The spec sheet says STC 42. The client brief requires STC 38. You are 4 points above the requirement, the partition is signed off, the joinery contractor orders 400 linear metres of full-height glass partition system, and the fit-out proceeds. Nine months later, the CEO is complaining that confidential board meetings can be heard in the open-plan area outside the boardroom.
This is not unusual. It is the predictable outcome of a widespread misunderstanding of what glass partition STC ratings actually represent and what they conceal. This article gives you the full picture — the physics, the test methods, and the four failure mechanisms that manufacturers do not explain on their product literature.
How STC Ratings Are Measured
The STC (Sound Transmission Class) rating for a partition is measured in a laboratory under ASTM E90 — the Standard Test Method for Laboratory Measurement of Airborne Sound Transmission Loss of Building Partitions and Elements. The test procedure:
- The partition is built into a test opening between two reverberation rooms (source room and receiving room)
- A broadband sound source generates a diffuse field in the source room
- Sound pressure levels are measured in both rooms at 16 one-third octave bands from 125 Hz to 4000 Hz
- The Transmission Loss (TL) at each frequency is calculated: TL = L₁ − L₂ + 10 log(S/A), where S is the partition area and A is the absorption in the receiving room
- The STC contour (per ASTM E413) is fitted to the TL curve to determine the single-number rating
This is legitimate. The test method is standardised, reproducible, and internationally recognised. The problem is that it tests a mechanism — direct transmission through the partition — that is rarely the dominant path in a real building.
The Four Flanking Paths That Destroy Your STC
Flanking transmission is sound that reaches the receiving room via paths other than through the partition itself. In typical glass partition installations, four flanking paths dominate:
1. The Ceiling Plenum Gap
The most common and most damaging flanking path. A glass partition that terminates at the suspended ceiling — rather than extending to the structural soffit above — leaves an open or partially open path through the ceiling plenum. Sound travels over the top of the partition, through the plenum, and re-enters the receiving room via the ceiling tiles on the other side.
Ceiling tiles in a suspended grid system typically have a Ceiling Attenuation Class (CAC) of 25–35. This means they attenuate sound by 25–35 dB in the plenum-to-room transmission path. If the plenum gap flanking path contributes sound at 45 dBA and the ceiling provides 30 dB attenuation, the flanking contribution to the receiving room is 15 dBA.
With the direct path through the partition performing at TL = 42 dB, the partition path would produce a receiving room level of approximately 25 dBA. The flanking path at 15 dBA dominates. The effective STC of the combined system is controlled by the flanking, not by the glass.
Fix: Full-height partition from floor slab to structural soffit, with acoustic sealant at all joints. Where the soffit is a suspended ceiling, the partition must penetrate the ceiling plane and the acoustic seal must be maintained at the plenum boundary.
2. The Floor Track Connection
Glass partition systems use a floor track that is typically screwed or adhesively fixed directly to the structural floor slab. If the floor track is in rigid contact with the slab, structure-borne vibration — particularly low-frequency sound — is conducted through the track, into the slab, along the slab, and up through the floor track on the other side of the partition.
This structure-borne flanking path is particularly significant for low-frequency sound (125–500 Hz), which propagates efficiently through concrete and steel. The TL of a glass partition at 125 Hz is already limited by the mass-law (typically 20–25 dB for 6 mm glass). If the flanking path at 125 Hz contributes an additional 8 dB reduction, the effective TL at 125 Hz drops to 12–17 dB — which is the performance of almost no partition at all.
Fix: Resilient isolation tape or flexible acoustic sealant under the floor track. The track must not be in direct rigid contact with the slab.
3. HVAC Ductwork Penetrations
Ductwork that penetrates the partition plane — or supply/return diffusers that serve both sides of the partition from the same duct run — create a direct acoustic path. The sound attenuation of an unlined duct section over 3 m is approximately 0.5–1 dB/m for 300 mm duct at 500 Hz. A shared duct run of 10 m provides approximately 5–8 dB attenuation — far below the 38–42 dB that the glass partition is specified to provide.
Even where the ductwork technically penetrates the partition via fire dampers and acoustic sleeves, the detail is frequently installed without proper acoustic sealing around the sleeve perimeter. A 5 mm annular gap between the sleeve and the partition framing, over a sleeve perimeter of 800 mm, provides an effective flanking aperture with transmission loss close to 0 dB.
Fix: Separate supply and exhaust systems for enclosed spaces. Where penetrations are unavoidable, specify acoustically rated duct sections (with sound power attenuation certificates) and seal all penetration perimeters with acoustic sealant.
4. Structural Frame Vibration
In steel-frame and concrete-frame buildings, sound excites structural vibration in the floor/ceiling slab and the column grid. This vibration travels through the structure, bypassing the partition, and is re-radiated as airborne sound on the other side. The glass partition intercepts only the airborne transmission path; it is transparent to structure-borne paths.
This mechanism is most significant in lightweight steel-frame buildings (curtain walled commercial towers) and becomes relevant at low frequencies (below 250 Hz) where structural radiation efficiency is highest. Vibration isolation of the partition — soft seating, resilient floor track, and acoustic breaks at ceiling junction — reduces this mechanism.
