Hacker News new | past | comments | ask | show | jobs | submit login
The iPhone 12 Mini Makes Me Sick (Literally) (ieee.org)
106 points by TN567 53 days ago | hide | past | favorite | 29 comments



Display flicker caused this person to experience nausea. The flicker is from the use of PWM for driving the OLED display brightness. “Fixed” by switching to a device with an LCD display.


Not exactly flicker, but I have one monitor that makes an audible whining noise when certain spreadsheets are open and the window is sized above a certain % of the screen space.

It's maddening, but IT won't swap it out for me.

I say certain spreadsheet, but really it's something about the formatting and contents of a single page in that sheet. Open in Excel, noise. Snipping tool snip of the screen, also noise. That page printed to PDF and opened, however, does not make noise.


Life must have been rough when fluorescent lighting was the norm.


The standard thing was to turn off the fluorescent lighting back when everybody (tech support, sysops, programmers) was almost entirely looking at a screen.

I got very used to calling "lights" before turning them on or there would be moans and groans.


I had to quit one of my first jobs over it, the combination of cheap fluorescent lighting and no natural light caused me to have constant headaches.


You’re luckier than me, I get headaches from natural light.

If I’m outside on a sunny day without sunglasses I am pretty much done for the day. I have no clue why I am this way or if this is a common issue for folks out there. I should look into this more, and if anything can be done about it to reduce the nausea and headaches I get from being in the sunlight.

One of the things I’m looking forward to seeing is if wearing contact lens is better than wearing glasses.


I was pretty far down the road with a neurologist due to frequent migraines back in the day. Unrelated, I quit my job during that time. Suddenly, when using an LCD in a mostly naturally lit environment, instead of a CRT under fluorescent lights, my migraines completely vanished and haven’t come back once in a decade. Tada!


I thought I was imagining things in my head.


Another fix is to use DC dimming.


Is there a list somewhere of midrange to high-end phones, Apple or Android, that have LCD or exceptionally high quality OLED displays? I find I am very sensitive to the "Bayer matrix" style subpixel arrangement in OLED panels and I freak out anytime I scroll someone else's OLED phone screen. (It might also be the flicker, as this article suggests, though I'm not usually sensitive to PWM artifacts, at least not when I've done LED driver work before.)

Also, 240 Hz is criminally low for a PWM frequency. Individual LEDs are trivial to PWM with a 240 kHz carrier these days, though obviously that's not a whole matrix's worth.


>exceptionally high quality OLED displays.....

The highest quality should be Samsung Galaxy S21 Ultra Screen. The problem is the sub pixel alignment on OLED means we have to push PPI density and resolution to a ridiculous level to compensate, which in turns requires higher GPU performance and energy requirement. I sometimes I wonder if it even makes sense for these trade offs.

I know we are still far away from it happening but I am really looking forward to microLED display.


I have all the problem described in this article. And I have definitely rant [1] about OLED for a very long time since the launch of iPhone X. The PWM problem is quite well known on Internet forums but again mainstream media doesn't want it as a story (yet). I also ran away from Malls or shops that use crappy LED that have their flickering level at a low rate. ( Not to mention their utter crap colouring ) I did once try to call for quality standards of LED but obviously no one cares. It was the first time I felt what disability were like when the your surroundings assumes everything is perfect or without disability.

I also think those who are susceptible to PWM are also latency sensitive. Which means ProMotion / 120Hz panel is definitely a thing I can felt. And micro jank are definitely observable. ( I discovered that in the early days of Firefox, that was before Mozilla took jank / micropause in their UI seriously )

> I had had motion issues with an earlier phone at one point, and knew to turn off motion effects to solve that problem.

Unfortunately this is not widely known. I have friends who have very mild symptoms they didn't know it was those motion causing it.

>Am I going to be shut out of future phone upgrades? I don’t see many of the top-of-the-line models coming out with LCD displays.

What the article didn't state, was that OLED has an economy of scale and roadmap that makes LCD pretty much irrelevant from a BOM cost perspective. i.e It isn't even about picking a lower end iPhone SE or iPhone 11 for LCD Screen. Soon OLED will take over that segment as well. And you can bet those OLED screen will have even more PWM than the current best iPhone. And remember, iPhone already have the best OLED screen in terms of PWM levels.

That is why I am praying there will be iPhone SE Plus with a LCD Screen. And iPad Air stays with LCD.

[1] https://news.ycombinator.com/item?id=18505704


More on the technical side of this, solving the PWM LEDs issue isn't actually as easy as it may seem to begin with.

It comes down to

1) Analog is hard, and so is high speed PWM

2) Some engineers (and online engineering resources) believe a few hundred Hz and above is imperceptible

3) It's very, very difficult to explain to someone that something is flickering, it bothers some people, but you might not be able to see it.

