Hacker News new | past | comments | ask | show | jobs | submit login

f.lux author here, still slogging through the article. It is hard to understand the light levels used because they use "power" (uW, mW) from a laser, and not "irradiance" (uW/cm^2 or mW/cm^2), so which area they have concentrated that light over is hard to understand. All I can see is it is from a laser, so the irradiance could be extremely high.

The human lens filters most light at the peak of the given spectrum for free retinal (383nm), and so once you get to 450nm like an LED, the hazard data in the visual range is 100x less sensitive, see Fig 1 here:


It bears repeating that computer screens have <5% the hazard-weighted irradiance of a blue sky. (Can't make a direct comparison with lasers.)

Not sure the conversion to white light is correct, and it is unclear to me right now if this much retinal is available in vivo.

Thank you for your input, but I mostly want to thank you for f.lux. It has been one of my favorite things for a long time. It's silly how long it's taken companies to provide a native blue light filter and your product just kicks ass.

If you guys ever wanted to figure out your app on a smart TV (or apple TV), I know there'd be a lot of thankful people!

Also, I'd love to contribute to the ongoing dev of f.lux (I don't think I've had to pay for it?) How might I?

Check out the FAQ for a PayPal donation link:


Huge thanks to the author from me as well!

Thanks for telling me :) Paypalling now

thanks for the link. I have been using flux for years and recommending it to my friends too.. just donated a small amount..

On the AppleTV go to Settings > General > Accessibility > Display Accommodations > Light Sensitivity.

Then you can assign it to the Accessibility Shortcut and turn it on or off by triple pressing the Menu button.

Can't wait to do this tonight! This will help so much

Hook your TV up to a computer running f.lux or similar! I've used a HTPC for the past decade and more, wouldn't go back.

For my Samsung TV I use movie mode and also change colour temperature to warm.

I have the "blue light filter" on by default, on my android mobile. Don't know how effective it is, but it will at least help a bit.

On my Google Pixel, this is under Settings > Display > Night Light.

In Android Pie, it's also one of the 9 shortcuts you get when you pull down from the top twice on your home screen, right next to airplane mode.

I just want to remind everyone that LED backlights emit blue light that isn't eliminated by f.lux or similar utilities.

You might want to refer to our measurement project for some numbers:


Also, you can be below threshold for circadian responses just by being very dim.

Looking at your measurement, if I understand it correctly, you are saying that OP is incorrect and flux does in fact reduce the blue light?

Can you explain this more? If LED backlights emit multispectrum light (to get a white), and the LCD display filters those spectrum via RGB selection (say, letting less blue through, and more red/green), wouldn't this effectively filter the blue spectrum?

I know very little about this so I'm genuinely curious.

So if a new macbook pro has:

"15.4-inch (diagonal) LED-backlit display with IPS technology; 2880-by-1800 native resolution at 220 pixels per inch with support for millions of colours"

What possible laptop alternatives are out there that are better for eyes?

If you are curious, the smallest doses where "blue light hazard" has been observed in humans (short-term exposure) are around 10 J/cm^2. International standards use around 2 J/cm^2 for a safety margin.

One example of this is hiking in fresh snow for hours, no sunglasses, meaning exposures well above 10,000 cd/m2 to "white light".

Computer screens are way dimmer - an hour of iPad use is .0365 J/cm^2.

http://www.nature.com/eye/journal/v30/n2/full/eye2015261a.ht... is a good summary.

For long-term exposure (AMD risk) we have very little data, so understanding the mechanisms (such as above article does) can help.

Do those numbers account for pupil dilation?

Hijacking this for a moment to say thanks for f.lux :) I've been running it for years and it's one of those things where you don't really notice the effect until you turn it off and then EGADS MY EYES.

Windows 10 has this built-in, so does macOS.

I assume you mean the night light button in the action center?

If so I would argue f.lux does a lot more. For one thing you can choose the hue it goes to, but also unlike the Windows 10 one which (AFAIK) is simply on/off, f.lux will adjust based on a schedule which means that you don't typically notice the effects (as the change happens slowly). This, combined with the ability to change damn near every parameter it uses, makes f.lux significantly better.

