How Is an Acoustic Measurement Actually Performed? A Step-by-Step Guide

An acoustic measurement is a structured, on-site process that uses a calibrated Class 1 sound level meter, a loudspeaker source, and (for impact noise) a tapping machine to capture reverberation time (RT60), background noise level, and sound insulation between rooms, all following international standards such as ISO 3382, ISO 16283, and ISO 717. Measurements are taken at multiple positions in each room, broken down by octave band, and compared against the target for that room’s use — the result is data, not an opinion. This guide walks through exactly what happens during a real acoustic survey, room by room and instrument by instrument.

If you’ve never had an acoustic measurement done before, it’s reasonable to picture someone waving a decibel app on their phone for five minutes and calling it a report. The reality is considerably more procedural, and the difference matters: a measurement done to the wrong methodology, at the wrong positions, or without the right calibration produces numbers that look precise but don’t hold up to scrutiny — and can’t be used to design a treatment plan with any confidence. This article covers the equipment, the step-by-step process for each measurement type, why the details (positions, frequency bands, lab vs. field) matter, and what we ask clients to prepare before we arrive. As the reference guide for our what a professional acoustic survey includes and reading an acoustic report articles, this is the piece to bookmark if you want to understand the methodology behind every number that ends up in your report.

The Equipment: What’s Actually in the Case

A proper acoustic measurement rig is a small, specific toolkit, not a single app or gadget. On a typical survey we bring:

• A Class 1 sound level meter. “Class 1” (per IEC 61672) is the precision-grade accuracy tier required for compliance and design work — Class 2 meters (and phone apps) are acceptable for rough estimates but not for a report a client will act on financially. The meter measures sound pressure level (SPL) in dB and dBA, logs data over time, and — paired with the right software — resolves it into octave or third-octave frequency bands.
• An acoustic calibrator. A small handheld device that fits over the microphone and outputs a known, fixed reference tone (typically 94 or 114 dB at 1 kHz). Every meter is calibrated against this reference immediately before and after a measurement session; if the reading drifts outside tolerance, the data from that session is considered invalid. This single step is one of the most commonly skipped corners in low-cost “measurements,” and it’s non-negotiable in ours.
• An omnidirectional loudspeaker source. A specialized speaker (dodecahedron-shaped, in the classic design) that radiates sound roughly equally in all directions, used to generate the steady noise or sine-sweep signal needed for reverberation time and sound insulation testing. A domestic Bluetooth speaker won’t do — its directional radiation pattern biases the results depending on which way it happens to be pointed.
• A tapping machine. A mechanical device with five weighted hammers that drop in sequence onto a floor at a fixed rate, generating a standardized impact excitation. This is the source used exclusively for impact/footfall noise testing (see below) — there is no substitute for it; footsteps or a dropped object don’t produce a repeatable, standardized signal.
• Measurement microphones and a positioning system. Calibrated microphones on stands or tripods, moved through a defined pattern of positions per room rather than left in one spot — more on why that matters shortly.
• Building plans / a laser distance meter, used to record room volumes and surface areas, which feed directly into the Sabine calculation and the report’s predicted-treatment section.

None of this equipment is exotic or expensive by industrial standards, but each piece exists because a shortcut at that step — an uncalibrated meter, a directional speaker, a stopwatch instead of a tapping machine — introduces an error that’s invisible in the final number but very real in the room.

Before the Meter Comes Out: Planning the Survey

A measurement session is planned before anyone sets foot on site. We review the building’s floor plan (or take one on arrival), decide which rooms and adjacencies are actually in scope, and — critically — choose measurement positions in advance rather than picking a convenient spot once inside. For reverberation time and background noise, this typically means several source and several receiver positions spread across the room (away from walls and corners, at typical head height, avoiding direct line-of-sight to the source at very short distances), with results averaged across positions. A single-point measurement can be badly skewed by a local reflection, a nearby vent, or standing-wave nulls that are entirely unrepresentative of how the room sounds overall — this is exactly the kind of detail a full professional acoustic survey is built around, beyond just the raw numbers.

