How to Measure RT60: Interrupted Noise vs Impulse Response vs MLS

Reverberation time (T60 or RT60) is the most frequently measured parameter in room acoustics. It appears in every compliance framework — WELL v2, BB93, ANSI S12.60, DIN 18041 — and forms the basis for assessing whether a room is acoustically fit for its intended function. Yet the method used to measure RT60 significantly affects the result, and choosing the wrong method, the wrong number of positions, or the wrong evaluation range can produce data that is misleading or non-compliant with the applicable standard.

ISO 3382-2:2008 (Acoustics — Measurement of room acoustic parameters — Part 2: Reverberation time in ordinary rooms) defines three measurement methods. This guide describes each method, its advantages and limitations, the equipment required, and the situations where each should be used.

What RT60 Actually Measures

RT60 is defined as the time, in seconds, for the sound pressure level in a room to decay by 60 dB after a sound source is abruptly stopped. A room with RT60 = 0.5 seconds at 500 Hz takes half a second for a 500 Hz sound to decay from its steady-state level to one-millionth of its original energy (60 dB = factor of 1,000,000 in energy, or 1,000 in pressure).

In practice, achieving a clean 60 dB decay range in a measurement is difficult. Background noise in most rooms is only 30–50 dB below the steady-state sound level generated by the test source, which limits the usable decay range. For this reason, ISO 3382 defines three evaluation ranges:

• T60: Evaluated from -5 dB to -65 dB below the initial level (full 60 dB range). Requires extremely high source levels and very low background noise. Rarely achieved in ordinary rooms.
• T30: Evaluated from -5 dB to -35 dB below the initial level (30 dB range), then extrapolated to 60 dB by doubling the measured decay time. The standard method for most rooms.
• T20: Evaluated from -5 dB to -25 dB below the initial level (20 dB range), then extrapolated by tripling. Used when background noise is too high for T30, or as a cross-check.

Early Decay Time (EDT)

In addition to T60 (and its T20/T30 approximations), ISO 3382-1 defines Early Decay Time (EDT), evaluated from 0 dB to -10 dB below the initial level and multiplied by 6. EDT corresponds more closely to the subjective perception of reverberance because the ear is most sensitive to the first 10 dB of decay. In well-diffused rooms, EDT and RT60 are similar. In rooms with non-uniform absorption (e.g., a room with a highly absorptive ceiling and reflective walls), EDT may be significantly shorter than RT60, indicating that the early sound field decays faster than the late reverberant field.

EDT is measured from the impulse response and cannot be obtained from the interrupted noise method (Method 1).

Method 1: Interrupted Noise

Principle

The interrupted noise method is the oldest and conceptually simplest approach. A loudspeaker generates broadband noise (typically pink noise, which has equal energy per octave band) at a steady level in the room. The noise is abruptly switched off, and the decay of sound pressure level is recorded at one or more microphone positions. The decay curve is analyzed to determine the time for the level to drop by the required range (20 dB for T20, 30 dB for T30).

Procedure per ISO 3382-2

1. Place an omnidirectional loudspeaker at the source position. The source should generate a sound pressure level at least 35 dB above the background noise level in each octave band (for T20 evaluation) or 45 dB above (for T30).

1. Place a measurement microphone at the receiver position. The microphone must be omnidirectional and meet at least IEC 61672 Class 2 accuracy (Class 1 preferred).

1. Turn on the noise source and allow the sound field to reach steady state. ISO 3382-2 recommends a minimum excitation time equal to half the expected reverberation time (e.g., at least 0.5 seconds for a room with expected RT60 of 1.0 second, though longer is better).

1. Abruptly switch off the source (the switch-off time must be less than one-tenth of the expected reverberation time).

1. Record the decay of sound pressure level at the microphone.

1. Repeat the measurement multiple times at each position. ISO 3382-2 requires a minimum of 3 decays per source-receiver combination for the interrupted noise method. More repetitions reduce the variance of the result.

1. Average the decay curves (or the individual T values) to obtain the RT60 at each octave band.

Equipment Required

• Broadband noise source (loudspeaker + amplifier + noise generator or software)
• Class 1 or Class 2 sound level meter with data logging, or a measurement microphone connected to a recording system
• Octave-band or one-third-octave-band analysis capability (built into the sound level meter or performed in post-processing software)
• Timer/recording system to capture the decay

Advantages

• Conceptually simple and easy to perform without specialized software
• Widely understood by building surveyors and facilities managers
• Does not require complex signal processing — the decay curve is directly visible on the sound level meter
• Can be performed with relatively basic equipment

Limitations

• Low signal-to-noise ratio: The source must be very loud to achieve sufficient decay range above the background noise. In noisy environments (occupied offices, spaces near busy roads), T30 may not be achievable.
• Multiple repetitions needed: Each measurement is a single decay, subject to random variation. At least 3 repetitions are needed per ISO 3382-2, and 5 or more are recommended for reliable results. This makes the method time-consuming.
• Cannot measure EDT: The interrupted noise method does not produce an impulse response, so EDT and other parameters (C50, C80, D50, STI) cannot be derived.
• Source directivity issues: The loudspeaker must be reasonably omnidirectional to excite the room uniformly. Directional sources produce position-dependent results.

