At the same thickness and the same mounting, PET felt, mineral wool and fiberglass absorb sound within about 0.10 of each other; at about 1 in laid flat, published lab data puts all three between NRC 0.65 and 0.75. So choose the core on everything else. Mineral fiber cores lead on fire results and fiberglass is the lightest per inch, PET felt leads when the core must also be the finished face, be handled without loose mineral fibers or carry recycled plastic content, and foam is the outlier the building code largely keeps off walls and ceilings.
This comparison is for designers, specifiers and owners deciding what goes inside a panel. It uses published lab reports and maker data sheets, names the test and mounting behind every number, and keeps to the core. How thick a panel should be is covered in how thick acoustic panels should be, the face fabric in choosing and cleaning acoustic panel fabric, and fire code routes in what a Class A fire rating does and does not cover.
The short comparison
Each core does the same acoustic job; they differ in fire results, density, how they are built into a panel, what they shed when cut, and what they are made from. The table summarizes the data cited in the sections below.
| Property | Fiberglass board | Mineral wool (stone wool) board | PET felt |
|---|---|---|---|
| What it is | Glass fibers with a thermosetting binder | Stone wool fibers in a rigid board | Polyester fibers compressed into a felt |
| Absorption at about 1 in, Type A | NRC 0.70 (3 pcf), 0.75 (6 pcf) | NRC 0.65 to 0.75 (25 mm, three grades, ISO 354 data) | NRC 0.65 to 0.75 (6 to 9.67 pcf) |
| Example density | 3.0 or 6.0 pcf | 6 pcf | 6 to 9.67 pcf |
| Example E84 indices | Flame spread 5, smoke 5 (unfaced) | Flame spread 0, smoke 0 | Flame spread 15, smoke 250 (one maker) |
| Needs a fabric face? | Yes, for looks and to contain fibers | Yes | No; felt can be the finished face |
| Edges | Resin-hardened or framed | Resin-hardened or framed | Cut edges can stay exposed |
| Recycled content (example) | 53 percent recycled glass (one maker) | Ask for the document | 60 percent recycled PET (one maker) |
Every figure in that table belongs to a named product. Treat it as a guide to what to ask for, not as a rating for a category.
What “mineral wool” means on a data sheet
The terms overlap. OSHA’s summary of the IARC evaluation lists glasswool, rockwool and slagwool as separate fibers, and US data sheets usually call the rock-based product stone wool. Both the fiberglass board and the stone wool board cited here state compliance with ASTM C612, the standard for mineral fiber block and board insulation, under different types. That is why some specifications say “mineral fiber” and accept either. If your project means stone wool specifically, say so.
What thickness each core comes in
Availability shapes the choice before acoustics does. The Owens Corning sheet lists its 3 pcf board from 3/4 in to 4 in and its 6 pcf board from 1/2 in to 2-1/2 in. The Rockboard 60 sheet lists 2 in through 4 in. The PET catalog listing offers 1/2 in, 1 in and 2 in, and the UFGS template lists felt panels as thin as 9 mm and PET panels as thin as 6 mm. So a thin, shaped panel points to felt, and a 3 or 4 in panel points to mineral fiber.
Absorption at the same thickness and mounting
When thickness and mounting match, the three cores land close together. The differences that matter acoustically come from depth and air space, not from the fiber.
The figure compares three cores about 1 in thick, laid flat against the test surface. The fiberglass data comes from the Owens Corning 700 Series data sheet (ASTM C423, Type A). The stone wool data comes from an SRL Technical Services report for Rockwool, measured to ISO 354, with an NRC the lab calculated from it. The PET felt data comes from Riverbank Acoustical Laboratories report RAL-A23-084, an accredited ASTM C423-22 test in Type A mounting.
What the numbers say
At 500 Hz the three cores read 0.63, 0.69 and 0.67. At 250 Hz, where speech and room boom start to matter, they read 0.26, 0.28 and 0.19, all weak. The three single numbers are NRC 0.70, 0.75 and 0.75. Other products of the same thickness fall in the same band: the denser 6 pcf fiberglass board in the same data sheet is NRC 0.75, three 25 mm stone wool grades in the SRL report run from 0.65 to 0.75, and a 1 in, 7.5 pcf PET panel in a maker’s catalog listing is NRC 0.70.
