain and ease falling asleep.

> Type "blue light circadian rhythm study" on google and you basically only find studies supporting that blue light has a negative impact.

But the moment you actually evaluate the studies, you'll find that almost all

* don't research blue light specifically, but light at all

* use unrealistic settings, like 4 to 8 hours bedtime reading (really)

* don't properly control for other known sleep affecting factors - and have small sample groups so they can't rely on the law of large numbers to solve that issue

* only measure melatonin, not actual sleep.

If you filter those out, the result ist basically zero quality studies.

I'm aware that many people swear that this works, but there really is not much scientific support for the theory.

Just for comparison: Full sunlight has up to 30000lux of blue light alone, while a very bright phone screen, set to all white, manages to output at most 300lux of blue light. There may be a real effect here, but it has to be weak, else people in scandinavia would have severe problems with their circadian rhythms.

> Full sunlight has up to 30000lux of blue light alone

Well yeah that's the whole point, blue light during the day is perfectly fine, that's how your body knows it's time to get shit done, sitting in front a screen after sunset during X hours _will_ fuck up your rhythm and blocking blue light will moderately mitigate the issue.

> else people in scandinavia would have severe problems with their circadian rhythms.

They do, just like you do when we switch to winter time. It's not a "problem" though, it's just our body's natural response. The problem arise when you shift your rhythm so bad that it impact your life / work schedules.

> If you filter those out, the result ist basically zero quality studies.

It's 100% a fact that blue light is the light which impact the production of melatonin the most, and that melatonin directly regulate your circadian rhythm.

Even on a cloudy winter day the sun's luminocity is about 20 times higher than a very bright phone display.

My point is that it is not reasonable to assume a strong physical reaction if the artifical light is so much weaker.

Note I'm not denying that there isn't a small effect! But it is probably dwarved in comparison to a change in eating habits, anti stress methods like meditation, sports and so on.

> It's 100% a fact that blue light is the light which impact the production of melatonin the most, and that melatonin directly regulate your circadian rhythm.

Then why don't most studies study effects on the actual circadian rhythm? My guess is because the results there are less clear with actual sleep (i.e. they actually studied both but only published the "good" results - a sad side effect of publication bias).

BTW the same is true for light therapy for prevention of winter depression: There probably is an effect, but it is so weak that rigorous meta analysises don't show real evidence.

My overall point is: There seldom are strong effects with most treatments of this kind. If there were, we already would have found out in pre science times. People are doing artifical lighting for thousands of years (think campfires, not LED lights).

Campfires are not artificial light. This is about blue light after all. You need to show some data to back your statements besides hand-waving all existing studies in one swoop.

Campfires are artifical as in they are not the sun, and people were exposed to their radiation for millenia. A fire obviously emits most of its electromagnetic radiation in the IR band, but it the emission goes down right into UV. This is not exactly rocket science, just point a UV/Vis spectrometer at a fire and see for yourself.

I'm not quite sure why you think I should provide data - I'm not the person postulating the effect of blue light. But anyway. You can find examples of the sun's illuminance all over the net, for example Wikipedia has a list here: https://en.wikipedia.org/wiki/Illuminance

Illuminance usually only covers the VIS part of the EM spectrum, so all of about 380-700nm. So 1/4 to 1/3 of that is the cold light part.

The radiation emmited from a heat source can be determined using Planck's law. A campfire is about 1400K. Here you can see that the higher the temperature, the more the maximum radion shifts into smaller wavelengths: https://www.spektrum.de/lexika/images/geo/f4f1653_w.jpg

So obviously the sun with about 6000k has way more UV and blue light parts, but a campfire is still hot enough that the absolute amount of blue light emitted trumps every phone screen. Yes, probably more than 95% will be IR, but than again the overall luminocity of a campfire is about 10kW.

For light therapy and winter depression I refer you to the excellent chochrane review: https://www.cochrane.org/CD011269/DEPRESSN_light-therapy-pre...

As to my criticisms regarding most studies about the effects of light: You can look them up yourself, most of them apply to the studies linked at f.lux's research links: https://justgetflux.com/research.html

For example the top two studies there:

"Exposure to Room Light before Bedtime Suppresses Melatonin Onset and Shortens Melatonin Duration in Humans" used 8h room light and didn't check for blue light at all. Neither did they look at actual sleep patterns.

"The human circadian system adapts to prior photic history" : 6.5h of exposure, no blue light specifically, only melatonin, no actual sleep measurments.

Again: I'm not the person making claims here. I would be delighted to see some proper studies showing the effects of blue light.

You are indeed making claims, and worse, generalizing a “lack of evidence” from a handful of unrelated studies you picked. It’s kinda obvious that not all of the studies linked in the flux page are about blue light specifically, but how electronic device usage affects health. The effect of blue light is widely accepted in the scientific community and not some kind of fringe theory.

On your campfire rant, first, luminosity is an astronomy measure, what matters for humans is the perceived light (lumens); nobody stares at a campfire directly and the little light bounced from the surroundings has very little blue; on the other hand, you literally stare non-stop for hours at the blue light from your phone, which makes the amount it emits a lot more relevant.

Well then, feel free to provide the research. You can then also send it directly to UK's General Optical Council because they actually sue companies who claim health benefits from blue light filtering [1] because there is no evidence for real-world effects. Or the EU's Scientific Committee on Health, Environmental and Emerging Risk, because they, too, say that there is no scientific base for the claim of real-world effects [2, page 90]. Or Cochrane review, who say that there is no health effect in case of macula degeneration [3].

