Two cabin-noise figures can both be honest and still contradict each other, because a decibel reading only means something once you know the weighting applied, the flight phase measured, and whether the number is a peak or an average. The most traceable in-cabin data for the A330-300ER comes from a study reporting cruise averages of 69.9 to 72.8 dB(A), a mean of 71.3 dB(A).
The widely shared 100–104 dB figure has no identifiable measurement basis, and certification values such as EPNdB describe exterior noise at defined ground points — not what you hear in your seat.
Ask two passengers on the same Airbus A330 how loud the cabin was, and you can get answers that differ by thirty decibels. One slept through the flight. The other will swear it was a chainsaw at 35,000 feet.
Both may be describing a real measurement. That is the part most cabin-noise coverage skips.
A sound level on its own is close to meaningless. Weighting, measurement position, flight phase and peak-versus-average decide what a number represents, and a figure quoted without them can be made to mean almost anything — which is roughly how the argument over cabin noise got so loud.
What follows is not a verdict on whether your next flight is dangerous. The verified record does not support a dramatic claim either way. It is a guide to reading any noise figure you meet, including the ones that arrive with a study attached.
Why two honest decibel readings can look like a contradiction
A decibel is a ratio, not a fixed quantity of loudness. That is why one cabin can produce a 71 dB reading and a 104 dB reading without anyone lying.
Start with weighting. A-weighting adjusts a reading to reflect the fact that human hearing is less sensitive at the very bottom and very top of the frequency range. C-weighting keeps more low-frequency energy, which matters on an aircraft because engine and airflow noise sit low in the spectrum. One in-cabin A330 study reports the same cruise interval averaging 85.6 dB(C) — roughly fourteen decibels above its A-weighted figure for that interval.
Then comes the detector setting. Leq averages sound energy across a stated interval, so a whole-flight Leq smooths out the noisy stretches. Lmax records the highest level inside the measurement window. A thrust increase can spike one and barely move the other.
Flight phase does the rest. Takeoff, cruise and approach produce genuinely different noise profiles, so a figure with no phase attached is a figure you cannot place.
Six metrics get swapped for one another constantly. The table below separates them by what each one actually measures.
| Metric | What it measures | Weighting or corrections | Relevant phase or window | Typical value range where verified |
|---|---|---|---|---|
| dB(A) Leq | Time-averaged sound level weighted toward human hearing sensitivity | A-weighting; Leq averages sound energy over the stated interval | Cabin measurement over a segment or whole flight | A330-300ER cruise averages of 69.9–72.8 dB(A) across three reported trips; 71.3 dB(A) mean |
| dB(A) Lmax | Highest A-weighted level recorded during the stated measurement segment | A-weighting; maximum rather than energy average | A defined cabin segment or flight phase | A330-300ER table values are reported in the study, but a defensible general range was not independently established here |
| dB(C) Leq | Time-averaged sound level with C-weighting, retaining more low-frequency energy than A-weighting | C-weighting; Leq averaging | Same measurement interval as the source study’s cabin readings | The A330-300ER study reports an average C-weighted Leq of 85.6 dB(C) |
| EPNL | Effective perceived noise level used in aircraft certification | Perceived-noise weighting plus tone and duration corrections | A certification event at takeoff, lateral/flyover, or approach measurement points | UK CAA A330 entries include 98.1 takeoff, 100.3 lateral, 88.0 flyover, and 97.0 approach EPNdB for one configuration |
| EPNdB limit | Certification limit associated with the approved aircraft configuration | EPNL framework and applicable certification corrections | Regulatory certification rather than cabin exposure | One UK CAA A330 configuration lists a 103.7 EPNdB applicable limit |
| Peak or maximum dB | Largest instantaneous or short-window reading, depending on instrument setting | Weighting and detector setting must be specified; neither is supplied by the number alone | Usually a transient event such as thrust increase, landing, or turbulence | The circulated 100–104 dB aircraft-cabin claim remains unverified; its source lacks aircraft, location, weighting, and duration |
| Source: Applied Acoustics paper copy; UK Civil Aviation Authority; Accident Compensation Corporation (New Zealand) | ||||
What the A330 study actually reports
The most traceable in-cabin dataset on this aircraft is a paper titled “Assessment of in-cabin noise of wide-body aircrafts,” posted online on 6 May 2022. It covers five wide-body types; the A330 portion rests on three commercial A330-300ER trips.
Cruise averages across those trips came out at 72.8, 71.2 and 69.9 dB(A) — a spread of 2.9 dB and a combined mean of 71.3 dB(A).
Those are time-averaged cabin readings taken across a cruise, not peaks.
What the record retrieved here does not carry is a confirmed 75.7 dB(A) takeoff figure, or the paper’s DOI, or its peer-review status. That is not a small omission when the paper is being cited as the study that settles the question.
Superdeals cover departures from North America, Europe and Australia.
Flight deals most people never see
Air Traveler Club monitors thousands of routes for pricing anomalies that traditional search engines miss. With ATC Superdeals members save up to 80% on each flight vs. regular fares.
Too good to be true? Check it out yourself. It's free.
Delivered weekly. Unsubscribe anytime.
The 100–104 dB figure has no traceable origin
Search for the alarming version — cabin noise aboard an A330 at 100 to 104 dB — and the trail runs a short distance and stops. The only compilation located that carries that range ties it to an undated entry on an NPC website.