Where the cores differ, band by band
The curves are not identical. In this set, the PET felt is the weakest of the three at 250 Hz (0.19) and the strongest at 1,000 Hz (1.01), and the stone wool slab is the strongest at 4,000 Hz (1.09). Density moves the result too, within a material. At 1 in, the Owens Corning 6 pcf board reads 0.73 at 500 Hz and 1.06 at 1,000 Hz against 0.63 and 0.88 for the 3 pcf board. The three 25 mm stone wool grades in the SRL report span NRC 0.65 to 0.75, and at 100 mm five slab grades span 0.95 to 1.15. A 1 in, 6 pcf PET panel on a Riverbank preliminary data sheet (A23-085) reads NRC 0.65 against 0.75 for the 9.67 pcf felt in the figure. None of these differences is as large as the step from 1 in to 2 in.
How small is small?
The SRL report gives its own expanded uncertainty for the single-number weighted coefficient as plus or minus 0.07, and an accredited US report on a fabric over a board, discussed in our fabric article, lists between-lab reproducibility of plus or minus 0.14 at 500 Hz. Differences of 0.05 between cores tested in different labs are inside that spread. They are not a reason to pick one material over another.
At 2 in, the same story
| 2 in core, published data | 125 Hz | 250 Hz | 500 Hz | 1,000 Hz | 2,000 Hz | 4,000 Hz | NRC |
|---|---|---|---|---|---|---|---|
| Fiberglass, 3 pcf, Type A | 0.15 | 0.71 | 1.14 | 1.16 | 1.02 | 1.00 | 1.00 |
| Fiberglass, 6 pcf, Type A | 0.18 | 0.86 | 1.16 | 1.16 | 1.03 | 0.97 | 1.05 |
| Stone wool, 6 pcf, mounting not stated | 0.28 | 0.75 | 1.11 | 1.08 | 1.05 | 1.07 | 1.00 |
| PET felt, 7.5 pcf, Type A | 0.31 | 0.58 | 1.08 | 1.11 | 1.04 | 1.04 | 0.95 |
The fiberglass rows are from the Owens Corning sheet, the stone wool row from the ROCKWOOL Rockboard 60 data sheet issued in April 2026, and the PET row from the catalog listing above. Going from 1 in to 2 in moved every core far more than switching between cores did. That is the thickness question, and it is answered in detail in how thick acoustic panels should be.
Why core comparisons go wrong
Most claims that one core “absorbs better” compare different thicknesses, mountings, densities or test standards. Line those up before you compare, and the gap usually closes.
Four things to match
- Thickness. A 2 in board against a 1 in felt is a thickness comparison, not a material one.
- Mounting. Type A data against data with an air space behind the panel says more about the air than the core. Specifying acoustic panels explains how to match the tested mounting to the installed detail.
- Test standard. ASTM C423 reports NRC and SAA; ISO 354 reports coefficients that are often summarized as a weighted coefficient. The SRL report gives both a weighted coefficient and an NRC it calculated, which is helpful but not the same as a C423 test.
- Facing. A board tested bare is not the same as a board with a foil or film facing. In the same Owens Corning sheet, a 1 in board’s coefficient at 4,000 Hz falls from 0.94 unfaced to 0.21 with a foil and kraft facing, while its NRC stays at 0.70.
Density is not a quality grade
Higher density is often presented as better. In the data above it helped a little at 1 in and hardly at all at 2 in: the 6 pcf fiberglass board is NRC 1.05 at 2 in against 1.00 for the 3 pcf board. Density does change weight, stiffness and impact resistance, which may be the real reasons to choose it. Ask why a supplier proposes a given density, and expect an answer about the panel’s use, not only its NRC.
When a data sheet leaves something out
The Rockboard 60 sheet gives ASTM C423 results but does not name the mounting. The PET catalog listing gives a coefficient table headed as a demonstration but does not name the test standard. Neither omission makes the data wrong. Both mean the number cannot be compared like for like until the maker supplies the full report, and the NRC calculator is only as good as the band data you enter.