4) Some websites do review devices to see if they flicker. RTINGS for monitor and TV, NotebookCheck for Notebook, Tablet, Phones.

So the way LEDs, and OLEDs by extension work, is that they require the current to be limited. Usually, in electronic circuits, voltage is usually the independent variable, and current is the dependent variable, since it's affected by things like resistance, temperature change, and other factors. All displays currently commonly found use LEDS or OLEDs, whether as a backlight or for the display itself.

The problem with controlling the current, is because it "depends" on all these other factors, it has to be dynamically controlled. There's a control loop, where the circuit has to sense the current, then change the voltage, in order to keep that current constant. You could always just use a resistor, but that wastes a ton of power and cannot be used for anything more than an indicator LED. Many smaller LED drivers do this with some analog IC, but for higher power and efficiency, typically this involves switching, inductors, and capacitors. These components, to simplify a complicated matter, cause a delay in response. You can only react so fast, so you can't change the current you want too quickly, or else it will go unstable and weeble wobble. This limitation makes it hard to PWM these circuits very fast. Furthermore, it's difficult to build an analog circuit that can sense both a huge amount of current and a small amount of current accurately enough, which limits how much you can dim LEDs with this technique.

As for the second half of the problem, LEDs, particularly white ones, work by emitting blue light, and using chemical phosphors (or, more recently, quantum dots) to turn that into white light. The response of these phosphors is sometimes not very linear with respect to how much blue light comes in. If you simply dim the LED, the color might change (think, get pink or greenish).

This gives rise to the two main types of dimming used for LEDs currently: Analog (turn down the current) or PWM (turn the full current on for less time). There's also Hybrid, where you Analog dim down to 10%, and then PWM it down to however low you need, since you can't analog the current down any more.

Analog doesn't flicker, but it suffers from the LED-color-shifting problem, the can't-dim-too-far problem.

PWM suffers from the it-flickers problem, the analog-responds-slowly problem, and the it's-hard-to-turn-something-on-and-off-very-fast problem.

The problem with PWM, is that there are two dimensions - carrier frequency (how fast it's switching) and resolution (how big the increments are). Most digital circuits are limited by many factors in how fast they can switch - divide it down by how fast you want the PWM to run, and you'll see that it's often a trade off between frequency and resolution. Often, especially for something like an OLED display, where you have to actually update all of your pixels, of which there may be millions, this could be a challenge. Thus, in many designs, frequency is compromised first, because it's never advertised (but they do advertise bit depth!), and because of problem 2) and 3) above. Of course, with PWM, you STILL need the aforementioned analog feedback stuff, to varying extents, but that also limits how fast you can switch LEDs.

In the case of simpler LED displays, such as the large ones you may see in stores, the LEDs might not actually individually even have a driver, and PWM is a necessary evil due to the fact that a single driver is being rapidly switched between all of the pixels really quickly - the switching speed caps your PWM frequency.

In the case of illumination LEDs, not only do you have to deal with the aforementioned problems, but you also need to deal with the fact that inherently, the AC power grid flickers at 60hz, and to get rid of it you need capacitors and inductors. Both are expensive, and large numbers of LED light bulbs, particularly those filament kinds that are very space-limited for the electronics, flicker terribly because they omit those parts. If you look at illumination LED drivers, most quote 2kHz as the absolute-max not-to-exceed flicker frequency. Naturally, usually the limit is not what people ride, and many chips that do not support over 490Hz exist. Also, the Arduino default is 490Hz. Ugh.

In the case of phones and tablets, and OLED screens in general, some progress is being made. Xiaomi (Pocophone?) released a DC-dimming software update a while ago for some phones, which allow the user to choose whether to trade off flickering for the leds-shift-their-colors problem. However, they do warn the user that the color of the screen will be worse, and the "dirty screen effect" or poor uniformity effect will be noticeable and many customers may assume the screen was broken if this was default.

Often, users note that OLED screens flicker, even on their highest brightness setting. That is often because OLED screens actually have a hidden "Turbo" mode built in. Try to shine a bright light at the sensor on a OLED phone or tablet, and there is a chance you'll catch it "clicking" into overdrive brightness. Most phones don't let you do this manually by default because it burns the OLED screen in really quickly, but there used to be a tweak for rooted devices that let you toggle it on all the time manually.

So what can be done about it?

As mentioned in the article, DisplayMate (the publication of Raymond Soneira from the article) apparently doesn't test flicker yet. But, for monitors and TVs, RTINGS does measure that, and NotebookCheck does, for phones and tablets and laptops.