What we ask the client to provide:

• Access to every space in scope, including the adjacent room for insulation testing (not just the room of complaint) and, ideally, the space above/below for impact testing.
• Quiet conditions during measurement — HVAC, music, and normal occupant activity should be paused where practical, since background noise contaminates both the LAeq reading itself and, at high enough levels, the tail of a reverberation decay curve. We schedule around this rather than fighting through it.
• Basic building information — floor plans, construction details if known (wall type, floor buildup), and which rooms are the actual pain points versus which are being measured as a baseline or reference.
• A point of contact on site who can grant access to locked rooms, plant rooms, or roof areas if flanking paths need to be traced.

With planning done, the actual measurement sequence usually runs in the same order on every job, because each test builds on information from the last.

Step 1: Background Noise (LAeq)

We start with background/ambient noise because it’s the baseline everything else gets measured against — and because it tells us immediately whether conditions are good enough to proceed with the more sensitive tests. Using the Class 1 meter, we log the A-weighted equivalent continuous sound level (LAeq) over a representative period at each measurement position, with HVAC and other services running in their normal operating state (unlike the “quiet” condition used for RT60 and insulation tests, background noise is measured under normal occupied/operating conditions, since that’s the noise floor people actually experience). This single reading answers questions ranging from “is this office too noisy to concentrate in” to “does this classroom meet the ANSI/ASA S12.60 background-noise limit” (see our classroom acoustics standards article for that specific threshold) — and it’s a required input for interpreting every other measurement, since a room can’t show a clean, extrapolatable decay curve if the ambient floor is already high.

Step 2: Reverberation Time (RT60)

With background noise established and the room otherwise quiet, we measure reverberation time per ISO 3382 methodology, using one of two standard techniques:

1. Interrupted noise method. The omnidirectional loudspeaker fills the room with steady broadband noise until levels stabilize, then the source is cut abruptly. The measurement microphone captures the decay, and the resulting curve is analyzed per octave band.
2. Impulse response method. A sine sweep (or, less commonly today, an impulsive source) is played through the loudspeaker and captured by the microphone, producing a room impulse response that software decomposes into the same decay information in a single, faster capture. This is now the industry-default technique on most of our surveys.

Because a real room rarely produces a clean, uninterrupted 60 dB drop before background noise obscures the tail, the actual decay measured is usually the first 20 or 30 dB of drop-off — called T20 or T30 — and mathematically extrapolated to what a full 60 dB decay would take. This is standard, accepted practice, not an approximation to be suspicious of; it’s built into the ISO 3382 methodology itself. What does matter is that the extrapolation is done per octave band (typically 125 Hz through 4 kHz) rather than as a single blended number, because a room’s absorption is rarely flat across frequency — we cover exactly why that single-number-vs-frequency distinction matters below, and go deeper on the metric itself in what is reverberation time (RT60).

Step 3: Airborne Sound Insulation Between Rooms

If the brief includes a sound insulation question — “how much does the neighboring office hear through this wall” — we move to a two-room measurement. The omnidirectional loudspeaker generates steady broadband noise in the source room, and the meter logs the resulting sound level both there and in the receiving room on the other side of the partition, at multiple positions in each. The difference between the two, corrected for the receiving room’s own absorption (so a more absorptive receiving room doesn’t unfairly flatter the wall), gives the field sound level difference per frequency band.

This is rated one of two ways depending on the client’s market and which standard the project follows: DnT,w (per ISO 16283 for the field measurement procedure, rated to a single number per ISO 717) in most international and European contexts, or a field STC (ASTM-based) rating in US-oriented projects. We explain how these two rating families relate — and where they diverge — in STC vs. Rw vs. NIC, and cover the broader airborne-vs-impact distinction in airborne vs. impact noise. One point worth flagging up front: field-measured values are almost always a few points worse than a wall assembly’s laboratory rating (see the lab-vs-field section below) — that gap is normal, not a sign of a defective wall, and it’s exactly why field testing exists as its own discipline.

Step 4: Impact Sound Insulation (The Tapping Machine)

Airborne testing tells you nothing about footfall — a floor that blocks voices well can still transmit every heel-strike from the room above, which is why impact testing uses an entirely different source. The tapping machine is placed on the floor of the room above and run through its standard cycle (five weighted hammers striking the floor in sequence at a fixed rate), while the meter measures the resulting sound level in the room below. This produces the field impact insulation rating — L’nT,w internationally (again per ISO 16283/717) or field IIC in US practice.

