f.lux vs. Night Shift in macOS 10.12.4
-
I have two things to say about Apple's copy of f.lux.
#1 Amount
Night Shift's defaults are pretty gentle, and for most people they won't reduce the impact of a screen by very much. Here's what Night Shift does before bedtime:
And here's what f.lux does before bedtime:
With both, you're going to see orange, and most people won't understand that there can be a tremendous amount of blue-green light in that orange light. Apple is claiming to help people sleep, but their displays are really far from achieving "circadian darkness".
So right now, about half the press about this is saying you don't have to worry about "blue light" anymore before bed, and this is just incredibly untrue.
We've spent the time to understand how light affects sleep because we want our software to actually work for almost everyone. But the numbers from Apple's efforts just aren't enough to help sleep.
To use an example that explains why I'm pretty convinced they're not doing enough, consider that the warmest Night Shift setting makes an iMac show more light at night than the iPad that Harvard studied back in 2012.
This means you should really dim your screen at night a whole lot, because a big screen makes a lot of light.
The "Color" vs. the "Amount" of light
Our circadian system is actually not reacting to small changes in "color". Instead, it is mostly reacting to the "amount" of light. Our eyes are extremely good at distinguishing little shades of color from each other, but this is a different system than the one that drives circadian rhythms.
And so if you rely on what makes your eyes feel relaxed, it's not always going to connect up with your sleep timing - maybe you'll feel like reading later because it feels better. This is the same assumption we made in 2008! But now people are trying to guess that what "looks good" will always help you sleep better. This really only gets you a little bit of the way to understanding what's going on in the body. So, you should know that "the name of the color" isn't a very good way to measure light, and the "amount of blue-green light" is a lot better way to think about it.
With screen filters, we've seen people compare products that remove 10% of the bright light at night with ones that remove 99%, and they seem to think they're about the same. This idea that everyone is doing basically the same stuff in this area, is what we're up against in the public discussion, and it's pretty hard. If you're interested in this topic, you should be pretty critical and ask for the numbers always.
So starting there, f.lux removes 4-5x as much by default as Night Shift does, and yet it seems like most people will continue to think the two are similar, when they aren't in the same range.
#2: timing
The second idea we've been working on with f.lux is that everyone can benefit from slightly different timing of bright and dark, and getting control of the clock in your brain is pretty hard to get right. This "timing" is actually different than picking the times the lights turn on and off - it's about understanding how those lights affect you. So it's your timing, not the timing of the lights. You can't really guess about this - you need some careful measurement over days and weeks.
In this regard, most people we've talked to believe that the only way that light is disrupting their sleep is in the moment, right now. And so if they turn out the lights a little bit before bed, they'll sleep better. But this is not the big picture at all - how our bodies understand day and night (and the seasons) is a very complicated process, and it varies a lot from person to person and throughout the day.
A few of these ideas are already in the Mac f.lux build, but not so much in Night Shift.
We're building a new set of tools around these ideas, and we really hope some of our users are going to understand that what we're doing with f.lux is pretty different, and there's no substitute for trying to understand how human biology works.
We're working every day on this, making f.lux better all the time. We really hope you'll help us out with what we're doing next.
-
Thank you!
-
@herf when you are comparing to Night Shift, are you using the default mode? How much blue light does it remove if you put it on the warmest setting?
-
@cjmegatron81 The full dataset is here for iPad Pro:
https://fluxometer.com/rainbow/#!id=iPad Pro Night Shift/defaultThe Night Shift UI implies that the defaults will help you sleep better, which I disagree with. To be specific, their defaults remove <30% of the impact blue light has on your body. (Your body wants a "dark signal" not a "slightly less bright" signal.) Also, in software we know that a huge majority of people stick with the default settings.
The strongest settings do appear to be better. However, the Mac version is also brightening the screen on the warmer settings (which cancels out most of the benefit), so the iPad data are somewhat better than the Mac. We'll post data when we've updated a newer machine.
-
@herf is it possible for me to run tests on my machine to find out myself? I have a late 2015 iMac.
-
@cjmegatron81 We have a way to do this coming up very soon. Most of the f.luxometer measurements are done on a Photo Research PR-655 Spectra Scan, you can see the current project at https://fluxometer.com/rainbow/
-
@lorna thanks!
-
I have compared Night Shift with f-lux for the past couple of days, or should I say evenings and nights, and I do not hesitate to say that f.lux is much better than Night Shift. So, I happily donated another 5$ this year. Keep up the good work!
-
Very nice comparison. Are you guys planing to compare the night mode in Windows 10 with f.lux?
-
@AlixSPQR Thank you for your kind words and your support, it means a lot.
-
@meinmoises Yes, we post all new data at the f.luxometer project: https://fluxometer.com/rainbow
-
@lorna You're welcome! :slight_smile:
-
One (and the only one) thing that Night Shift as of now does better is that there is no white areas color corruption on fullscreen YouTube videos.
-
@bazilio_lg said in f.lux vs. Night Shift in macOS 10.12.4:
One (and the only one) thing that Night Shift as of now does better is that there is no white areas color corruption on fullscreen YouTube videos.
People are reporting that it works in f.lux too since they fixed the bugs we reported to them... Love to know more about your system if it's working in one and not the other.
-
@lorna sorry, I might have old information because since latest macOS update I have never actually watched a YouTube video in fullscreen with f.lux enabled.
Did a quick test right now, no issues so far, woo-hoo! -
@herf Could you please include some axis titles on your color charts? What do the grey areas signify? Is uW, Micro watts? I take it the colors shown are what are displayed on the screen, not what colors are being blocked.
What do you use to output the graph? I'm curious enough to do some tweaking on my own screen.
-
More info is at https://fluxometer.com/rainbow/
The screenshots do show scale as uW/cm^2/nm (spectral irradiance) in the top corner, but for spiky spectra that is not the most useful number.
We might be more interested in the total irradiance for each photoreceptor, which you can find in the "non-visual" tab on our site. It shows the "equivalent illuminant E" irradiance, as well as total power.
Our models and sources are on the pages too.
-
This post is deleted!