When to Use

The interrupted noise method is appropriate for straightforward RT60 measurements in unoccupied rooms with low background noise. It is the method most commonly used by building control inspectors and non-specialist surveyors for compliance checking.

Method 2: Impulse Response Method

Principle

The impulse response method captures the room's complete response to a short, sharp sound (an impulse). The impulse response contains all information about how the room modifies sound over time: the direct sound, early reflections, and the late reverberant decay. From the impulse response, RT60 (and T20, T30, EDT, C50, C80, D50, STI, and many other parameters) can be extracted through signal processing.

The impulse can be generated by several means:

• Balloon pop: A large latex balloon (typically 300mm diameter when inflated) is burst near the source position. This produces a broadband impulse with reasonable energy across 250–4000 Hz but limited energy at 125 Hz and below.
• Starter pistol / blank gun: Produces a very sharp, loud impulse with good energy across the full frequency range. Requires appropriate safety precautions and may not be permitted in all venues. Certain jurisdictions require firearms licensing even for blank-firing pistols.
• Wooden clapper boards: Two flat boards slapped together produce a sharp impulse. Energy is concentrated in mid and high frequencies; low-frequency content is limited.
• Spark source: An electrical spark gap produces a very short, repeatable impulse. Used in laboratory settings and scale model measurements.

Procedure per ISO 3382-2

1. Place the source (balloon, pistol, or other impulsive source) at the designated source position, at least 1.5 meters from any surface.

1. Place the measurement microphone at the receiver position, at least 1.5 meters from any surface and at least 2 meters from the source (to ensure the direct-to-reverberant energy ratio allows clear identification of the decay).

1. Ensure background noise is stable and at least 35 dB (T20) or 45 dB (T30) below the peak level of the impulse.

1. Trigger the impulse and record the complete response at the microphone. The recording must continue until the decay has reached the background noise floor.

1. Process the recorded impulse response to extract the decay curve. This is done using the Schroeder backward integration method (Schroeder, 1965), which converts the noisy raw decay into a smooth, monotonically decreasing energy decay curve (ETC — Energy Time Curve).

1. Fit a straight line to the appropriate portion of the decay curve (-5 to -25 dB for T20, -5 to -35 dB for T30) and extrapolate to 60 dB.

1. Repeat at multiple source and receiver positions per the requirements of the applicable standard.

The Schroeder Method

The Schroeder backward integration is the standard signal processing technique for extracting RT60 from an impulse response. Instead of looking at the raw amplitude decay (which is noisy and irregular), the method integrates the squared impulse response backward from the end of the recording:

E(t) = integral from t to infinity of h²(tau) d(tau)

Where h(t) is the impulse response. This produces a smooth, monotonically decreasing curve from which the decay rate can be reliably measured. The method was published by Manfred Schroeder in 1965 and is referenced in ISO 3382-2, Annex B.

Equipment Required

• Impulsive source (balloon, starter pistol, or clapper)
• Measurement microphone (omnidirectional, Class 1 or Class 2)
• Recording device (digital audio recorder, laptop with audio interface, or measurement-grade sound level meter with recording capability)
• Analysis software capable of Schroeder integration and octave-band filtering (DIRAC, REW, ARTA, or equivalent)
• Calibrator for microphone level calibration

Advantages

• Single event captures everything: One balloon pop gives you RT60, EDT, C50, C80, D50, and the raw data for STI calculation. The interrupted noise method only gives you RT60.
• Fast in the field: Each measurement takes seconds (one pop, one recording). Multiple positions can be measured in an hour.
• Provides EDT: The impulse response is the only way to measure Early Decay Time, which is required by ISO 3382-1 and which correlates better with subjective reverberance than RT60.
• Schroeder integration reduces noise: The backward integration method produces a smooth decay curve even from a single impulse measurement, reducing the need for multiple repetitions.

