A closed window cut road traffic noise by a median 27.8 dB(A) in 102 Swiss homes, but tilting it open dropped that to 15.8 dB(A) and opening it fully to 10.0 dB(A) (Locher et al., International Journal of Environmental Research and Public Health 2018). The interlayer matters more than most specifiers assume: laminating one pane of vacuum glazing improved insulation by 8 to 9 dB at its coincidence dip (Nugroho, Cabrera and Kocer, J. Phys. Conf. Ser. 2018). In the lab, five premium windows tested at NASA Langley all landed between OITC 28 and 30, despite ratings of OITC 28 to 35 (PARTNER Project 1.5 Report 2008). We aggregated data from the National Research Council Canada, the FAA and the PARTNER Center of Excellence, Pilkington technical datasheets, the European Environment Agency, three US federal regulations (14 CFR 150, 24 CFR 51 and 23 CFR 772), and peer-reviewed studies in Building and Environment, Sustainability and IJERPH. For the wall side of the envelope, see our roundup on why STC-rated partitions underperform in the field, and for the exposure side, our data on aircraft noise around airport communities.
TL;DR
- Closed windows cut road noise indoors by a median 27.8 dB(A), tilted windows 15.8 dB(A), open windows 10.0 dB(A) (Locher et al., IJERPH 2018).
- Two 3 mm lites with a 12 mm air space rate STC 26 to 30 but only OITC 22 to 26 (FGIA, Designing for quiet 2026).
- Each doubling of the air space adds about 3 STC points to sealed double glazing (NRC Canada, CBD-240 1988).
- A 12.8 mm acoustic laminated pane rates Rw 40 against Rw 34 for 12 mm float glass (Pilkington Optiphon Datasheet 2020; Pilkington Noise Control with Glass 2019).
- Laminating one pane of vacuum glazing raised Rw from 35 to 38 and added 8 to 9 dB at the coincidence dip (Nugroho et al., J. Phys. Conf. Ser. 2018).
- Measured window STC came in up to 4 dB below the manufacturer rating (PARTNER Project 1.5 Report 2008).
- An integrated roller shutter box capped a window at Rw 34, against Rw 38 to 43 without one (Caballol Bartolome et al., Building and Environment 2026).
- A PVC window lost 3 dB of traffic-noise insulation at -20 C versus standard conditions (Konstantinov et al., IOP Conf. Ser. 2020).
- Standard US homes deliver about 20 dB of noise level reduction, but 25 to 30 dB is needed at DNL 65 to 75 (14 CFR Part 150, Appendix A).
- About 92 million Europeans are exposed to road traffic noise at or above 55 dB Lden (EEA, Environmental noise in Europe 2025).
1. Ratings: STC, OITC and Rw Measure Different Noise
The single number on a window spec sheet depends on which noise it was weighted for. STC was built to rate speech between rooms, so it ignores the low bands where trucks, trains and jet engines carry their energy. OITC reaches down to 80 Hz, and the European Rw is usually quoted with spectrum terms C and Ctr that subtract points for bass-heavy sources. The same everyday double-glazed unit drops 4 points when it moves from STC to OITC: STC 26 to 30 becomes OITC 22 to 26 (26 - 22 = 4, 30 - 26 = 4) (FGIA, Designing for quiet 2026).
The practical lesson: an STC number is the wrong figure to compare against a facade requirement written for traffic or aircraft. A specifier comparing two products needs the same metric, measured over the same frequency range, for both.