No aircraft is named. No measurement position, weighting, flight phase or duration. Nothing indicates whether it is a peak or an average.
That does not make the figure false. It makes it unusable — a different problem, and a far more common one.
Certification decibels were built for airports, not seat rows
EPNdB values in a type-certificate data sheet answer a regulatory question: how much noise does this configuration make at defined points on the ground? The UK Civil Aviation Authority publishes them in its type-certificate data sheet for noise, UK.TC.A.00044, keyed to the exact variant, engine fit and mass.
One A330-301 with General Electric engines, at a 192,000 kg maximum takeoff weight, is listed at 98.1 EPNdB at takeoff, 100.3 laterally, 88.0 in flyover and 97.0 on approach, against an applicable limit of 103.7 EPNdB.
Read those as cabin comfort and the aircraft sounds deafening. Certification engineers would tell you the sheet answers a question asked by airports, not by anyone in seat 34A.
How a raw sound level becomes a certification number
Certification converts a physical measurement into an administrative one. Engineers record the sound at fixed exterior positions — takeoff, lateral, flyover and approach — and then convert each reading into perceived-noise units, adjusting for any prominent tones and for the duration of the event. Two aircraft with identical raw readings can land on different EPNdB values if one has a whine.
That sensitivity shows up in the paperwork. Certification values are keyed to the exact variant, engine fit and mass, so a different configuration of the same airframe family carries its own certificate.
None of it describes the sound field inside a cabin at cruise. Certification records are held by EASA and national authorities precisely because they are regulatory documents, and regulators are answering for the people on the ground.
What this means for you
Four questions defuse most cabin-noise claims: What weighting? Which flight phase? Peak or average? Measured where — cabin, flight deck, or outside the fuselage? A figure that answers all four is worth arguing with. One that answers none is decoration.
On risk, the reference points are unglamorous. The UK’s Noise at Work regulations set their first action level at 85 dB(A) for a working day. That is an occupational threshold written for crews, not a cabin-comfort standard, and the exact clause was not retrievable in this session — treat the figure as directional.
What the record supports is modest. A wide-body cabin measured as a cruise average sits well below that line, and a momentary peak during a thrust change is not sustained exposure.
What it does not support is a single dB number you can bank for your next flight.
Reporting by
Key terms
- A-weighting
- A-weighting is a frequency correction applied to a sound measurement so that it roughly tracks the ear’s reduced sensitivity at the extremes of the audible range. It is one of several standard weighting curves used in acoustics, and it is the one behind the “dB(A)” label on most cabin-noise figures. For a traveler, the weighting is the first thing to check, because two figures for the same flight can differ sharply depending on whether the low-frequency energy was counted.
- C-weighting
- C-weighting is a frequency correction that retains more low-frequency energy than A-weighting does. It is applied where the low end of the spectrum matters, such as engine and airflow noise inside an aircraft. A C-weighted figure is not simply a louder version of an A-weighted one; it is a different measurement of the same sound, which is why the two should never be compared directly.
- Leq
- Leq is the time-averaged sound level, expressing sound energy accumulated over a stated interval. It smooths out short noisy stretches, so a whole-flight Leq sits below the loudest moment within it. A Leq figure answers a different question from a peak reading — it describes exposure across a segment rather than the loudest instant inside it.
- Lmax
- Lmax is the highest sound level recorded inside a stated measurement window. Unlike Leq, it captures a single loud moment rather than an average across the interval. For a traveler, an Lmax figure describes the worst moment of a flight while a Leq figure describes the flight as a whole, and a claim that quotes one without saying which is quoting neither clearly.
- EPNdB
- EPNdB, or effective perceived noise level in decibels, is the unit used in aircraft noise certification. It is built from exterior measurements at defined ground points, with added corrections for prominent tones and for how long the event lasts. Because an EPNdB entry is tied to a specific variant, engine fit and mass, it belongs to a configuration rather than to a cabin, and it is not a number a passenger can carry onto a flight.
Questions? Answers.
How loud is airplane cabin noise?
On the most traceable A330 data, a cruise averages around 70 dB(A): three commercial A330-300ER trips in a 2022 in-cabin study averaged 72.8, 71.2 and 69.9 dB(A), a mean of 71.3 dB(A). Those are time-averaged readings across a cruise segment, not peaks.
Can a cabin dB(A) reading be compared directly with an aircraft’s EPNdB certification figure?
No. A cabin dB(A) figure is an interior measurement weighted for human hearing. EPNdB values are exterior certification measurements with added corrections for tone and duration at defined ground positions. Different weightings, different locations, different purposes — neither can be compared with or subtracted from the other.
Is the widely shared 100–104 dB cabin noise figure for the A330 verified?
No. The figure’s origin and measurement basis could not be established. The one compilation carrying that range points to an undated website entry that names no aircraft, no measurement position, no weighting and no duration, so it cannot support a claim about A330 cabin peaks.
Does the Applied Acoustics study confirm a 75.7 dB(A) takeoff figure?
Not in the record retrieved here. The study’s A330 cruise averages are documented, but a 75.7 dB(A) takeoff value was not confirmed, and neither was the paper’s DOI or its peer-review status.