You can use those review sites to compare devices that do or do not flicker, and to what extent. If you use a device that has PWM dimming, turn the brightness up, possibly adding privacy shield screen protector to darken it if too bright. This reduces battery life, but because the LEDs are "on" longer, the "off" time will be less noticeable.

When you shop for something in a store, take your hand, stretch out your fingers in front of the screen displaying white, and then wave your hand quickly. If your fingers don't blur, but rather you see outlines, then it flickers. Same with light bulbs. Light bulbs that don't turn off immediately, but rather fade out, likely have bigger capacitors inside, which reduce flickering.

disclaimer I cannot guarantee that what is written above is correct.


> disclaimer I cannot guarantee that what is written above is correct

It's actually pretty good! I'd nitpick some minor details or change some emphasis but you're close enough that it's basically being a dick to nitpick at that level.

The only real mistake is this: "a resistor... wastes a ton of power". In fact it wastes exactly as much power as a current source, which is basically just a variable resistor that sets itself to the proper value. Of course, it's still useless for anything dynamic.


Do you know if this is related to why Apple’s TrueTone would make me feel nauseous too?

I’ve got a 2020 iPad Pro and was having great trouble using it until I accidentally found something online that suggest I turn of TrueTone. The effect was instantly noticeable. Before I could only look at the screen for a few minutes, now it is much more like my desktop where it takes hours to get eye strain.


The iPad Pro 2020 uses a PWM frequency of 60kHz or so, and only under 13% brightness according to NotebookCheck. This is unlikely to be a problem.

True tone, iirc, is a simple color temperature adjustment of the screen. I wonder what happens if you enable Night Shift manually and set it to how it looks with true tone on, and see if it triggers the same problem. You can try the same experiment with your computer. Perhaps it is that color scheme that is causing you the issues.

Apple iDevices have a feature called Color Filters in Accessibility settings, which allows you to manually change the color tint of the screen. It may be of interest to you to see if a light tint helps the screen be more suitable for you.


Thanks for the very informative answer!

Regarding point 2), with high enough frequency it does stop for even people very sensitive to this, doesn’t it?


Huh, interesting. I’ve noticed I get a bit of a headache looking at my iPhone X. Ironically I switched to iOS partly due to the latency sensitivity you mention.


This issue may only affect a minority of users, but it's a seriously frustrating issue for us. I'm currently unable to use any modern Mac, iPad, or iPhone (Same with PCs) due to whatever flickering they've got going on, it's actually ruining my life here (My job requires a computer). I understand that most people have no issues with flicker, but I hope the tech develops to a point where these companies don't need to rely on it.


People have had a lot of problems with modern LED screens:

https://forums.macrumors.com/threads/new-4k-imac-causing-mot...

https://ledstrain.org/

Personally, I feel weird already after 20 min checking out screens in an Apple store. I just hate the screens, cold colors, strange eye focus due to the anti glare garbage, PWM flicker.


The new screens are so addicting. I switched back to a 2011 MacBook Pro recently to stop using my laptop so much. MacOS and iOS offer no way to reduce contrast in their OS.


> PWM is one of two ways to dim a display; the other is by changing the voltage of every sub-pixel—the analog method. LG uses this approach to adjust brightness for its OLED TVs.

These sentences are confusing to me. Am I correct to understand this as saying that LG uses “the analog method” to dim the TV displays (and not PWM)?


I did find the PWM quite noticeable on my X when I first got it, but have totally got used to it. I can still sort of detect it at low brightness though. Didn’t consider that it might have a physical impact for some people but it makes sense, as it did make me feel a bit “weird” to look at.


Hmm interesting. Anyone knows if the display in the 12 series is similar, pwm wise, with the one in older oled phones? I'm just fine with my XS and wouldn't give up oled colours for anything.


PWM is awful, in displays, flashlights, headlights, taillights, and dash lights. I challenge the notion that PWM is more efficient than analogue dimming. PWM adversely affects enough of a percentage of a population that it should be eliminated and banned. In those that PWM does not affect, it causes two issues, one revealing callousness and the second a fallacy: 1) "PWM doesn't bother me, so I like PWM and I don't care about anyone that PWM bothers or hurts; and 2) "I can't detect PWM, therefore it does not hurt me."

To drive home the fallacy in 2), "I can not detect CO, therefore it does not hurt me."

PWM is pure evil.


226.2Hz flicker. that’s pretty low frequency… surprising it’s not in the KHz.


Similarly, those flashing ads on websites should be banned. Not only are they incredibly distracting, but they make me feel a little queasy.


Yeah, iPhone 12 sucks. :(


I'm surprised Apple didn't put a toggle in the accessibility settings




Applications are open for YC Winter 2022

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

Search: