How we test noise cancelling headphones

What this is

Perception, measured carefully

The investigation records lived experience and perception using a series of real world tests and replicable, standardised acoustic tests. The report is led by one tester’s documented experience, repeatable and described in full, so you can judge how well it maps onto your own ears.

The laboratory websites already measure decibels better than any independent tester could. This investigation measures the things their rigs can’t feel.

The pressure feeling

Why heavy cancellation can feel like a plane descending

If you’re autistic or otherwise sensitive to pressure, and you’ve been told your discomfort from noise-cancelling headphones is “all in your head,” you have reason to push back on that. This is sometimes why headphones feel like “airplane suck”.

There are two different things going on, and most headphone sites only talk about one.

The brain-guessing part. In daily life, there’s one situation where you lose deep bass sound but keep hearing higher sounds fine: when air pressure changes across your eardrum and stiffens it. Your brain has learned that pattern over your whole life. Cancellation copies the same hearing pattern by accident, because it’s very good at removing steady low sound and leaves higher sound mostly alone. Bass gone, treble still there. Your brain matches that pattern to “pressure change” and produces the feeling, even with no pressure change happening. This causes a mild, general fullness for a lot of people.

The part that’s actually physical. The lead reviewer had to look into this second kind as their experience didn’t fully match with the common explanations. We needed to explain pain that shows up only when cancellation is switched on, using the exact same tips or cushions you wear with it off. For that, there’s a second explanation, and it holds up better under that test. A sealed ear canal with an active speaker inside it acts like a tiny pressure chamber: the same speaker movement makes a much bigger pressure change in that small trapped pocket of air than it would out in the open (Bose patent). Cancellation output is mostly low-frequency sound, exactly the range that gets amplified this way. So an active speaker in a sealed ear canal can genuinely push and pull the trapped air, not just make sound you hear. Engineering patents call this a real “static pressure,” different from ordinary sound level, and note that this kind of pressure — the same kind used in a hearing test called tympanometry — can set off a protective ear-muscle reflex by itself, from the eardrum physically stretching, not from sound (a href-“https://image-ppubs.uspto.gov/dirsearch-public/print/downloadPdf/8774435”>patent on static pressure and the stapedius reflex). That’s the same reflex involved when a plane descends or a train enters a tunnel.

No one has measured this directly inside consumer ANC headphones and published it, as far as we could find, so we can’t say for certain how big the effect is in any given pair. But if your pain only happens with cancellation on, and it feels like a plane or a train, you may be feeling a small real pressure event, not a false alarm your brain invented.

The eight axes

What we measure

Every headphone is examined across eight sensory dimensions. Each maps to a way a nervous system can be caught off guard.

Noise. What it blocks and, just as important, what it doesn’t. Steady low rumble is where cancellation works. Sudden sharp sounds, voices, and clattering crockery are where it fails, and the gap between the marketing and the reality is part of the finding.

Pressure and internal sense. The pressure feeling above, split into seal and electronics, tracked over ninety minutes. Dizziness and queasiness logged honestly, because a small share of people get them and deserve the warning.

Feel. Cushions against the skin, the headband on the crown, the warmth that builds inside sealed cups over an hour, and the audible thud some headphones make when you touch the controls.

Squeeze and weight. Clamping force at a fixed head width, weight on the scale, and a separate check of what happens when glasses break the seal, which a great many readers need and no mainstream review tests.

The glitches. Three failures the labs don’t test. Whether cancellation roars in wind, whumps or cuts out over a bump, or struggles when a door slams. For a startle-sensitive person an unpredictable glitch from the coping device itself is its own kind of harm.

Social. Whether a headphone reads in public as ordinary headphones or as visible equipment, because wearing a coping tool shouldn’t carry a social cost, and for many people it does.

Visual and scent. Indicator lights that pulse at the edge of vision, and the smell of new materials. Small dimensions, not zero.

Where we test

Six places, twice each

A headphone behaves differently in a silent room and on a rush-hour train, so each is tested across the places sensory-sensitive adults actually use them. A quiet room first, where the pressure feeling and any hiss show with nothing to hide them. A kitchen and home workspace, where steady appliance hum is cancellation’s home ground. An open-plan office, where voices and keyboards expose its limits and the social question is sharpest. Public transport, the technology’s headline setting and the home of two of its glitches. A café and a supermarket, the honest-limits tests, where sudden sharp sounds defeat cancellation regardless of price, and where the report asks plainly whether a pair of earplugs would have served better.

Each setting is tested twice: once with nothing playing, which exposes the hiss and pressure at their most obvious, and once with audio playing, because that’s how most people wear them. The audio is steady pink noise rather than music, partly for consistency and partly because a person overwhelmed enough to reach for heavy cancellation often can’t tolerate music at all.

Keeping ourselves honest

One tester, stated plainly

This protocol didn’t start in a lab. It started with over twenty years of buying, returning, and living with this category of headphone, and finding that the returns were never explained by the spec sheet. The three-way split above, blocking, ICP, and clamp, is what that much personal history eventually made obvious. The measurement is new. The problem it measures isn’t.

One tester is a limitation, and pretending otherwise would be dishonest, so the investigation is built to expose its own biases rather than hide them. A guess about each headphone is written down before testing begins, so findings that merely confirm expectations can be spotted. The tester’s sensory load is noted at the start of every session, because a difficult day inflates every discomfort score. The first headphone tested is tested again at the very end, weeks later, to check whether the scores drifted as the testing wore on. Every objective number is taken with the same equipment in the same room, so the products compare fairly even where the absolute figure is rough.

The verdicts

Cleared, Caution, Flagged

Every Case File ends in one of three verdicts, never a score out of ten, because a single number hides exactly the variation that matters to a sensory-sensitive reader.

Cleared: no significant sensory problem for the profile described. It does what it claims without a hidden cost.

Caution: it works, with a trade-off that matters to some people and not others. The Case File says precisely who should think twice and why.

Flagged: a problem serious enough that a sensory-sensitive reader should approach with care. The reason is always stated and tied to a documented test.

A verdict is always given for a particular kind of nervous system, not for everyone. A headphone that earns Caution for pressure sensitivity may be entirely right for someone who doesn’t feel pressure at all. The point of the investigation is to tell you which of those people you are.

The investigation

What’s coming

The ANC Headphones Investigation tests eight over-ear noise cancelling headphones, from budget to premium, against everything above. The first Case Files publish at the end of June 2026, and each will be linked here as it goes live. For some people and some places a pair of earplugs is the better answer, and a forthcoming comparison will set the two technologies side by side, building on the earplug frequency attenuation findings already published.