Sound masking for offices and apartments: what actually works
The disruptive sound isn’t the loud one. It’s the conversation you can just make out — and that changes everything about how to fix it.
General information, not medical advice. Talk to a doctor about any health concern. Keep the volume low, especially for long sessions or sleep. For all-night listening, use a sleep timer.
There is a specific kind of noise that ruins concentration, and it is not the loud kind. It is the colleague two desks away having a perfectly ordinary phone call, at a perfectly ordinary volume, about something that has nothing to do with you.
This turns out to be the central finding of open-plan office acoustics, and it changes what you should do about it. The problem is not volume. It is intelligibility, and that distinction is the difference between fixing your sound environment and making it worse.
Speech you can understand is usually the most disruptive sound in an office, and it impairs performance on memory-heavy tasks even when you are not listening to it. Masking works not by drowning speech out but by making it unintelligible — which takes far less volume than people assume. Steady, featureless sound at a modest level beats loud sound, and music with lyrics is the one thing you should not reach for.
Why speech is the problem
Experimental psychology has a name for this: the irrelevant speech effect. Background speech that has nothing to do with your task measurably degrades cognitive performance, particularly on anything involving memory, reading or writing.1 Across open-plan office research, irrelevant speech is consistently reported as the most distracting sound in the environment.2
The reason is that language is not optional. Your brain processes speech it can make out whether you want it to or not, and the machinery it uses for that overlaps with the machinery you are trying to use for your own work. Steady, meaningless sound does not compete in the same way, which is why the general hubbub of a busy café can be easier to work in than a quiet office with one clear conversation.
A survey of more than a thousand office workers found that irrelevant speech increased annoyance, reduced self-assessed performance, and was associated with worse wellbeing, with these effects more pronounced in open-plan offices than in smaller shared rooms.3
Intelligibility, not loudness
Here is the counterintuitive part that acousticians build entire buildings around. What predicts disruption is not how loud speech is, but how understandable it is. The measure used is the Speech Transmission Index, running from 0 (unintelligible) to 1 (perfectly clear).
The international standard for open-plan offices, ISO 3382-3, formalises this with a metric called the distraction distance: the distance from a talker at which intelligibility falls below an STI of 0.5.1 Inside that radius, a voice is clear enough to be highly disruptive. Beyond it the disruption falls away, though research suggests speech only stops affecting performance at much lower intelligibility, around an STI of 0.2.2
This is why sound masking is engineered rather than simply switched on. The goal is to raise the noise floor just enough to pull speech below that intelligibility threshold, shrinking the distraction distance, without adding a new annoyance of its own.2
What to actually play
Steady and featureless wins. Commercial systems typically use filtered pseudo-random noise shaped to resemble ventilation sound: it is cheap, royalty-free and steady, and steady sound is easy to habituate to — exactly what you want from something running eight hours a day.2
Water gets mixed results. In a laboratory study, a spring-water sound was the best-rated speech masker.4 But a later three-week field trial in a real office found water sounds rated worse than conventional masking noise.2
Avoid lyrics entirely. Music with words reintroduces the exact problem you are solving. If you want music, keep it instrumental — see music for focus for what the evidence supports.
Resist the volume reflex. Turning masking up until you cannot hear anyone is usually unnecessary and makes the masking itself annoying. You need speech to be unclear, not inaudible.
The apartment problem is different
Domestic noise is not office noise. The disruptions are intermittent rather than continuous: a door, a bass line, footsteps overhead, a dog. The goal is usually sleep rather than concentration.
Masking may help some people, because a steady noise floor reduces the contrast of a sudden sound, and contrast is what jolts you awake. But it is worth being honest about the evidence: a systematic review of 38 studies found the quality of evidence that continuous noise improves sleep to be very low, which the authors noted sits awkwardly against how widely it is used.5
Masking is also poorly matched to low-frequency intrusion. Footfall and bass often travel through the building structure, and masking covers them poorly. Physical measures help more — sealing door gaps, and soft flooring for higher-pitched impact noise — though deep bass is hard to stop without structural work.
