ACOUSTIC TESTING STANDARDS
ISO 354 Acoustic Panel Testing: Method and Mounting
ISO 354 acoustic panel testing measures how much sound a material absorbs. The test is carried out in a reverberation room by comparing the rate at which sound decays with and without the sample present, and it produces an absorption coefficient, α, for each frequency band.
SCOPE OF THE STANDARD
What ISO 354 Measures
The standard specifies a method for measuring the sound absorption coefficient of materials used as wall and ceiling treatments, and the equivalent sound absorption area of objects such as furniture or space absorbers.
01
The Coefficient Runs From 0 to 1
A value of 0 means the surface reflects essentially everything; a value of 1 means it absorbs essentially everything. Two measurement methods are permitted: the interrupted noise method and the integrated impulse response method.
02
Weakly Damped Resonators Are Out of Scope
The standard is not intended for weakly damped resonators, whose behavior it does not describe well. Applying its method to a resonant device produces a number that does not represent how that device behaves in a room.
03
Results Are for Comparison and Design
Its results are intended for comparison and for design calculation — not as a guarantee of in-situ performance. Mounting, room geometry, and installation detail all change what a product actually does on site.
04
Absorption Is Not Insulation
It measures absorption, which is a different property from insulation. An absorbing panel reduces reflected sound inside a room. It does not stop sound passing through a wall.
05
The Edition Matters
Every standards reference should carry its edition. ISO 354:2003 supersedes ISO 354:1985, and several regional adoptions of the older text remain in circulation. Confirm the current edition before quoting a figure.
TEST CONDITIONS
The Conditions ISO 354 Requires
The standard is specific about how the test is run. These are the conditions a report can be checked against, and they are the first thing to look for when comparing two sets of absorption data.
HOW TO READ THIS TABLE
Ten conditions govern the room, the sample, and the measurement positions. If a report cannot answer these, the figure it carries is not comparable with any other figure — whatever its value.
ROOM, SAMPLE AND MEASUREMENT
ISO 354 Test Conditions
Ten Conditions
| Condition | Requirement |
|---|---|
| Reverberation room volume | Greater than 200 m³ (7,063 ft³) |
| Room shape | Longest diagonal below 1.9 × V^(1/3) |
| Sample size | 10–12 m² (108–129 ft²) |
| Sample width-to-length ratio | 0.7 to 1 |
| Temperature | Above 15 °C (59 °F) |
| Humidity | 30–90% |
| Microphone positions | Minimum 3; at least 1.5 m apart, 2 m from the source, 1 m from any surface |
| Loudspeaker positions | Minimum 2; at least 3 m apart |
| Position combinations | At least 12 microphone and loudspeaker combinations |
01
A Figure Without Conditions Is Not Comparable
Two absorption coefficients measured under different room, sample, or position conditions describe two different tests. Comparing them directly produces a false conclusion either way.
02
Check the Report Before the Number
Ask the report to state the room volume, sample area, temperature, humidity, and the microphone and loudspeaker arrangement. If those are missing, the number cannot be checked against anything.
WHY THE CONDITIONS MATTER
The Same Panel Can Produce Different Numbers
Room volume, sample size, mounting detail, and the microphone and loudspeaker arrangement all change the measured coefficient. A report that states its conditions can be reproduced and compared; one that does not cannot.
| Group | Conditions Included |
|---|---|
| Room | Volume, shape, temperature, humidity |
| Sample | Size, width-to-length ratio |
| Measurement | Microphone positions, loudspeaker positions, position combinations |
THREE GROUPS
Room
Volume above 200 m³, controlled temperature and humidity, bounded diagonal.
Sample
10–12 m², with a width-to-length ratio between 0.7 and 1.
Measurement
At least 12 microphone and loudspeaker combinations, spaced per the standard.
Send us the acoustic specification you are working to — we will confirm which test data we hold within 12 hours.
TWO FAMILIES, THREE NUMBERS
ISO 354, ISO 11654, and ASTM C423: Two Families, Three Numbers
This is where most specifications go wrong, and it is worth stating as a table. The physics is essentially the same in both families. The outputs are not interchangeable, and the single-number ratings are different quantities.
MEASUREMENT AND RATING
ISO Family
ISO 354 is the measurement standard and produces α per frequency band. ISO 11654 is the rating standard and produces αw, classes A–E, and shape indicators L/M/H. The weighted coefficient is derived by comparing the practical coefficient against a reference curve, which penalizes low-frequency weakness more heavily than a plain average does.
What It Produces
MEASUREMENT AND RATING
ASTM Family
ASTM C423 uses the same reverberation-room physics and produces both NRC and SAA. The Noise Reduction Coefficient is the arithmetic average of absorption at 250, 500, 1,000, and 2,000 Hz, rounded to the nearest 0.05. It weights every band equally, ignores 125 Hz and 4,000 Hz, and can exceed 1.0 because of diffraction at the sample edges.
What It Produces
Why They Disagree, With Numbers
A Single Number Is Only Comparable Within Its Own Family
One published example: a product with octave-band absorption values of 0.30, 0.75, 0.95, 1.00, 0.95, and 0.85 returns an NRC of 0.91 and an αw of 0.80, Class A. A specification written as “NRC ≥ 0.85” would reject that product — a Class A material, refused on an arithmetic technicality, because the plain average fell two points short while the weighted rating placed it in the top class. The lesson is not that one rating is better. It is that a single number is only comparable within its own family. Where a specification crosses families, compare the per-band data instead. Broadly, Class A and B correspond to an NRC above about 0.75, and Class C and D to between about 0.50 and 0.75. Treat any translation as indicative only — these are two different calculations, not two units for one quantity.
