T60 Reverberation time is an acoustic parameter that quantifies how long it takes for sound to decay by 60 decibels (dB) in an enclosed space. Sometimes the term "RT60" is used in lieu of "T60", but they refer to the same thing. In this article the term T60 will be used.
The T60 reverberation time is frequency-dependent and is normally plotted against 1/3 octave center frequencies as shown in Figure 1.
Figure 1: T60 reverberation decay times plotted versus 1/3 octave band center frequencies.
The T60 time is calculated by plotting the decay of individual 1/3 octave bands over time on a decibel scale. The time for a drop of 60 decibels is determined as shown in Figure 2:
Figure 2: T60 decay time determined by the interrupted noise source method for a specific 1/3 octave band.
Reverberation times are used in many different applications, including:
Room Acoustics: Rating the perceived acoustic quality of environments like concert halls, theaters, recording studios, office spaces, test-cells and vehicle cabins. Longer T60 times lead to more reverberant, "live" spaces, while shorter times result in "dry" or "dead" acoustics.
Sound Absorption of Materials: Rating the ability of a material to absorb sound by comparing the decay in a reverberant room with and without the material present.
Audio Systems: Optimizing audio system design and speech intelligibility.
This article has the following sections:
1. T60 Measurement Methods 1.1 Interrupted Noise Method (ISO 3382-2, ASTM C423) 1.2 Integrated Impulse Response (ISO 3382-1) 2. Key Settings 2.1 Conversion of Time Data to Octave Levels 2.1.1 Averaging Frame 2.1.2 Time Increment 2.2 Curve Fit Octave Level Decay 2.2.1 Decay Start and Distance to Noise Floor 2.2.2 Decay Curve Extrapolation (T20, T30) 3. Sound Absorption 4. Reverberation in Simcenter Testlab
1. T60 Measurement Methods
The two primary standardized methods for measuring T60 decay times are:
1.1 Interrupted Noise Method (ISO 3382-2, ASTM C423)
T60 decay times are determined using the interrupted noise method with the following steps:
A broadband noise signal is played through an omnidirectional loudspeaker to build up a steady-state reverberant sound field in the room
The source is abruptly switched off and the sound decay is recorded with measurement microphones
1/3 octave levels versus time are extracted from the microphone recordings
Each 1/3 octave decay curve is analyzed to extract T60 times
The decay time is governed by the volume and sound absorptive properties of the measurement space. Interrupted noise method is commonly used for precision T60 testing in laboratories.
1.2 Integrated Impulse Response (ISO 3382-1)
An impulsive sound source (e.g. balloon pop, starter pistol) generates a short, high-level broadband excitation
The impulse response of the enclosed space is captured by the microphones
Schroeder integration is applied to the squared impulse response to derive a smooth decay curve for T60 calculation
Both methods require averaging multiple measurements at different locations to characterize the space statistically.
This section contains a detailed explanation of the key settings for reverberation time analysis and how they influence the results. These settings are shown in the two Simcenter Testlab software menus below (Figure 3).
Figure 3: Simcenter Testlab software menus with key settings used to calculated T60 decay times.
There are two main steps to calculate T60 reverberation times from the microphone recording:
Conversion of Time Data to Octaves: The first step in calculating T60 decay times from a interrupted source recording is to convert the microphone time data to level versus time for each 1/3 octave band. The “Averaging frame” and “Time Increment” are two key settings used in this conversion.
Curve Fit Octave Level Decay: In the next step, a line is fit to the decay of each 1/3 octave band level versus time. This curve fit is used to determine the time it takes for the octave band level to decay 60 decibels. The “Decay Start”, “Distance to Noise Floor”, and “Decay Curve Extrapolation” parameters are used in the curve fit.
The settings are explained further in the next sections. More about calculating octaves in the knowledge article: Octaves in Human Hearing.
2.1 Conversion of Time Data to Octave Levels
In the first step for calculating T60 times from an interrupted source test, the microphone time data (green in Figure 4 below) is converted to 1/3 octave levels versus time curves (blue in Figure 4 below).
