Hackers use an ordinary light bulb to spy on conversations 80 feet away | Hacker News Reader

Hackers use an ordinary light bulb to spy on conversations 80 feet away

forbes.com

121 pointsharambae42 comments

Previously discussed here:

https://news.ycombinator.com/item?id=23510617 https://news.ycombinator.com/item?id=23498185

Reminds me of a 2014 work by MIT on extracting audio signals from vibrations of objects in a video [0] which I found was truly fascinating. I guess someone was paying attention to bring this to production mode out of the lab.

[0] http://news.mit.edu/2014/algorithm-recovers-speech-from-vibr...

Also interesting (though different): https://en.wikipedia.org/wiki/The_Thing_(listening_device)

What a beauty!

"The Thing" was designed by Soviet Russian inventor Léon Theremin, best-known for his invention of the theremin.

Indeed this is a very interesting artifact, the fact that it was designed by Léon Theremin, and the technical know-how that went into it.

> Because it was passive, needing electromagnetic energy from an outside source to become energized and activate, it is considered a predecessor of radio-frequency identification (RFID) technology.

If you want to even further back you can look at the side channel attacks shown in person at the CanSecWest conference in 2009 - https://cansecwest.com/csw09/csw09-barisani-bianco.pdf

The most interesting part might be the frequency analysis that was done after getting the signal. It was amazing to see the words get decoded in real-time.

That's why those good-ol'-fashioned windowless, inner, Faraday-caged rooms with concrete foundations separate from the rest of the building with their own isolated power-supplies are still probably the go-to place for the paranoid. I remember those in a gov contracting facility in the Boston area -- I'm sure there are more.

https://en.wikipedia.org/wiki/Sensitive_Compartmented_Inform...

Plenty of SCIFs with Windows out there. It all just depends on the context of your location.

Do you mean "windows" like made out of transparent glass or "windows" like the notoriously insecure operating system.

Both, from my experience

You forgot the dampened suspension and vacuum/air gap.

Let's not get carried away - that's for earthquake / nuclear shockwave protection, not information security.

Sorry, I thought was was implicit in "separate foundation" :)

> hooking up an electro-optical sensor to a telescope and pointing it at a light bulb, but that's the meat and potatoes of the sub $1,000 (£800) hardware hack at least. Fluctuations in air pressure on the surface of the hanging bulb are created by the sound of conversation, or music, and make a hanging bulb vibrate.

Isn't this the same technique used to listen in on a conversation using a glass window?

One thing that's not obvious to me from TFA is why an incandescent bulb is called out specifically. It seems to me that a hanging LED bulb would be just as vulnerable, as might a glass chandelier of sorts.

I also wonder whether Christmas trees with hanging ornaments are vulnerable. These are often placed in a prominent window, so it might be quite easy to start listening in on conversations in people's living rooms or business lobbies.

Maybe the filament is more susceptible to vibration and "flickering". But it may just be my interpretation, the text doesn't really spell it out as such.

A couple of other comments mentioned the filaments, but the article says nothing about clear bulbs. They are very much in fashion these days, but most incandescent bulbs are frosted, so it seems to me that if you want to use a laser on the filament, there would need to be some considerations around photons passing through the frosting and reflecting off the filament.

The bulb doesn't have to be clear, the filament vibration would make the light flicker visibly (for sensitive equipment that is) even if the glass isn't clear. So it may be that the filament is far more susceptible to vibration than the glass on any type of light bulb. As I said, this is my interpretation, the text doesn't make it clear to me.

I don't think it's about observing the filament directly, just that the filament basically being a spring is more susceptible to vibrations which would change the amount of light emitted.

I am totally guessing at that, but it seems reasonable.

The filament is in a vacuum...

Almost no incandescent light bulb has a vacuum inside. They are filled with low pressure inert gas (mostly argon sometimes mixed with nitrogen, some bulbs have krypton or xenon for lower temps) not only because it makes them safer to handle and harder to break but mainly because it slows down the decay of the filament by slowing down its evaporation. It bounces tungsten atoms back on the filament.

LED lights and fluorescent lights are vulnerable too per the Wired article at https://www.wired.com/story/lamphone-light-bulb-vibration-sp... , just with different degrees of sensitivity

That article seems to say that this technique works better with LED lamps vs. incandescent bulbs.