| Metric | Value | Source |
|---|---|---|
| Frequency range used by STC / Rw / OITC | 125 to 4,000 Hz / 100 to 3,150 Hz / 80 to 4,000 Hz | FGIA, Designing for quiet (Construction Specifier 2026), citing AAMA TIR-A1, ASTM E413 and ASTM E1332 |
| Two 3 mm lites, 12 mm air space | STC 26 to 30, OITC 22 to 26 | FGIA, Designing for quiet 2026 |
| 8 mm and 6 mm lites, 12 mm air space | STC 32 to 40, OITC 29 to 34 | FGIA, Designing for quiet 2026 |
| Same vacuum glazing sample (3V3) rated two ways | Rw 35 (C -1; Ctr -3) but STC 31 | Nugroho, Cabrera and Kocer, J. Phys. Conf. Ser. 2018 |
| Ctr change when the rating range is extended to 50-5,000 Hz | Falls 1 to 2 dB in most glazings tested | Caballol Bartolome et al., Building and Environment 2026 |
Context note: OITC itself is described as unsuitable where sound energy is dominated by frequencies below 63 Hz. The explainer in Construction Specifier attributes its typical values to AAMA TIR-A1 rather than to individually cited tests, and quotes the OITC upper bound as both 4,000 Hz and 5,000 Hz in different passages.
2. Single, Double and Triple Glazing: More Panes, Not Always More Silence
Adding glass mass helps, but with diminishing returns once the panes are locked together by a thin air space. A third pane mostly buys thermal performance: it splits the cavity into two narrow gaps, each with its own resonance, instead of creating one deep cavity. Two Pilkington units, one double and one triple, built around the same 8.8 mm acoustic laminate both rate Rw 42 (C -3; Ctr -8) (Pilkington Optiphon Datasheet 2020).
The table shows why: what a third pane adds is mostly extra weight, and that only pays off below the resonance. Where the budget allows, a thicker or laminated pane in a double unit usually moves the rating more than a third lite does.
| Metric | Value | Source |
|---|---|---|
| Single float pane, 4 mm vs 12 mm | Rw 29 vs Rw 34 | Pilkington, Noise Control with Glass (GL090) 2019 |
| Double glazing 4(12)4 vs 4(12)8 in a PVC window | Rw 38 vs Rw 43 | Caballol Bartolome et al., Building and Environment 2026 |
| Triple vs double glazing | About 3 dB better below resonance (about 250 Hz), consistent with a 50% mass increase; STC essentially the same with typical air spaces of 15 mm or less | NRC Canada, Canadian Building Digest CBD-240 (Quirt) 1988 |
| 8-16-Optiphon 8.8 double vs 6-12-4-12-Optiphon 8.8 triple | Both Rw 42 (C -3; Ctr -8) | Pilkington Optiphon Datasheet (GL103) 2020 |
| Triple-pane picture window, rated vs measured | Rated STC 47 / OITC 35; measured STC 46 / OITC 29 | PARTNER Project 1.5 Report 2008 |
Outlier note: the triple-pane picture window above kept its STC almost intact but lost 6 OITC points against its rating (35 - 29 = 6), the largest OITC gap of the five windows tested. Most recent available data for the NRC series: NRC Canada, CBD-240, 1988, and the underlying physics has not changed.
3. Air Gaps, Laminated Interlayers and Gas Fills
Of the three levers a glazing designer has, the interlayer gives the most decibels per millimetre. A 12.8 mm acoustic laminated pane rates Rw 40 against Rw 34 for 12 mm monolithic float glass, a 6 dB gain (40 - 34 = 6) for 0.8 mm of extra thickness (Pilkington Optiphon Datasheet 2020; Pilkington Noise Control with Glass 2019). The air gap works too, but the returns come in roughly 3-point steps per doubling, which is why conventional sealed-unit cavities cannot get far on spacing alone.
Gas fills are mostly a thermal decision. Argon makes no practical acoustic difference, and krypton adds at most about 1 dB. SF6, dense enough to help acoustically, is banned in large parts of Europe because of its climate impact. For noise-control materials that work beyond glass, see our roundup on acoustic metamaterials and soundproofing.