Example (fictional). Two quotes arrive for the same 2 by 4 ft wall panels. One offers a 2 in mineral wool core “NRC 1.00.” The other offers a 1 in PET felt “NRC 0.90.” The second figure comes from a test with a 400 mm air space behind the felt; laid flat, the maker’s own report shows NRC 0.70. The specifier asks both suppliers for Type A data at the thickness quoted, and the comparison becomes 2 in at 1.00 against 1 in at 0.70: a difference in depth, not in material. Both quotes are invented for this example.
Fire results: what each core’s report says
Mineral fiber cores post the lowest E84 indices, and stone wool is also tested as noncombustible; PET felt can reach Class A but with higher smoke figures. None of these reports describes a finished panel with fabric and adhesive on it.
| Core (named product) | ASTM E84 flame spread | ASTM E84 smoke developed | Other |
|---|---|---|---|
| Fiberglass board, unfaced | 5 | 5 | Values “reported to the nearest 5 rating” |
| Fiberglass board, faced | 25 | 50 | Same sheet, with facing |
| Stone wool board | 0 | 0 | Non-combustible under CAN/ULC S114 |
| PET felt panel (catalog) | 15 | 250 | Listed as Class A; test not named |
The fiberglass figures are from the Owens Corning sheet, which also notes the product “does not contain the fire retardant decabrominated diphenyl ether.” The stone wool figures are from the Rockboard 60 sheet. The PET figures are from the catalog listing. Under the classes in IBC section 803.1.2, all four are Class A: flame spread 0 to 25 with smoke 450 or less.
Smoke as well as flame
The smoke-developed index is where the cores separate most. The mineral fiber boards report 0 and 5 unfaced; the PET listing reports 250. All are within the code’s 450, but a project specification can set a lower limit. The VA master specification asks for “200 or less smoke density” for the complete composite system, which a panel whose core alone reports 250 could not meet. Check the project specification before choosing a core, not after.
Heat sources nearby
The fiberglass sheet gives an operating temperature range up to 450°F, and the stone wool sheet a hot surface performance of 1200°F. Those are insulation ratings, not permission to put an acoustic panel near a fireplace surround, a pizza oven or a hood. Clearances to heat-producing equipment are set by the appliance listing and the fire official, and the panel’s fabric and adhesive are usually the limiting parts.
Read the report, not the class
ASTM E84 warns in section 1.5 that materials that “melt, drip, or delaminate” can produce low flame spread indices “that do not relate directly” to materials that stay in place. For any core that can soften in heat, ask whether the lab recorded melting or dripping. Section 1.4 adds that supports under a specimen can lower its index. A core’s Class A is a fact about the core; whether the finished panel qualifies is a separate document, as our article on fire ratings for acoustic panels explains.
Added flame retardants
The federal UFGS template says felt acoustical panels “must be free of formaldehyde, heavy metals, and halogenated flame retardants,” and the public excerpt of BuildingGreen’s product guide lists flame retardants among the things to “watch out for” in acoustic panels. If a core reaches its fire class with an added retardant, ask which chemical it is.
Weight, rigidity and how each core becomes a panel
Fiberglass at 3 pcf is the lightest per inch of the products compared here; dense PET felt is the heaviest. Mineral fiber boards need a fabric face and hardened or framed edges, while felt can be cut, shaped and left exposed.
Weight per panel
Weight comes from density and thickness. A 2 by 4 ft panel 1 in thick has a core volume of 8 sq ft x 1/12 ft = 0.67 cu ft.
- Fiberglass at 3.0 pcf: 0.67 x 3.0 = 2.0 lb.
- Fiberglass or stone wool at 6 pcf: 0.67 x 6 = 4.0 lb.
- PET felt from RAL-A23-084, at 0.77 lb per square foot: 8 x 0.77 = 6.2 lb.
Those are core weights only; fabric, frame and hardware add more. Weight matters for adhesive mounting, for ceiling clouds hung on cables, and for the IBC requirement in section 803.14 that interior finish stay attached at room temperatures of 200°F for not less than 30 minutes.