Unlike airborne insulation, where lower transmitted level is better, impact ratings run in the opposite numeric direction for the ISO metric: lower L’nT,w is better (it directly represents the transmitted impact sound level), while IIC follows STC’s convention where higher is better. This is a common point of confusion when comparing reports across markets, and one we flag explicitly in every report we deliver — it’s exactly the kind of detail covered in reading an acoustic report.

Step 5: Open-Plan Office Parameters and Speech Intelligibility

Open-plan offices get their own dedicated measurement protocol under ISO 3382-3, because the question in an open floor isn’t “how loud is it here” but “how far does one person’s speech travel before it stops distracting their neighbor.” Using the loudspeaker as a simulated talker at a fixed position and a series of receiver positions moving away from it at set distances, we capture:

• D2,S — the spatial decay rate of speech (dB per distance doubling); higher means speech dies off faster with distance, which is what you want.
• Lp,A,S,4m — the A-weighted speech level at a 4 m reference distance.
• rD (distraction distance) and rP (privacy distance) — the distances at which the Speech Transmission Index (STI) drops below 0.50 and 0.20 respectively, derived from the same dataset.
• Lp,A,B — the average background noise level across the floor.

We go through what “good” looks like for each of these parameters, and how they translate into layout and treatment decisions, in open-plan office acoustics and ISO 3382-3. Related but distinct, Speech Transmission Index (STI) is also measured directly inside enclosed meeting and conference rooms — there the goal is the opposite of open-plan (you want high STI, meaning everyone is clearly understood) — using the same sine-sweep/impulse-response equipment but analyzed against the IEC 60268-16 intelligibility scale rather than a distance-decay curve; see speech intelligibility (STI) in meeting rooms for that side of the measurement.

Why Octave-Band Data Matters More Than the Headline Number

Every measurement described above can be, and usually is, collapsed into a single number for a summary table — one RT60 figure, one STC/DnT,w rating, one STI score. Those single numbers are useful for quick comparison and for meeting a code minimum, but they hide information. RT60, NRC, STC, and STI are all averages or weighted composites across a frequency range, and a room or wall can look acceptable on the headline number while having a serious problem at one end of the spectrum — most commonly at low frequency, where absorption is hardest to achieve and mass-based insulation is weakest.

A ceiling absorption treatment sized purely off a single-number RT60 target might leave a room with excellent mid-frequency decay and a persistent 125 Hz boom that the headline number never flagged. A partition wall can hit its rated STC while still transmitting bass-heavy music or HVAC rumble that a broadband rating doesn’t weight heavily. This is exactly why a proper report presents octave-band data (typically 125 Hz–4 kHz) alongside the single-number rating rather than instead of it — the octave-band curve is what actually drives the treatment design, while the single number is what goes on a compliance checklist. We designed our reporting format around this principle; see reading an acoustic report for how to interpret both layers together, and NRC vs. the sound absorption coefficient for the same single-number-vs-frequency-curve issue applied to absorptive materials specifically.

Lab vs. Field Measurements — Why the Numbers Differ

A recurring source of confusion for clients comparing a product datasheet against a report is the gap between laboratory and field ratings. A wall or floor assembly’s STC, IIC, or a material’s NRC is measured under controlled laboratory conditions — a purpose-built test chamber, isolated from flanking paths, with idealized construction and no real-world workmanship variables. Field measurements (ASTC/AIIC, or the ISO 16283 field DnT,w/L’nT,w methodology described above) are taken in a completed, occupied building, where sound doesn’t just travel through the tested partition but also flanks around it — through a shared ceiling plenum, ductwork, structural connections, gaps under doors, or back-to-back electrical boxes.

The practical result: field-measured insulation values typically read several points worse than the laboratory rating of the same nominal assembly. This isn’t a measurement error or a sign the wall was built incorrectly — it’s the expected, standard gap between an idealized lab test and a real building, and every credible report accounts for it rather than treating a lab datasheet number as a guarantee of field performance. It’s also the reason our surveys always measure the completed condition on site rather than relying on a manufacturer’s lab figure alone — see our sound insulation solutions service page for how this feeds into actual wall, floor, and door specification, and how weak links like an untreated door undermine an otherwise well-rated wall.