Figure 4: The first step in calculating T60 reverberation times, octave band level curves (blue) are calculated from a raw microphone recording (green).
The “Averaging Frame” and “Time Increment” are used for the conversion of the microphone time recording to octave levels versus time.
2.1.1 Averaging Frame
The averaging frame (orange arrow in Figure 5 below) controls the smoothing of the decay curves prior to the T60 regression. The averaging frame is specified in seconds.
Figure 5: The “Averaging Frame” (orange arrows) is the time window used on the raw microphone data (red trace top) at each increment (purple dots) to create the octave level versus time trace (green).In this case, the overlap and averaging frame are identical.
The microphone time data contained in each “Averaging frame” is condensed to a single point (purple dot) on the octave decay curve. Either “Linear” or “Exponential” averaging can be applied over a specified time window to reduce local fluctuations.
Shorter averaging frames retain more detail of the original signal but makes the regression more susceptible to noise (Figure 6).
Figure 6: Comparison of a 1/3 octave band decay with a finer averaging frame (0.015725 seconds) versus a longer averaging frame (0.0625 seconds). The longer averaging frame (green) reduces the noise in the curve relative to shorter averaging frame (blue).
On the other hand, making the averaging frame too long will distort the slope of the decay curve and needs to be avoided. For example, using a averaging frame that has a longer time than the actual decay time should be avoided. The optimal averaging frame depends on the decay duration and noise level. ISO 3382 suggests setting the averaging frame to 10-20% of the expected T60 time.
2.1.2 Time Increment
The time increment setting determines the time spacing of the data points (purple points) along the decay curve as shown in Figure 5.
The “time increment” is specified in seconds. In the Simcenter Testlab software, this setting is not part of the Reverberation settings menu. Instead it is part of the "Tracking Setup" menu in Signature acquisition or in the "Tracking and Triggering" menu in Signature Throughput processing.
The ISO 3382 standard recommends at least 5-10 points along the actual decay curve to have confidence in the result.
For acoustic material testing in a reverberant room, setting both the increment and averaging frame to 0.0625 seconds is a good starting point. A finer increment and averaging frame may be needed for highly absorbent spaces with fast decays (e.g. anechoic chambers, vehicle interiors).
2.2 Curve Fit Octave Level Decay
After 1/3 octave levels versus time have been calculated, the next step is to fit a curve (magenta line in Figure 8) to the decay of the octave band (blue line in Figure 8).
Figure 8: To determine the T60 decay time a curve is fit (magenta) to the octave band decay curves (blue) which are calculated from a raw microphone recording (green).
The settings “Decay Start”, “Distance to Noise Floor”, and “Decay Curve Extrapolation” are used to determine the section of data used to fit a curve to the decay in the octave band level.
2.2.1 Decay Start and Distance to Noise Floor
Two settings determine the time range over which the decay is calculated via the curve fit (Figure 9):
“Decay Start” defines the start point of the T60 decay range by excluding the initial 5 dB (default value) of the decay after the source cutoff.
“Distance to Noise Floor” setting specifies the end point of the T60 decay range relative to the background noise level (default value minimum of at least 10 dB).
Figure 9: “Decay Start” and “Distance to Noise Floor” settings for T60 reverberation time calculations.
The “Decay Start” determines the start point of the T60 decay range. This setting helps avoid the initial region of the decay curve immediately after the source shuts off. This region often contains a rapid, transient drop due to the direct sound and early reflections, which does not follow the decay.
The “Distance to Noise Floor” setting in Simcenter Testlab specifies the end point of the T60 decay range relative to the background noise level. It is typically set 5-10 dB above the noise floor to avoid corruption of the decay tail:
If set too low, the noise will artificially steepen the decay slope, resulting in underestimated T60 values.
If set too high, the usable decay range will be truncated, leading to poor regression fits and unstable T60 estimates, particularly at low frequencies.
