Hacker News Reader

Comment by akersten | Hacker News Reader

I've always wanted to ask an audiologist whether "active noise cancelling" type technologies actually pose an auditory risk. They kind of have to emit the same amplitude-at-the-ear of audio of what they're cancelling out, right? And sure, most of the time the waves cancel, but maybe the algorithm gets it wrong a couple milliseconds at a time and - while possibly not perceptible - are those high peaks of audio-cancelling-sound damaging?

View on HN

There is a running conspiracy theory that manufacturers found out about possible damage around 2018, and coincidentally all the popular models like the Sony WH1000 got firmware upgrades that perceivably downgraded the quality of ANC. I have a model from 2016 and can easily tell the difference from a newer one, now if that is the real reason or not... we’ll find out in the coming decade.

This theory was brought up during a previous thread on HN by @novok, in response to https://news.ycombinator.com/item?id=22080228

I wonder... would this be theoretically possible? If so, are there any peer-reviewed papers published?

I got tinnitus after using Sony WH1000 noise cancelling headphones in 2018. Damn...

I have a pair of those. I really like them except that they don't quite fit my ears and wearing them for more than a couple hours causes some soreness. If it weren't for that I'd probably wear them all the time. So, maybe that ended up saving me in the end.

The thing is, noise cancelling is a very simple physical process. If it "miss" you will ear it. So there is no invisible danger, it cannot be not perceptible, either it cancel, or it add to the noise. The worst that ANC can do is add noise to the noise. Plus, most ANC focus on low frequency sound since they are the easiest one to cancel.

If you're trying to cancel out a loud 100dB+ environment, that means the driver needs to be blasting at at least 100dB.

Any harmonic distortion in the headphones will be introduced relative this, as well as being in the higher frequencies where it's less noticable, but still may cause hearing loss.

I've used ANC IEMs instead of proper hearing protection around loud machinery before and although it may have felt quieter, my ears still "rang" afterwards so I was causing permanent hearing damage. People who think ANC "sounds quieter" and use it instead of proper hearing protection (or just turn the music too loud) will definitely suffer hearing loss, though not directly because of the ANC.

I am not an audiologist, but an interesting similar issue happens with our eyes. If you wear tinted lenses with no UV filter, your body will be happy that it looks darker, and open your iris more. Now since your iris is open more, you actually cause more UV damage to your eye without noticing. I remember this being a big story when polarized 3D movie glasses came out and people were keeping them on outdoors.

> If you're trying to cancel out a loud 100dB+

i dont think it's feasible to cancel out such loud noise using just ANC. A sealed, airtight headset (or those tight earplugs) would have to be used. If you're using ANC for such loud noises, you'd still hear the noise through your bones/face imho. For example, if a car beeped its horn at you while using ANC, i hear it.

ANC is merely for constant low background noise like airplane engine.

ANC doesn't ever provide perfect isolation. If you see advertisements ANC it will say "blocks 99% of sound" or "cancels 99.9% of noise" which is equivalent to a 20dB and 30dB drop respectively.

You're right though that it doesn't work as well at high volumes. The louder the noise, the more distortion in the microphone/amplifier/driver, so the efficacy of the noise cancelling is reduced. For example in this THD graph of the AirPod Pro's, at 100dB it adds approximately 3dB compare to the 90dBtest.[1] Going much louder and you'll start getting digital clippings as the built in amplifier won't have enough range.

> . For example, if a car beeped its horn at you while using ANC, i hear it.

Car horns are above 100dB, and at the frequency they are, APD only measures about 20dB of attenuation [2], so that only brings it down to 80dB, which is still very loud.

> ANC is merely for constant low background noise like airplane engine. What I find very neat is that even though only 20-30dB of isolation is measured, many noises really do "go away". If you play around with ANC you'll realize that as the background noise gets louder, the white noise hiss from the ear piece gets louder. This mostly the result of error in the system, but this increase in background noise has the effect of masking any external sounds that make it through. I believe it also moves a lot of that energy to higher frequencies where you don't notice it.

[1] https://www.rtings.com/headphones/1-4/graph#1625/4029 [2] https://www.rtings.com/headphones/1-4/graph#1625/4053

This assumes that active noise cancellation doesn't introduce energy in audible parts of the pressure spectrum.

This is no more a risk than normal audio you listen to doing the same.

Of course not. If your driver is trying to cancel 95dB and creating a harmonic then that harmonic has 100 000 times more energy than the same harmonic when playing at 80dB without noise cancelation.

> If it "miss" you will ear it. So there is no invisible danger, it cannot be not perceptible

not always true. Minor phase mismatch will create higher frequency sound. You wouldn't hear ultrasound, though how bad it is to have an ultrasound source in you ears i don't know.

