What does an acoustic survey of a school actually measure? Decibels are only the beginning

“This room is at 58 decibels. Is that a lot?” - the question asked almost every time a school receives its first measurement result. The honest answer is: there is no way to know. No way, because the number alone does not say what was measured or for how long, in an empty room or during a lesson, at which point in the space, with the windows and ventilation in what state, with what instrument - or, above all, what the measurement was for. So before answering whether 58 is a lot, it is worth asking something else: what actually goes into a professional acoustic survey of a school?

The question first, the meter second

A good acoustic survey starts not with equipment but with defining the problem. The reports schools bring usually sound something like this: “the teacher is hard to understand”, “everything booms in this room”, “you can hear every conversation in the corridor”, “the ventilation is very loud”, “you can’t hold a conversation in the canteen”, “in the room facing the street you have to keep the windows shut”, “teachers have to raise their voices a lot”, “you hear completely differently in different parts of the room”.

Each of these sentences points to a different technical problem and each calls for a different measurement. “Everything booms” leads to reverberation time. “You can hear the corridor” - to the insulation of the building fabric. “Loud ventilation” - to measuring services noise in an unoccupied room. “The teacher is hard to understand” may follow from any of these at once and usually requires several parameters to be examined together.

Hence the rule that orders everything else: you do not fit the problem to the meter, you fit the measurement method to the question you want answered. A measurement taken without that question produces a number nobody will know what to do with.

The decibel - a number that needs context

Sound pressure level, expressed in decibels, describes how strong the sound-induced disturbance of air pressure is at a given point. One thing matters: the decibel is a logarithmic scale, not an ordinary linear measure. It should not be read intuitively - 60 dB does not mean “twice as much as 30 dB”, and a difference of a few decibels can correspond to a far greater change in sound energy than the difference in numbers suggests. Adding and comparing these values follows its own rules.

The second thing worth understanding is the notation dB(A). The human ear does not receive all frequencies equally - it is far less sensitive to low tones than to mid-range ones. A-weighting is a filter that brings the measurement closer to this property of hearing by attenuating the lowest frequencies. That is why reports almost always show dB(A) rather than “bare” dB.

But even “55 dB(A)” still does not answer: over what period it was measured, what the source was, how much the level varied over time, where in the room the microphone stood and what was going on there at the time. Without that information it is a number, not a result.

LAeq - what an “average” sound level means

The parameter most often found in reports is LAeq - the A-weighted equivalent sound level over a defined period. Put simply: it is the value describing what steady level would carry, over the period studied, the same acoustic energy as actually occurred. It is not the arithmetic mean of the readings on the display - because of the logarithmic nature of the scale, loud moments weigh far more in this calculation than quiet ones.

Two lessons with a similar LAeq make this clear. In the first the level stays roughly constant: focused work, a steady hum. In the second it is quiet most of the time, but a dozen or so very loud events occur. The same number in the report, two entirely different profiles - and for the teacher running the lesson and the pupil trying to concentrate the difference is fundamental.

LAeq is therefore very useful, but it does not describe everything and should not be called “the school’s noise level”. Whenever an LAeq value is given, the period it refers to must be stated.

LAmax and single events

Where the problem is not general conditions but specific events, maximum levels become the interesting figure - LAmax. A slammed door, a chair dragged across the floor, an aircraft overhead, the bell, a ball hitting the wall of the sports hall, a machine switching on loudly. Events of this kind interrupt speech and the flow of work whatever the averaged level may be.

The difference between the two views is simple: LAeq describes exposure spread over time, LAmax a single peak. Both parameters are needed, but they answer different questions and do not substitute for one another. It is also best to avoid setting “safe / unsafe” thresholds from LAmax alone - such limits must rest on a normative or health basis, not on intuition.

Frequency - why one number can hide the problem

The sounds encountered in a school are not single tones - they consist of many frequencies at once. The low hum of building services occupies a completely different part of the spectrum from human voices, which in turn differ from high-pitched squeals or brief impacts. That is why a specialist analyses the result not only as a single number but in octave bands or narrower third-octave bands - that is, in successive frequency ranges separately.

The reason is practical. Two situations with almost identical dB(A) values can have entirely different frequency characteristics, and therefore a different cause and a different solution. An absorbing material effective in the mid frequencies may change little against the low hum of ventilation. So the question “how loud?” sometimes has to be supplemented by: “which part of the spectrum is the problem in?”

