Glass absorption coefficient
Ordinary window glass absorbs α 0.35 at 125 Hz and almost nothing above it — NRC 0.15. That first number is larger than 50 mm mineral wool (0.15), larger than 50 mm acoustic foam (0.08), and larger than a heavy curtain (0.14). The window you were planning to cover is one of the better bass absorbers in an untreated room — and one of the worst surfaces you own everywhere else. Coefficients come straight from our sourced absorption database. We don't sell panels or glazing.
The short version
- Glass runs backwards. α 0.35 at 125 Hz falling to 0.04 at 4 kHz — every porous absorber does the opposite.
- It beats thin treatment in the bass. At 125 Hz it out-absorbs 5 of the 5 common thin treatments below.
- Thicker glass absorbs less. Heavy plate is 0.18 against 0.35 for an ordinary pane — mass moves the resonance out of the band.
- NRC hides all of it. NRC 0.15 averages 250–2000 Hz and never looks at the one band where glass works.
- Sealed double glazing has a hole in the speech range near 245 Hz — that is an isolation figure, and it is computed below.
Glass and hard-surface absorption coefficients by octave band
| Material | 125 Hz | 250 Hz | 500 Hz | 1k Hz | 2k Hz | 4k Hz | NRC |
|---|---|---|---|---|---|---|---|
| Ordinary window glass | 0.35 | 0.25 | 0.18 | 0.12 | 0.07 | 0.04 | 0.15 |
| Heavy plate glass | 0.18 | 0.06 | 0.04 | 0.03 | 0.02 | 0.02 | 0.05 |
| Gypsum board (12 mm on studs) | 0.29 | 0.10 | 0.05 | 0.04 | 0.07 | 0.09 | 0.05 |
| Brick, unglazed | 0.03 | 0.03 | 0.03 | 0.04 | 0.05 | 0.07 | 0.05 |
| Marble or glazed tile | 0.01 | 0.01 | 0.01 | 0.01 | 0.02 | 0.02 | 0.00 |
Random-incidence values measured per ISO 354 / ASTM C423, rendered from the same rows that feed the database and its CSV and JSON downloads. Planning estimates — check a manufacturer's datasheet for a specific unit.
Glass absorbs bass, not treble — the opposite of foam
This is the whole point of the page. Read the 125 Hz column against the 2k Hz column for the same materials:
| Material | 125 Hz | 2 kHz | Overtakes glass at |
|---|---|---|---|
| Ordinary window glass (this page) | 0.35 | 0.07 | — |
| Mineral wool, 50 mm, on wall | 0.15 | 1.00 | 250 Hz |
| Open-cell acoustic foam, 50 mm | 0.08 | 0.95 | 250 Hz (tie) |
| Fabric-wrapped panel, 50 mm | 0.16 | 1.00 | 250 Hz |
| Carpet, heavy, on foam pad | 0.08 | 0.71 | 500 Hz |
| Curtains, heavy, deep folds | 0.14 | 0.70 | 250 Hz |
Every one of those materials is bought to absorb sound, and at 125 Hz 5 of the 5 are beaten by a pane of ordinary window glass. The crossover column is computed by comparing the two rows band by band, not asserted: each thin absorber overtakes glass somewhere between 250 Hz and 500 Hz and then leaves it far behind — 50 mm mineral wool reaches α 1.00 at 1 kHz where glass is at 0.12.
So the two are not competing. Glass contributes a little absorption in the one band that thin porous panels cannot reach, and nothing in the range those panels are actually for.
Why: panel resonance, not porosity
Glass has no open pores, so it cannot absorb the way wool or foam does — there is nowhere for air to move and lose energy to friction. What a window does instead is flex. A pane held in a frame is a limp panel with air behind it, and low-frequency sound drives it into motion; some of that energy is lost in the pane and the frame instead of being reflected back into the room. It is the same mechanism that makes gypsum board on studs the other unexpected bass absorber in a small room — glass is at α 0.35 at 125 Hz where drywall is at 0.29.
Because the mechanism is resonance rather than porosity, adding mass makes it worse, which is the counter-intuitive result in the table above. Heavy plate glass is α 0.18 at 125 Hz against 0.35 for an ordinary pane: a heavier panel resonates lower and more reluctantly, so less of the 125 Hz band is absorbed. The thicker window is the better barrier and the poorer absorber.
