Flicker Detector Lets You Hear What You Can’t See

Have you ever looked at modern LED lighting and noticed, perhaps on the very edge of your perception, that they seemed to be flickering? Well, that’s because they probably are. As are the LEDs in your computer monitor, or your phone’s screen. Pulse width modulation (PWM) is used extensively with LEDs to provide brightness control, and if it’s not done well, it can lead to headaches and eyestrain.

Looking to quantify just how much flashing light we’re being exposed to, [Faransky] has created a simple little gadget that essentially converts flashing light into an audio tone the human ear can pick up. Those LEDs might be blinking on and off fast enough to fool our eyes, but your ears can hear frequencies much higher than those used in common PWM solutions. In the video after the break, you can see what various LED light sources sound like when using the device.

The electronics here are exceptionally simple. Just connect a small solar panel to an audio amplifier, in this case the PAM8403, and listen to the output. To make it a bit more convenient to use, there’s an internal battery, charger circuit and USB-C port; but you could just as easily run the thing off of a 9 V alkaline if you wanted to build one from what’s already in the parts bin.

Who knows? If you carry this thing around long enough, you might even hear the far less common binary code modulation in action (but probably not).

57 thoughts on “Flicker Detector Lets You Hear What You Can’t See

  1. This is the kind of device I did not knew I needed.
    I’m gonna build it.

    Maybe the other way round is also interesting? Making sound visible, for example infra- and ultrasound?

    1. Making sound visible is an interesting one. The low frequency cutoff of our eyes prevents a simple modulation of the light which is why we need a light to sound device to detect it.

      The audio frequency would need to be expanded to cover the visible light spectrum.

      Wow when I started typing this I thought the solution would be simple but as I’ve been typing and thinking it’s gets harder and more complicated.

      If you just want to detect a specific frequency but dont have to convey much information a simple “color organ” type device would suffice.

      I’m thinking something like a waterfall display would come close – works for RF but again the amount of information that can be translated is limited.

      1. I haven’t checked to see if my thoughts are accurate; but I believe the RF spectrum humans can see, is broader that the audio spectrum, we know exists, that humans can’t hear. I have to feel that heterodyning could create such a color organ. Pretty much how a waterfall display is created,even if it takes multiple stages.

      1. The process already exists, called heterodyning. We human’ can’t directly hear the sounds marine mammals and other life forms make. Nature films covert those sounds to what we can hear.

    1. Yeah, using this on a monitor is kinda dumb. It’s gonna pick up the screen’s refresh rate way more strongly than the LED PWM.

      LED PWM “strain” is mostly hypochondria and neurotic paranoia anyway. Or a reason for boomers to complain about the new light bulbs and justify hoarding incandescents because that’s what they do for some weird reason. This gizmo is just gonna fuel that fire and make more people hallucinate symptoms about harmless lights.

          1. It’s getting rarer and rarer now that switching power supplies have become so much cheaper, and I think you can no longer get Energy Start certification for an LED bulb that lacks power factor correction. Not sure – a lot of LEDs have PFC, and any “smart” bulb is not going to PWM at line frequency.

            The worst bulbs I’ve seen as far as PWM frequency are Philips Hue RGB at around 600 Hz – which is more than OK for human eyes but REALLY bad for photographers.

          2. Unfortunately 120Hz flickering, while not obviously visible to the naked eye, causes some people (me) eye strain, headaches, a general feeling of unwellness, etc. I only buy brand name lightbulbs (GE, Philips, etc) because I can be sure they won’t try to kill me.

      1. Lol, I hoarded incandescent bulbs but not because of flicker, which I’ve never found to be an issue with any properly constructed LED bulb. The issue is, or rather used to be, the dimming characteristics. An incandescent lamp will change its color temperature characteristically when dimming. It’s only now that some companies have begun selling LED bulbs that mimic this behavior so I can finally switch over too. As a light quality purist I beg to differ between that and the “5Gs” panic that gives some people headaches. I also blame that to 99% on nocebo effects.

        1. Interestingly enough, dimming via full on/off flicker (no filtering) will often give less colour temperature shift than dimming by voltage, since the amount of light hitting the phosphor is consistent during the on-time, regardless of the duty cycle.

        2. I’ve been looking for LED bulbs that change color temperature while dimming and can’t find them. Can you post any links to good ones you’re found? It’s also the reason I haven’t upgraded our ceiling lights yet.

      2. Most studies says otherwise regarding flicker. especially for people with adhd and similar conditions. lights that flash under a couple of khz is still registered in your eye receptors as flicker. that’s measurable. It affects you a bit like background audio noise. Some people don’t mind, some get a headache, and others get nothing done if in such environment. (open offices för example)

        1. Some are poorly constructed. Yes, if you build an LED that visibly flickers, it will have an effect. Decent, properly constructed LED lights do not flicker visibly. If it annoys or disturbs you, don’t buy from the bottom of the line. The “corner of you viewfield check” works well to determine whether there is possibly problematic flicker.

