The Acoustic Audit: What Device Noise Figures Actually Tell You

Almost every product in this category is described as quiet. Very few manufacturers explain what they measured, at what distance, on which setting, or against what background. Those four omissions matter more than the number itself, because the value that actually determines whether a device is audible in the next room is not the one printed on the box.

This audit sets out what the measurements mean, where published figures mislead, and how to reach a defensible conclusion about your own building rather than a vendor’s conclusion about theirs.

Why “whisper quiet” is not a measurement

Quiet is a comparison, and a comparison requires a reference. Without one, the word carries no information. The three reference points that matter are:

• The ambient noise floor in the room where a listener is.
• The distance between the device and the listener’s ear, including the structure in between.
• The frequency distribution of the sound, because a wall does not treat all frequencies equally.

A manufacturer who publishes one number and no conditions has told you nothing you can act on. A manufacturer who publishes a range, a distance, a setting and a method has told you something usable, even if their device is not the quietest one available.

Sound pressure level, and the reference points that make it meaningful

Sound pressure level is a logarithmic ratio, which is why it behaves in a way that surprises people. A 10 dB increase represents roughly a doubling of perceived loudness and about ten times the acoustic power. A 3 dB change is the smallest difference most people reliably detect under normal conditions.

Environment	Approximate level	What it tells you
Acoustic treatment, studio	15 to 20 dB	Below most residential conditions
Quiet bedroom at night, nothing running	22 to 28 dB	The hardest case for discretion
Typical bedroom, distant traffic	30 to 35 dB	The realistic baseline for most homes
Refrigerator or air conditioning nearby	35 to 40 dB	Useful masking
Air purifier on high, or a fan	38 to 45 dB	Substantial masking
Normal conversation	55 to 60 dB	Not a realistic night-time floor

The practical implication is that the same device can be perfectly discreet in one room and clearly audible in another, without anything about the device changing.

Distance: the factor that gets left out

In a free field, sound pressure falls by approximately 6 dB for every doubling of distance from the source. Move from 30 cm to 60 cm and you lose about 6 dB. Move from 30 cm to 1.2 m and you lose roughly 12 dB.

Real rooms are not free fields. Reflections and reverberation reduce the fall-off, and in a small, hard-surfaced bathroom the level barely drops at all over a metre. The principle still holds: proximity matters enormously, and any device used right against a shared wall is at a disadvantage regardless of its rating.

This is also why a figure quoted without a distance cannot be compared with another figure quoted without a distance. Ten centimetres and one metre are not the same measurement, and in the free-field model the difference between them is around 20 dB.

What a wall actually attenuates

Published partition ratings come from laboratory assemblies tested in ideal conditions. Real buildings perform worse, because of construction tolerances, services penetrations and flanking paths that bypass the wall entirely.

Separating structure	Indicative field performance	Practical note
Hollow-core interior door	15 to 20 dB	Often the weakest link in the assembly
Interior stud wall, one layer each side	25 to 30 dB	The most common apartment partition
Stud wall with cavity insulation	30 to 35 dB	Meaningfully better than an uninsulated one
Double layer board one side	35 to 40 dB	Diminishing returns, substantial weight
Brick or block, 100 to 150 mm	40 to 45 dB	Residential party wall territory
Concrete, 200 mm and above	45 to 55 dB	Discretion stops being the constraint

The table is a starting point, not a verdict on your building. Flanking paths are the most common reason a well-built wall fails in practice: back-to-back electrical outlets, shared ductwork, an open ceiling void, or a gap under a door can each claw back more attenuation than the wall itself provides.

The frequency problem the labels omit

This is the part of the discussion that almost never appears in product marketing, and it is the part that determines the answer.

A typical lightweight partition attenuates high frequencies very effectively and low frequencies poorly. This is not a quirk of cheap construction; it follows from mass law and coincidence effects. It is also why you can hear a neighbour’s bass line but not their television dialogue through the same wall.

The consequence for product selection is counterintuitive. A device whose output is dominated by low-frequency vibration can be more audible through a wall than a device that measures louder but produces mostly high-frequency content, because the low-frequency energy is not being stopped. The sensation of depth that people describe as rumbly is, acoustically, the property that transmits best through a building.

Muffling behaves the same way. A duvet or a folded towel absorbs high-frequency content efficiently and low-frequency content poorly. Covering a buzzy device therefore helps considerably, while covering a strongly rumbly device helps only a little. If discretion is the priority, motor character deserves as much attention as the decibel figure. Our article on motor architecture explains why the two profiles exist.

Masking: the intervention that costs nothing

Because audibility is a relationship between transmitted level and ambient level, raising the ambient level is functionally equivalent to lowering the transmitted level. It is simply much cheaper.

Moving a room from a 28 dB floor to a 38 dB floor buys 10 dB of masking, which is typically more than any product change would deliver. A fan, an air purifier, a white noise source or a television at low volume all work on the same principle. The trade-off is that the masking source has to be running before and during use, not started afterwards.

Physical measures come second, and are worth doing in this order: a rug, soft furnishings, a bookcase loaded with books against the shared wall, a towel along the bottom of a door, and sealing obvious gaps around services. Each is inexpensive, and each is measurable with a phone sound level application.

Practical protocol

1. Establish the ambient floor in the room someone else occupies, at the time of day in question.
2. Identify the weakest structural element between the two rooms. It is usually a door, a vent or a gap rather than the wall.
3. Enter the device’s stated level, the measurement distance, the structure and the ambient floor into the Noise Threshold Calculator to get an estimated margin.
4. Change one variable at a time and note the effect. Product character and ambient masking will move the result more than small differences in decibel figures.
5. If the margin stays above roughly 10 dB under your hardest conditions and discretion matters, the honest conclusion is that the product is not suited to that room.

Where this gets misreported

The recurring errors are worth naming, because they appear in most category coverage:

• Comparing numbers measured at different distances. This is the most common defect, and it invalidates the comparison entirely.
• Treating the maximum setting as representative. Many people use a device well below maximum, so a figure taken at the loudest setting understates the discretion of normal use.
• Assuming quiet means weak. Noise correlates with housing isolation and frequency profile, not with output capability. Several of the lowest-noise devices we cover deliver more usable low-frequency energy than louder competitors.
• Treating one unit’s measurement as universal. Unit-to-unit variation within a production run is real. Ranges are honest; single values claim a precision that does not exist.
• Ignoring the listener’s room. Two identical devices in identical rooms produce different outcomes depending on what is already making noise in the adjacent one.

For products ranked specifically on measured noise bands, see the quietest devices ranking. For the wider question of keeping a routine private in shared housing, start from the complete privacy guide.

Related

• Privacy at Home: the complete guide - the four layers of discretion in shared housing
• Discreet storage - containment, and why storage choice affects material life
• Shared bedroom strategies - privacy when you share the room itself
• Noise Threshold Calculator - run the estimate for your own building