Reverberation time - how long sound stays in a room

Reverberation is the overlapping reflections of sound from walls, ceiling and furnishings that persist after the source has fallen silent. Reverberation time (RT) describes the rate of that decay - the longer it is, the more successive syllables overlap. We write more fully about the phenomenon and its effects in the article on reverberation and speech intelligibility.

Measurement reports rarely show “RT” on its own - most often the symbols T20 and T30 appear. This is a frequent source of confusion, so let us be explicit: they do not mean 20 or 30 seconds. They are ways of determining reverberation time from a defined section of the decay curve - a fall of 20 and of 30 decibels respectively - from which the result is extrapolated to the full decay. In practice the entire fall referred to in the classic definition can rarely be recorded.

A professional measurement of reverberation time requires a defined procedure: a suitable signal source, several source and microphone positions as prescribed by the method used, analysis in frequency bands and controlled conditions in the room. The basic methodological document here is ISO 3382-2, on measuring reverberation time in ordinary rooms [1]. The current status and edition of this document should be checked in the official ISO catalogue before citing it in documentation.

An empty room and a full room are not the same

The presence of people changes a room’s acoustics - pupils, the teacher, their clothes, bags and equipment absorb sound. Open or closed doors, lowered blinds and the arrangement of furniture matter too. The same room measured on Friday after lessons and on Wednesday at ten will give different results.

Measurement in an empty room assesses the properties of the room itself according to a defined, repeatable procedure - and compares them with requirements, which usually refer to exactly those conditions. It is repeatable, because it does not depend on how many pupils happened to turn up.

Measurement during normal lessons shows something else: the actual sound level, how the space is used, the teacher’s voice level, the background the class generates and the variability across a lesson.

There is no hierarchy here - not a better and a worse measurement, but two different ones answering two different questions. An example of research in real conditions is the work of Bradley (2008), covering detailed measurements in 41 working primary classrooms in Canada [2]. It shows how many parameters have to be gathered to describe the conditions for speech communication in a room in use. Caveat: average values obtained in Canadian schools are not requirements for Polish ones and should not be treated as such.

SNR - does the voice have the advantage over the background

SNR (signal-to-noise ratio) describes how far the level of the speech signal exceeds the interfering background at a given place and moment. The intuition is simple: if the teacher’s voice only slightly tops conversations, ventilation and sounds from the corridor, receiving the message becomes harder. The greater the signal’s advantage, the better the conditions.

SNR is not, however, the only condition for good communication. What a pupil actually receives is also shaped by reverberation, distance from the speaker, the character of the noise (steady hum disturbs differently from background speech), the listener’s age, the type and predictability of the message and where they sit in the room.

The systematic review by Murgia and colleagues (2023), covering 23 publications meeting the inclusion criteria, shows that SNR and STI were among the most frequently analysed parameters in research on speech intelligibility in primary classrooms [3]. The very fact that both appear in parallel in this literature is telling: they describe different aspects of the same question and are usually studied together.

STI - a speech transmission index, not a “percentage understood”

STI (Speech Transmission Index) is a number from 0 to 1 describing how faithfully a room and the conditions in it transmit those features of the speech signal that carry its legibility. It takes account of background noise, reverberation and the degradation of signal modulation - the loss of the rapid changes in loudness in which much of the information about individual speech sounds is encoded.

One simplification must be firmly rejected. An STI of 0.60 does not mean a pupil understands 60% of the words. It is not a percentage of anything - it is a technical indicator describing signal transmission conditions, derived from measurement, not from a test on people. Actual comprehension depends additionally on the individual child’s hearing, their knowledge of the language and the topic, their attention and the context of the utterance.

The basic methodological document here is IEC 60268-16, on the objective rating of speech intelligibility by the STI [4]. Before citing it, check the current version and any amendments. The quality categories it contains are also not automatically requirements for Polish schools - that needs an appropriate Polish basis.

Several parameters together say more than one

A good example of why it is worth measuring several things at once is the work of Rabelo and colleagues (2014) [5]. The study covered 18 classrooms in 9 public schools in Belo Horizonte and 273 pupils with a mean age of about 9.4. Equivalent sound level, reverberation time T30 and the STI were measured in the rooms, while the pupils took a speech intelligibility test.

Setting three parameters against an actual test result gives a picture none of them would give alone: in quieter rooms, with shorter reverberation and higher STI, pupils performed better - though in practice these parameters change together and none explained the result on its own. Caveats: a sample of 273 children is not representative of all schools, and values obtained in Brazilian schools do not translate directly into Polish requirements.