Worked example: a ceiling cloud
Suppose a design calls for 4 by 4 ft clouds, 2 in thick, hung on four cables each. The core volume is 16 sq ft x 2/12 ft = 2.67 cu ft. At 3 pcf the core weighs 8.0 lb, at 6 pcf 16.0 lb, and at 7.5 pcf 20.0 lb, or 2, 4 and 5 lb per cable before fabric, frame and hardware. The numbers are hypothetical but the ratio is the point: the same cloud can weigh two and a half times as much depending on the core. Hanging points, structure and the install method all follow from that, which is one reason installation support starts with the panel schedule.
Edges and frames
The VA’s master specification for sound-absorbing wall units offers edge construction as a “chemically hardened core with no frame,” an aluminum or steel frame, or a PVC frame. The UFGS template offers a “formulated resin edge hardener,” an aluminum frame or a plastic edge track for fabric-covered panels. A mineral fiber board needs one of these to keep its edges square and contained under the fabric. A felt panel does not: the UFGS describes felt panels as “homogeneous felt construction” or felt bonded to a rigid PET or tackable board, with square, chamfered or CNC-cut edges.
Impact and pins
The Owens Corning sheet describes its 6 pcf board for areas where “high abuse resistance is required,” and the VA template offers standard, high-impact and tackable core types. PET felt panels are often sold as tackable, as the catalog listing cited here is, and the UFGS asks for “non-irritating, tackable” PET panels. In a classroom, a corridor or a restaurant at chair-back height, ask which core type the supplier proposes and why.
Felt as the finished face
A felt panel can be its own surface: cut to shapes, grooved, printed or layered without fabric. That is a design advantage and a care consideration, since the face you see is the core. The felt panels page shows the options, and fabric-wrapped wall panels describes the wrapped alternative.
Moisture, mold and where cores live
All three cores tolerate normal indoor humidity, and the mineral fiber data sheets report standard moisture and fungi tests. Ask for the same data from a PET supplier rather than accepting a general claim.
The Owens Corning sheet reports water vapor sorption of less than 5 percent by weight at 120°F and 95 percent relative humidity (ASTM C1104) and says the board meets the fungi resistance requirements of ASTM C1338. The Rockboard 60 sheet reports water vapor sorption of 1.5 percent by volume and a pass on fungi resistance under the same two methods. The PET catalog listing describes its panels as “resistant to mold and fungi” without naming a test.
Ask for the method, not the adjective
“Resistant to mold” is an adjective; a pass under ASTM C1338 is a result. The two mineral fiber sheets above give both a fungi resistance result and a water vapor sorption figure under ASTM C1104, so a PET supplier can be asked for the same two tests on its felt. If the answer is a brochure line rather than a report, record it as undocumented in your comparison.
Where moisture is a real question
Panels near dish areas, entrances that see rain, pools or unconditioned spaces deserve a direct question to the maker about humidity limits and cleaning, for the face as well as the core. Whatever the core, keep panels out of the air above cooking lines and steam wands, a point covered for cafes in coffee shop acoustics.
Caring for a felt face
When felt is the finished surface, the core is what people touch and what collects dust. Follow the felt maker’s care instructions, dust gently rather than scrubbing, and test any spot cleaner on an offcut first. A wrapped panel’s care is the fabric’s care, which is covered in choosing and cleaning acoustic panel fabric.
Fibers, dust and handling
Fiberglass and mineral wool are fibrous, and cutting them makes dust; PET felt contains no glass or stone fibers. For the occupants of a finished room, a sealed fabric wrap and hardened edges keep mineral fibers inside the panel.
OSHA’s page on synthetic mineral fibers reproduces the 2002 IARC evaluation, which lists glasswool, rockwool and slagwool in Group 3, defined there as “Cannot be evaluated as to its Carcinogenicity to Humans.” That is a classification of evidence, not a statement that the fibers are harmless to work with. Anyone cutting boards should follow the maker’s safety data sheet.
What this means in practice
- Fabrication. Fabric-wrapped panels with mineral fiber cores are usually cut, edge-hardened and wrapped before they reach the site. Felt can be trimmed on site with a knife.
- Exposed edges. A mineral fiber core should never be left with raw edges in an occupied room. Felt edges can stay exposed by design.
- Tackable walls. If the wall will take pins, specify a tackable core type, which the VA template offers, or a felt panel.