NRC 0.15 hides all of it
NRC averages the 250, 500, 1000 and 2000 Hz coefficients and rounds to the nearest 0.05. Glass's only meaningful absorption sits at 125 Hz — the one band NRC does not look at — so the rating comes out at 0.15, and heavy plate glass at 0.05, the same as unglazed brick (0.05). By NRC alone a window and a brick wall are indistinguishable. They are not: at 125 Hz the window absorbs 0.35 and the brick 0.03. If you are comparing hard surfaces, read the 125 Hz coefficient and ignore the NRC — the reason why is on the NRC page.
What one real window is worth, in sabins
Coefficients are ratios; what a room responds to is absorption area, A = S × α, measured in sabins. Here is a single 2 m² window against the same 2 m² covered in each thin treatment:
| 2 m² of… | 125 Hz | 250 Hz | 500 Hz | 1k Hz | 2k Hz | 4k Hz |
|---|---|---|---|---|---|---|
| Ordinary window glass | 0.70 | 0.50 | 0.36 | 0.24 | 0.14 | 0.08 |
| Mineral wool, 50 mm, on wall | 0.30 | 1.10 | 1.80 | 2.00 | 2.00 | 2.00 |
| Open-cell acoustic foam, 50 mm | 0.16 | 0.50 | 1.20 | 1.80 | 1.90 | 1.96 |
| Fabric-wrapped panel, 50 mm | 0.32 | 1.10 | 1.90 | 2.00 | 2.00 | 1.96 |
| Carpet, heavy, on foam pad | 0.16 | 0.48 | 1.14 | 1.38 | 1.42 | 1.46 |
| Curtains, heavy, deep folds | 0.28 | 0.70 | 1.10 | 1.44 | 1.40 | 1.30 |
Sabins, computed from the coefficients above at 2 m². The window supplies 0.70 sabins at 125 Hz — more than the same area of 50 mm foam (0.16) or 50 mm wool (0.30) — and 0.14 at 2 kHz where the foam supplies 1.90. Feed either figure into the RT60 calculator as a surface and the per-band effect on reverberation time follows directly.
Absorption is not isolation — and double glazing is where that bites
Everything above is about sound inside the room. The question most people actually arrive with is about sound coming in, and glazing has a well-known failure mode there that is worth computing rather than asserting.
Two panes with a sealed air gap behave as a mass–air–mass system: the two glass leaves are the masses, the trapped air is the spring. At its resonant frequency the assembly transmits sound more readily than a single pane of the same total weight. The standard approximation is:
f₀ ≈ 60 √( (m₁ + m₂) ÷ (m₁ · m₂ · d) )
with surface masses m in kg/m² and cavity depth d in metres. Taking soda-lime glass at a nominal 2500 kg/m³, a 4 mm pane is 10.00 kg/m² and a 6 mm pane 15.00 kg/m²:
| Build-up | Cavity | m₁ + m₂ (kg/m²) | f₀ |
|---|---|---|---|
| Sealed double glazing, 4 mm + 4 mm | 12 mm | 20.00 | 245 Hz |
| Sealed double glazing, 4 mm + 6 mm | 16 mm | 25.00 | 194 Hz |
| Secondary glazing, 4 mm + 4 mm | 100 mm | 20.00 | 85 Hz |
| Secondary glazing, 4 mm + 6 mm | 150 mm | 25.00 | 63 Hz |
| Secondary glazing, 6 mm + 6 mm | 200 mm | 30.00 | 49 Hz |
Computed from the formula above, not from a datasheet. The pattern is the useful part: sealed thermal double glazing puts its weak point near 245 Hz, in the middle of the speech range, which is why a well-insulated modern window can still let a conversation or a television through clearly. What moves f₀ down is the gap, not the glass — secondary glazing with a 200 mm cavity reaches about 49 Hz, below most of what you are trying to keep out. A sealed unit optimised for heat and a secondary pane optimised for noise are different products, and the arithmetic above is the reason.
Assumptions stated so they can be checked: sealed cavity, limp panels, no absorption in the reveal, nominal glass density 2500 kg/m³. Real units vary with laminated interlayers, gas fill and frame construction, all of which shift f₀ and add damping. Sources: Bies & Hansen, Engineering Noise Control; Cox & D'Antonio, Acoustic Absorbers and Diffusers.
What this means for treating your room
- Don't cover the window for bass. You would be removing 0.70 sabins of 125 Hz absorption to add 0.30 back.
- Do treat it for echo. Above 500 Hz glass is nearly as reflective as marble tile — heavy curtains are the standard answer, and they leave the pane free to keep flexing.
- Count the glass in your surface budget. A window wall is often 10–20% of a small room's surface area, with a curve unlike anything else in the room.