Insulation - does sound pass from one space to another

So far we have talked about what happens to sound inside a room. That is only half the subject. The sound insulation of the building fabric describes something quite different: how much sound passes between spaces. Conversation from the classroom next door, the buzz of the corridor, traffic outside the window, sounds from the floor above - that is a problem of the building fabric, not of reverberation, and it requires different measurements and different solutions.

Testing airborne sound insulation usually requires measurements on both sides of the element: in the source room, where the signal is generated, and in the receiving room, where how much of it passed through is measured. Field measurement methods are described in ISO 16283-1 [6], and single-number rating of the results in ISO 717-1 [7]. The status of both documents should be verified before citing them.

It is worth abandoning the intuition that insulation comes down to the thickness of a wall. The actual result also depends on the door (usually the weakest link in a school), gaps and leaks, workmanship, how elements are joined, flanking transmission paths - through a shared floor slab or a ventilation duct, for instance - and service penetrations. A perfect wall with a draughty door gives you the result of the draughty door.

Impact sound is a different problem again

So far we have spoken of airborne sound - sound that travels through the air and only then sets a building element vibrating. A separate category is impact sound, produced by direct contact with the structure: footsteps on the floor above, impacts on the floor, furniture being moved, movement in the corridor above a room. It travels through the building structure itself, so solutions that help against airborne sound are often powerless against it.

Field measurement of impact sound insulation is described in ISO 16283-2 [8]. The practical conclusion is single: “you can hear the class next door” and “you can hear footsteps from above” are not the same measurement problem, and one test should not be expected to answer both questions.

Services noise

A separate source of sound in a school is the building’s technical equipment: mechanical ventilation, air conditioning, pumps, fans, equipment in specialist rooms. Not every steady hum from services is undesirable by definition - it can even be masking to a degree. The problem arises when it raises the background level enough to reduce the teacher’s advantage over the surroundings, that is, when it worsens the SNR discussed above.

In the Polish system, requirements for permissible sound levels in rooms - including sound from a building’s technical equipment - are set by the standard PN-B-02151-2:2018-01 [9]. The document covers various types of room, not only school ones, and its current status should be checked in the catalogue of the Polish Committee for Standardization. It is a paid document, which is why we do not reproduce its tables or values.

Polish requirements: reverberation, speech intelligibility, insulation

The Polish system for assessing building acoustics separates three groups of issues, and three different parts of the same standard correspond to them.

Reverberation conditions and speech intelligibility in rooms, together with testing guidance, are covered by PN-B-02151-4:2015-06 [10]. The standard’s scope covers reverberation conditions, audibility, speech intelligibility and whether a room can be used for its intended purpose - so it differentiates requirements according to what a space is for.

Sound insulation of building elements is covered by PN-B-02151-3:2015-10 [11]. Particular care is needed when citing it: the standard has been amended, and quoting a historical version as current is a common error. The current edition with all amendments should be checked with the PKN.

The same editorial rule applies in all three cases: standards are paid documents protected by copyright. We can give their number, title and scope and explain in our own words what they cover - we cannot reproduce tables or values. Anyone who needs a specific figure for a specific type of room must obtain the current text of the standard or commission an assessment from someone who has it. The sentence “every classroom in Poland must have exactly X seconds”, without stating the type of room, its volume and the current version of the document, is simply untrue.

A Polish Standard, ISO and foreign guidance are not the same thing

Three different kinds of document appear in acoustic documentation, and confusing them leads to real misunderstandings.

A Polish Standard is a document within the Polish standardisation system, prefixed PN. ISO and IEC standards are international documents, most often describing methods of measurement or rating - they say how to measure, not what value must be achieved. Design guidance, in turn, consists of documents that apply or are used in a particular country.

A good example of the third category is the UK’s Building Bulletin 93 - the Department for Education’s design standard for school acoustics [12]. It is valuable as a comparison because it covers internal noise, insulation, reverberation conditions and separate, stricter requirements for rooms intended for pupils with particular hearing needs. It is not, however, Polish law and its values do not apply in Polish schools. It can be a source of inspiration; it cannot be cited as a basis for assessing compliance in Poland.

A separate question is whether a standard is the same as law. The answer is neither “standards are mandatory” nor “standards are voluntary and irrelevant”. What a particular document means depends on whether and how it has been invoked - in regulations, in technical conditions, in design documentation, in a contract or in the terms of a tender. Establishing the legal status in a specific case requires checking the current regulations; this article explains measurements, it does not give legal advice.