The Coincidence Dip: Glass's Hidden Weakness
Even in an idealised laboratory test that eliminates all flanking, glass partitions have an intrinsic acoustic weakness that manufacturers rarely highlight prominently: the coincidence dip.
Every plate material has a critical frequency f_c at which the bending wavelength in the plate equals the acoustic wavelength in air. At this frequency, the plate "couples" efficiently with the sound field — sound passes through with much less attenuation than the mass-law would predict. The critical frequency for glass is:
f_c = c² / (1.8 × c_L × t)Where c = speed of sound in air (343 m/s), c_L = longitudinal wave speed in glass (~5200 m/s), t = glass thickness (m).
For common glass thicknesses:
| Glass Thickness | Critical Frequency f_c |
|---|---|
| 4 mm | 3,750 Hz |
| 6 mm | 2,500 Hz |
| 8 mm | 1,875 Hz |
| 10 mm | 1,500 Hz |
| 12 mm | 1,250 Hz |
The critical frequency for 6 mm glass — by far the most common single-pane glass used in office partitions — is approximately 2,500 Hz. This is directly in the most important speech frequency band. The octave band containing 2,500 Hz (the 2 kHz octave band) has a weighting coefficient of 0.19 in the STC contour — one of the highest-weighted bands. A significant TL dip at this frequency substantially depresses the STC.
The measured TL of 6 mm glass at octave bands:
| Octave Band (Hz) | 125 | 250 | 500 | 1k | 2k | 4k |
|---|---|---|---|---|---|---|
| 6 mm glass TL (dB) | 18 | 24 | 28 | 32 | 25 (coincidence dip) | 35 |
| Mass law prediction | 19 | 25 | 31 | 37 | 43 | 49 |
The mass-law predicts 43 dB at 2 kHz. The measured value is 25 dB — an 18 dB coincidence deficit. At 4 kHz (above f_c), performance partially recovers. The STC contour procedure partially accounts for this by allowing up to 32 dB total unfavourable deviations, but the dip at 2 kHz is so significant for speech intelligibility that the practical reduction in perceived sound insulation is greater than the STC number suggests.
The Laminated Glass Solution
Laminated glass (two glass plies bonded by a polyvinyl butyral (PVB) or ionoplast interlayer) damps the coincidence resonance. The PVB layer has viscoelastic properties that dissipate bending wave energy at and around f_c. The result is a shallower, broader coincidence dip rather than the sharp trough seen in monolithic glass.
For 10.38 mm laminated glass (4/0.38/6 — 4 mm glass + 0.38 mm PVB + 6 mm glass):
| Octave Band (Hz) | 125 | 250 | 500 | 1k | 2k | 4k |
|---|---|---|---|---|---|---|
| Laminated TL (dB) | 22 | 27 | 33 | 38 | 35 | 40 |
| Monolithic 6 mm TL (dB) | 18 | 24 | 28 | 32 | 25 | 35 |
| Improvement (dB) | +4 | +3 | +5 | +6 | +10 | +5 |
The 10 dB improvement at 2 kHz from laminated glass is the most important number on this table. At the frequency where speech intelligibility is highest, laminated glass performs 10 dB better than monolithic glass of equivalent mass. The STC improvement is approximately +6 to +8 points.
Insulated Glazing Units (IGUs): Double Trouble
Insulated glazing units — double-pane systems with a gas fill between two glass lites — are commonly specified for their thermal performance. Their acoustic performance is more complicated, and manufacturers frequently market them at their optimistic STC values without adequate caveats.
The acoustic benefit of an IGU depends on:
- Cavity resonance frequency. An IGU with 6 mm cavity depth has a mass-air-mass resonance at approximately 500–700 Hz, where TL drops sharply below either pane's individual mass-law performance. A 12 mm cavity has resonance around 350–500 Hz. These resonances can reduce TL by 10–15 dB in the affected frequency range.
- Coincidence dip interaction. If both panes have the same thickness, their coincidence dips stack at the same frequency, doubling the TL deficit. For good acoustic performance, the two panes of an IGU should have different thicknesses to stagger their coincidence frequencies. A 6/12/8 IGU performs significantly better than a 6/12/6 IGU at the 2–4 kHz range.
- Gas fill. Argon fill (the standard thermal fill) provides minimal acoustic benefit. Krypton provides modest improvement. For pure acoustic performance, air gaps outperform gas fills because gas fill slightly increases cavity stiffness and raises the resonance frequency into a more acoustically sensitive range.
| Configuration | STC |
|---|---|
| Monolithic 6 mm glass | 26 |
| 6/12/6 IGU (equal panes, argon) | 26–28 |
| 6/12/8 IGU (unequal panes, air) | 31–33 |
| 6/12/6 with PVB inner pane | 33–35 |
| 8/16/10 laminated IGU | 38–42 |
The common misconception that double glazing is acoustically superior to single glazing is not universally true. A 6/12/6 IGU achieves essentially the same STC as 6 mm monolithic glass in many configurations, because the cavity resonance deficit cancels out the mass-doubling advantage.