Edit: apparently windows does have hue adjustment and sunset/sunrise activation/deactivation available. TIL.

Both macOS and Windows 10 has on/off now, scheduling, less/more warm adjustment. Windows 10 even has sunset to sunrise based on location.

EDIT: HN is becoming more like Reddit every day, I see. Ridiculous. Down-voted for pointing out the existence of helpful features. EDIT: might have been my fault.

Probably you were down-voted because of you starting with "What you are saying is wrong." People probably didn't like your tone and the original comment was only somewhat inaccurate. There still are more features in f.lux than the built in options.

And I assume you understand that since you edited that line out.

Correct. I realized that opening line set the wrong tone and was unnecessary.

I know these features exist in some places. But nobody had it before f.lux. And it's still - imo - one of the best implementations.

f.lux screwed up how my MBP displayed graphics so I'm just using the built-in tool now. Unfortunate.

IIRC that's a bug in the Intel graphics driver or something. I ran into that in the past as well and it's not directly f.lux's fault.

There was nothing wrong with your tone.

Night Shift on iOS and macOS also has a sunset-to-sunrise setting.

You'll almost certainly get down-voted on HN for whining about being down-voted, and for comparing it to Reddit.

Just checked. Also Mac OSX has Sunset to Sunrise based on location.

I've been a flux user for years, but sadly I had to switch to the Windows 10's version. Flux causes a 1 frame lag every 60 frames which was driving me crazy. Only fix was to put it into Safe Mode, but Flux is really janky in that mode; especially when alt tabbing. The Windows 10 version also has no lag while shift colors. It's a shame. I'd love to have kept supporting flux.

Our 2010 version polled once a second to get around a lot of bugs (Intel GPUs back then would forget their color settings on standby and lots of other things).

Since our update last year (v4) f.lux stopped doing this and does a lot of other tricks to reduce impact on framerates too.

Too bad neither f.lux nor the Windows 10 version properly support color profiles. f.lux has some support, but it sucks. And the Windows 10 version just outright has NONE. :\

These are relatively new features, and they arguably wouldn't have been added if not for the popularity of flux ;)

Does that matter? It's built-in now, and that's why I commented. To raise awareness. If built-in features have you covered there is no point going third-party.

I understand where you're coming from but they don't have feature parity. Flux allows me to target much lower color temperatures than MacOS' built in "Night Shift" feature. That alone makes Flux worthwhile.

Does it take a long time to get used to that? For me, Night Shift at full warm makes some websites difficult to use due to the change in color. I can't imagine going even warmer.

out of curiosity, how low are you going in terms of temperature?

It's nice that they are there for people who won't install 3rd party software to get the extra features, but f.lux is still better.

I would argue that if some software is under a free (libre) license, it's worth using over [non-free] first party stuff. This doesn't apply to f.lux, but does to redshift.

I would argue that it is only because of f.lux(and it popularizing the blue=bad idea) that MS/Apple/Google put this ability in their products. The technical ability to do such a thing has existed for ...decades, but no one ever did it until after f.lux came out.

So does Fedora, courtesy of recent versions of GNOME with "Night Light" feature.

Ubuntu got that too with the switch to Gnome. Before that, I used the open source https://github.com/jonls/redshift

ChromeOS has this feature for chromebooks now as well under the experimental flags section in case those users miss f.lux too. You can go to chrome://flags and search for Night Light to enable. As others have mentioned, I would rather use f.lux but (and my memory is hazy on this) I think the f.lux creator said the ChromeOS folks were not providing something he needed to get it working right.

I wonder from whom they took the idea? f.lux has been around for YEARS (2009)

Also, this was done on cell lines in vitro, only one of which was derived from a non-diseased human retinal epithelium (ARPE-19). Cell lines have undergone genetic change that allows them to replicate indefinitely, so they are fundamentally different from normal cells.

I presume that they used a laser because they anticipated that it would potentially take years to see a result from a non-coherent blue light source, far exceeding their available funding, patience, and ability to keep uncontaminated cell cultures going.

Applying a blue laser with unknown irradiance to a retinal cell line from a single human genetic source in vitro is very, very far from proving a link between blue light and diseases such as macular degeneration and RP.