| Metric | Value | Source |
|---|---|---|
| STC gain per doubling of sealed-unit air space | About 3 points | NRC Canada, CBD-240 1988 |
| Krypton / argon fill | Up to 1 dB / no practical improvement | Pilkington, Noise Control with Glass 2019 |
| 4(12)4 vs 4(12)3+3 vs 4(12)4+4 laminated, same window | Rw 38 vs Rw 41 vs Rw 43 | Caballol Bartolome et al., Building and Environment 2026 |
| Vacuum glazing, monolithic vs one laminated pane | Rw 35 vs Rw 38; 8 to 9 dB better at the coincidence dip | Nugroho, Cabrera and Kocer, J. Phys. Conf. Ser. 2018 |
| 12.8 mm acoustic laminated pane vs 12 mm float | Rw 40 (C -1; Ctr -3) vs Rw 34 | Pilkington Optiphon Datasheet 2020; Pilkington GL090 2019 |
Context note: the laminated vacuum glazing in the Nugroho study jumped from STC 31 to STC 38 while Rw rose only 3 dB, a reminder that the two ratings can move very differently for the same physical change. The full measurement set is in the open-access conference paper. The manufacturer values in this table are measured to EN 717-1, with 90% argon fill in the insulating units.
4. Resonance and Coincidence Dips: Where Glazing Loses Its Rating
Every glazing unit has two weak frequencies. At the mass-air-mass resonance, the two panes and the air between them bounce as a spring system and insulation collapses. At the coincidence frequency, bending waves in the glass match the incoming sound and the pane becomes nearly transparent to it. The PARTNER windows sold as ultra-high-performance had up to 7 dB less transmission loss below 125 Hz than the cheaper high-performance windows (PARTNER Project 1.5 Report 2008), which is exactly the band aircraft and heavy trucks occupy.
Thicker glass moves the coincidence dip down into the range where speech and traffic live, so a heavier pane can lose at the very frequency a designer hoped it would help. The low-frequency problem is also a measurement problem: whether two windows differ below 100 Hz depends on which standardized method is used, as the last row of the table shows.
| Metric | Value | Source |
|---|---|---|
| Coincidence frequency rule of thumb | About 12,000 Hz divided by thickness in mm: about 3,000 Hz (4 mm), 1,500 Hz (8 mm), 1,000 Hz (12 mm) | Pilkington, Noise Control with Glass 2019 |
| Coincidence dip, 2 mm vs 18 mm glass | About 5 kHz vs the 500 Hz band (18 mm glass has lower TL than 4 mm near 500 Hz) | NRC Canada, CBD-240 1988 |
| Mass-air-mass resonance of typical factory-sealed units | 200 to 400 Hz | NRC Canada, CBD-240 1988 |
| Estimated mass-air-mass resonances of five tested windows | 85 to 208 Hz | PARTNER Project 1.5 Report 2008 |
| Vacuum glazing with dissimilar panes (3V5) vs equal panes (3V3) | Both Rw 35; dip depth not significantly reduced | Nugroho, Cabrera and Kocer, J. Phys. Conf. Ser. 2018 |
| Effect of window type below 100 Hz, lab method vs in situ (receiving room only) | F(1,28) = 4.35, p = 0.046 vs F(1,28) = 0.06, p = 0.81 | Caballol Bartolome et al., Building and Environment 2026 |
Data conflict: the two glazing references disagree on how asymmetric panes should be. NRC CBD-240 names roughly a 2:1 thickness ratio as most effective, while Pilkington recommends panes that differ by about 30%. The Building and Environment 2026 paper tested two PVC windows of 1.83 m2 across five glazings from 50 to 5,000 Hz.
5. Lab Ratings vs Installed Windows and Curtain Walls
A glazing rating describes a sealed unit in a test opening. A building gets a frame, gaskets, operable sashes, shutter boxes, mullions and weather. Adding an integrated roller shutter box capped the same PVC window at Rw 34, against Rw 38 to 43 without one (Caballol Bartolome et al., Building and Environment 2026), wiping out the entire benefit of upgrading the glass. In the PARTNER tests, measured STC fell up to 4 dB short of the manufacturer rating, and the report notes that a published rating may belong to a similar model with different glass or geometry.
