For rooms where people mostly talk, such as offices, meeting rooms, dining rooms and living rooms, acoustic panels 1 to 2 inches thick do most of the useful work, and 2 inches is the usual sweet spot. Published lab data for fiberglass board laid flat on a hard surface shows why: 1 inch absorbs about a quarter of the sound at 250 Hz, 2 inches absorbs most of it, and only 4 inches absorbs about half the sound at 125 Hz. An air gap behind a thin panel shifts its absorption toward lower pitches, but it does not add fiber, and every inch of gap counts against the 4 inch protrusion limit on walking routes.
This article reads the manufacturer lab tables that forum threads argue about, explains the mounting codes printed on every test report, and ends with a checklist for choosing a depth. It does not compare core materials; PET felt, mineral wool or fiberglass covers that. Where panels should go is the subject of where to put acoustic panels.
Why thickness decides how low a panel reaches
A porous panel absorbs sound by forcing air to move back and forth through its fibers, and friction turns a little of that motion into heat. Thickness matters because of where in front of a wall the air actually moves.
Right at a hard wall, air in a sound wave can hardly move, because the wall stops it. Movement grows with distance from the wall and peaks about a quarter of a wavelength out. Fibers sitting where the air barely moves do little work.
Low pitches have long wavelengths. Taking the speed of sound as 1,125 ft per second, about right for room-temperature air, the numbers look like this:
| Octave band | Wavelength | Quarter wavelength |
|---|---|---|
| 125 Hz | 9.0 ft | 27.0 in |
| 250 Hz | 4.5 ft | 13.5 in |
| 500 Hz | 2.25 ft | 6.75 in |
| 1,000 Hz | 13.5 in | 3.4 in |
| 2,000 Hz | 6.75 in | 1.7 in |
| 4,000 Hz | 3.4 in | 0.8 in |
Read the table and the forum argument starts to make sense. A 1 in panel fixed flat to drywall is already deep enough to matter for the top two bands and far too shallow for the bottom two. Even a 4 in panel reaches only about a seventh of the way to the 125 Hz quarter wavelength.
The quarter-wave rule is a direction, not a cutoff. A panel still absorbs below the frequency it points to, only less, and the lab tables in the next section show how much less. The air gap calculator does this arithmetic for any thickness and gap, so it is not repeated here.
Two things thickness does not change are worth stating early. A thicker panel absorbs more of the sound inside a room, but it does not stop sound passing through the wall behind it; that is a separate job measured with separate tests, explained in NRC vs STC. And thickness is only part of the material story: density, fiber type and any facing on the board change the curve too, as the Owens Corning table below shows.
What the lab tables show for 1, 2, 3 and 4 inch boards
The cleanest evidence comes from material makers that test one board at several thicknesses with the same mounting. In the data sheet for its 700 Series fiberglass board, dated August 2026, Owens Corning lists ASTM C423 coefficients for boards from 1 to 4 in thick, all in Type A mounting, described on the sheet as “material placed against a solid backing.”
Here is the same data as a table, with the 3 in board added. The unfaced Type 703 board has a nominal density of 3.0 lb per cubic foot.
| Thickness | 125 Hz | 250 Hz | 500 Hz | 1,000 Hz | 2,000 Hz | 4,000 Hz | NRC |
|---|---|---|---|---|---|---|---|
| 1 in | 0.05 | 0.26 | 0.63 | 0.88 | 0.97 | 0.94 | 0.70 |
| 2 in | 0.15 | 0.71 | 1.14 | 1.16 | 1.02 | 1.00 | 1.00 |
| 3 in | 0.34 | 1.09 | 1.25 | 1.15 | 1.05 | 1.00 | 1.15 |
| 4 in | 0.56 | 1.26 | 1.22 | 1.11 | 1.07 | 0.96 | 1.15 |
Three patterns stand out:
- The step from 1 to 2 in buys the most. At 250 Hz the coefficient rises from 0.26 to 0.71, and at 500 Hz from 0.63 to 1.14.
- The step from 2 to 4 in mostly buys 125 and 250 Hz. At 125 Hz the coefficient nearly quadruples, from 0.15 to 0.56. From 500 Hz up, the 2 in and 4 in boards are within about 0.1 of each other.