All this demonstrates is that coherent blue light is capable, under certain circumstances, of interacting with retinal to produce cell damage.

Any relationship to human disease on the basis of this paper is pure speculation. Advising people to wear "blue blocker" sunglasses or avoid cellphone screens (or take Vitamin E!) is pure fantasy made up by the scientific press to sell your retinas.

Great analysis of the paper’s experimental protocol.

Isn’t the other “sniff test” that we haven’t seen a huge increase in age related retinal diseases since the introduction of computer monitor and handheld devices?

If these sources had such a big impact, it would be obvious from incidence numbers.

The problem with this is that the desktop computer being widely used is probably the late 1990s (and we should be isolating for screens that emitted the white light/blue light mentioned. Assuming that it takes daily exposure over a number of years say 20 years, we should probably start seeing that within the next 5-10 years from tech industry workers. It may be as severe as tobacco or as benign as fluoride, we'll see early indications, but we won't know until much later. I think everyone just likes it when it feels better for your eyes to filter out blue light.

I wanted to leave a note for others to keep in mind that f.lux can make the computer screen TOO comfortable to read in the evening. Which can lead to eye strain.

I used f.lux for 2 years. I found that I would stay up way past my bedtime, glued to the computer screen, and end up with correspondingly tired & fucked up eyes.

So I got rid of f.lux, and after that, I would go to bed at a more proper time. When the screen starts irritating my eyes, it's a direct indication that it's time to close the computer and get the hell to sleep. Ever since that, my eyes have done a lot better, because I no longer push to stay up way past my bed time.

I also tend to have eye strain if I'm computing very late at night, and take it as a signal to go to bed. I've described this to people, and it seems it doesn't affect everyone. It could also just be that there's a difference in awake time, blink frequency, or some other factor. I'll often go to bed more because my eyes are giving up on me and starting to burn rather than any sort of mental fatigue. I don't always keep normal daylight hours, so read this more as staying awake a long time than something specific to daylight

The pain signals of eye strain doesn't seem to hit right away, so I can keep reading past the point where I really should stop. The pain creeps up gradually but it's really too subtle if I'm stuck on an engaging story. But I don't have that problem if I turn off f.lux, that way my discomfort is more immediate and I stop.

What you described is exactly how I've been feeling for last 3-4 years. The pain signal kicks in the next day, sometimes even a day after. So yeah you can't tell at the time you're working late how the eyes will feel the next days. Sometimes it also triggers a headache and that spoils the whole day.

So the only solution I have figured is to be regimented about my screen time. I love working late at night and now have to tell myself to turn the work off. That kinda sucks.

Btw, It happened both with my previous laptop with LCD screens and the current LED screen.

For a few months I got hooked to Mobile gaming, and that just multiplied the problem! So I drastically cut down on my mobile time and that helps. My mobile had an LCD screen btw; so LED or LCD doesn't seem to be a factor for this particular problem.

I also ended up getting prescription glasses (really mild correction) to reduce eye strain when staring at computer screens. I only wear the glasses when looking at computer or phone (ideally, but not always).

I also found retina display helped reduce eye strain.

I suffer from chronic photophobia thanks to Uveitis. Flux allowed me to keep my job.

I'm running f.lux on my work PC (tones down a bad otherwise VERY bright-blue monitor).

Thank you for reading/writing out your interpretation of the article so clearly (highlighting points others are likely to trip on not coming from a related background)

Have we fixed the “optimized for electronics store demo” problem with monitors yet?

Part of my ritual of setting up a new monitor is to turn down the brightness and increase the contrast. They almost always default to very bad values. Brightness, temperature and green/blur balance are all off for an office setting.

In the worst case the backlight was capable of outputting twice as much light as I really needed.

When I got my first LCD monitor, the first thing I did was turn the brightness down to 0 and the contrast to slightly above; at that point it was already plenty bright enough, and after a few years when I noticed the brightness begin to fade I increased it 2-3 notches.