VERIFICATION CHECKLIST
How to Read an Acoustic Test Report
This checklist applies to any supplier, and it applies to us. Eight questions separate a documentable absorption figure from a marketing statement, and none of them require an acoustics background to ask.
PART 01
Which Standards Were Used
A report should state both the measurement standard and the rating standard. ISO 354 measures; ISO 11654 rates. ASTM C423 produces NRC and SAA. Where the measurement standard and the rating standard are mixed between families, the figure is not comparable with anything else.
Which standard was the measurement carried out to
— ISO 354 or ASTM C423?
Which standard produced the single number —
ISO 11654 for an αw, or C423 for an NRC?
If any of these are missing, the figure is not
comparable — whatever it says.
A single number is only comparable within its own family. Where a specification crosses families, compare the per-band data instead.
PART 02
Mounting, Sample, and Data
The mounting condition changes the measured coefficient more than any other single factor. A panel tested on a direct fix and the same panel tested on an air gap are two different products acoustically, and a figure without its mounting type cannot be applied to a project.
What mounting type was used, and does it match how
the panel will be installed on your project?
What sample size was tested, and did it sit inside
the standard’s range?
The full per-band data, 125 Hz to 4,000 Hz — not
only the headline figure.
Use the acoustic panel specification template when you write the clause you will hold a supplier to, so the questions above are built into the document rather than asked afterwards.
Two More Requests
Ask for the Laboratory and the Product Identity
Ask which laboratory carried out the test, and on what date — and whether the report is current for the product as supplied. Then ask whether the report names your profile, at your thickness and finish. A figure from a different profile, a different thickness, or a superseded report does not transfer to the board you are buying. If any of these are missing, the figure is not comparable — whatever it says.
TEST METHOD
The Reverberation Room Method
The principle is straightforward. A calibrated sound source is switched off, and the decay of sound in the room is measured. The same measurement is taken first in the empty room and then with the sample present, and the difference between the two decay rates gives the sample’s contribution.
DECAY COMPARISON
Empty room vs sample present
Sound Source Off
Decay Curve
Empty Room
Sample Present
Measure the Empty Room
The reverberation time is measured in the empty room first. This establishes the baseline decay rate, which depends on the room’s own surfaces, volume, temperature and humidity. A calibrated source is used, and the decay is measured across the frequency bands.
Baseline DecayMeasure With the Sample Present
The same measurement is repeated with the sample mounted in the room. Because the sample adds absorption, the sound decays faster than it did in the empty room, and the difference between the two decay rates gives the sample’s contribution.
Faster DecayDerive the Equivalent Absorption Area
The Sabine absorption coefficient follows from the equivalent absorption area, using the room volume, the speed of sound, the measured reverberation time, the sample area, and the power attenuation coefficient of air at the measured temperature and humidity.
A = 55.3 × V ÷ (c × T)Compare Only Like With Like
The answer depends on the room as much as on the sample. Results are only comparable between laboratories when the standard’s conditions are met on both sides — room volume, sample area, temperature, humidity, and the microphone and loudspeaker arrangement.
Conditions Must Match
WHAT TO TAKE AWAY
The Answer Depends on the Room as Much as on the Sample
The full expression uses the room volume, the speed of sound, the measured reverberation time, the sample area, and the power attenuation coefficient of air, which depends on temperature and humidity and is specified separately. The practical consequence is that a figure without its room conditions cannot be transferred to a project. Confirm the clause numbering for this calculation against the current edition before publishing the formula.
Room Conditions
→ not optional
ISO 354 FAQ
Frequently Asked Questions
Common questions about ISO 354 acoustic panel testing, the difference between standards families, mounting condition, and how to check a test report.
ISO 354 is the international standard for measuring sound absorption in a reverberation room. It sets out how a material’s absorption coefficient is measured across frequency bands, along with the room conditions, sample size, and equipment positions the test requires. It measures absorption of sound, not transmission through a wall.
No. ISO 354 produces an absorption coefficient per frequency band. The Noise Reduction Coefficient is defined in ASTM C423, in a separate standards family. The two describe the same physical property but calculate their single-number ratings differently, which is why the values can disagree.
Both are reverberation-room measurement methods and the underlying physics is essentially the same. They differ in the family they belong to and in what they output: ISO 354 feeds ISO 11654, which produces a weighted coefficient and an absorption class; ASTM C423 itself produces the NRC and the Sound Absorption Average.
No, though they are often compared. NRC is a plain arithmetic average of four octave bands, rounded to 0.05. αw is weighted against a reference curve and is more sensitive to low-frequency weakness. A product can return an NRC of 0.91 and an αw of 0.80, which is a Class A result.
Because an air gap behind the panel lets it absorb more at lower frequencies. The standard defines several mounting types, from directly on a hard backing to a 400 mm air gap on battens. Published guidance puts the spread for one panel at roughly 20–50% across those arrangements, which is why the mounting condition has to travel with the figure.
Ask which standard the measurement followed, which standard produced the single number, what mounting type was used, what the sample size was, for the full per-band data, and for the laboratory and date. If any of those is missing, the figure cannot be compared with anything.