- From 1,000 Hz up, thickness hardly matters. Every thickness reads 0.88 or more, which is why thin panels already tame the hiss and clatter of a hard room.
Coefficients above 1.00 are not a typo. The significance section of ASTM C423 explains that diffraction usually makes a specimen’s apparent area larger than its geometric area, so a highly absorptive specimen can measure above unity. Treat anything above 1.00 as absorbing essentially everything that reaches it.
A second maker, a second standard, the same pattern
A 2025 lab report by SRL Technical Services for Rockwool UK tested three mineral wool slab grades at 25 mm and 100 mm, about 1 in and 4 in. The lab used ISO 354, the international reverberation room method, with Type A mounting: specimens laid directly against a room surface with no air space behind them. The report gives one-third octave results; the figures below are the bands centered on 125, 250 and 500 Hz, plus the NRC that SRL also calculated to ASTM C423.
| Slab | Thickness | 125 Hz | 250 Hz | 500 Hz | NRC |
|---|---|---|---|---|---|
| RW3 | 25 mm | 0.07 | 0.27 | 0.62 | 0.70 |
| RW3 | 100 mm | 0.54 | 1.16 | 1.24 | 1.15 |
| RW5 | 25 mm | 0.08 | 0.28 | 0.69 | 0.75 |
| RW5 | 100 mm | 0.63 | 1.18 | 1.13 | 1.10 |
| RWA45 | 25 mm | 0.10 | 0.25 | 0.60 | 0.65 |
| RWA45 | 100 mm | 0.51 | 1.25 | 1.23 | 1.15 |
Different fiber, a different lab and a different test standard, yet the thin slabs land within a few hundredths of the 1 in fiberglass at 250 Hz, and the thick slabs land close to the 4 in board. That agreement is the strongest reason to trust the general rule: depth, more than brand, decides how low a porous panel reaches.
Why the NRC on a spec sheet hides the difference
The noise reduction coefficient flattens exactly the frequencies where thickness matters. As Intertek summarizes ASTM C423, NRC is the average of the coefficients at 250, 500, 1,000 and 2,000 Hz, rounded to the nearest 0.05. The sound absorption average (SAA) averages the one-third octave bands from 200 to 2,500 Hz and is rounded to 0.01.
Neither number includes 125 Hz. In the Owens Corning table, the 3 in and 4 in boards share an NRC of 1.15, yet at 125 Hz one reads 0.34 and the other 0.56. The denser Type 705 board on the same sheet shows the same thing: its 3 in and 4 in versions both carry an NRC of 1.10, with 125 Hz values of 0.45 and 0.60.
NRC also saturates. Once three of its four bands are at or above 1.00, extra depth barely moves the average, even when the low end improves a great deal. The rating does show the step from 1 in to 2 in clearly (0.70 to 1.00 for Type 703), which is the step most speech rooms care about.
A detail from the reports themselves: a 2023 Riverbank Acoustical Laboratories report describes NRC as “defined from previous versions of ASTM C423,” while it points to ASTM C423-17 for the definition of SAA. Spec sheets, including the two quoted here, still list NRC, so you will keep seeing it; by definition it leaves out the 125 Hz band.
The practical rule follows. Compare panels on the full octave-band table, and look at 125, 250 and 500 Hz before you look at any single number. The NRC calculator turns a band table into NRC and SAA so you can check a supplier’s rounding, and specifying acoustic panels covers how to write the table into a specification.
How much depth rooms for talking need
Most rooms people complain about are rooms for talking, and the reverberation targets that exist for them are set in the middle bands. In the LEED v4.1 acoustic performance credit, Table 2 gives reverberation limits “at 500 Hz, 1000 Hz, and 2000 Hz”: under 0.6 seconds for conference rooms and under 0.8 seconds for open-plan offices without sound masking and for hotel meeting or banquet rooms.
In those three bands, the 2 in board in the Owens Corning table already reads 1.02 or more. A 4 in board adds almost nothing to the numbers a target like this checks. That is the core of the decision: for speech rooms, the step from 1 in to 2 in changes the result you are graded on, and the step from 2 in to 4 in mostly changes bands the target does not mention.