What the Spec Sheet Should Show — and Usually Doesn't
When evaluating glass partition products, the following information is essential for an informed acoustic decision:
1. Full TL curve at octave bands, not just STC. Request the ASTM E90 test report showing TL at 125, 250, 500, 1000, 2000, and 4000 Hz. A product marketed as STC 42 may have TL = 25 dB at 2 kHz. Another product at STC 40 may have TL = 34 dB at 2 kHz. For speech intelligibility, the second product is substantially better despite the lower STC.
2. OITC rating in addition to STC. The Outdoor-Indoor Transmission Class (OITC, ASTM E1332) emphasises lower frequencies (40–4000 Hz) that are important for speech and exterior noise. OITC is always lower than STC; a product with STC 42 and OITC 31 has a significant low-frequency weakness, while STC 42 and OITC 36 indicates more balanced performance.
3. AIIC (Apparent Impact Insulation Class) or field STC data. Very few manufacturers publish field test data alongside laboratory data. Request data from completed installations with the same system, full height to structural soffit, and ask what field STC was measured.
4. Plenum compatibility detail. The test configuration used in the laboratory should be documented. If the lab test was performed with the partition terminating at a suspended ceiling (not at the structural soffit), the lab STC includes some ceiling attenuation and is not purely the partition's performance.
Realistic Performance Expectations by System Type
| System | Lab STC (typical) | Field STC (typical, with flanking mitigation) | Field STC (typical, without flanking mitigation) |
|---|---|---|---|
| 10 mm monolithic glass, floor-to-suspended ceiling | 30 | 28–32 | 22–26 |
| 10 mm monolithic glass, floor-to-structural soffit | 30 | 27–30 | 22–26 |
| 6/12/6 IGU, floor-to-suspended ceiling | 28 | 26–30 | 20–24 |
| 10.38 mm laminated glass, full-height | 37 | 34–38 | 28–33 |
| 12.76 mm laminated glass, full-height | 40 | 37–41 | 30–36 |
| Double-leaf 6/100/6 (separated panes) | 50–55 | 46–52 | 38–44 |
| 16 mm laminated glass + resilient frame | 46 | 42–46 | 36–42 |
The jump from single-leaf to double-leaf (separated glazed systems) is the most significant acoustic step change available in glass partition design. Double-leaf systems — two independent glass panels separated by a 50–150 mm air gap — eliminate the mass-air-mass resonance problem, avoid stacked coincidence dips, and break the structural coupling between the two faces. Field STC 46–52 is achievable with careful installation.
The Minimum Specification for Reliable Speech Confidentiality
For a glass-walled meeting room or enclosed office where speech confidentiality is required (AI < 0.05, ASTM E1130 Confidential class), the minimum reliable specification is:
Glass: 12.76 mm laminated glass (4.76/0.76/8 or 6/0.76/6 construction), minimum. This provides STC 40–42 in the laboratory with a reduced coincidence dip at 2 kHz. For high-performance applications, double-leaf 6/50/6 system (STC 48–52 laboratory).
Head detail: Partition extends to structural soffit (not suspended ceiling). Acoustic sealant (minimum 10 mm depth) at all joints. Perimeter neoprene gasket between partition frame and structure.
Floor track: Resilient acoustic tape under track, minimum 6 mm neoprene durometer 40–50. No direct rigid contact between track extrusion and concrete slab.
HVAC: Independent supply and return serving each enclosed space. No shared ductwork. All penetrations (power, data, pipe sleeves) sealed with non-hardening acoustic mastic.
Door/opening: Acoustic door meeting STC 38 minimum with seals on all four edges (top, bottom, and both sides). The weakest element in a glass partition assembly is almost always the door.
Check your proposed specification against these requirements using AcousPlan's sound insulation calculator, which models the combined effect of glass TL, flanking contributions, and door/opening performance, and reports the predicted field STC against IBC 2021 §1207 requirements.
Summary
Glass partition STC ratings are real numbers from real tests. They are just not numbers from your building. The ASTM E90 laboratory test eliminates the four flanking paths that typically dominate sound transmission in real glass partition installations: plenum gaps, floor track vibration, ductwork penetrations, and structural frame transmission.
The coincidence dip in 6 mm monolithic glass at 2–2.5 kHz reduces the transmission loss precisely at the most speech-critical frequencies. Laminated glass with a PVB interlayer reduces this dip by 8–10 dB and represents the minimum glass specification for any partition where speech privacy matters.
Achieving the manufacturer's laboratory STC in the field requires: full-height partition to structural soffit, resilient floor track isolation, sealed HVAC penetrations, and an acoustic door with perimeter seals. Omitting any one of these elements typically reduces the field STC by 8–12 points.
The rule of thumb: to achieve field STC 38 (the minimum for Normal office privacy), specify a system with laboratory STC 48+, full-height to structural soffit, and a properly sealed acoustic door. Then verify with field measurement (ASTM E336) after fit-out and before handover.