The good news from a 2025 laboratory study: with careful mounting, the frame costs only 1 to 2 dB against the glass alone, and the installed windows beat the conservative estimates of EN 14351-1. The installation method, not the glass, is where field performance is won or lost.
| Metric | Value | Source |
|---|---|---|
| Measured vs rated STC, five windows | Up to 4 dB lower (Window A: rated STC 45, measured 41) | PARTNER Project 1.5 Report 2008 |
| IGU alone vs installed window (IGU 1) | Glass Rw 46; window Rw 45 with expansion tape, Rw 44 with foam | Sustainability 2025, Effect of Window Structure and Mounting on Sound Insulation |
| Installed window vs EN 14351-1 conservative estimate (IGU 1) | Measured Rw 44 to 45 vs estimate Rw 42 | Sustainability 2025 |
| PVC window (4-16-4-14-4 IGU) at -20 C vs standard conditions | 3 dB lower (R_Atran 22 dB vs 25 dB) | Konstantinov, Verkhovsky and Lyabakova, IOP Conf. Ser. 2020 |
| Typical larger windows, curtain wall and storefronts | OITC 25 or higher, STC 29 or higher | FGIA, Designing for quiet 2026 |
Curtain wall note: in curtain-walled offices the facade also becomes a flanking path between rooms. A 2025 study of five Italian office buildings with floor-to-ceiling glazing found glass-fin partitions reaching at most DnT,w 44 dB, short of the 45 dB UNI 11367 baseline, and operable partitions at only 33 and 35 dB (Forum Acusticum / Euronoise 2025); the certification side is covered in our roundup on LEED and WELL acoustic compliance. A simulation of a curtain wall with an 8 mm / 16 mm / 8 mm acoustic laminated glazing found the combined result tracked the airborne path through the mullion closely (Caniato, Bettarello and Gasparella, IBPSA Building Simulation 2019). The cold-weather figure comes from a climate-chamber test in the IOP Conference Series paper, which attributes the loss to seal tightness and glass deflection.
6. What Real Facades Deliver in Occupied Homes
Field data from lived-in homes is the reality check on every rating above. Across 102 Swiss homes exposed to road traffic, closing the window was worth a median 27.8 ± 4.4 dB(A), against 15.8 ± 2.7 dB(A) tilted and 10.0 ± 2.9 dB(A) fully open (Locher et al., IJERPH 2018). The jump from tilted to closed is about 12 dB (27.8 - 15.8 = 12.0), more than most glazing upgrades in section 3 can deliver on their own.
Small details move the result: missing gaskets cost more than the choice of frame material helped. The study also found that the Swiss federal default of 25 dB(A) for closed windows was conservative compared with what homes actually achieved.
| Metric | Value | Source |
|---|---|---|
| Outdoor-indoor difference, open / tilted / closed (median ± SD) | 10.0 ± 2.9 / 15.8 ± 2.7 / 27.8 ± 4.4 dB(A) | Locher et al., IJERPH 2018 |
| Vinyl or metal frames vs wooden frames, closed window | +1.9 dB(A) | Locher et al., IJERPH 2018 |
| No window gaskets | -2.4 dB(A) | Locher et al., IJERPH 2018 |
| Each additional window in the room | -0.8 dB(A) | Locher et al., IJERPH 2018 |
| Noise level reduction of standard US residential construction | About 20 dB | 14 CFR Part 150, Appendix A, Table 1, Note 1 |
Most recent available data: Locher et al., 2018, still the most detailed open dataset on window position in occupied homes. Its median closed-window value sits inside the 25 to 30 dB band that US airport rules ask for (section 7), which suggests a well-sealed modern window is close to the requirement before any specialist glazing is added. The comparison is indicative only: a road-traffic dB(A) difference and an aircraft-noise NLR are not measured the same way.
7. Facade Insulation Requirements Near Airports and Highways
Regulators rarely specify a window rating. They specify an indoor outcome or an outdoor-to-indoor reduction, and leave the facade to deliver it. Where communities allow homes inside DNL 65 to 75 dB contours, 14 CFR Part 150 calls for 25 to 30 dB of noise level reduction, 5 to 10 dB more than the roughly 20 dB a standard house provides (14 CFR Part 150, Appendix A). The FAA’s funding rule works backward from a 45 dB interior target, tightened from 50 dB in the original 1986 AIP Handbook.
