How to Measure and Improve Sound Transmission Class

By Marcus Chen · April 30, 2026

1. Project overview: what, where, who, and why

In March, SonusGearFlow was brought into a tenant-improvement buildout for a post-production suite inside a mixed-use building in Culver City, California. The project was a 1,850 sq ft facility on the second floor: a picture editorial room, a VO booth, a small mix room (nearfield-focused), and a machine closet. The client was a boutique post house expanding from a single-room operation into a space that could handle concurrent sessions without schedule collisions.

The “why” was straightforward: the space sat above street-level retail and shared a demising wall with a Pilates studio. The post house had already signed a five-year lease. They needed predictable isolation to keep VO takes clean during daytime foot traffic and to avoid being the reason the adjacent tenant complained at night.

The team included the client’s project manager, a general contractor (GC) with commercial TI experience, an MEP engineer, and our role covering acoustic design, measurement, and construction verification. The client requested a measurable target tied to a standard metric they could explain to ownership and insurers. We aligned on Sound Transmission Class (STC) as the primary rating for airborne isolation, with a secondary emphasis on low-frequency performance and flanking control that STC alone does not fully capture.

2. Challenges and requirements at the outset

At kickoff, three constraints defined the project:

Existing base construction: The building had metal stud partitions with a single layer of 5/8" gypsum each side in most areas, lightweight concrete deck, and a suspended T-bar ceiling. The demising wall to the Pilates studio was a 3-5/8" metal stud wall, 16" o.c., single 5/8" Type X each side, batt insulation inconsistently installed. That assembly typically tests around STC 34–38 when well-built, but field conditions often reduce it.
Space and schedule: Ceiling heights were limited. The client wanted to preserve at least 9'-0" finished height in editorial and 8'-6" in the booth. Construction had an 11-week window to substantial completion, with the client moving in immediately afterward.
Noise sensitivity: VO and dialog editorial demand quiet backgrounds. The client requested that speech and moderate music playback in one room not be intelligible in the adjacent rooms, and that outside tenant activity not intrude into recording. They cited a prior space where “voices leak through at night” and wanted to avoid a repeat.

We translated these into measurable requirements:

Partition performance targets: Interior partitions between editorial/mix and the VO booth: target STC 55 or better. Demising wall to Pilates studio: target STC 60 in the frequency range where intelligibility matters, while acknowledging that true low-frequency isolation would depend on mass and flanking control beyond STC.
Field verification: Perform pre-construction baseline field testing and post-construction field testing using ASTM E336 procedures for field airborne sound insulation, reporting field STC (FSTC). For the client, the pass/fail would be based on the post-construction FSTC values and a narrative on low-frequency behavior.
Flanking risk management: Address ceiling plenums, door sets, penetrations, and return air paths, which routinely defeat high-STC walls if not controlled.

3. Approach and methodology chosen

We used a three-part methodology:

Measure the baseline: Before demolition, we ran quick field tests on the existing demising wall and a representative interior partition. This established realistic starting points and identified flanking paths.
Design for STC with buildable assemblies: We specified wall assemblies with known lab STC ratings, then adjusted for field realities. For example, a lab-rated STC 63 wall might produce FSTC 55–60 depending on workmanship and flanking.
Verify during construction: We scheduled two on-site “hold points”: (a) after framing and MEP rough-in but before gypsum, and (b) after first layer of gypsum before final sealing and second layer. This let us catch bridging, back-to-back boxes, unsealed edges, and ceiling plenum shortcuts while they were still cheap to fix.

For measurement, we used a calibrated field kit:

NTi Audio XL2 analyzer with Class 1 measurement microphone (M2230) and calibrator (NTi Audio Calibrator 94/114 dB).
Omnidirectional dodecahedron loudspeaker with power amplifier for test noise (pink noise, 1/3-octave band analysis).
Tripod stands and laser distance tool for consistent mic positioning and documented source/receiver geometry.

While STC is a single-number rating derived from transmission loss (TL) across 125 Hz to 4 kHz, we also logged the full 1/3-octave spectra and noted deficiencies below 125 Hz, because adjacent tenant bass-heavy music and footfall-induced structure-borne components were expected.