2.2.2 Decay Curve Extrapolation (T20, T30)
Simcenter Testlab allows calculating T60 decay times even when the full 60 dB decay is not present in the data. The T60 time can be estimated from a 20 dB decay time (T20) or 30 dB decay time (T30) and linearly extrapolated to a full 60 dB decay time (Figure 10).
Figure 10: Left – Data contains a full 60 dB of decay in octave curve. This is the ideal case, but depending on the test environment, it may not be possible to achieve 60dB decay. Right – Data only contains 30 dB of decay in octave curve, which is often the case in practice. Linear extrapolation from 30 dB of decay can be made to T60 decay time.
In the standards and Simcenter Testlab software, extrapolation to the 60 dB decay times can be done from either a 20 dB (T20) or 30 dB (T30) decay:
T30 strikes a good balance between decay range and noise immunity and is the preferred method in most standards. T60 = 2*T30
The T20 method is generally not recommended unless the noise floor severely limits the decay dynamic range. Using a shorter decay range makes the measurement more susceptible to background noise and decay curve anomalies. T60 = 3*T20
In all cases, it is important to inspect the decay curves visually in Simcenter Testlab to verify their quality and adjust the settings if needed. Conducting measurements at multiple positions and averaging the results spatially will further improve the T60 time accuracy and repeatability.
3. Sound Absorption
One application where T60 measurements are used is determining the sound absorption coefficients of materials. Absorption coefficients indicate how effective a material is at absorbing or reducing sound as a function of frequency.
A typical absorption plot is shown in Figure 11 below.
Figure 11: Absorption coefficient (ranges from 0 to 1) versus frequency plot for an acoustic material. The higher the absorption number, the more sound that is absorbed (zero is no sound absorbed, one is all sound absorbed).Typically the coefficients are determined at the center frequencies of the 1/3 octave bands.
Absorption coefficients range from a value of 0 to 1:
Absorption coefficient of zero indicates no sound is absorbed.
Absorption coefficient of one indicates all sound is absorbed.
The sound absorption properties of a material can be measured in a reverberation chamber. An example measurement setup is shown in Figure 12 below.
Figure 12: The sound absorption of an acoustic material can be measured in a reverberation chamber using an acoustic source and microphones.
T60 decay times are measured in a reverberation chamber to calculate the absorption of an acoustic material using the following procedure:
Measuring T60 of a large, reverberant test chamber with hard, reflective surfaces
Introducing a sample of the test material into the chamber and re-measuring T60
Calculating the absorption coefficient from the change in T60 using Sabine's equation (Equation 1):
Equation 1: Sabine’s equation for determining absorption coefficients from T60 times.
In Equation 1:
α is the absorption coefficient at a given 1/3 octave frequency
0.161 is a constant with units of seconds over meters to keep α unitless (constant is 0.049 if using feet instead of meters)
V is room volume (meters cubed)
S is the area of the sample (meters squared)
T1 is the empty room T60 decay time (seconds)
T2 is the T60 with the material sample present (seconds)
This formula is repeated across the 1/3 octave frequency bands to get the absorption spectrum of the material.
Two standards that detail measuring absorption are ISO 354 and ASTM C423. Simcenter Testlab also has a dedicated software module for calculating absorption in rooms called “Sound Absorption Testing using Room methods” that is contained in the “Testlab Acoustic” folder.
4. Reverberation in Simcenter Testlab (Interrupted Mode)
Simcenter Testlab can be used to calculate T60 decay times both in online acquisition and in offline processing.
4.1 Online T60 Acquisition
Connect the Simcenter SCADAS hardware to a PC running Simcenter Testlab software. Using a reverberation chamber designed to have a diffuse sound field, hook up the SCADAS to an acoustic source and microphones as shown in Figure 13.
Figure 13: Reverberation chamber setup to measure T60 decay times using SCADAS hardware, Simcenter Testlab software, microphones, and acoustic source.