- Homes and children’s rooms. Some owners prefer a material without glass or stone fibers. That is a preference, not a safety finding, and it is a reasonable one to plan around.
Recycled content, emissions and chemicals of concern
All three cores can carry recycled content, and each maker documents it differently. Ask for the document, the emissions report and the version of the method it used, rather than a label.
Recycled content
- Fiberglass. The Owens Corning sheet reports an average of 53 percent recycled glass content, 31 percent pre-consumer and 22 percent post-consumer, certified by SCS Global Services.
- PET felt. The PET catalog listing gives “60% PET-recycled fiber” and “40% PET-virgin fiber.” The UFGS template’s bracketed default for PET panels is 50 percent post-consumer recycled content, which a specifier is told to verify before keeping.
- Stone wool. We did not find a recycled-content figure in the documents read for this article, so ask for it.
For context on where recycled PET comes from, the US EPA reports that “the recycling rate of PET bottles and jars was 29.1 percent in 2018,” per its plastics material data page. A felt’s recycled content is a claim about its fiber source, which the maker should document.
Environmental and health product declarations
The Owens Corning sheet lists an environmental product declaration certified by SCS Global Services and a Health Product Declaration for the 700 Series board. An environmental product declaration reports life-cycle impacts under published rules, and a health product declaration lists contents and their hazards. Ask stone wool and PET suppliers for the equivalent documents and compare like with like: the same functional unit, the same rule set and a current date.
Emissions and chemicals to watch
The UFGS template requires PET and felt panels to be “formaldehyde-free and low-emitting” and to comply with the California Department of Public Health Standard Method, which it cites as Version 1.2 (2017). CDPH has since published Version 1.3, with new criteria effective January 1, 2027, so name the version you accept. The PET catalog listing says its panels meet CDPH Version 1.2. BuildingGreen’s excerpt lists PFAS stain repellents, flame retardants, formaldehyde and PVC as things to watch for; the first of those usually arrives with the fabric, which the fabric article covers.
We do not describe any core as “non-toxic” or “eco-friendly,” and you should be wary of suppliers who rely on those words instead of reports.
Foam: the budget outlier
Open-cell foam panels are sold widely as acoustic treatment, but the IBC restricts foam plastics as interior finish, including when they are covered with fabric. For commercial interiors, that usually settles the question before acoustics does.
In the 2021 Seattle Building Code, Chapter 8, which adopts the 2021 IBC with local amendments, section 802.7 says foam plastics “shall not be used as interior finish except as provided in Section 803.4,” and that this applies “both to exposed foam plastics and to foam plastics used in conjunction with a textile or vinyl facing or cover.” Section 803.4 points to the conditions in section 2603.9. A committee statement in the NFPA 5000 first revision record for the 2017 cycle notes that foam plastics “are not allowed to be tested to ASTM E84” in that code, which leaves room-corner testing.
Where foam still appears
Foam turns up most in home studios and other small private rooms. Even there, check the product’s fire documents and ask the local building official, since the foam plastic provisions are written for interior finish generally, not only for commercial buildings. For any business, school or public space, plan on a mineral fiber or PET core instead, and treat foam as a different product category rather than a cheaper version of the same one.
If foam is already on the walls
Owners sometimes inherit foam panels in a cafe, studio or office. Before adding more, ask the fire official how the existing foam is viewed, and keep the fire documents for whatever goes up next. Replacing foam with mineral fiber or felt panels at the same thickness keeps the acoustic result in the same range, based on the data above, and takes the foam question off the table.