          1. Disagree.
            A casual corner of your field of view check would work for the very worst offending lights, but not the slightly faster ones or necessarily the insufficiently filtered ones.
            The difference between perceptible and visible.

          2. The problem isn’t single lightbulbs, but an office full of LEDs with PWM drivers that are slightly detuned, so you get the sum and difference of the waveforms as a result.

            Because of the persistence of vision effect, if the two lights happen to be on at the same time, it looks brighter than when the light pulses are interleaved. As the oscillators drift continuously, you get a slight stroboscopic effect which can throw you off, even cause nausea and epileptic fits in some people.

        2. I think the problem of “open offices” is another one. If I am right, that “open offices” are this halls of cubicles, then the problem is missing privacy and (acoustic) background noise – massive sources of distraction.

      3. LCD screens don’t flicker at the refresh rate, because they don’t fade and rescan the way CRTs did. The pixels just update directly to their new values, without the 17mS of fade in between.

        The thing is, lots of people are bothered by the flicker, and it’s generally very easily dealt with. designs with excessive flicker tend to be the cheapest of the cheap, and there’s really no excuse for visible flicker on things like room lighting.

        Part of what makes people dislike the “new light bulbs” is that a lot of them are really awful, whether it’s flicker, or terrible phosphors giving really bad colour rendering.

        To those of us with some knowledge on the subject, the solution is simple. Use bulbs that have better phosphors and a decent power supply design in places where you care, or simply stick with incandescent in places where efficiency is a low priority, and colour accuracy and flicker are critical.

        Yeah, there’s some “boomers” who use that as an excuse to resist change for no good reason, but that doesn’t make it any less true in other cases.

    1. “Dodo says:
      January 28, 2020 at 3:44 pm

      Would it not work better with a photodiode or phototransistor and some gain? Most solar cells have a very high capacitance (massive junction…) which really limits their BW.”

      Just replacing the solar cell with a phototransistor that goes higher in frequency (e.g. >100 KHz range) will not really help because you can not hear beyond say 10-20 KHz (depending on how old you are). Plus, the audio amplifier is also limited to the human audible bandwidth range. What you need to do is add a variable frequency oscillator and a mixer to downconvert what the phototransistor “sees” to the audible range, sort of like a how radio receiver works. The cheapest way to do this these days in terms of parts is by using something like a microcontroller with digital signal processing (DSP) capability to form most of what would become a software designed radio (SDR) for light. [An IQ direct down converting SDR can also be fairly easily made into a light transmitter. This would make the light receiver into a light transceiver.]

      A phototransistor like one from the Vishay BPW85 series might be a reasonable choice. It costs $0.72 in unit quantity, has a 2 us rise and 2.3 us fall time (~180 KHz Fc @ Vs = 5 V, Ic = 5 mA, Rl = 100 Ohms), is pretty sensitive over most of the human visual spectrum (works over 450 to 1080 nm), is quite linear over two or three orders of magnitude of incident irradiance (remember, the light receiver will have manual and/or automatic gain control), and a half power beamwidth of +/-25 degrees. However, the 3 mm diameter T-1 LED package tells us the physical aperture (light collecting area) is quite small compared to the much larger solar cell.

      1. The problem with a downconverter is that you don’t gain bandwidth, so listening to any really high frequencies would need a lot of retuning of the circuit. Realistically you’re probably not going to get more than 8kHz or so at most that’s in a range of hearing that you’re going to be able to hear anything useful at. (maybe double that if you only care if there’s sound present or not)
        once you’re up in that 100kHz range, you would really just be scanning with the dial, and the pitch of the audio wouldn’t tell you anything useful (unless you did it digitally so you know your offset), since you would only be able to hear 96-104kHz (Unless you want it really crude, and allow it to alias for a bit more range reflected at the ends of the scale)

        1. Good discussion here, and good call on the BPW85. Small diode? Use a lens, focus on it.

          Since you don’t really care about the quality of the flicker, but just its presence, you could get by with a divide-by-n downconverter instead. Diode, op-amp, into a binary counter, and then tap off the /8 or /16 outputs into an audio amp.

          Tuning the op-amp to work with your brightness levels and get the thresholds right for the counter might need a potentiometer and some practice. But this seems super doable.