The phase mismatch can be temporal - circuitry reacting either a bit too fast or too slow - as well as spatial - the sound wave front most probably aren't parallel to the speakers frontal plane.

Phase mismatch won’t change frequency

A noise-cancelling system can be designed to filter out and not produce high-frequency sound, to not worsen the problem. I'm not sure if properly designed noise cancelling can produce increased pressures felt by any parts of the eardrum or ear canal or outer ear.

What is the mechanism for this? From what I remember of signal processing, linear combination of waves of a specific frequency can only result in a combined signal of the same frequency but different phase and amplitude.

It's not just phase mismatch. There is error in the frequency domain, both from the digital processing going on and the physical characteristics of the driver.

This is anecdotal but I do feel a sense of pressure when using the Bose QC35. This pressure lasts for 30-60m when I’ve used them all day. I also have my ears popping after using them.

This is one the most popular topics in the Sony vs Bose war. For reasons I haven’t seen any explanation for, Sony headphones have much less of that vacuum feeling.

Anecdotally I strongly agree. I notice no such feeling at all when using my WH1000-XM3, but it's very perceivable when wearing the QC35 or QC25.

I have tinnitus, but I've had it for way longer than I've been using noise cancelling headphones or in-ear earbuds.

I noticed the same. Disabling or lowering noise cancellation in-app fixed everything.

I too feel it. Then again it might make sense that even if we don't hear it, the pressure is still there. As conservation of energy. After all active noise cancelling is destructive interference. Which has funny meaning that the energy is actually doubled. As we produce noise of equal power but opposite phase. Now does the energy end up in your ear or the headphone or outside? Or is it divided between these.

The energy is dissipated as heat in the speaker coils. Energy is removed from the air. Think of it as using brakes to slow down a car. Yes, energy is consumed to do so (that's why the rotors get hot), but the car doesn't have 'hidden' energy, just because it was both accelerated, then decelerated. It really is at rest. The energy it embodied while in motion has been converted into other forms.

Yes. This can cause sinus issues.

Unless the active noise cancelling gets it spectacularly wrong, the high peaks won’t be higher than the peaks in the to-be-cancelled noise.

So, you’d have to argue that short, sudden peaks are more damaging than continuous sounds. There is adaptation to sound volume (https://asa.scitation.org/doi/10.1121/1.1906477), but I think you’d need more than that.

I find active cancelling headphones to be very unpleasant. I've switched to custom molded in-ear monitors which do an even better job of eliminating external sound (musicians wear them on-stage to protect their hearing), sound better, and don't bother me the way active cancelling earphones do.

I have no idea if active headphones are "bad", just that they make me feel very uneasy.

The downside is that Custom In Ear Monitors are pricy (> $1000). When you have them made, you can often choose how many db of ambient sound attenuation you want.

Ten years ago I was interested in buying Bose noise cancelling headphones. Someone warned me that they had read somewhere on the internet that noise cancelling creates problems with your hearing.

At the time I couldn’t find anything on Google about this, but I still bought the headphones anyway.

Using the headphones the lingering thought of that statement bothered me. Using the headphones something didn’t feel right.

Flash forward to now, people are talking about this so I can’t find relevant info on this without the well (Google) being poisoned by armchair doctors.

I still feel noise cancelling is not healthy for ears.

I feel like WiFi is damaging my brain /s

My Sony MXH100 headphones gave me tinnitus..

These "peaks" should all be within what you are able to hear, because there's no reason for having the headset produce cancelling waves with supersonic frequencies.

If you want to worry about sudden sounds, there are probably more dangerous sources of them out there.

That's only the produced sound, right? If the produced sound is at a frequency that coincides with the original sound (as opposed to being half a wavelength off), wouldn't it amplify the sound and make it up to 2x? Apologies if I misunderstood your comment.

I am not an audiologist... but I don’t think it would result in 2x the sound. I believe a sound originating from 2 sources with the same amplitude results in either phase cancellation or the same resulting amplitude, not 2x. This is similar to what happens with line arrays.

I don't think so? This is classic wave interference right?

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

You are thinking about the volume. As another replier said, two interfering frequencies produce an interference wave equal to their difference, so you can't have supersonic waves from something only capable of producing 20kHz sound

Yes they cause a number of problems. I experience nasty sinus headaches after prolonged use in loud environments (where the speaker driver is having to match the high amplitude of external noise).

Mistakes in the cancelling might make peaks outside of human audible ranges that cause problems? I'm speculating here but wondering if anyone is looking into this.

Reply on news.ycombinator.com