- Don't read the NRC. At NRC 0.15 glass and brick look identical and behave differently where it matters.
- Separate the two problems. Reverberation inside is an absorption question; traffic noise outside is an isolation one, and the f₀ table is the isolation half.
Plan the rest of your room
Windows are a fixed constraint rather than a lever — the useful move is to account for them and treat what you can reach. Size that treatment with the acoustic panel calculator, then check it against a reverberation target with the RT60 calculator, which lets you enter the glass as its own surface and see the per-band effect of a large window directly. If the complaint is boom rather than echo, find the problem frequencies first with the room mode calculator.
Frequently asked questions
What is the absorption coefficient of glass?
For ordinary window glass the sourced octave-band coefficients are α 0.35 / 0.25 / 0.18 / 0.12 / 0.07 / 0.04 at 125 / 250 / 500 / 1000 / 2000 / 4000 Hz, giving an NRC of 0.15. The curve runs the opposite way to every porous absorber: highest in the bass and falling steadily with frequency. Heavy plate glass is lower again at 0.18 at 125 Hz, NRC 0.05.
Does glass absorb sound?
Only bass, and only because the pane flexes. At 125 Hz ordinary window glass is α 0.35 — higher than every one of the 5 thin treatments compared on this page, including 50 mm mineral wool (0.15) and 50 mm acoustic foam (0.08). Above 250 Hz it absorbs almost nothing (α 0.18 at 500 Hz, 0.07 at 2 kHz) and behaves as the hard reflector everyone assumes it is.
Why does thicker glass absorb less bass than thin glass?
Because the absorption comes from the pane resonating, and resonance is governed by mass. Ordinary window glass is α 0.35 at 125 Hz; heavy plate glass is 0.18 — roughly half. Adding mass lowers and damps the panel's resonant frequency, moving it away from the 125 Hz measurement band, so the heavier pane scores lower even though it is the better sound barrier. Absorption and isolation pull in opposite directions here, exactly as they do for gypsum board.
Does double glazing block sound?
Less than most people expect, and there is a computable reason. Two panes separated by a sealed cavity form a mass-air-mass resonator with a frequency f₀ ≈ 60 √((m₁+m₂)/(m₁m₂d)). For typical sealed double glazing (4 mm + 4 mm, 12 mm cavity) that lands near 245 Hz — inside the speech range, which is where the unit is at its worst at keeping noise out. Widening the gap is what fixes it: secondary glazing with a 150–200 mm cavity pushes f₀ down to roughly 49 Hz, below most of the noise you are trying to exclude. Thermal double glazing is optimised for heat, not for sound.
Why does a room with big windows sound echoey?
Because glass is reflective everywhere above the bass. At 500 Hz to 4 kHz — where speech intelligibility and flutter echo live — window glass runs α 0.18 down to 0.04, close to marble tile (0.01 at 500 Hz). A large window is effectively a hard wall for the whole range you hear as "echo", while quietly helping in a band you were not complaining about.
Should I put acoustic panels over my window?
Rarely, and not for the reason usually given. Covering a 2 m² window with 50 mm mineral wool would add 1.80 sabins at 500 Hz where the glass gives 0.36 — a real gain — but it also removes 0.70 sabins of 125 Hz absorption and replaces it with only 0.30. You would be trading away bass absorption you already own for mid-band absorption you could have obtained on any other wall. Treat the reflective surfaces you can reach first, and use heavy curtains if the window itself is the problem.
Related material guides
- Acoustic foam absorption coefficient — NRC by thickness and foam type — and why it barely touches bass.
- Mineral wool absorption coefficient — Thickness and an air gap beat density — the panel-builder’s material.
- Carpet absorption coefficient — Tames echo, not bass — and why it is not soundproofing.
- Curtains absorption coefficient — Fullness and an air gap are the levers — “soundproof” curtains do not block noise.
- Gypsum board (drywall) absorption coefficient — The one surface in the room that absorbs bass rather than treble — and NRC 0.05 hides it.
- Cork absorption coefficient — Rated the same as bare brick — sold as treatment, measured as a hard surface.
- Concrete absorption coefficient — Four different materials share the name — and painting the block costs you most of it.
Sources for the coefficient rows on this page: Standard architectural-acoustics coefficient tables (ISO 354 / ASTM C423 measured).
What the numbers mean: what is the sound absorption coefficient? · New here? Start with Acoustic Treatment 101 · Browse all material absorption data · See our methodology & sources.