Start Simcenter Testlab Signature Acquisition. After the software starts, click on the white icon in the upper left corner (Figure 14):
Figure 14: After Simcenter Testlab Signature starts, click on the white icon in the upper left corner to start a new project.
After pressing the white icon, the software will talk with the SCADAS hardware and set the channel list appropriately based on the hardware present. By default, the new project will be called “Project1”. Choose “File -> Save As” to give the project a name.
From the main Simcenter Testlab menu, under “Tools -> Add-ins” turn on:
Time Recording during Signature Acquisition (16 tokens)
In the “Channel Setup” worksheet, define the microphones to be used during the test as shown in Figure 15.
Figure 15: “Channel Setup” worksheet with four ICP microphone setup. In this example four microphones are used but more or less can be used as needed.
In the “Tracking Setup” worksheet, set the duration and increment for the measurement in the right side of the screen as shown in Figure 16.
Figure 16:In the Tracking Setup worksheet, set the "Duration" and "Increment".
The duration must be sufficient to capture the high level time when the source is on and the complete decay of the sound once the source turns off. In most applications, 15 seconds will suffice. If the increment is set to the same as the averaging frame, no data will be missed. The increment should be set to ensure at least 5 to 10 data points are acquired during the sound decay.
In “Acquisition Setup” worksheet, turn on sources in the middle right side as shown in Figure 17.
Figure 17: On the right hand side of the “Acquisition Setup”, in the “Source Control” area, turn on the sources and set the signal type to “Random”. ISO standards recommends using pink noise.
In the Acquisition Setup, the bandwidth setting dictates the highest frequency that can be measured. Set this appropriately for the frequency range desired to be measured.
In the source control area, do the following:
Connect the output of the SCADAS to the amplifier for the acoustic source.
Set the signal type to “Random” and level to 1 Volt. (pink noise under “More”)
Turn the amplifier all the way down.
Press the “Start Source” button.
Turn the amplifier up slowly until the desired sound level is achieved (minimum = Noise floor (dB) + 10 dB + planned decay (20, 30, or 60 dB) + 5 dB. Be careful not to over drive the acoustic source and amplifier.
In “Online Processing” worksheet, go to “Real-time Octaves” tab. Then go to “Reverberation” tab (Figure 18).
Figure 18: In the “Online Processing”, choose the “Real-Time Octaves”
Under “Time Averaging Parameters”:
Specify an Averaging method (linear or exponential) for smoothing the decay curves.
Select a “Decay Curve Extrapolation” : T60, T20, or T30.
Specify “Distance to Noise Floor” (default 10 dB) to avoid end of decay effects, and “Decay start” (default 5.0 dB) to exclude early decay exclusions after source cutoff.
Use “Measure” worksheet to acquire the data (Figure 19):
Figure 19: To start the measurement, first press the "Arm" button and then the button with arrow symbol. Note that the Source Control area should say "Reverberation".
Note that the source control area should say “Reverberation”. This means that the source will be turned on and then shutoff automatically during the measurement for the interrupted mode.
Press the “Arm” button and then “Start” button (button with arrow symbol) to take data.
4.2 Reverberation Data Viewing
In the Simcenter Testlab Navigator, the T60 reverberation spectrum can be displayed in 2D graphical layout as shown in Figure 20.
Figure 20: T60 Reverberation times per octave band.
To display the T60 times:
Go to the folder called “Reverberation” in the Run. Find the “Time” folder.
Change the Y-axis format to “Amplitude”
Change the X-axis format to “Octave”
Change the line style to “Block Outlines”
Note that even if T20 or T30 method was selected to calculate reverberation time, the reported results in the Time folder are the T60 times. The software automatically does the extrapolation to report the T60 results.
Decay curve plots help assess the signal-to-noise ratio and regression quality in each band as shown in Figure 21.
Figure 21: An important tool in Simcenter Testlab for validating the T60 calculation is to overlay the octave level versus time (red curve) with the corresponding decay regression (green). In this case, eight points (green crosses on green line) were used in the calculation of the T60 decay time.