That does not make thick panels pointless. It means their extra depth is spent below 250 Hz, which matters in rooms where low sound is the complaint. Choosing a target for your room is its own question, answered in what reverberation time your room should have.
The table below gives starting points, not rules. They follow from the lab data above and the bands each room’s complaints usually sit in.
| Room | Usual complaint | Starting depth | Go deeper when |
|---|---|---|---|
| Meeting room or private office | Echo on calls, voices bouncing between glass and drywall | 1 to 2 in on walls, plus ceiling treatment if the ceiling is hard | Low hum or a boomy video system is part of the complaint |
| Open office | Voices carrying across desks | 2 in wall panels; ceiling absorption does most of the work | Panels hang high or on the ceiling, where depth is free |
| Restaurant, cafe or bar | A roar that builds as the room fills | 2 in wall panels at seated ear height, clouds or baffles overhead | Loud music with heavy bass is a large part of the noise |
| Living room or open-plan home | Echo and clatter off hard floors and glass | 1 to 2 in, counted alongside rugs and soft furniture | The room doubles as a music or movie room |
| Home office | Sounding echoey on calls | 1 to 2 in near the desk and on the wall behind it | Rarely needed |
| Home theater or music room | Uneven, boomy low end | 4 in or more, deep air spaces, corners | Usually the starting point, not the exception |
The last row is a different problem from echo, and it needs depth measured in whole inches of fiber or air. Acoustic panels for a home theater covers it. For the other rows, the money and wall depth saved by staying at 2 in is usually better spent on more square feet of panel.
A worked example: the same wall area in three thicknesses
The fastest way to see what depth buys is to hold everything else still and change only the thickness. The example below does that with a Sabine estimate in three octave bands.
Example (fictional). A meeting room measures 24 by 16 ft with a 9 ft ceiling, so its volume is 3,456 cu ft. In the imperial Sabine formula, RT60 = 0.049 × V ÷ A, the top line is 0.049 × 3,456 = 169.3. Assume the room already has 130 sabins of absorption at 125 Hz, 120 at 250 Hz and 110 at 500 Hz. These are assumptions for the example, not measured values. Before treatment it rings for 169.3 ÷ 130 = 1.30 s at 125 Hz, 169.3 ÷ 120 = 1.41 s at 250 Hz and 169.3 ÷ 110 = 1.54 s at 500 Hz.
The plan adds eight panels, each 2 by 4 ft, for 64 sq ft. Their coefficients are taken from the Type 703 table above as stand-ins, with anything above 1.00 counted as 1.00. For simplicity the example ignores the small absorption of the drywall the panels cover.
| Panel depth | Sabins added at 125 / 250 / 500 Hz | RT60 at 125 Hz | RT60 at 250 Hz | RT60 at 500 Hz |
|---|---|---|---|---|
| None | 0 / 0 / 0 | 1.30 s | 1.41 s | 1.54 s |
| 1 in | 3.2 / 16.6 / 40.3 | 1.27 s | 1.24 s | 1.13 s |
| 2 in | 9.6 / 45.4 / 64.0 | 1.21 s | 1.02 s | 0.97 s |
| 4 in | 35.8 / 64.0 / 64.0 | 1.02 s | 0.92 s | 0.97 s |
Check one row: at 250 Hz the 2 in panels add 64 × 0.71 = 45.4 sabins, the total becomes 120 + 45.4 = 165.4, and 169.3 ÷ 165.4 = 1.02 s.
Three lessons come out of the example. Going from 1 in to 2 in shortens the decay at 250 and 500 Hz by 0.22 and 0.16 seconds. Going from 2 in to 4 in leaves 500 Hz exactly where it was, improves 250 Hz by a tenth of a second and makes its biggest change at 125 Hz.
The third lesson is about area. To match the 35.8 sabins the 4 in panels add at 125 Hz using 2 in panels, you would need 35.8 ÷ 0.15 = 239 sq ft of them, almost four times the wall area. When low frequencies matter, depth is the efficient way to get them; when they do not, area is cheaper than depth.
The real numbers for a finished panel differ from bare board data. Fabric, frame, edge details and panel size all change the result, so a room plan should cite a test of the actual construction, or of the core at the same thickness and mounting, and say which it is.
Can an air gap replace thickness?
