What sound masking is
Sound masking is the introduction of a carefully engineered background sound into a space to make conversations harder to follow, without eliminating sound altogether. It does not block, cancel or silence anything; it simply makes speech blend into the background.
The problem it solves
Sound masking is one of the most consistently misunderstood acoustic technologies in the workplace. It is not noise-cancellation, it is not white noise, and it does not block sound. To see why it works, it helps to start with the everyday problem.
In a quiet office, a conversation a few desks away is not just audible, it is followable. You can make out the words, and once you can make out the words, your brain pays attention whether you want it to or not. That is the real cost of overheard speech: not the volume, but the meaning. Sound masking tackles exactly this. By gently raising the steady background sound in a room, it narrows the gap between that background and any overheard speech, so the words slip below the point where they can be consciously followed. Why it is intelligibility, not loudness, that determines privacy is explained in full in The Science of Speech Privacy.
What you actually experience
In plain terms, a well-set-up masking system feels like a slightly more present, neutral background sound: a little like soft airflow, but smoother and more even. A well-calibrated masking system is barely noticeable. Most people who do perceive it would identify it as a low-level ventilation sound, and both research and installation experience show that the majority of occupants don’t even notice it once the system has been commissioned. The privacy benefit, though, is there from the moment the system is switched on: from day one, people nearby find it noticeably harder to follow your conversations, even as they cease to notice the masking itself.
A useful way to hold the idea: a good masking system is designed to be noticed only in its absence. If people are constantly aware of it, it has probably been set too loud. This is also why masking is different from simply blocking sound; that distinction has its own article, Sound Masking vs Soundproofing, coming shortly.
Where it fits, and where it does not
Masking is powerful, but it is not a cure-all, and being honest about that matters. It works best as one of three levers, alongside reasonable partitions and sound absorption, rather than as a substitute for them. Used alongside fit-for-purpose partitions and a sensible layout, it is a strong component of a privacy strategy. Used on its own in an acoustically poor space, its contribution is reduced, because it would have to be pushed uncomfortably loud to compensate.
Installation, reassuringly, is quick and undisruptive: for a typical small to medium office it takes about a day, with no structural work. For regulated sectors, masking can also form part of a broader, documented approach to protecting confidential information; our sector-by-sector compliance guides cover that picture in detail.
Inside the technology
The rest of this article is the technical reference: how masking lowers the relevant metric, how the signal is engineered, how it is set, and what a real installation involves.
1. Masking lowers the Speech Transmission Index, it does not cancel sound
Sound masking — a carefully shaped background sound that raises a room’s noise floor to reduce speech intelligibility — does not block or cancel sound. The distinction matters. Active noise cancellation, as used in headphones, generates an opposing sound wave to neutralise a specific signal. Masking does nothing of the sort. It raises the level of background noise, the ‘noise floor’ of the room, so the difference between background and overheard speech is reduced. When that gap narrows sufficiently, the Speech Transmission Index, STI — a measure, from 0 to 1, of how much of what is said can actually be understood, defined by IEC 60268-16 and used in ISO 9921 — falls, and the brain can no longer parse the speech as intelligible. In the terms of BS EN ISO 3382-3:2022, this shortens the distraction distance: the distance from a talker at which the STI falls to 0.5.
2. The masking signal is engineered, not random
The sound produced by a masking system is not a generic hum, a fan, or randomised white noise. It is a signal deliberately shaped to the frequency range of human speech, approximately 300 Hz to 4 kHz. The shaping matters because it is speech intelligibility, not broadband noise, that needs to be reduced. Adding high-frequency hiss or low-frequency rumble would create discomfort without addressing the problem. A well-specified masking signal targets the frequencies where speech carries meaning, particularly the consonant range, and leaves frequencies outside that band largely untouched.
The Routledge Handbook of High-Performance Workplaces (2024), Chapter 9, ‘Acoustic privacy’, by Yadav and Cabrera (CC-BY-NC-ND 4.0), drawing on acoustic measurements across 36 open-plan office floors and a separate occupant survey of 426 participants, identifies that speech distracts specifically through its roughly 4 Hz syllabic fluctuation, or fluctuation strength (a concept set out in Zwicker and Fastl, Psychoacoustics, 1999). Hongisto (2005) models how this varying intelligibility maps onto measurable losses in work performance. A well-engineered masking signal addresses this mechanism directly, by raising the noise floor in the frequencies where that fluctuation is most audible.
3. Level is critical: too low has no effect, too loud creates a new problem
Sound masking systems have a commissioning phase in which the output level is set for each zone of a building. This is not a set-and-forget step. Too low, and the masking signal does not raise the noise floor enough to lower the STI: it simply adds a faint hum that occupants notice without benefit. Too loud, and the masking signal itself becomes a distraction, and may make normal conversation effortful.
In practice, a masking signal is typically between 38 and 40 dB(A) (the unit of loudness weighted to match human hearing), though the exact level depends on the specific office and can be lower. This sits below the occupied-level targets that BS ISO 22955:2021 associates with good acoustic quality (about 48 dB(A) for limited-collaboration spaces up to 55 dB(A) where the office receives the public). These figures are reported via the Routledge chapter (Yadav and Cabrera, Ch.9), which cites the standard. Professional commissioning by a trained installer is therefore not optional: it is what separates a system that works from one that does not.