For any given octave band, overlay the octave band level versus time with the decay regression:
In the “Real time octaves -> Time -> Sections -> Octave” folder find and plot the 1/3 octave band of interest.
In the “Reverberation -> Decay regression” folder, find the corresponding decay regression for the same 1/3 octave band for the same microphone.
This allows the user to see if the software found an acceptable location and number of points on the decay curve to do the T60 decay time calculation.
Some additional considerations for accurate T60 testing in Simcenter Testlab are:
5.1 Warning: Accuracy Failure Due to Background Noise
Possible warning messages include “The decay curves cannot be estimated accurately for the following octave bands” or “Accuracy fails because of the background noise in the following octave bands” can appear when doing a T60 calculation in both online and offline mode (Figure 23).
Figure 23: Simcenter Testlab warning message when calculating T60 decay times. In this case, octave bands 3150, 4000, and 5000 are suspect.
The warning message flags octave bands that cannot be calculated correctly for two major reasons:
Range: There is not enough range in the decay between the high level limit and the noise floor.
Increment: The time increment was not fine enough to get points on the decay curve.
Pressing “Yes” calculates all octave bands possible, but the octave bands where the calculation is not possible will be set to value of 0.
To overcome the warning several things can be tried:
Switch the “Decay Curve Extrapolation” to T30 or T20 from T60 as shown in Figure 24.
Figure 24: Switching “Decay Curve Extrapolation” to T20 or T30 from T60 will permit measurement in spaces where 60dB decay couldn’t be achieved.
Increase the volume of the acoustic source to increase the high level.
Make sure the reverberation chamber is appropriately sealed from external noise to lower the noise floor.
Use higher dynamic range microphones or more sensitive microphones to lower the noise floor.
Lower the “Decay Start” and “Distance to Noise Floor” from the default values of 5 dB and 10 dB respectively. (I wouldn’t recommend this, violates ISO standard)
Make the time increment finer. The decay may be fast relative to the spacing of the data points over time such that there are not enough points on the decay to determine the slope accurately.
If the above suggestions fail, may also want to restrict which octave bands are used in the calculation.
Adjustments will depend on the 1/3 octave band. View the highest and lowest octave bands overlaid with the decay regression for guidance on how to change the settings. Processing the data several times in offline mode successfully can be used to inform online acquisition settings for subsequent acquisitions.
5.2 All Decay Times Nearly Identical for Every Octave Band
If the decay times for each octave in the reverberation spectrum are nearly identical, this may indicate that a different issue (Figure 25).
Figure 25: Reverberation spectrum with nearly identical decay times for each 1/3 octave band.
Identical decay times across all octave bands are unlikely. There is a good possibility that the reverberation spectrum is measuring the decay of the time-domain 1/3 octave filters being used (which is the same for each octave), not the decay of the noise level of the 1/3 octave in the measurement environment.
For time-domain filters, even if the signal instantaneously drops to zero, the corresponding filtered signal will not. Filters utilize small amounts of past time data (the exact amount depends on the filter parameters), so it takes some time for the filter to respond to the change in the signal. In other words, the filter itself has a decay. More about filters in the knowledge article: Introduction to Filters: FIR versus IIR
If a test environment has a quicker decay than the 1/3 octave filter decay (Figure 26), the reverberation times will reflect the settings (averaging frame, exponential versus linear, etc) used in the calculation of the octaves.
Figure 26: If the decay associated with the 1/3 octave filter (yellow) is longer than the decay of the test environment (green), it is impossible to measure the T60 decay time properly.
When the test environment decay is faster than the decay of the octave filter, the decay times are identical because the real decay of the test environment is not being measured. The averaging frame needs to be made shorter.
A common reason might be testing in a very absorptive environment. Examples include anechoic chambers, vehicle interiors, etc. This situation can arise when octave filter settings for a reverberant environment are used in a very absorptive setting.