I wonder if they're talking about direct line of sight to the LED chip. Most of the LED "bulbs" in my house use a frosted plastic housing around the LED chip, which I would guess to interfere with the sound recovery.

>> One thing that's not obvious to me from TFA is why an incandescent bulb is called out specifically.

Because they produce light continuously. An led, CFL, or other will actually flicker at 120Hz.

An incandescent bulb still has 120Hz power going in, but because it depends on heating to produce light and the filament has some thermal capacity, this is largely smoothed out.

> An led, CFL, or other will actually flicker at 120Hz.

In 60Hz countries. Also, if it has a constant current driver there will be PWM noise from the regulator.

I have noticed LED bulbs keep glowing for a short period of time after turning them off:

https://www.quora.com/Why-do-LED-bulbs-keep-glowing-even-aft...

I imagine there is still some flicker, but it seems to be smoothed out pretty well on LEDs also.

I've seen that too. I think they're using a phosphor of some kind to convert UV wavelengths to visible light. That would have some persistence for those wavelengths.

Maybe it’s because an LED bulb is much heavier?

The technique used to listen in on a conversation using the window is an active technique (an IR laser is emitted towards the window to get the bounce), so it can be detected, and requires LOS to the window + to the reflected path if you're not hitting the window completely perpendicularly. For example you can't hit a window a floor above from ground level, and architectural treatments that result in angled windows can reflect light away in directions that are more difficult to observe.

You can point your telescope to any reflection or light (even fluorescent ones) on a room, and the induce vibration will be enough to create a signal to listen.. I accidentally figured that out when I was 13 or so, and tried to build a led walkie-talkie.. I kept hearing other voices, and that is what it was..

already submitted here, from the real source not the lousy forbes article : https://news.ycombinator.com/item?id=23498185

Source: https://ad447342-c927-414a-bbae-d287bde39ced.filesusr.com/ug...

Modern LED fixtures and lamps may be more resistant to this. Who has a hanging incandescent bulb anymore? That being said, a person looking for bugs may not notice a otherwise normal looking hanging incandescent fixture.

A lot of bars, cafes, etc use decorative bulbs with thin, easily vibrating wires inside (even though LED-based). I can imagine such a bulb at a home, too.

While the title is about lightbulbs this looks like the same principle as using a laser to measure the vibrations on the surface of a window in order to listen to the audio inside a room (something that's been science fact since at least the early 90s). I don't see why the laser or optical sensor couldn't be measuring vibrations of something inside the room: a clear line of sight and a mylar balloon might be the perfect tool for such passive surveillance trick.

"Ok, we want to know Raganwald's age. Point a laser at the mylar Happy Birthday balloons and let's listen in on the singing. Take notes! Are you one, are you two, are you three..."

Um, chief?

"Shhh! I'm listening! Are you fifty-one, are you fifty-two..."

Chief!

"Be quiet, damn you. Ok, he just turned 58. Did you get that?"

Yes, chief. I got that. When I pointed the laser at the balloons, I could clearly see a "five" balloon, and an "eight" balloon, and while Raganwald has some very old-fashioned ideas about programming, I don't think he's eighty-five.

"Good work. Now make me another coffee."

This is the comment I'd like to end my HN career on. Goodnight for now and very pleasant dreams.

This is very unlikely to be practical in any way. Just ask yourself how many times you've seen a lightbulb in the last five years. Sure, you've seen plenty of light fixtures, lots of neon tubes, some LED bulbs, etc, but an old fashioned light bulb, not enclosed in any type of lamp shade?

That aside, how are you going to discretely point a telescope at the light bulb from a distance of 25 meters? You might as well point a parabolic microphone at them and get a better signal.

Come to Brooklyn or LA or any place where hipsters congregate. Exposed “Edison” style bulbs are a bit of a trend even residentially.... you can buy them at Home Depot and Walmart, so it’s mainstream. Granted it’s still the vast minority of fixtures.

I've seen a fair number of naked lightbulbs in the last few years in hipster restaurants, coffee shops, etc. If done right, it can actually be a pleasing aesthetic.

I think your general point is right though, it's not terribly practical.

Half of those are actually LEDs.

https://en.wikipedia.org/wiki/LED_filament

A different site, but the topic was posted before (and made to front page): https://news.ycombinator.com/item?id=23498185

We need a device like a mask or helmet which can encrypt our voice before put it to the internet. Just like a VR glasses.

Reply on news.ycombinator.com