What Gets Measured, With Which Instrument, To Which Standard

From Raw Numbers to a Usable Report

None of the readings above are the end product of a survey — they’re the raw material for the report. A proper acoustic report doesn’t stop at “RT60 measured 1.1 s” or “field DnT,w measured 42 dB”; it states the applicable target or code threshold, the gap between measured and target, the diagnosed cause (surface finishes, a weak partition, a flanking path), and a remediation design with a predicted post-treatment value calculated from the same measured room data — absorption specified in added m² sabins, insulation improvement modeled from the identified weak link. We cover how to read and act on that full report structure in reading an acoustic report, and how a properly measured, room-specific design differs from a generic panel package in measured solution vs. catalog panels — a distinction that only exists because the measurement was done properly in the first place.

Measurement also underpins compliance work beyond a single room’s comfort: green-building certifications like LEED, WELL, and BREEAM require documented acoustic testing to award their acoustic credits (see acoustics in green buildings), and regulatory sound insulation requirements for new-build and converted residential and hospitality projects are verified the same way (see sound insulation building regulations). In every one of these contexts, the measurement methodology described above — calibrated equipment, standardized sources, multiple positions, octave-band analysis — is what makes the resulting number defensible, whether it’s going in front of a certifying body, a building inspector, or simply a facilities director deciding whether to approve a treatment budget.

Request a Measured Acoustic Survey

If you’re weighing a treatment decision — or need documented numbers for a compliance submission, a tenant dispute, or a renovation brief — the fastest way to get a reliable answer is a proper on-site measurement rather than an estimate based on how a room “feels.” Our team brings calibrated Class 1 equipment, follows the ISO 3382/16283/717 methodology described in this guide, and delivers octave-band data alongside the headline numbers so the resulting treatment plan is sized to your room’s actual gap, not a generic assumption. See what a professional acoustic survey includes for what a visit involves end to end, browse our projects for examples of measurement-led work, or explore room acoustics and sound insulation solutions for the treatment side of the process. Request a site survey or acoustic measurement and we’ll schedule a visit and walk you through exactly what we’ll measure and why.

FAQ

What equipment is needed for a proper acoustic measurement?

At minimum: a Class 1 sound level meter (per IEC 61672), an acoustic calibrator to verify it before and after use, an omnidirectional loudspeaker source for reverberation and insulation testing, and a standard tapping machine for impact/footfall testing. Class 2 meters and phone-based decibel apps can give a rough estimate but lack the accuracy and traceable calibration needed for a report used to size treatment or demonstrate compliance.

How long does an on-site acoustic measurement take?

It depends heavily on scope — a single room’s RT60 and background noise can be captured in under an hour, while a multi-room insulation survey (source room, receiving room, and impact testing between floors) with several rooms in scope typically takes half a day to a full day. Planning measurement positions and required access in advance, as described above, keeps the on-site time efficient.

Why do measured (field) values differ from a product’s lab rating?

Laboratory ratings (STC, IIC, NRC) are measured in a purpose-built test chamber with no flanking paths and idealized construction. Field measurements happen in a real, completed building, where sound also travels around the tested element — through ceiling plenums, ductwork, gaps, or shared structural connections. As a result, field-measured insulation values (ASTC, AIIC, DnT,w, L’nT,w) typically read several points worse than the equivalent laboratory number, which is expected and standard, not evidence of a defect.

Do I need to empty or quiet the room before a measurement?

For reverberation time and sound insulation testing, yes — HVAC, music, and normal activity should be paused where practical, since background noise contaminates the decay curve and the insulation readings. Background noise (LAeq) itself, by contrast, is measured under normal operating conditions on purpose, since that’s the noise floor people actually experience day to day.

What’s the difference between RT20/T30 and RT60?

RT60 is defined as the time for sound to decay by a full 60 dB, but real rooms rarely produce a clean 60 dB drop before background noise obscures the tail. In practice, the first 20 dB (T20) or 30 dB (T30) of decay is measured and mathematically extrapolated to what a full 60 dB decay would take — this is standard ISO 3382 practice, not an approximation to distrust.

Can I measure acoustics myself with a phone app?

A phone app can give a rough, order-of-magnitude sense of loudness, but it isn’t calibrated, its microphone isn’t measurement-grade, and it can’t run the standardized source signals (steady broadband noise, sine sweeps, or tapping-machine impacts) that ISO 3382/16283 methodology requires. It’s fine for a gut check on whether a room feels loud; it isn’t sufficient for sizing a treatment budget, satisfying a code requirement, or resolving a tenant dispute, all of which need a Class 1 instrument and proper methodology.