Choosing a core by room
Pick by the room’s fire class, the look you want, who handles the panels and what the owner values; absorption at a given thickness will be close either way. The table gives a starting point.
| Room | Usually leans to | Why |
|---|---|---|
| Restaurant dining room, fabric-wrapped wall panels | Mineral fiber core | Low E84 indices help with fabric-faced panels, which may face the code’s textile route |
| Cafe or office with shaped or grooved felt features | PET felt | The felt is the finish; no wrap or edge treatment needed |
| Open office ceiling clouds | Either; weight decides | Lighter fiberglass eases hanging; felt clouds need no frame |
| School corridor or classroom tack wall | PET felt or a tackable facing over mineral fiber | Pins and knocks at body height |
| Home office or nursery | PET felt, by preference | No glass or stone fibers; can be trimmed at home |
| Recording or practice room at home | Mineral fiber at greater depth | The 2 in data above absorbs far more at 250 Hz than any 1 in core |
Example (fictional). A home office is 10 by 12 ft with a 9 ft ceiling, 1,080 cu ft. Suppose its surfaces and furniture add up to 60 sabins at mid frequencies, so its reverberation time is 0.049 x 1,080 / 60 = 52.92 / 60 = 0.88 s. Six 2 by 4 ft panels cover 48 sq ft of drywall, taken here at 0.05. With 1 in fiberglass at NRC 0.70, they add 48 x (0.70 - 0.05) = 31.2 sabins, for 52.92 / 91.2 = 0.58 s. With 1 in felt at NRC 0.75, they add 48 x (0.75 - 0.05) = 33.6 sabins, for 52.92 / 93.6 = 0.57 s. The core changed the estimate by about a hundredth of a second; the decision rests on looks, handling and documents. The room and its absorption are invented for this example, and using NRC as a mid-frequency coefficient is a planning shortcut.
Mixing cores in one project
Nothing stops a project from using more than one core: mineral fiber in fabric-wrapped wall panels, felt for shaped features and clouds. Keep the documents in step. Each core needs its own absorption and fire reports, the finish schedule should name which core goes where, and the extra stock and replacement plan should follow the same split.
A simple decision rule
- Set the thickness and mounting from the acoustic need.
- Set the fire class and code route from the room.
- Decide whether the core is also the finished face.
- Then choose the core that meets steps 1 to 3 with the best documents for emissions, recycled content and handling.
Example (fictional). The site’s example dining room plans 14 fabric wall panels, 2 by 4 ft, at seated ear height, at a planning coefficient of 0.80, which is an assumption rather than a test result. Its owner asks whether to use felt instead. At 1 in and Type A, the published data above puts mineral fiber and PET felt at NRC 0.70 to 0.75, so absorption is not the deciding factor. The fabric-wrapped panels may fall under the code’s textile route, where a mineral fiber core’s low indices give more margin. The owner keeps mineral fiber cores for the wall panels and considers felt only for a shaped feature by the host stand. The room is not a Fibson project.
What to ask for, and what to do next
Ask every supplier the same questions, whatever core they propose. The answers will usually make the choice for you.
- Which core, at what thickness and density?
- The full ASTM C423 report for that core at that thickness, with the E795 mounting named.
- The ASTM E84 report for the core, and separately for the finished panel with its face and adhesive.
- Whether the lab recorded melting, dripping or delamination.
- Edge construction: hardened, framed, or exposed felt.
- Weight per panel and the mounting hardware for it.
- Moisture and fungi test results.
- Recycled-content documentation and the emissions report, with the CDPH method version.
- Any added flame retardants, by name.
Comparing the answers
Put every supplier’s answers in one table, one row per question, and fill the gaps with “not documented” rather than leaving them blank. Compare absorption only at the same thickness and mounting, compare fire only on reports that describe the finished panel, and compare weight per panel rather than per cubic foot. A core that wins on paper but leaves three rows undocumented is a weaker choice than one that is slightly heavier and fully documented.
Red flags in core data
- An NRC with no thickness, no mounting or no test standard.
- Coefficients above 1.00 presented as “absorbs more than 100 percent of sound”; ASTM C423 explains them as a diffraction effect that makes a specimen’s apparent area larger than its real area, not as extra absorption.
- A fire class with no flame spread and smoke indices behind it.
- “Non-toxic”, “eco” or “green” in place of an emissions report or a recycled-content document.
- A felt and a wrapped mineral fiber panel compared at different thicknesses.
Fibson Acoustics designs and makes acoustic panels to order with mineral wool, fiberglass or PET felt cores, and has not yet tested its finished panels. Its room plans name the core maker’s report behind every figure, as described on materials and safety. Re-run the band data from any report through the NRC calculator, and see wood slat acoustic panels if the face you want is timber rather than fabric or felt.