Limitations

• Source repeatability: Balloon pops and pistol shots are not perfectly repeatable. Balloon size, inflation pressure, and burst position affect the spectral content of the impulse. ISO 3382-2 notes this and recommends at least 2 measurements per source-receiver combination when using impulsive sources.
• Limited low-frequency energy: Balloons and clappers produce insufficient energy below 125 Hz for reliable low-frequency RT60 measurement. Starter pistols are better but still limited compared to loudspeaker-based methods.
• Safety and permissions: Starter pistols and even balloon pops can cause alarm in occupied buildings, hospitals, and public spaces. Advance notification and coordination are essential.
• Processing complexity: Extracting parameters from the impulse response requires specialized software and knowledge of signal processing. The method is not suitable for untrained personnel.

When to Use

The impulse response method is the preferred method for professional acoustic surveys where multiple parameters are needed (not just RT60) and where the acoustician has the equipment and expertise for signal processing. It is the standard method used by acoustic consultants worldwide.

Method 3: Maximum Length Sequence (MLS) and Swept Sine

Principle

MLS and swept sine methods use a loudspeaker to play a known test signal — either a pseudo-random binary sequence (MLS) or an exponentially swept sine wave (ESS, also called a logarithmic sweep or chirp) — and record the room's response through a microphone. The impulse response is then extracted mathematically by correlating the recorded signal with the known input signal (for MLS) or by deconvolution (for swept sine).

These methods achieve very high signal-to-noise ratios because the test signal contains much more energy than a single impulse. A 10-second swept sine distributes the same total energy as a very loud impulse but over a much longer time, making it far less disruptive and more repeatable.

MLS vs Swept Sine

MLS (Maximum Length Sequence): A binary pseudo-random noise signal of length 2^N - 1 samples (commonly N = 16, giving 65,535 samples or about 1.5 seconds at 44.1 kHz). The impulse response is extracted by circular cross-correlation with the MLS signal. This method was widely used in the 1990s and 2000s.

Swept Sine (ESS): An exponentially swept sine wave, typically spanning 20 Hz to 20,000 Hz over 5–30 seconds. The impulse response is extracted by convolving the recorded signal with the time-reversed version of the excitation signal (deconvolution). This method has largely replaced MLS in modern practice because it has superior immunity to time-variance (e.g., temperature changes during the measurement) and naturally separates harmonic distortion from the linear impulse response.

ISO 3382-2:2008 accepts both methods.

Procedure

1. Set up a loudspeaker at the source position and a microphone at the receiver position, as for the interrupted noise method.

1. Connect the loudspeaker and microphone to the measurement system (typically a laptop with a calibrated audio interface running measurement software such as DIRAC, REW, ARTA, EASERA, or WinMLS).

1. Calibrate the system: verify microphone sensitivity, set playback level to achieve adequate signal-to-noise ratio without clipping, and verify the audio interface clock.

1. Play the test signal (MLS or swept sine) through the loudspeaker and simultaneously record the response through the microphone.

1. Process the recording to extract the impulse response via correlation (MLS) or deconvolution (swept sine).

1. Analyze the impulse response using Schroeder backward integration to obtain RT60 (T20, T30), EDT, and other parameters at each octave band.

1. Repeat at multiple positions as required.

Equipment Required

• Omnidirectional loudspeaker (dodecahedron loudspeaker per ISO 3382-2 is ideal; for ordinary rooms, a quality studio monitor on a stand is acceptable)
• Measurement microphone (omnidirectional, calibrated)
• Audio interface (at least 2 channels: 1 output to loudspeaker, 1 input from microphone)
• Measurement software (DIRAC, REW, ARTA, EASERA, or equivalent)
• Amplifier for the loudspeaker
• Microphone preamplifier (often built into the audio interface)

Advantages

• Highest signal-to-noise ratio: The swept sine method routinely achieves 80–100 dB of decay range, enabling T30 and even true T60 measurement in most rooms. This is unmatched by the other two methods.
• Perfect repeatability: The test signal is generated digitally and is identical every time. There is no variation between measurements due to the source (unlike balloon pops).
• Lowest disruption: A swept sine sounds like a rising tone and is far less startling than a balloon pop or pistol shot. It can be used in occupied buildings with advance notice.
• Full parameter extraction: Like the impulse response method, MLS/swept sine provides the complete impulse response for extraction of EDT, clarity, STI, and all other ISO 3382-1 parameters.
• Low-frequency performance: A loudspeaker can generate controlled low-frequency energy that impulse sources cannot match. RT60 at 63 Hz and 125 Hz is reliably measurable with a swept sine and a subwoofer-capable loudspeaker.

Limitations

• Equipment cost and complexity: A measurement-grade dodecahedron loudspeaker, calibrated microphone, audio interface, and analysis software represent a significant investment (roughly $5,000–$20,000 for a professional kit). This is the domain of specialist acoustic consultants.
• Setup time: The loudspeaker, amplifier, audio interface, and software must be set up and calibrated at each site. Allow 15–30 minutes for setup before measurements begin.
• Sensitivity to time variance (MLS only): MLS requires the system to be time-invariant during the measurement. Temperature changes, air movement, or even a door slowly closing can corrupt the result. Swept sine is much more robust to time variance.
• Loudspeaker non-linearities (MLS only): MLS is sensitive to harmonic distortion in the loudspeaker, which appears as artifacts in the extracted impulse response. Swept sine naturally separates distortion components.

