A workplace can register as fairly quiet on a sound level meter and still feel unpleasant to sit in all day, because loudness is only one part of what makes an office comfortable acoustically. What the sound is made of, how predictable it is, and how long it lasts all shape whether a room actually feels comfortable to work in.
A single decibel number tells you how much sound energy is in a room, not how that room feels to the people sitting in it.
Research across 64 office buildings and 1,340 workstations assessed this directly: measured sound level, and even more sophisticated single-number ratings like Room Criteria, Noise Criteria, and Balanced Noise Criteria, were not significantly correlated with how satisfied occupants were with the acoustics - physical building attributes were. Every conventional way of boiling sound down to one number failed to predict how people actually felt about it.
In that same study, cutting the number of distributed noise sources in a space from over 40% down to under 2% raised satisfaction by around 21 percentage points.
ASHRAE's design guidance for building acoustics splits sound perception into two separate questions: how loud a sound is relative to normal activity, and the quality of that sound, meaning whether it reads as a rumble, a hum, a hiss, or a tone. A sound can sit within an acceptable loudness range and still draw complaints because of what it sounds like.
The same conflation problem that affects metrics also affects evidence. Much of the widely quoted research on office environments compares open-plan against enclosed offices, which differ in shared air, density, temperature, visual distraction, and privacy all at once, and then attributes the result to noise. Isolating sound and measuring it on its own terms matters as much in the evidence as it does in the building.
One study did exactly that. A repeated-measures study tested the same participants in the same room under two conditions, changing only the soundscape playing in the background: a typical open-plan noise level and a quieter private-office level. With everything else held constant, there was no significant difference in an immediate proofreading task between the two conditions.
However, heart rate and skin conductance response (physical markers of the body's stress response), along with self-reported mood, were all significantly worse in the louder condition. Acoustic discomfort showed up in stress and mood well before it showed up in a performance score, and neither would have appeared on a decibel reading alone.
The same loudness can read completely differently depending on where its energy sits in the frequency range. A steady low rumble from the HVAC system, a mid-range roar, and a high-pitched hiss can all measure the same overall level while producing very different reactions from the people who have to sit near them.
ASHRAE's guidance notes that even at moderate overall levels, background noise dominated by very low frequencies can leave some occupants with a sense of oppressiveness after several hours in the space, a complaint that a short visit to the room would not surface.
ANSI/ASA S12.2, the standard for evaluating room noise, includes a dedicated Room Noise Criteria (RNC) method built specifically for low-frequency fluctuating noise like this, on top of its standard A-weighted approach, because a flat dB number doesn't capture it.
A controlled experiment verified this by holding a fixed decibel level and changing only the frequency spectrum. Two ventilation noises played at the same 40 dBA produced measurably different effects: the one weighted toward low, rumbling frequencies (31.5 to 125 Hz) blunted the normal afternoon decline in cortisol among participants who were generally sensitive to noise, and led to worse proofreading and slower verbal reasoning across the group, at a volume most sound meters would call identical to the other noise.
The goal for any indoor space is to maintain a well-balanced sound spectrum, with no single frequency range standing out, to ensure that sound fades into the background. An unbalanced spectrum will continually announce itself no matter how quiet the meter says the room is.
Pink noise, the kind built into many sleep and focus machines, is a familiar example of a balanced spectrum in action. Its energy tapers off steadily at higher frequencies instead of sitting flat like white noise, which is part of why it reads as an even, rain-like hush rather than a rumble, hum, or hiss demanding attention.
A 2026 sleep-lab trial played continuous pink noise alongside simulated traffic noise, and found that the well-balanced spectrum measurably reduced both the sleep disruption and the short-term metabolic stress markers that the traffic noise caused on its own.
Two sounds at the exact same volume can have very different effects on focus, depending on whether that volume holds steady or keeps shifting. One listening study compared a continuous speech-like noise against a variable speech-like noise, and found that the variable version was more impairing on a memory task and rated more annoying than the continuous version.
That finding lines up with the largest analysis available on this question. A pooled analysis of 70 separate cognitive-disruption experiments found that a single measure, fluctuation strength (how much a sound's level and spectrum vary over time), explained 55% of the variance in how disruptive a sound turned out to be, correctly predicting the outcome in 63 of the 70 experiments.
That pattern shows up constantly in a real office. A colleague's footsteps, a door closing, and a phone buzzing are all individually brief and quiet, but together they create a stream of small, unpredictable shifts that a single steady hum from ventilation never would. A room can hold a constant sound level and still feel far more disruptive because the swings, not the average, are what the brain keeps reacting to.
Research measuring fluctuating noise against a steady reference sound put a number on that penalty. At an identical A-weighted level, a fluctuating sound is about as annoying as a steady one 11 to 12 decibels louder, and that penalty shows up with as little as 2 dB of modulation. Tonal sounds carry a similar 8 to 12 dB penalty, and impulsive sounds like a bang or slam add about 8 dB more.
In practice, that means a 45 dBA room with constant swings can end up feeling like a 56 dBA room to the people sitting in it.
A short burst of noise and the same noise running for eight hours are not the same problem. In the same repeated-measures study on open-plan sound, the exposure period was only 8 to 10 minutes per condition, and the researchers noted that this short duration likely explains why cognitive performance held steady even as stress and mood measures worsened.
They pointed out that real workdays involve continuous exposure, not a single short lab session, and expected the physiological and emotional effects to compound over a full day rather than level off.
ASHRAE's design guidance points to the same pattern from the building side: a complaint about an oppressive, low-frequency-heavy background often only surfaces after several hours of exposure, not during a quick walkthrough. A sound that is easy to tolerate for a five-minute meeting can easily wear an employee down by 4:00 pm, which is exactly the kind of effect a one-time spot measurement is built to miss.
What makes an office comfortable acoustically is a combination of four factors playing out together, not any single decibel reading: loudness, frequency mix, variability, and duration. None of these show up in a spot-check measurement taken at one moment in time; acoustic comfort is a combination of all four, unfolding over the course of a whole workday.
That is the core argument for measuring sound the same way you would measure indoor air quality: continuously and across several parameters to get the complete picture.
Capturing all four means going beyond a single averaged decibel figure. A useful read on a room's acoustic comfort needs the overall level, how that level is distributed across frequencies, and how much both of those swing over the course of a day.
Understanding what actually shapes acoustic comfort is the first step. Acting on it means being able to see how loudness, sound character, and variability behave in your own space over real working hours, not just during a single walkthrough.
To learn more about how you can optimize acoustic comfort to support employee experience, health, and well-being, reach out to our team to talk to an IEQ expert.