Why measurements are taken in more than one place

Conditions within a single room need not be uniform. It sounds one way at the teacher’s desk, another in the front row, another at the back, different again by the window, by the corridor door, directly beneath an absorbing element or near a piece of equipment. The differences can be larger than the size of the room would suggest.

That is why measurement procedures prescribe an appropriate number of microphone and source positions - their number and layout depend on the method and on what is being measured, so we give no universal figure here; the relevant normative document defines it. The rule worth remembering is simpler: one reading from one place does not describe a whole room.

Why the measurement conditions have to be recorded

A result without a description of the conditions is in practice useless. Sound measurement documentation covers the date and time, the room, the microphone and source positions, the number of people present, the state of doors and windows, whether ventilation was running, the furnishings present, the duration of the measurement, the instrument used and its calibration, the procedure applied and any unusual circumstances.

Without this information four things become impossible at once: repeating the measurement, comparing it with another, checking the effect of a refurbishment and establishing where a particular value came from at all. A school that in two years wants to check whether the works helped starts from scratch without these data.

Measuring equipment

Professional apparatus is not simply “a better microphone”. A measurement set-up usually comprises a sound level meter, a measurement microphone with known characteristics, an acoustic calibrator, a signal source and - where the method requires it - a measurement loudspeaker with defined properties, together with software for analysing the results.

Requirements for the meters themselves are defined by IEC 61672-1, which sets out accuracy classes for measuring instruments among other things [13]. This does not mean that every measurement in a school always requires one particular class of instrument - the requirements depend on the method and the purpose. The right formulation is: the equipment must be appropriate to the specific method, the scope of the survey and the requirements of the document the assessment is being made against.

What a good survey looks like

The description below is a high-level view of the process - it shows the logic, not a procedure for carrying out an expert assessment yourself.

1. Identifying the problem. What exactly users report, and in what situations.
2. Defining the scope. Which parameters have to be measured to answer that question.
3. Preparing the measurements. Choosing rooms, conditions and methods appropriate to the scope.
4. Taking the measurements. Following the proper procedures, with appropriate, calibrated equipment.
5. Analysis. Not only the value, but the frequency distribution, place, time and context.
6. Reference to criteria. If the aim is a compliance assessment - comparison with the document appropriate to that type of room and that type of measurement.
7. Interpretation. What is probably causing the reported problem.
8. Recommendations. What kind of action has a chance of helping.
9. Verification. After a larger intervention - follow-up measurements checking whether the intended effect was achieved.

The last point is the one most often skipped, and it is what distinguishes an investment that has been checked from one merely declared a success. We write about it more fully in the article on the effectiveness of acoustic treatment.

A measurement is not a diagnosis of the cause

A number says how things are - it does not always say why. A long reverberation time usually follows from the properties of the room’s surfaces, and there the link is fairly direct. But a high sound level recorded during a lesson may originate in the type of activity, the number of people in the room, how the work is organised, noisy furnishings, long reverberation amplifying every conversation, or noise from outside - most often in several of these at once.

That is why professional interpretation means combining several pieces of information, not reading one. Users’ observations are a good complement to technical data - on how to collect them and what not to conclude from them, we write in the article on how pupils perceive their school’s acoustics.

An example of an assessment combining both layers is the Polish study by Polewczyk and Jarosz (2020), in which measurements of reverberation time, sound level and STI taken before and after an intervention were set against questionnaires from teachers and pupils [14]. The value of this approach is that technical parameters and users’ experience explain one another: the measurement says what changed, the questionnaire whether the change matters in practice.

What a readable report should contain

There is no single mandatory template, but a good report should let the reader - including one without technical training - establish: what was measured and why, in which rooms and when, by what method and with what equipment, under what conditions, what the results were, what they were compared with, how they should be interpreted, what the limitations of the measurement are, what problems were identified and what actions are worth considering.

Two extremes are equally troublesome. A report consisting solely of tables of numbers without interpretation shifts all the analytical work onto a school with no way of doing it. A report reduced to “good / bad” labels without showing the basis of the judgement makes it impossible to check where that judgement came from or to repeat the survey later.

What cannot be concluded from a measurement alone

An acoustic measurement describes the environment - and only the environment. It does not diagnose ADHD or attention disorders. It does not detect hearing loss. It does not measure stress or identify a voice disorder in a teacher. It does not establish mental health problems. It does not predict any particular child’s educational outcomes.