When to Use

MLS/swept sine is the gold standard for professional acoustic measurement. It should be used for:

• Commissioning measurements for certification (WELL, BREEAM, LEED)
• Performance space measurements (concert halls, theatres, recording studios)
• Any measurement where true T60 (not just T20 or T30) is needed
• Any measurement where low-frequency RT60 (63 Hz, 125 Hz) is critical
• Research and detailed acoustic analysis

Measurement Positions: How Many and Where

ISO 3382-2 specifies minimum requirements for source and receiver positions based on the required measurement precision:

Survey Method (lower precision)

• Minimum 1 source position
• Minimum 2 receiver positions per source
• At least 2 source-receiver combinations
• Receiver positions at least 2m from the source and 1m from any surface
• Result: ±20% uncertainty in spatial average RT60

Engineering Method (standard precision)

• Minimum 2 source positions
• Minimum 3 receiver positions per source
• At least 6 source-receiver combinations
• Receiver positions at least 0.7 sqrt(V/T) from the source (where V is volume and T is expected RT60), minimum 2m
• At least half the receiver positions more than 1.5m from any surface
• Result: ±10% uncertainty in spatial average RT60

Precision Method (highest precision)

• Minimum 2 source positions
• Minimum 3 receiver positions per source
• At least 12 source-receiver combinations (6 per source)
• Same distance requirements as engineering method
• Result: ±5% uncertainty in spatial average RT60

Source and Receiver Heights

• Source height: 1.5m above the floor (representing a standing speaker's mouth height)
• Receiver height: 1.2m above the floor (representing a seated listener's ear height)
• These heights are specified in ISO 3382-2 and should not be varied without justification

Common Measurement Errors

Error 1: Insufficient Signal-to-Noise Ratio

If the background noise is too close to the decay tail, the measured decay curve bends upward at its lower end (the decay appears to slow down as it approaches the noise floor). This produces T20 or T30 values that are too long. The solution is to increase the source level, reduce background noise (turn off HVAC during measurement if the standard permits), or use a swept sine method with longer averaging.

Error 2: Source or Receiver Too Close to Surfaces

Placing the microphone within 0.5m of a wall produces results influenced by the local sound field rather than the diffuse field average. The RT60 measured at this position may be significantly different from the spatial average. ISO 3382-2 specifies minimum distances from surfaces for this reason.

Error 3: Measuring in a Furnished vs Unfurnished Room

RT60 changes significantly when furniture is added. Upholstered chairs, curtains, carpets, and even stacks of books contribute absorption. Measuring before furniture is installed produces longer RT60 values than the room will exhibit in service. Measuring with furniture represents the operational condition but may not match the condition tested by the applicable standard. Always note the room condition in the measurement report.

Error 4: Ignoring Temperature and Humidity

The speed of sound, and therefore the frequency-dependent air absorption, varies with temperature and humidity. ISO 3382-2 requires that temperature and humidity be recorded during measurement. In large rooms (volume above 500 m³), temperature changes of 5 degrees Celsius or more between design calculations and measurements can shift high-frequency RT60 values by 10–15%.

Error 5: Reporting T20 as T60

T20 and T30 are extrapolations of the decay rate measured over a limited range. They are valid approximations of T60 only if the decay is exponential (a straight line on a dB vs. time plot). In rooms with coupled volumes (e.g., a room connected to a corridor through an open door), the decay curve may show a double slope — a fast initial decay followed by a slower late decay. In such rooms, T20 (based on the early part of the decay) will be shorter than T30 (which includes the slower late decay), and neither accurately represents the true T60. Always report which evaluation range was used, and note any discrepancy between T20 and T30.

Which Method Should You Choose?

For building control compliance checks (verifying that a completed room meets its RT60 target): the interrupted noise method with a portable loudspeaker and sound level meter is adequate and accessible.

For acoustic consultant surveys (commissioning, design verification, troubleshooting): the impulse response method with balloon pops or a starter pistol provides fast, comprehensive data at moderate cost.

For precision measurements (performance spaces, research, certification requiring full parameter sets): the swept sine method with a dodecahedron loudspeaker and calibrated measurement chain is the gold standard.

In all cases, follow ISO 3382-2 requirements for source-receiver positions, evaluation ranges, and reporting. The measurement is only as good as the method allows, and the method is only as good as the execution permits.