4. Step-by-step execution narrative

Week 1–2: Baseline testing and site investigation

We conducted baseline E336-style measurements between the future mix room area and the Pilates demising wall. With existing finishes in place, the measured FSTC was 36. The spectrum showed predictable weakness at 125–250 Hz, and we found two major leakage paths:

The demising wall stopped at the suspended ceiling, with the plenum shared above both tenants. That meant sound could go up and over the wall.
Electrical boxes were back-to-back in several stud bays with no putty pads.

We also walked the slab/deck and noted that the Pilates studio had a hard floor and frequent impact activity. That would not be solved by STC upgrades alone, but airborne isolation still needed to be robust to reduce airborne music and instructor voice spill.

Week 2–3: Assembly selection and detailing

For the demising wall, we recommended a buildable, high-performing assembly without requiring a full room-within-room:

New independent stud wall built inboard of the existing demising wall: 3-5/8" steel studs, 16" o.c., with a 1" air gap from the existing wall (no rigid connections).
Fill stud cavity with 3-1/2" mineral wool (e.g., Roxul/ROCKWOOL Safe’n’Sound).
Two layers of 5/8" Type X gypsum on the studio side with Green Glue damping compound between layers (2 tubes per 4x8 sheet equivalent).
All perimeter edges sealed with non-hardening acoustic sealant; staggered seams between layers; no back-to-back boxes.
Extend the new wall to the underside of the structural deck (not stopping at T-bar ceiling). Where the deck geometry required, we used deflection track and fire-rated sealant in coordination with the firestop sub.

For interior partitions (mix room to editorial, editorial to VO), we used double-stud or staggered-stud assemblies depending on space:

VO booth: double-stud wall, two separate 3-5/8" stud frames with a 1" air gap, mineral wool in both cavities, double 5/8" gypsum both sides with damping compound on at least one side (booth side prioritized).
Between editorial and mix: staggered stud on a 6" track to save space, mineral wool, double 5/8" gypsum each side, damping compound on one side.

Week 4–6: Rough framing verification (first hold point)

At framing, we inspected for mechanical bridging. Two issues came up that would have undermined the design if left alone:

Stud contact: In three locations, the new independent wall was inadvertently screwed into the existing demising studs through the air gap. We had the GC remove the fasteners and add clearance blocks to keep the 1" separation consistent.
Duct support: The HVAC contractor had planned to hang a supply duct using straps attached to both the existing wall top track and the new wall top track. We revised the support plan so the duct was supported from the deck with vibration-isolating hangers, preventing a rigid connection across the isolation assembly.

We also coordinated with electrical to avoid back-to-back boxes. Where outlets were unavoidable on both sides of a partition, we required offset placement with at least one stud bay separation, putty pads on all boxes, and sealed conduit penetrations.

Week 6–8: First gypsum layer and interim checks (second hold point)

Once the first layer of gypsum went up, we performed an interim air-leak inspection rather than a full STC test. This is where many projects win or lose: a high-STC assembly can behave like a low-STC assembly if it leaks.

We used a simple but effective method: play broadband noise in one room and sweep the perimeter with a measurement mic and headphones, listening for “hot” points. We found leakage at:

Deflection track corners where sealant was skipped.
One pipe penetration oversized by ~3/4", left unsealed.
A door frame rough opening where drywall did not meet the jamb continuously.

All were corrected before the second layer, when access would be harder.

Week 8–10: Doors, ceiling, and flanking control

Doors were the next critical item. The client initially wanted standard commercial hollow-core doors. We pushed back with numbers: even a great wall is undermined by a weak door. We selected:

VO booth door: 1-3/4" solid-core wood door with full perimeter acoustic seals and automatic door bottom (Pemko or equivalent), heavy-duty hinges, and a latch that pulls tight against the seals.
Mix room and editorial doors: solid-core with perimeter seals, with threshold details coordinated to avoid trip hazards.

For the ceiling, the biggest flanking risk was the shared plenum. Over the most sensitive partitions, we specified that walls continue to the deck, and where that was impossible (around existing beams and MEP), we added gypsum “ceiling breaks” and sealed soffit returns to prevent sound from traveling over the top through open ceiling space.

HVAC was treated as both a noise source and a transmission path. We added lined duct sections (1" duct liner, 5–8 ft where space permitted), specified low-velocity grilles, and avoided straight “line of sight” return paths between rooms. Where a transfer path was unavoidable, we used lined transfer boots with two 90-degree turns to increase attenuation.