Also please don't get personal. My campfire illustration was not a rant, it was an illustration of humans using light after sunset. It is a fact that they did so for millenia. It is also a fact that a 1300K hot fire emits blue light and even UV light.

And luminosity (lux) is just luminous flux (lumens) per square meter. Astronomers use that because luminosity is the measure used to compare different light sources. If comparing e.g. a ceiling light and a phone screen one uses luminosity because that makes both light sources easier to compare. It is also the measure most research on light effects work with. You could of course multiply the 1000lux of a bright phone display with its surface area in square meters if you prefer to calculate in lumens.

[1] https://www.aop.org.uk/ot/industry/high-street/2017/05/26/bo...

[2] https://ec.europa.eu/health/sites/health/files/scientific_co...

[3] https://www.cochranelibrary.com/cdsr/doi/10.1002/14651858.CD...?

It is important first to say that sleep optimization is not the only goal with these things, and understanding the timing of melatonin (not just melatonin suppression) is how we find out "when" the circadian clock is set.

The circadian system is loosely coupled to sleep, but it affects many other parts of the body. So if your rhythms are more than 2-3 hours out of sync you will definitely have more trouble sleeping, but this is not the only effect, which is why we work with people studying other topics also.

But overall, yes, I think we'll learn that regular room lighting has at least as big an effect as screens, and that's because in terms of "hours" people spend a lot of them in general illumination, but often only check their phone a few times, or watch a dark movie, which isn't very bright. It's only some of us that code all night in front of a bright screen.

Well there are several topics here: first, it is not true that people thinking about visual performance will find the right schedule for non-visual (circadian, alertness). The two are simply different concentrations, and we have chosen to focus our research and work on non-visual responses.

Treating sleep-maintenance insomnia is often done by increasing light levels at night to move rhythms later, and seasonality often requires different patterns of light based on the time of year. This is one of the reasons we have been asking people not to copy what we said in 2009 - some people need more light at night during the winter, not less, and timing needs vary greatly.

Our core audience for f.lux is younger people, and on average they need less light at night - the normal distribution of chronotypes has a very long tail in the direction of night owls, and their timing tends to be most at odds with society's requirements for wake times.

So I would say the product as it is today tends to be a better fit for the night owl half of the distrubution, and not extreme early types. But our main area of study is to generalize these schedules and lighting changes to be "right" for nearly everyone. It is rather difficult work, and you have to study the effects over years.

In response to this part of your comment, f. Lux allows you to specify programs like Netflix, VLC, Photoshop where it will automatically turn off 'you're paying $13/mo for Netflix to watch shows with multi-million dollar budgets and dedicated colorists, and you start watching Orange is the New Black and you make it yellow,'

Maybe there hasn't been any clinical trial, but this article suggests there were quite a few studies on it: https://www.health.harvard.edu/staying-healthy/blue-light-ha...

Not the cancer thing, but the blue light effects. Well actually it talks a little about cancer as well :p

It does mention that it is very preliminary:

Some studies suggest a link between exposure to light at night, such as working the night shift, to some types of cancer, diabetes, heart disease, and obesity. That's not proof that nighttime light exposure causes these conditions; nor is it clear why it could be bad for us. But we do know that exposure to light suppresses the secretion of melatonin, a hormone that influences circadian rhythms, and there's some experimental evidence (it's very preliminary) that lower melatonin levels might explain the association with cancer.

I'm just commenting to save what might be the saltiest comment I've ever seen on HN. I bet you're very fun at parties.

Please don't cross into personal attack, even if another comment was bad for HN.


The first part and general tone of the comment might be weirdly negative but their main point is valid

Quoting something as questionable as "For instance, cancer progresses much faster in summer than in winter" as a fact, without even a citation while signaling expertise is not exactly great.

I certainly wouldn't call GP's comment friendly, but I found it informative. Especially when you're trying to discuss things scientifically, I care much more about truth than kindness (not that you shouldn't try to be nice). I think replies criticizing flaws in comments are some of the most underappreciated tidbits of HN, because they are drilling into the details which are what really matters.

You can directly save comments by clicking on the timestamp. Then click "favorite".

Favorited comments show up publicly. You may not want to appear to exclusively enjoy barbed comments.

> I have no qualms about people using f.lux, if you want to take your $4,600 laptop and $1,400 phone, on which people spent decades perfecting color reproduction, and you're paying $13/mo for Netflix to watch shows with multi-million dollar budgets and dedicated colorists, and you start watching Orange is the New Black and you make it yellow, that's your prerogative. So you don't need my opinions, so don't downvote just because you like f.lux and hate opinions.

You do know there are lots of toggles and shortcuts to tweak or disable flux when doing something color sensitive, right? There's even an explicit `Movie Mode` that does the best it can (not much tbh) without noticeably affecting color reproduction for the average viewer.

I disable flux when watching movies, playing games, or editing (color) images. It's not that hard.

That said, I don't run it as a substitute for medical advice. I run it because monitors are fucking bright and I enjoy working at night with dimmer lights that match my environment. I'd use an eink monitor at night if they were affordable, large enough, and software supported them better.

Thank you for your comment, you have just saved me from taking in fake news. Any tone you take is fine when you're preventing the spread of misinformation, so keep being you.

It's a lovely article; in substance and style. I purchased a very cheap (£50) win10 tablet a few years ago, and content like this makes me very happy.

Applications are open for YC Winter 2023

Guidelines | FAQ | Lists | API | Security | Legal | Apply to YC | Contact