Measurement
How a room is measured
What a microphone and a sweep can tell you about a room, which graphs matter, and why a room isn’t finished until it has been measured.
7 min read5 sourcesChapter 20 of 24
Schematic
A sine sweep is played through the loudspeaker and recorded at the seat. Deconvolving the recording against the sweep gives the room’s impulse response — one click, and everything the room did to it. Every other graph is computed from that.
Animated illustration of the principle. Not measured data.
Opinion is cheap in acoustics; everyone has one and the room does not care. A measurement replaces opinion with evidence. It takes a calibrated microphone, a test signal and about an hour, and it is the only way to know whether a room meets its target, whether the treatment did what it was meant to, and what to do next.
The sweep and the impulse response
Modern measurement plays a logarithmic sine sweep — a tone rising smoothly from below 20 Hz to above 20 kHz — through the loudspeaker under test, and records it at the listening position. Mathematically comparing what was played with what was recorded yields the impulse response: the room’s complete answer to a single instantaneous click. It contains the direct sound, every reflection, the reverberant tail and the ringing of every mode.
From that one recording, software derives everything else. The frequency response, the decay in each band, the arrival time of each reflection, how long each bass note rings. One measurement, many views.
The graphs that matter
Frequency response shows level against frequency: the peaks and dips at the seat, including the modal peaks in the bass. It is the graph everyone looks at first, and the least complete on its own.
The waterfall or spectrogram adds time: it shows not just where the peaks are but how long they ring. A peak that decays quickly is a small problem; one that hangs on for 400 ms is the boom you hear. This is the graph that shows whether a bass trap worked.
The energy–time curve shows arrivals against time: the direct sound, then each reflection as a spike. It identifies which surface a problem reflection came from by its delay, and confirms that treatment removed it.
RT60 per band, from the decay slope in each octave, shows whether the room’s reverberation is balanced across the range or tilted.
One seat is not a room. A theatre is measured at every primary seat, because the bass at the centre seat and the bass at the end of the row can be a different note.
Before and after
The value of a measurement doubles when there are two. A baseline before any work, and a verification after, made with the same equipment at the same positions, is the only defensible evidence that a room was improved rather than decorated. It is also how a design is checked against its target — and adjusted, because no prediction survives a real room untouched.
This is why a case study without measurement files is a description, not a result.
What it needs
A calibrated measurement microphone with an omnidirectional capsule, an audio interface, software such as REW or a dedicated analyser, and a way to position the microphone repeatably at ear height in each seat. A calibrator for the microphone if absolute levels matter. And a quiet room — measurements are made with the air conditioning off where possible, because the noise floor sets how far down the decay can be followed.
For integrators and architectsThe numbers, the standards, and what to specify
- Standard
- ISO 3382-2 for reverberation time in ordinary rooms: source and receiver positions, T20 / T30 evaluation, octave or third-octave bands, and averaging requirements.
- Sweep length
- Longer sweeps improve signal-to-noise at low frequencies; 256k samples or more at 48 kHz is typical for bass work. Repeat and average where the noise floor is high.
- Microphone position
- Ear height at each primary seat, pointed upwards for an omni capsule, clear of the seat back. Log positions so the after-measurement can repeat them.
- Windowing
- Frequency responses are gated to separate direct from room sound; state the window. Unwindowed responses in small rooms are dominated by modal ripple and are hard to compare.
- Deliverables
- Raw measurement files (e.g. .mdat), exported per-band RT60, frequency response at each seat, waterfall for the bass range, and an ETC for each front loudspeaker. Files, not screenshots.
What to write into the brief
- A baseline measurement before design, a verification after, both to ISO 3382-2 at named seats, with raw files delivered.
Sources
- 01ISO 3382-2:2008 — Measurement of room acoustic parameters, Part 2: Reverberation time in ordinary roomsInternational standard for measuring RT (T20 / T30).
- 02ISO 3382-1:2009 — Measurement of room acoustic parameters, Part 1: Performance spacesDefines EDT, clarity (C50, C80), definition and strength for halls.
- 03ISO 16283-1:2014 — Field measurement of airborne sound insulationField results include flanking paths.
- 04ASTM E336-25a — Measurement of airborne sound attenuation between rooms in buildings (field)
- 05Kuttruff, Room Acoustics, 6th ed. (CRC Press, 2016)Standard textbook on room acoustics theory.
External links open in a new tab. Figures quoted in this chapter are taken from these sources; a Science of Sound measurement is only ever presented as one, with its file, date and instrument.
Keep reading