5. Technical decisions and trade-offs made

Several trade-offs shaped the final system:

Independent wall vs. resilient channel: Resilient channel can test well in the lab but is sensitive to installation errors (short-circuiting with screws, fixture loads). Given schedule and the number of trades, we chose independent framing on the demising wall to reduce risk.
Damping compound cost vs. predictability: Green Glue added material cost and a bit of labor complexity. We used it selectively: always on the studio side of critical walls, not everywhere. That kept the budget in check while improving mid-band TL where STC is most influenced.
Wall-to-deck full height vs. ceiling height: Going to deck reduced flanking but risked losing ceiling height and complicating firestopping. We prioritized full-height walls at the demising boundary and around the VO booth, accepting more detailing effort to preserve ceiling height elsewhere.
STC as a metric vs. low-frequency reality: The client wanted a single number. We provided it, but documented low-frequency behavior separately, explaining that STC doesn’t grade isolation below 125 Hz and that Pilates impact noise is often structure-borne, not airborne.

6. Results and outcomes with specific details

Post-construction testing occurred in week 11, after doors, seals, and final paint, but before furniture fully loaded the rooms. We conducted ASTM E336-style tests with multiple mic positions and averaged levels to reduce spatial variance.

Measured results:

Demising wall (mix room to Pilates studio side): FSTC 58. The 125 Hz band still showed a dip compared to mid-band performance, but speech intelligibility through the wall dropped dramatically. During an active Pilates class, conversational voice in the mix room was no longer masked by instructor cues, and during VO tests the ambient intrusion was reduced to a low, non-distracting presence.
Editorial to VO booth partition: FSTC 61. With editorial playback at ~75 dBA Leq in the editorial room, the booth measured below audibility for speech content, with only faint low-frequency energy detectable when placing an ear to the door—expected and acceptable.
Mix room to editorial partition: FSTC 54. This was the closest to the minimum target. Investigation found that one return-air path had higher-than-expected coupling. We added additional duct liner and a lined elbow at the return, improving subjective isolation even though we did not re-test for a revised FSTC number.

Operational outcomes after move-in:

VO sessions could run during daytime retail activity without re-takes due to audible neighbor speech.
Adjacent tenant complaints did not occur during the first two months of operation, even with occasional client playback in the mix room in the evening.
The client reported a measurable workflow benefit: two sessions could run concurrently (editorial playback and VO recording) without schedule coordination.

7. Lessons learned and what could be done differently

Start flanking control earlier: The biggest early discovery was the shared plenum. If we had been involved before lease signing, we would have flagged that the existing demising wall stopping at ceiling was a major risk. Early awareness could have reduced redesign churn.
HVAC paths deserve STC-level attention: The mix-to-editorial result (FSTC 54) was good, but it highlighted that return paths and transfer grilles can dominate real-world isolation. In a repeat project, we would model duct/transfer attenuation earlier and avoid shared returns wherever possible.
Door detailing must be treated as a system: A solid-core slab alone is not enough. The frame, seals, undercut, threshold, and latch alignment must be coordinated. We avoided major mistakes, but it took repeated field checks to ensure the automatic door bottoms were correctly adjusted after flooring.
Interim leak checks save money: The “noise sweep” after the first gypsum layer caught issues that would have been expensive after finishes. This step is easy to schedule and consistently pays off.

8. Takeaways applicable to other projects

Measure baseline conditions, even briefly: A quick field test and a plenum inspection can prevent false assumptions. Knowing that the starting demising wall was around FSTC 36 helped justify the needed scope.
Design to a field target, not a lab brochure: Lab STC numbers are optimistic compared to field conditions. If you need FSTC 55, choose assemblies that lab-test higher and control flanking.
Control air leaks obsessively: Sealant at perimeters, putty pads on boxes, tight penetrations, and careful door sealing often matter as much as adding mass.
Prefer robust assemblies when multiple trades are involved: Independent walls and full-height partitions are less fragile than solutions that fail when a single screw short-circuits the system.
Document and schedule acoustic hold points: Two site visits—pre-gypsum and mid-gypsum—are a practical minimum for ensuring the design becomes real performance.
Explain what STC does and doesn’t cover: STC is a useful, communicable metric for airborne isolation, but low-frequency and structure-borne issues need separate attention. Your client will thank you when the first complaint involves bass or impact noise.

This project finished on time within the 11-week construction window, met the client’s isolation goals where it mattered most, and—equally important—left behind a repeatable process: baseline measurement, realistic field targets, buildable assemblies, and verification at the moments when fixes are still cheap.