  2. I could see our Xmas tree lights flickering when looking straight at them.

    If your phone has a high speed camera (240fps or more) the light flicker is very easy to see. I’ve been filming LED lights for awhile, very susceptible to the flickering of cheap LEDs

  3. I used to work with infrared remotes and we had made what we called a “buzz box” that used basically just an IR receiver and an audio amp to convert the IR light into an audible signal (yes a binary signal). This was usually used just to verify when something was being transmitted over IR but many people got to where they could often tell what protocol was being used just by the sound. We were using an IR only receiver because we wanted to block out the annoying flickering lights from so many other sources all around us. It was a very effective tool.

    1. The camera of your smartphone is very good with that. New ones have a IR filter on the back camera, but usually not on the the cheaper front camera. So put your phone is selfie mode and you will know if you still have battery in that non-functioning remote control.

      1. This was back in the late 90’s, there were no smart phones. :-) The buzz box also had added advantages, it was very sensitive, it didn’t have to be directly in front of the remote and it converted the IR to sound which freed up your hands and eyes so you could be doing something else like working on a receiver and still know what was happening.

        1. Depends how far away from the lens you hold the remote. I just tested all the remotes sitting next to me, a slightly interesting discovery: my XBox One media remote has two slightly laterally off axis IR LEDs. I had wondered how it managed to have such a good coverage.

      1. That’s really cool!

        I picked up something like this (https://www.aliexpress.com/item/32973957673.html) when I was working on a project involving radios and wasn’t sure if I was broadcasting anything. It’s sort of a cool toy, but just emits a tone when it detects radio, it doesn’t give you any information about the content.

        I looked for a PCB for that Techlib circuit, but couldn’t find one. Maybe I’ll put one together when I get a minute.

  4. I seriously need this, someone rear ended me (don’t text and drive) giving me TBI with permanent light and particularly frequency sensitivity, and as a result i have developed a 6th sense for light with a frequency that gives me problems. If i keep exposure to a minimum (under 3 hours a week with sunglasses) I can usually avoid migraines, and this device might be a life saver for me.

    1. There are TBI specialists out there who can help with this. My wife is one. She has tons of patients who show up with similar or identical symptoms, and after a couple weeks or months of a combination of vision therapy and migraine meds, they usually end up with way less headaches and way less tendency to get them from bright light flashes or strobes.
      It’s also worth mentioning that in the aftermath of a TBI incident, you are much more likely to get another one, and the consequences of that are much worse. Brain injuries are more than cumulative, more like multiplicative. Be extra careful for several months.

      1. At this point, it’s been almost three years since the accident, and my neurologist has said the effects are permanent at this point. I sought treatment immediately after the accident (I have a background as a sailing instructor and recognized the symptoms immediately) and again when new symptoms appeared two weeks after the accident. I have been through vision therapy, and neuro-feedback therapy, as well as tried several different medications, all old antidepressants with off-label migraine uses. Honestly, the side effects from the meds were worse than the symptoms. I always was careful with my head, the accident was my first and only head injury. If there are more options I should be exploring, I am all ears.

  5. I was responsible for our community Christmas tree lighting. We switched to LEDs but, being aware of the flickering problem, I made a DC power supply to drive them. Simple as a bucking transformer, bridge rectifier and capacitor for 120 VDC out. The flickering of cheap LED lamps drives me crazy. Lamps with coatings or better drive circuits pretty much solve that personal problem. I’ve been thanked by others in our town for the improved tree lights.

  6. Grow some balls and just put your tongue to it. If it lays you flat out on your ass, it is powered by DC. If you flutter on your way down it is AC. If it does not kill you, count the flutters before you hit the floor for a rough idea of the frequency. Kids these days… They probably don’t know how to tell DC polarity with potato either.

  7. Flickering LED light fixtures is such a problem!
    I get so nauseous working under non constant voltage driver led lights, might as well work under a strobe light… ugh…
    For the guy who asked about warming effect led bulbs- I (an electrician) always try to steer people towards Philips warm glows. They are a smooth light. They warm up, not as much as incandescent but some. And I find they dim really well- use a Lutron ariadni 153 dimmer.

  8. A dollar store solar light or your junk box for a solar cell and case, battery powered speaker (amp) and a earphone cord for parts. Connect the solar cell to the plug end of the earphone cord (after removing the extra guts). The solor cell will give 1-2 volts about 26 ma max too the amplifier. Solder the cord to the solar cell leads. It has a ready made housing tie a knot before soldering through a hole. Add batteries to the speaker/amp plug in the solar cell and you are ready.
    Cheep Led Lights are very noisy, good computer monitors are quiet, LED tv’s can be very noisy. Persistence of vision keeps us from seeing the Blink but wave your hand fast and if you see stroboscopic blinking you’ve got eye strain – PWM. A filter cap, DC power on leds are Much easier on the eyes.
    PS The frequency response of ears/amp/speaker will effect your results.

Leave a Reply

This site uses Akismet to reduce spam. Learn how your comment data is processed.