In Europe the scale of the problem is the headline: road traffic alone puts tens of millions of people above the reporting threshold, and the WHO guideline is stricter still. That gap is the market for acoustic facades in dense cities.
| Metric | Value | Source |
|---|---|---|
| FAA interior design target for AIP-funded insulation | 45 dB DNL in habitable rooms (50 dB in the 1986 AIP Handbook) | FAA Program Guidance Letter 12-09 (2012) |
| Required NLR where homes are allowed at DNL 65 to 70 / 70 to 75 | 25 dB / 30 dB | 14 CFR Part 150, Appendix A |
| Part 150 insulation eligibility cut-off | Noncompatible development existing as of October 1, 1998 | FAA Program Guidance Letter 12-09 (2012) |
| HUD interior / exterior noise goals | 45 dB DNL / 55 dB DNL | HUD, 24 CFR 51.101 |
| HUD extra attenuation, DNL above 65 to 70 / above 70 to 75 | 5 dB / 10 dB (above 75 dB is Unacceptable) | HUD, 24 CFR 51.103 and 51.104 |
| FHWA Noise Abatement Criteria, residential exterior / sensitive interior | 67 dB(A) Leq(h) (Category B) / 52 dB(A) Leq(h) (Category D) | FHWA, 23 CFR 772, Table 1 |
| Europeans exposed at or above 55 dB Lden: road / rail / aircraft | About 92 million / 18 million / 2.6 million | EEA, Environmental noise in Europe 2025 |
| Transport noise exposure when WHO guideline levels are applied (road: 53 dB Lden) | About 145 million people, over 30% of the population | EEA, Exposure of Europe’s population to environmental noise (indicator, 2025) |
Context note: the FAA math is spelled out in Program Guidance Letter 12-09: a 25 dB NLR at 70 DNL or a 30 dB NLR at 75 DNL yields the 45 DNL interior target. The HUD and FHWA criteria are published in 24 CFR Part 51 and 23 CFR Part 772, and the European exposure counts in the EEA’s Environmental noise in Europe 2025 report.
Summary: Curtain Wall and Acoustic Glazing by the Numbers
| Metric | Value | Source |
|---|---|---|
| Outdoor-indoor difference, open / tilted / closed window | 10.0 / 15.8 / 27.8 dB(A) median | Locher et al., IJERPH 2018 |
| Two 3 mm lites, 12 mm air space | STC 26 to 30, OITC 22 to 26 | FGIA, Designing for quiet 2026 |
| Double vs triple unit with the same 8.8 mm acoustic laminate | Both Rw 42 (C -3; Ctr -8) | Pilkington Optiphon Datasheet 2020 |
| STC gain per doubling of air space | About 3 points | NRC Canada, CBD-240 1988 |
| 12.8 mm acoustic laminated vs 12 mm float | Rw 40 vs Rw 34 | Pilkington Optiphon Datasheet 2020; Pilkington GL090 2019 |
| Laminated vs monolithic vacuum glazing at coincidence dip | 8 to 9 dB better | Nugroho, Cabrera and Kocer, J. Phys. Conf. Ser. 2018 |
| Measured vs rated window performance | STC up to 4 dB lower; OITC 28 to 30 for all five windows | PARTNER Project 1.5 Report 2008 |
| IGU alone vs installed window | Rw 46 vs Rw 44 to 45 | Sustainability 2025 |
| Window with integrated roller shutter box | Maximum Rw 34 (vs 38 to 43 without) | Caballol Bartolome et al., Building and Environment 2026 |
| PVC window at -20 C | 3 dB lower insulation | Konstantinov et al., IOP Conf. Ser. 2020 |
| Glass-fin office partitions behind glazed facades | DnT,w at most 44 dB vs 45 dB baseline | Forum Acusticum / Euronoise 2025 |
| Required NLR at DNL 65 to 75 vs standard construction | 25 to 30 dB vs about 20 dB | 14 CFR Part 150, Appendix A |
| FAA interior insulation target | 45 dB DNL | FAA Program Guidance Letter 12-09 (2012) |
| HUD extra attenuation at DNL 65 to 70 / 70 to 75 | 5 dB / 10 dB | HUD, 24 CFR 51.104 |
| FHWA interior criterion for sensitive uses | 52 dB(A) Leq(h) | FHWA, 23 CFR 772 |
| Europeans exposed to road noise at or above 55 dB Lden | About 92 million | EEA, Environmental noise in Europe 2025 |
| Europeans above WHO-based transport noise thresholds | About 145 million, over 30% of the population | EEA, Exposure of Europe’s population to environmental noise (2025) |
Methodology and Sources
- National Research Council Canada, J.D. Quirt, Canadian Building Digest CBD-240: Sound Transmission Through Windows (1988), read via an archived copy of the NRC digest series.