Keep bedroom sound low, no louder than background conversation, and place the machine in a corner of the room rather than beside your head.6 This matters more for infants: a 2014 study in Pediatrics found some infant sleep machines can exceed safe levels at full volume, so keep them well away from the cot and at low settings.7
The pattern underneath all of this is the same: the sounds that damage concentration are the ones carrying meaning. Fix the meaning and the volume mostly takes care of itself.
Frequently asked questions
Because your brain processes intelligible speech automatically, using resources that overlap with reading, writing and memory. This is known as the irrelevant speech effect, and it degrades performance on memory-heavy tasks even when you are not consciously listening.
No. Masking aims to make speech unintelligible rather than inaudible, which requires far less level than people assume. A common benchmark is a Speech Transmission Index below 0.5, though speech keeps disrupting some work until it is much less clear (around 0.2).
Steady, featureless sound you stop noticing — filtered noise shaped like ventilation is the commercial standard. Water sounds get mixed results: best in one laboratory study, rated worse than conventional masking noise in a real office. Avoid anything with lyrics, which reintroduces the problem.
Not really. Low frequencies often travel through the building structure, and masking works poorly against them. Physical measures such as sealing door gaps help more, though deep bass is hard to stop without structural work.
Sources
- Kostallari, K., Parizet, E., Chevret, P., Amato, J.-N. & Galy, E. (2020). Irrelevant speech effect in open plan offices: comparison of two models explaining the decrease in performance by speech intelligibility and attempt to reduce interindividual differences of the mental workload by task customisation. Applied Acoustics, 161, 107180. (On the ISO 3382-3 distraction and privacy distances, and the STI = 0.5 threshold used to limit effects of speech on work performance.) doi.org
- Hongisto, V., Varjo, J., Oliva, D., Haapakangas, A. & Benway, E. (2017). Perception of water-based masking sounds — long-term experiment in an open-plan office. Frontiers in Psychology, 8, 1177. (Field study: employees rated conventional pseudo-random masking better than water sounds. Also on filtered pseudo-random noise as the standard masking signal, habituation to steady-state sound, and the evidence that irrelevant speech impairs cognitive performance.) frontiersin.org
- Di Blasio, S., Shtrepi, L., Puglisi, G. E. & Astolfi, A. (2019). A cross-sectional survey on the impact of irrelevant speech noise on annoyance, mental health and well-being, performance and occupants’ behavior in shared and open-plan offices. International Journal of Environmental Research and Public Health, 16(2), 280. (1,078 respondents; effects more pronounced in open-plan than shared offices.) pmc.ncbi.nlm.nih.gov
- Haapakangas, A., Kankkunen, E., Hongisto, V., Virjonen, P., Oliva, D. & Keskinen, E. (2011). Effects of five speech masking sounds on performance and acoustic satisfaction: implications for open-plan offices. Acta Acustica united with Acustica, 97(4), 641–655. (Spring-water sound was the best masker in a 54-person laboratory study.) (author-deposited copy on a repository; the underlying study is peer-reviewed) researchgate.net
- Riedy, S. M., Smith, M. G., Rocha, S. & Basner, M. (2021). Noise as a sleep aid: a systematic review. Sleep Medicine Reviews, 55, 101385. (38 studies; GRADE quality of evidence for continuous noise improving sleep rated very low.) pubmed.ncbi.nlm.nih.gov
- Can white noise really help you sleep better? (clinician-reviewed explainer on safe bedroom volume and speaker placement). Harvard Health Publishing. health.harvard.edu
- Hugh, S. C., Wolter, N. E., Propst, E. J., Gordon, K. A., Cushing, S. L. & Papsin, B. C. (2014). Infant sleep machines and hazardous sound pressure levels. Pediatrics, 133(4), 677–681. (At 30 cm, every machine tested could exceed 50 dBA, the recommended limit for infants in hospital nurseries; three could exceed 85 dBA.) pubmed.ncbi.nlm.nih.gov