4. Masking works best with reasonable partition performance
Sound masking is a powerful tool, but it is not a substitute for partitions. Privacy depends on both how much sound the structure blocks and how much background noise covers what passes through. A space with very poor partition performance requires a very high masking level to achieve adequate privacy, which may push the background above a comfortable threshold. The most effective outcomes, the shortest distraction distances and the lowest STI, combine a partition specification that is fit for purpose with a masking level set to close the remaining gap. Helpful context here is the spatial decay rate of speech, D2,S (how quickly speech level falls with each doubling of distance), and reverberation time (how long sound takes to fade in a room): absorption that improves both makes masking’s job easier.
For new builds and refurbishments, acoustic performance should be specified at design stage alongside the masking system, not as an afterthought. For existing offices, a modest improvement to partition seals and door closers, combined with a calibrated masking system, often delivers a significant improvement at reasonable cost.
5. Installation is fast and non-disruptive
A standard sound masking installation, for a typical small to medium office, usually takes about a day and requires no structural work. Speakers are installed above a suspended ceiling (or surface-mounted where no ceiling void exists), connected to a central controller, and commissioned zone by zone.
A relevant example from our own work: Sound Directions installed a masking system in the meeting suite of a UK law firm, completed out of hours, with before-and-after sound-level data collected to verify the improvement. The brief was driven by client confidentiality requirements under paragraph 6.3 of the SRA Code of Conduct, the same obligation that applies to most UK solicitors. The full case study is coming to this site shortly.
Sound Directions’ core small-office masking system covers spaces up to 5,000 sq ft (465 m²), supports one to two zones, with a single controller driving up to 8 masking speakers, and is compatible with all standard ceiling types. Different types of space, a private office and a corridor for instance, are given their own controllers so each can be set correctly. Commissioning is managed via a touch-screen controller.
6. People stop noticing the masking signal
The effects of well-commissioned masking are not dramatic. Masking works subtly, as it should. A well-calibrated masking system is barely noticeable: most occupants who do perceive it would identify it as a low-level ventilation sound, and both research and installation experience show that the majority stop noticing it once the system has been commissioned. This is the expected and intended outcome. A masking system that people are constantly aware of has probably been set too loud.
The effect on speech privacy, however, is immediate from the point of commissioning. Occupants in adjacent spaces or open areas will find it noticeably harder to follow conversations from day one, even as they cease to notice the signal itself. For a comparison with US-style metrics, note that American sources often describe privacy using Speech Privacy Potential, SPP (= STC + NC), combining a partition’s Sound Transmission Class (STC, how much sound a partition blocks) with the Noise Criterion (NC, a rating of background noise level) on a 60 to 90 scale; it is a useful intuition but not the standard the UK designs to.
Frequently Asked Questions
Is sound masking the same as white noise?
No. White noise contains equal energy at all frequencies across the audible spectrum: it sounds like broadband static and is perceptually harsh at the levels needed for speech privacy. A masking signal is shaped specifically to the speech frequency range (approximately 300 Hz to 4 kHz), which makes it far more effective at reducing intelligibility and considerably more comfortable for occupants. The terms are sometimes used interchangeably, but they describe very different signals.
Will I hear the masking system?
A well-calibrated system is barely noticeable. Most people who do perceive it describe it as a low-level ventilation sound, and the majority stop noticing it once the system has been commissioned. A correctly commissioned system should feel like a slightly more present ambient environment, not like an audible machine. If occupants are consistently aware of the system, that is a sign the level may need adjustment.
Can sound masking be installed in a listed building?
In most cases, yes. Because masking speakers are typically installed above a suspended ceiling or mounted unobtrusively on existing surfaces, the installation is usually reversible and does not affect the fabric of the building. However, listed building consent requirements vary, and you should confirm with your local planning authority or conservation officer before proceeding. A specialist installer will be familiar with these constraints.
How long does a sound masking installation take?
For a typical small to medium office, about one working day. Speakers are fitted above ceiling tiles or surface-mounted, cabling is routed through existing voids, and the system is commissioned and calibrated zone by zone on the same day. There is no structural work, so the space remains usable with minimal disruption.
What does a sound masking installation cost?
Less than most people expect, because there is no building work involved. For a single room, the simplest system – one controller and two masking speakers – starts at around £800 including VAT, and a single controller drives up to 8 speakers. Larger or mixed spaces cost more, mainly because different types of space, a private office and a corridor for instance, need their own controllers so each can be set correctly. If your office has 50 people or fewer, working out the cost is straightforward: the calculator at speechprivacy.co.uk will turn your room dimensions into a requirement and an indicative price, or you can contact us directly if you would like a hand with it. If your office is larger than that, it is likely we will need to talk it through with you first.
Find out whether sound masking is right for your office
Sound Directions has specified and installed sound masking in UK offices and regulated environments for over 25 years. If you would like to understand whether masking could improve speech privacy in your space, or if you are working on a specification and want a technical conversation, we would be glad to help.
If your office has 50 people or fewer, it is straightforward: visit speechprivacy.co.uk and use the calculator to work out your requirements, or contact us directly if you would like help. If your office is larger, it is likely we will need to talk to you about it – and speechprivacy.co.uk has information that will help you with your decision.
Sources and further reading
Yadav, M., and Cabrera, D. (2024). ‘Acoustic privacy’. In Routledge Handbook of High-Performance Workplaces, Chapter 9. Routledge. CC-BY-NC-ND 4.0. Open access.
Hongisto, V. (2005). A model predicting the effect of speech of varying intelligibility on work performance. Indoor Air, 15(6), 458–468 (paywalled).
Zwicker, E., and Fastl, H. (1999). Psychoacoustics: Facts and Models (2nd ed.). Springer (paywalled).
Standards (official catalogue pages, paywalled): BS EN ISO 3382-3:2022; BS ISO 22955:2021; IEC 60268-16:2020; ISO 9921:2003; BS 8233.