I suspect part of it is also planned obolescence --- when the backlight is running at full brightness, its lifespan is greatly decreased. Now the aforementioned monitor is approaching 15 years of age and brightness has reached 20/100, so there's still plenty of life left in it.

Leds often dim faster when young. Something else will most likely fail before the leds are too dim. On the other hand flickering can be more obvious if you lower brightness of a led. Brightness is usually emulated by turning the led on and off very fast.

Most/all modern monitors have presets, one of which is "store demo". All you've got to do is pick the calibrated preset.

Hey since I have you here, please, can you add support for proper color calibration, profiling & characterization to f.lux? Or at least outline in more detail how f.lux handles color profiles?

I wouldn't mind having to create multiple display profiles for this, but right now, flux seems to only support non-LUT based, Matrix based calibrations and only if they have VCGT headers, and then only based on those headers.

But this seems highly problematic.

Take a look at this page here for why:


I assume you based the current mechanics of f.lux when it comes to handling color profiles based on what the quickgamma homepage says about this:


To quote:


Apple has been using the LUT since many years for its operating system built in color management. In fact, Apple has registered the private ICC Profile tag video card gamma table (vcgt) with the International Color Consortium (ICC). This tag is used in a dedicated Monitor Profile and holds the gamma values, which are then loaded into the LUT by a loader program.

Hardware based Monitor Calibration Systems also use the vcgt tag, which in this case contains the correction data to achieve the monitor calibration. Because Windows did not provide a LUT loader, each calibration program had to provide its own LUT loader program. ...


Things changed when Windows 7 came along and provided its built in calibration loader, which not only loads the LUT at system start up, but also takes care of Standby and Hibernate.


As you might notice, the two stories from those pages seem to conflict?

Right now, when I use Windows to load my icc profile, I get correct colors. Same if I use the DisplayCAL profile loader.

However, if I start f.lux during the day (where f.lux shouldn't mess with the colors at all) I get semi-incorrect colors - they end up "more correct" than when I have no color profile loaded at all, but still VERY incorrect.

Surely you don't expect color profiles to work when f.lux is operating? I'm not sure what the point would be for color profiles when you're explicitly opting for distorted color reproduction with f.lux.

Is what you want more like "don't mess with the operation of color profiles when f.lux is disabled or otherwise shouldn't be adjusting the screen?"

>Surely you don't expect color profiles to work when f.lux is operating?

1. Don't call me Shirley.

2. I in fact do, in a way:

>I'm not sure what the point would be for color profiles when you're explicitly opting for distorted color reproduction with f.lux.

I opt for __linear distortion__ of a non-linear function. What f.lux ought to do consists of shifting the whitepoint coordinates of the current color profile. Which has the point of only distorting colors in a way where my brain can adapt to it.

But that's not exactly what f.lux does - it already DOES try to account for color profiles, but it does so incorrectly.

Hi herf, your software is really nice but one thing that kept me from using it (and adopted since Apple's 'night shift', though I don't use all days) is that I enjoy photography. Probably you heard this a few times, but one thing that I'd love to have is this off for photos (I'd switch from being an eventual user to being a full-time user).

I do color sensitive work now and then too. All you need to do is right-click the icon and select "disable while running [active program]".

I was very happy when I found f.lux. Finally something replacing my own tool I wrote for my Atari.

The main benefits with f.lux is you can easily set brightness and on/off with keys in any program or game, no need to navigate with mouse in menus.

You can just disable it for an hour. That's what I do at least.

Herf - it would be great if you guys could have an official position on this paper and to really help us understand what flux is doing or not doing. I tend to see you folks as the vanguard pushing the bigger tech companies to adopt your technology - I for one tell EVERYONE to download Flux.

Ever since f.lux came into my life, my long screen starring hours have stopped causing me headaches and bad sleep. My phone and my computer have night mode on, all the time! Thank you for f.lux!

I'd like to thank you as well. I jailbroke my iPhone just to get flux on there (before Night Shift), and still prefer it over Night Shift.

how does eye color factor into this? I read that objects with certain colors reflect rays of the same wavelength.

Thank you for many years of f.lux. How would you say it stacks up against Nightshift in iOS and macOS natively?

f.lux is amazing; thanks for your work!

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