Hacker News Reader

Comment by throitallaway | Hacker News Reader

PWM. This is widely used method for dimming.

It's also widely known to cause headaches. I swapped out all my PWM bulbs and couldn't be happier. More info here: https://flickeralliance.org/

I have a video projector, and if I flick my eyes across the screen I can perceive the separate R G B frames. But just watching it, it's fine.

That's DLP and it's because DLP uses a color wheel at 3x the frame rate.

I know. I was just pointing out that flicker doesn't affect me, for which I'm glad. I was a bit concerned before I bought it, due to the complaints about the flicker.

Not only DLP. Some laser projectors use a color wheel to get the same effect; other laser projectors use multiple lasers, but pulse the colors individually and you get a similar effect; probably there are some laser projectors that do all colors simultaneously too, just as there are some DLP projectors with 3 DLP surfaces and prisms/etc to split and combine the light. My laser projector does the pulsing, I find it slightly less distracting than I remember color wheels being, but it's very similar; thankfully the family has gotten used to it; I only notice it on content that encourages a lot of fast eye movement.

Scanning displays like a laser projector or a CRT will flash any individual spot but the illumination is almost continuous, so people have less issues with them.

I'm talking about laser illuminated panel projectors used for viewing film or video content, not scanning lasers used for light shows. Scanning a laser beam is a lot more difficult than scanning an electron beam, so rasterised displays from a scanning laser are limited compared to CRTs.

Ah, those are just wider-gamut conventional projectors.

I believe the colour wheels now spin twice as fast which may reduce the effect somewhat. But for me it's not a very nice effect which is why I use a 3 chip projector. Unfortunately 3 chip DLP is prohibitively expensive (cinemas use them), but JVC DLA projectors are good. Although it looks like the newer 4K models are also prohibitively expensive :/

I can not stand the DLP projectors, so I got an Epson 3LCD, and have no issues with it, not crazy expensive either.

LCDs have typically been considered inferior for home cinema usage. The only advantage of them is the price and that they have no colour wheel. DLA is a sort of middle ground that is technically an LCD but works more like DLP. None of this matters if your room isn't set up right or you're projecting in a not completely dark room, though.

The interstimulus interval (ISI) for vision is much longer than most flicker rates or frame intervals in displays and projectors. However flicker can be perceived through temporal aliasing. For lighting, even simple motion in the scene can reveal flicker. Waving your spread fingers in front of your eyes is a sure way to detect flicker.

What you're describing is likely saccadic masking, where the brain suppresses visual input during eye movements. It "freezes" perception just before a saccade and masks the blur, extending the perception of a "frame" up to the point in time of the sharp onset of masking. That's how you get a still of a partially illuminated frame instead of the blended together colors.

I’m no expert in this, but if you're curious, check out the Wikipedia pages on interstimulus interval, saccadic masking, chronostasis, and related research.

IKEA has, at least in Sweden, started publishing how much their lights flicker. I found that it is actually an upper limit for their lights. The dimmable ones flicker less for most of their dimmable range for example.

Now, I only measured two bulbs, but I am pretty darn happy with those results. I also opened one of their chargers and haven't looked elsewhere for chargers since. The thing was even more well built than my apple charger (the 45w sjöss is actually quite crazy. The 30w had some issues)

Yes, I get those as well from some cheap LEDs. Though if you flicker LEDs fast enough, it's fine.

How could the same modulation achieve "dimming" and "fully bright" simultaneously?

The apparent brightness is caused by the ratio of on/off and is called the duty cycle. 50% brightness would mean that half the time the light is on, and half the time the light is off.

If the cycling of the light on and off is done at say 10kHz it's perceived as a dim light.

For traditional lightbulbs, a PWM signal actually makes the light noticeably dimmer, and has a very simmilar effect to simply reducing the current. This is because the mechanism for them is heating up the filement, which happens on a much slower time scale then the PWM duty cycle.

In practice, traditional dimmers are not quite PWM as they do not generate a square wave. Instead they generate a sin wave with portions of each cycle clamped to 0.

LEDs already need driver circutry to condition the relativly high AC voltage into a stable lower voltage DC. Dimmable LEDs create a stable DC power supply from the chopped up AC power, then use the width of the active portions of the AC as a signal to drive their own dimmer logic.

That's a good PWM explanation but I think GP was asking a different question. Walter claimed that if the PWM was high enough frequency you would perceive it as full brightness while using less power.

It obviously wouldn't work that way when lighting a room, but what about for an LED indicator light that you look at directly? I don't know enough to form an opinion.

Without any more information on frequencies and measurements it's impossible to know. My guess is either A) They did not perceive the dimming, and/or B) The frequency was high enough that the inductive and capacitive effects of the circuitry became relevant and was filtering the PWM signal to DC signal that still drew the same amount of power as it did when fed the full voltage DC signal

you dont. But your eyes perceive brightness non-linear, which means to your eye, 80% of brightness is very close to 100%

The measurable brightness of the LED is a straightforward sum of the times it's on and the times it's off. Shift the ratio so it's off more than it's on and it gets dimmer.

The "fully bright" part is a consequence of human vision. Your brain is making sense of limited, noisy information coming from your eyes. A flickering light appears brighter than the light in steady state, lots of still images shown in rapid succession look like they're moving, as far as you can tell you can perceive the full range of colour out of the corner of your eyes, and the dress could be either colour.

Also (and I think more important, but I might be wrong), we don't perceive light intensity linear but logarithmic. A 50% duty cycle does appear brighter than half as bright. A 90% duty cycle might be only barely perceptible.

Perhaps less widely known is that if you market a commercial product that actually uses PWM to modulate LED intensity, you're liable to be litigated against by Philips for patent infringement?

Wild...yeah, I know. Heard that from a buddy in Austin over a decade ago. Vaguely recall that he had to redesign using some sort of current driver instead to avoid the legal encumbrance.

Probably no longer true; patents only last 20 years.

To be fair, it's probably naive to blindly assume that a company like Philips, with a long history of litigious behavior, isn't playing patent continuation games.

Reply on news.ycombinator.com