Hacker News Reader

Comment by Retr0id | Hacker News Reader

I'm very sceptical of their claims that ~780Hz is in some way special, especially the way they represent it graphically. Playing a frequency sweep while wearing WH-1000XM3 headphones, I don't notice any particular drop-off there.

Near where I live, heavy goods vehicles are fitted with reversing indicators that make a "cshh cshh cshh" sound i.e. pulsed white-noise. White noise like that is the hardest for ANC to cancel. Sample: https://www.youtube.com/watch?v=A3Wt1_51EVA

View on HN

Right? This feels like an "arms race" similar to scraping vs. anti-scraping; countermeasures will be developed, likely due to the action of actors entirely disconnected from what you're doing, but to block something else in the ecosystem... and you'll need to re-engineer your approach entirely. Rinse and repeat.

(The amount of innovation in anti-anti-scraping that's resulted from "sneaker bots" - automated scalping of limited-edition shoe releases - is astounding, and somewhat relevant here in how an environment can become adversarial in ways that impact broad ecosystems. I suppose the equivalent here would be environmental ads that seek to penetrate noise-cancellation in a similar way.)

I suppose, though, that all this is good news for a company that wants to turn your bicycle bell into a subscription product!

I don't see why this would become an "arms race." There's no particular competitive value in filtering out this ONE sound.

I think there's a broader indication of an arms race between noise cancellation systems and things that want to be heard, like advertising. And this just-happening-to-exist bandpass that the DuoBell is depending on could easily become collateral damage in that fight.

I was going to make a joke about advertisers working in some kind of ultrasonic modulation to their audio so it breaks ANC (I'm aware this wouldn't really work) but then thought, whats more likely, advertisers doing that, or advertisers partnering with 80% of ANC chip makers to just let them by-pass with specific tone markers...

Then we'll be hacking our headphones with specific 3d printed clip-ons that involve a particular brand of coffee filters that happen to attenuate the "clear freq" enough for the headphones to miss it.

"Dad, why do our coffee filters advertise that they can run fast fourier transforms?"

"Well, kid, back in the year 2026, there was this bicycle bell.."

The construction site next door is using those vehicles, and they're also a lot more pleasant throughout the day. It's easier to tune out white noise than beeping. The first cshh is a little louder than the others, which is a nice design touch.

Speak for yourself, I can tune out a steady beep much easier than the sound of a seagull being strangled to death. (That's what the ones around here sound like anyway.)

On a more serious note: the loud beeping backup alarms were DESIGNED to be annoying and difficult to miss. I would not be surprised in the least if a study showed these "less annoying" backup alarms correlating to a higher number of children being run over by reversing vehicles.

There have been studies and those resulted in the less annoying backup sounds. These sounds are essentially harsh white noise, which has one significant difference to the beeping: it's level drops off differently with distance, meaning you can blast it louder and people who are really in the wrong spot will notice better it means them, while people who are not meant will not be annoyed or fatigued by it. Two noise sources combine different than two tonal sources and the human ear can locate broadband sources better than single tones.

This was developed especially for use in backup heavy environments like harbors where workers started ignoring constant beeps.

There's also another difference: beeps can reflect coherently off of surfaces, causing directionality confusion in a dense environment. White noise is much less likely to have odd interference patterns, maximizing our ability to localize the sound.

Me too, as soon as I saw this I put on my Anker Soundcore noise-cancelling earbuds and fired up my AAS Multiphonics CV-3 software synthesizer. Sweeping a sine oscillator, there was zero difference in perceived volume in the 750-780 Hz range.

Right. From the article:

Through acoustic testing, the research team identified a narrow frequency band – a “safety gap” – capable of penetrating ANC headphone filters. This range lies between 750 and 780 Hz.

Is there a standard specifying this "safety band"? Is whatever Apple does for AirPods a de-facto standard?

On my wh-1000xm2, wh-1000xm3, wf-1000xm4 and lastly wf-1000xm5, there is a quite high frequency pitch (usually coincides with some public transport beeps, and some accidental squeaking of doors) that toggles ANC to transparent mode automatically. I remember reading something about this on Sony's support website.

Also triggered by baby screams unfortunately

An evolutionary adaptation to ensure parents wearing headphones don't ignore their babies.

We don't ignore him. I wear headphones while soothing him.

I have a new idea for a bike horn.

Please do not duct tape a baby to your handle bars.

My XM4's always do that at the beeps from the cash register, although I always attributed that to their volume rather than frequency. My theory was that they refuse to produce sound loud enough to cancel the beeps for safety reasons.

I suspect as much too. If there's a failure to match the noise perfectly then the headphones are just going be be blasting a loud sound into your ears. And if it matches the frequency correctly but lines up with the sound instead of being out of phase, then it's acting as an amplifier!

That diagram is pure marketing nonsense. The real chart is on page 10 of their paper[1]. It shows a modest ~3dB less attenuation around 800Hz across several brands.

[1] https://cdn.skoda-storyboard.com/2026/04/Skoda-DuoBell-Resea...

The actual research paper shows it's pure nonsense

https://cdn.skoda-storyboard.com/2026/04/Skoda-DuoBell-Resea...

As expected ANC headphones cancel less noise at low frequencies so I guess the 780Hz is a trade off between high enough frequency to be a bell and low enough frequency to get attenuated a little bit less than high frequencies.

The research paper is pretty poor quality and this is mainly a marketing exercise.

It's fairly easy to test out: https://onlinetonegenerator.com/

Anecdotally, bells have always come through fairly clearly for me. They filter out lower tones, not higher + sine waves. Nothing about this adds up to more than any normal $5 bell, especially rotating ones which hammer repeatedly.

Reply on news.ycombinator.com