- Robinson, Bernhard and Mongeau (Purdue University), Passive Sound Insulation: PARTNER Project 1.5 Report, FAA/NASA/Transport Canada PARTNER Center of Excellence (2008), tests at NASA Langley.
- Pilkington, Noise Control with Glass, technical bulletin GL090 (2019).
- Pilkington, Optiphon acoustic performance datasheet GL103 (2020), manufacturer data to EN 717-1.
- Caballol Bartolome, Navacerrada, de la Prida, Diaz Chyla and Pedrero, Comparative laboratory analysis of window sound insulation across glazing types in the 50-5000 Hz range, Building and Environment 301 (2026).
- Effect of Window Structure and Mounting on Sound Insulation: A Laboratory-Based Study, Sustainability 17(15) 6892 (2025).
- Locher et al., Differences between Outdoor and Indoor Sound Levels for Open, Tilted, and Closed Windows, IJERPH 15(1) 149 (2018).
- Nugroho, Cabrera and Kocer, Improving the airborne sound insulation properties of vacuum insulating glazing, J. Phys. Conf. Ser. 1075 (2018).
- Konstantinov, Verkhovsky and Lyabakova, Sound insulation of PVC windows at negative outdoor temperatures, IOP Conf. Ser. Mater. Sci. Eng. 896 (2020).
- Rich Rinka (FGIA), Designing for quiet: How fenestration impacts building acoustics, Construction Specifier (2026), citing AAMA TIR-A1, ASTM E413 and ASTM E1332.
- Caniato, Bettarello and Gasparella, Simulation of the transmission loss of curtain wall with mullions, IBPSA Building Simulation (2019).
- Airborne sound insulation in five office buildings according to Italian regulations and the WELL Building Standard, Forum Acusticum / Euronoise (2025).
- FAA, Program Guidance Letter 12-09: Eligibility and Justification Requirements for Noise Insulation Projects (2012).
- 14 CFR Part 150, Appendix A, Land Use Compatibility (eCFR).
- HUD, 24 CFR Part 51, Subpart B, Noise Abatement and Control (eCFR).
- FHWA, 23 CFR Part 772, Procedures for Abatement of Highway Traffic Noise (eCFR).
- European Environment Agency, Environmental noise in Europe 2025 (EEA Report 05/2025).
- European Environment Agency, Exposure of Europe’s population to environmental noise, indicator (2025).
- Data watch: several anchor sources are old because window acoustics is settled physics with little new open field data: NRC CBD-240 (1988), the PARTNER window tests (2008), FAA PGL 12-09 (2012) and the Locher field study (2018) are the most recent available data for their questions, and the FAA folded most of PGL 12-09 into the 2014 AIP Handbook. Pilkington values are manufacturer laboratory data, and the PARTNER report shows field and test-lab results can trail such ratings by several dB. EEA figures conflict slightly: the 2025 report counts about 92 million people above 55 dB Lden for road traffic, while the October 2025 indicator gives about 89 million, and the report carries a March 2026 corrigendum to some health estimates. The Building and Environment 2026 Rw values were read from a manuscript copy and should be checked against the final table; the FGIA article gives two different upper bounds for OITC (4,000 Hz and 5,000 Hz). Laboratory curtain wall ratings with mullions remain scarce in open literature, so curtain wall figures here are typical ranges rather than measured averages.
Last updated: October 18, 2026. We update this roundup quarterly, and the next refresh is expected when the European Environment Agency publishes its next Environmental Noise Directive exposure indicator update.