Partly. A gap moves the panel’s fibers out from the wall, toward where the air moves more, so the panel works at lower frequencies than it would flat. A gap does not add fiber, though, and the fiber is where the sound’s energy is lost.
The idea is old and well documented. A 1980 NIOSH compendium of noise control materials puts it plainly: “For a porous type absorber the space between it and the wall will increase the absorption somewhat as the space is increased.” The same report, read here through a Wayback Machine capture because the CDC library blocks automated requests, adds: “Absorption coefficients should be based on the particular mounting method intended.”
The mounting standard says the same thing in its own words. The significance section of ASTM E795 begins: “The sound absorption of a material that covers a flat surface depends not only on the physical properties of the material but also on the way in which the material is mounted over the surface.”
The forum question: a 2 in panel on a 2 in gap, or 4 in flat
This is the comparison that keeps threads going, and the quarter-wave rule cannot settle it. In both cases the panel’s face sits 4 in from the wall, so the rule points both to about 844 Hz at the face. Measured at the middle of the panel, the spaced 2 in panel even looks better: its middle sits 3 in out, against 2 in for the flat 4 in board.
What the rule ignores is how much fiber the sound passes through. The flat 4 in board has fiber across its whole depth; the spaced panel has 2 in of fiber and 2 in of air. For the 4 in board flat, the data exists: 0.56 at 125 Hz and 1.26 at 250 Hz in the Owens Corning table. For the same board at 2 in on a 2 in gap, we did not find a neutral published test. The spaced-mount data we did find came from panel sellers, and none of it paired a flat test and a spaced test of one product.
So the honest planning rule is this. Count the gap only if the supplier can show a test at that mounting. Otherwise plan with flat-mounted data and treat whatever the gap adds as margin, not as part of the estimate.
When a gap is worth it
- Where depth is free. On ceilings, on high walls above 80 in and in spots no one walks past, a gap costs nothing in protrusion and moves absorption lower at no extra panel cost.
- Ceiling clouds. A cloud hung below a hard ceiling has a deep air space behind it by design. Its coefficients must come from a suspended test, which the next section explains how to spot.
- Reusing thin panels. If a room already has 1 in panels and a low-frequency complaint, spacing them off the wall is a cheaper first step than replacing them.
When a gap is not worth it
- On walking routes, where the 4 in limit usually leaves no room for one. The depth budget is worked through below.
- In speech rooms already at 2 in. The 2 in board is at or above 1.00 from 500 Hz up; a gap mostly adds low-frequency absorption the room’s target does not measure.
The air gap calculator shows how a given gap shifts the quarter-wave frequency. Treat its answer as the direction of change and the test report as the size of it.
How to read the mounting line in a test report
Every absorption report names a mounting, and that line decides whether its numbers apply to your wall. ASTM E795 notes that reports “may refer to these mountings by type designation instead of providing a detailed description.” Four designations cover most of what panel buyers see:
| Designation | What it means | Where its numbers apply |
|---|---|---|
| Type A (ASTM E795) | Material placed against a solid backing, with no air space | Panels fixed flat to a wall or ceiling |
| Type A (ISO 354) | Specimen directly against a room surface, held with adhesive or fasteners, no air space | The same, for reports tested to the international standard |
| Type C-25 | In one Riverbank test, slats over 25 mm (1 in) wood furring, with the cavity filled with PET felt | That wall build-up, cavity and infill included |
| Type E-400 | Specimen spaced so its exposed face is 400 mm (about 15.7 in) from the test surface | Suspended ceilings and similar deep mountings, never a panel flat on a wall |
The Type E definition is worth knowing precisely. In Riverbank report RAL-A22-451, the lab explains that the numeral suffix is defined in ASTM E795-16 “as the distance in millimeters from the exposed face of the test specimen to the test surface, rounded to the nearest integer multiple of 5.” The suffix counts the specimen’s own thickness, not only the air.
The Type C-25 row comes from Riverbank report RAL-A23-086, a 2023 test of a wood slat wall over a filled cavity, run to ASTM C423-22 with mounting per E795-23. The assembly measured an NRC of 0.90 and an SAA of 0.88, with 0.29 at 125 Hz and 0.55 at 250 Hz. Those numbers belong to the whole build-up: slats, felt backing, furring and the PET-filled cavity. Leave out the cavity infill and you have a different assembly, so the report no longer applies.