Scientific research can show statistical associations between particular acoustic conditions and how groups of people function - and we write about those associations in the other articles in this series. That is, however, an entirely different category of knowledge from a meter reading in a specific room. Confusing them leads to conclusions no measurement justifies.

Summary

Let us return to the opening question: “this room is at 58 dB - is that a lot?” Having worked through the parameters, it is clear why the only honest answer is: first you have to know what exactly was measured. An averaged level or a peak. An empty room or a lesson in progress. Ventilation on or off. At the board or in the back row. Each of these answers changes the meaning of the same number.

A good acoustic survey does not produce a single number to describe a whole school. It lets you break the problem into parts, measure the right parameters and understand what in a given room really needs improving. That is less striking than one figure on the front page of a report - but it is the only approach a decision can rest on.

References

1. ISO, ISO 3382-2 Acoustics - Measurement of room acoustic parameters - Part 2: Reverberation time in ordinary rooms. Katalog ISO.
2. Bradley J.S., Evaluation of Acoustical Conditions for Speech Communication in Working Elementary School Classrooms, „The Journal of the Acoustical Society of America” 2008. DOI: 10.1121/1.2839283; PMID: 18397014.
3. Murgia S., Webster J., Cantor Cutiva L.C., Bottalico P., Systematic Review of Literature on Speech Intelligibility and Classroom Acoustics in Elementary Schools, „Language, Speech, and Hearing Services in Schools” 2023; 54(1): 322-335. DOI: 10.1044/2022_LSHSS-21-00181; PMID: 36260411.
4. IEC, IEC 60268-16 Sound system equipment - Part 16: Objective rating of speech intelligibility by speech transmission index, wyd. 2020. IEC Webstore.
5. Rabelo A.T.V., Santos J.N., Oliveira R.C., Magalhães M.C., Effect of Classroom Acoustics on the Speech Intelligibility of Students, „CoDAS” 2014; 26(5): 360-366. DOI: 10.1590/2317-1782/20142014026; PMID: 25388068.
6. ISO, ISO 16283-1 Acoustics - Field measurement of sound insulation in buildings and of building elements - Part 1: Airborne sound insulation, wyd. 2014. Katalog ISO.
7. ISO, ISO 717-1 Acoustics - Rating of sound insulation in buildings and of building elements - Part 1: Airborne sound insulation, wyd. 2020. Katalog ISO.
8. ISO, ISO 16283-2 Acoustics - Field measurement of sound insulation in buildings and of building elements - Part 2: Impact sound insulation, wyd. 2020. Katalog ISO.
9. PN-B-02151-2:2018-01 Akustyka budowlana. Ochrona przed hałasem w budynkach. Część 2: Wymagania dotyczące dopuszczalnego poziomu dźwięku w pomieszczeniach, Polski Komitet Normalizacyjny, 2018 (dokument odpłatny; status do sprawdzenia w katalogu PKN).
10. PN-B-02151-4:2015-06 Akustyka budowlana. Ochrona przed hałasem w budynkach. Część 4: Wymagania dotyczące warunków pogłosowych i zrozumiałości mowy w pomieszczeniach oraz wytyczne prowadzenia badań, Polski Komitet Normalizacyjny, 2015 (dokument odpłatny; status do sprawdzenia w katalogu PKN).
11. PN-B-02151-3:2015-10 Akustyka budowlana. Ochrona przed hałasem w budynkach. Część 3: Wymagania dotyczące izolacyjności akustycznej przegród w budynkach i elementów budowlanych, Polski Komitet Normalizacyjny, 2015 (dokument odpłatny; norma posiada poprawki i zmiany - aktualne wydanie należy sprawdzić w katalogu PKN).
12. Department for Education (Wielka Brytania), Acoustic design of schools: performance standards (Building Bulletin 93), luty 2015. Dokument PDF (GOV.UK).
13. IEC, IEC 61672-1 Electroacoustics - Sound level meters - Part 1: Specifications, wyd. 2013. IEC Webstore.
14. Polewczyk I., Jarosz M., Teachers' and Students' Assessment of the Influence of School Rooms Acoustic Treatment on Their Performance and Wellbeing, „Archives of Acoustics” 2020; 45(3): 401-417. DOI: 10.24425/aoa.2020.134057.

This article is for information and education only. It is not a set of instructions for taking measurements, and it does not replace a professional acoustic survey or technical advice.