E795 also warns that some mountings use “special fixtures or minor deviations from normal practice” that are “for laboratory testing only.” The fixture exists to reproduce a real installation in the lab; it is not a mounting detail to copy.
Test size is part of the mounting too
Labs test a large field of material. Intertek’s summary says ASTM C423 recommends a specimen 8 ft wide by 9 ft long for products that cover a plane surface, and the SRL report used about 10.8 square meters laid on the floor, at least 1 meter from any edge. A handful of 2 by 4 ft panels spread over a wall is a different situation, with more exposed edges per square foot.
ASTM C423 itself says its coefficients “should be used with caution” because real areas are usually larger than the test specimen and real rooms rarely have a diffuse sound field. That caution applies to every thickness equally. How spacing and edges change a panel group’s effect belongs to acoustic panel placement.
Five things to check on any report
- The method and edition, for example ASTM C423-22 or ISO 354:2003.
- The mounting designation, and whether it matches how your panel will hang.
- The specimen: product, thickness, density, facing and size.
- The lab, report number and date.
- The full band table from 125 to 4,000 Hz, not only NRC or SAA.
The 4 inch limit on walking routes
Depth has a hard ceiling wherever people walk past at body height. Under section 307.2 of the 2010 ADA Standards, objects with leading edges more than 27 in and not more than 80 in above the floor may protrude no more than 4 in into a circulation path.
The Access Board’s guide to protruding objects explains why: 27 in is “the maximum specified for cane detection,” so objects lower than that can be found with a cane, and above 80 in they clear headroom. The guide also says these requirements “apply to all circulation paths and are not limited to accessible routes,” and that they apply “equally to circulation paths in rooms and spaces,” not only to corridors. A dining room aisle or an office walkway counts.
For panels, the hardware and any gap count toward the 4 in, not only the panel. That turns the thickness question into a depth budget.
Read the figure as a set of rules for walls along a walking route:
- 1 in and 2 in panels fit comfortably. A 2 in panel on a slim cleat projects about 2.5 in, leaving room for hardware tolerance and a thicker fabric wrap.
- A 2 in panel on a 2 in gap uses the whole budget. Any cleat depth, frame lip or out-of-plumb wall pushes it over.
- A 4 in panel does not fit at body height. Even with no hardware at all it reaches the limit exactly.
Deep panels and spaced mountings still have places to go. Hang them above 80 in, use the ceiling, or put them where no circulation path runs, such as behind fixed seating. A panel whose bottom edge sits at or below 27 in is within cane range, and the guide allows such objects to protrude any amount, though a low, deep panel in a dining room aisle invites knocks and kicks.
Overhead, the rule turns around. The same guide states: “Headroom clearance of at least 80″ high is required along all circulation paths.” A ceiling cloud with a deep air space above it is measured from its lowest point, hardware included. Depth spent above the cloud costs nothing; depth spent below it eats into headroom.
Run every wall panel along a route through the ADA protrusion checker before ordering. It applies sections 307.2 and 307.4 and is not legal advice; the local building official decides how the rules apply to your project.
Weight, density and building depth into the wall
Thicker panels weigh more, and heavier is not the same as better. The Owens Corning sheet gives nominal densities of 3.0 lb per cubic foot for Type 703 and 6.0 for Type 705. A 2 by 4 ft board 1 in thick holds 0.67 cubic feet of material, so the core alone weighs about 2.0 lb in 703 and 4.0 lb in 705.
The same arithmetic gives about 4.0 and 8.0 lb at 2 in, and 8.0 and 16.0 lb at 4 in. Fabric, frame, edge hardening and fixings come on top. Heavier panels call for sturdier fixings, and on a ceiling a heavier cloud may need more hanging points; the hardware is chosen for the panel as built.
Density does not buy absorption in a straight line either. At 1 in, the denser 705 reads higher than 703 at 500 and 1,000 Hz (0.73 and 1.06 against 0.63 and 0.88). At 4 in the order reverses at 250 Hz: 1.10 for 705 against 1.26 for 703. The core that suits a room depends on what it must absorb and how it must wear, which choosing a panel core covers.
Facings matter as much as density. On the same data sheet, 1 in Type 703 with a foil and kraft facing reads only 0.21 at 4,000 Hz, against 0.94 unfaced. That is why panel fabric has to be acoustically open: a face that reflects sound throws away the absorption a thick core would provide.
When the room needs depth and the route has none
Designers have a third option besides thin panels and high panels: build the depth into the wall. A furred-out wall section or a niche can hold a full-depth absorber whose face sits flush, or nearly flush, with the surrounding wall, so it projects little or nothing into the aisle. The Type C-25 slat wall above is a tested example of the idea: a 1 in cavity filled with fiber behind an open face, tested as one assembly.
If you take that route, the absorption data has to come from the assembly as built, and the wall detail belongs on the drawings early. Custom fabrication covers made-to-size panels for recesses, and fabric-wrapped wall panels describes how cores, fabric and mountings come together.
Common claims about thickness, checked against the data
Buying guides and forum threads repeat a handful of claims about thickness. Each one below is checked against the lab tables quoted earlier in this article.
| Claim | What the data says |
|---|---|
| “Thicker is always better.” | Only at low frequencies. From 1,000 Hz up, the 1, 2, 3 and 4 in Type 703 boards all read 0.88 or more. |
| “A 1 in panel does nothing.” | It absorbs most sound from 1,000 Hz up (0.88 to 0.97) and 0.63 at 500 Hz. It is weak at 250 Hz and below. |
| “An NRC of 1.15 means it handles bass.” | NRC starts at 250 Hz. Two boards with the same 1.15 rating read 0.34 and 0.56 at 125 Hz. |
| “A 2 in panel on a 2 in gap equals a 4 in panel.” | The quarter-wave rule cannot tell them apart, and no neutral published test we found compares them on one product. Plan with tested data for the mounting you will use. |
| “Denser cores absorb more.” | Not reliably. At 4 in, the 6.0 pcf board reads 1.10 at 250 Hz and the 3.0 pcf board reads 1.26. |
| “Thick enough panels will soundproof a wall.” | No. Absorption tests (ASTM C423) and sound transmission tests are separate families; panels do not soundproof a wall at any thickness. |
| “Coefficients over 1.00 are marketing.” | They come from diffraction at specimen edges, which ASTM C423 and the SRL report both describe. Count them as 1.00 when planning. |
Two habits keep these claims from steering a purchase. First, ask which mounting any quoted number comes from, because ASTM E795 makes the point that one material can absorb differently depending on how it is mounted. Second, ask for the band, not the rating: “what does it absorb at 125 and 250 Hz, flat on the wall?” is a question a supplier with real test data can answer in one line.
The claims also tend to come from studio and home-theater contexts, where low frequencies are the point. In a dining room or a meeting room, the same advice can push buyers toward panels that are deeper, heavier and harder to place than the room needs, while leaving less budget for the square footage that actually shortens the decay at 500 to 2,000 Hz.
A checklist before you choose a depth
Work through these in order. Most rooms stop at step four.
- Name the complaint in plain words. Voices, echo and clatter point to the middle and upper bands. Boom, hum and muddy bass point below 250 Hz.
- Find out which bands your target is measured in. If it is a speech-room target set at 500 to 2,000 Hz, a 2 in core usually covers it.
- Ask for the full band table from 125 to 4,000 Hz for the core or panel at the thickness you are considering, not only the NRC.
- Check the mounting line. Type A data supports a panel fixed flat. A spaced panel needs data from a spaced mounting, or its gap counts only as margin.
- Run the depth budget for every panel on a walking route between 27 and 80 in, hardware and gap included.
- Compare sabins per square foot of wall, band by band, before choosing a thicker panel over more panels of the same depth.
- Decide where depth is free: ceilings, high walls, corners and spots no one walks past are where 4 in cores and air gaps earn their keep.
Fibson designs and makes acoustic panels to order and plans each room before anything is ordered. Each room plan names the thickness and mounting of every panel group and the published test report behind each coefficient, with its method, edition and mounting type; Fibson has not yet tested its finished panels. If you would like those decisions worked out for a specific room, send us the room with its size, photos and what bothers you about it.
