Transmission Loss Simulation vs Real-World Results

By James Hartley · April 29, 2026

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

In Q3 of 2025, our team at SonusGearFlow documented and supported an isolation upgrade for a two-room audio facility in a mixed-use building in Austin, Texas. The space consisted of a control room (approximately 6.2 m × 4.6 m × 3.0 m) and a live room (7.5 m × 5.4 m × 3.0 m), separated by a shared wall and flanked on the corridor side by a tenant office suite. The studio operator had a recurring complaint history: guitar cabinets and drum sessions were audibly leaking into the corridor and the adjacent office, especially late afternoons when the building was quiet.

The stakeholders were typical for projects of this size: a studio owner/operator (who also engineered most sessions), a general contractor, the building’s property manager, and our acoustics consultant responsible for simulation, detailing, and verification. The project’s goal was not “perfect silence”—that isn’t realistic in a retrofit without structural rework—but a measurable reduction in corridor and neighbor intrusion, specifically in the 125 Hz to 1 kHz range where most complaints were occurring. The owner wanted to keep the studio operational throughout the work, with no more than five days of full downtime.

The central question driving the case study was simple: how closely would our transmission loss (TL) simulations match real-world performance once flanking paths, workmanship tolerance, and existing building conditions were taken into account?

2. Challenges and requirements at the outset

The facility had been assembled quickly during a previous lease cycle. The existing corridor wall was a single-stud 2×4 assembly with 16 mm gypsum board each side, nominally insulated, and terminated into a suspended ACT ceiling. The live room ceiling was also suspended with 600 mm × 600 mm mineral fiber tiles. Doors were hollow-core with basic perimeter seals. Electrical penetrations were plentiful: wall plates, conduit stub-outs, and a shared back-to-back outlet box between the control room and corridor.

Requirements at kickoff were documented in plain language and then translated into acoustic targets:

Neighbor impact: corridor level during a drum session should be low enough that conversations at 2–3 m from the studio wall were not disrupted.
Operational constraint: keep recording possible on most days; schedule noisy demolition after hours.
Budget: USD $42k hard cap including materials, labor, and rework contingency.
Verification: provide a before/after report using field measurements (ASTM E336-style methodology) and note any deviations from predictions.

From an engineering perspective, the immediate challenge was that the existing wall did not run slab-to-slab. The ACT ceiling created a classic flanking path: sound could go over the top of the wall into the plenum, across, and down into the corridor. A second challenge was the building’s mechanical system: a shared return plenum meant duct-borne and plenum-borne paths were likely, and the studio had limited access to the central HVAC trunk.

3. Approach and methodology chosen

We treated the project as two parallel tracks: predictive simulation to guide assembly choices, and pragmatic field investigation to map the actual dominant paths. For simulation, we modeled wall TL using a combination of manufacturer lab data and mass-air-mass estimates for double-leaf options. We used a layered calculation approach consistent with how most studio retrofit decisions are made: target the weak links first (doors, ceiling flanking, penetrations), then increase wall mass/decoupling if needed.

For field verification, we planned:

Baseline measurements of airborne sound isolation between live room and corridor, and between control room and corridor.
Spot diagnostics using a handheld intensity probe and a calibrated measurement mic to identify “hot” zones (around outlets, door perimeters, ceiling grid intersections).
Post-install measurements matching speaker and mic positions as closely as practical, including third-octave analysis to interpret frequency-dependent improvements.

Equipment choices were straightforward and intentionally replicable for project managers working with typical consultant toolkits:

Source: dodecahedron loudspeaker with 1 kW amplifier (pink noise), plus a secondary check using a powered 12” loudspeaker for low-frequency emphasis.
Measurement: Earthworks M30 measurement microphone with calibration file, and a Class 1 handheld SPL meter for spot checks.
Analysis: FFT/third-octave software for logging, with time-averaged measurements for stability.

4. Step-by-step execution narrative

Week 0 (site walk + baseline): We started with a two-hour walk-through and a review of existing drawings (which were incomplete). We documented the corridor wall height, ceiling grid layout, and door details. Baseline measurement was done after business hours to reduce background noise. We placed the dodecahedron in the live room at 1.5 m height, ran pink noise, and measured corridor levels at three positions: directly outside the live room wall, at the adjacent office entry, and at the corridor midpoint.

The baseline field results were consistent with the complaints: isolation was acceptable above 1 kHz but weak in the low-mid band. Using a simplified field metric (level difference corrected for room absorption), we saw a corridor intrusion peak around 160–250 Hz—typical of lightweight construction and flanking through ceiling plenums.

Week 1 (design decision + procurement): Simulation work began with two candidate assemblies for the corridor wall:

Option A: add a second layer of 16 mm Type X gypsum on the studio side with viscoelastic damping compound between layers; seal all perimeter edges; leave studs as-is.
Option B: build an independent staggered-stud or double-stud wall in front of the existing wall, insulated cavity, double gypsum on the new leaf; isolate from ceiling grid as much as possible.

Our TL simulation suggested Option A might add roughly 6–10 dB in the midband but would not solve the over-the-top ceiling flanking. Option B promised better broadband improvement, especially 125–500 Hz, but had higher cost, reduced corridor width at door returns, and a schedule risk due to framing and door/trim modifications.

We chose a hybrid plan: implement Option A on the wall surface, and simultaneously address the ceiling flanking by extending the wall to the structural deck with hard lid barriers (gypsum on framing above the ACT ceiling) and sealing the plenum path.

Week 2 (preparation + demolition): Work was staged to keep the studio functional. Demolition was limited to the corridor side ceiling tiles adjacent to the studio wall and selective removal of baseboard and trim for sealing. We discovered the insulation in the existing stud cavity was inconsistent—some bays were empty. That immediately explained why simulation based on “nominal insulated wall” would likely overpredict performance.

Week 3 (construction + sealing): The critical path tasks were completed in this order:

Plenum barrier installation: Above the corridor wall line, we framed a lightweight extension to the underside of the concrete deck using 25-gauge metal studs, installed 16 mm Type X gypsum both sides where accessible, and packed the cavities with 48 kg/m³ mineral wool. The key was continuity: no gaps at deck intersections, and fire-rated acoustic sealant at all perimeters.
Cavity remediation: We opened sections of the studio-side gypsum only where needed to confirm insulation coverage and added mineral wool in empty bays. This was messy but high value.
Mass addition + damping: On the studio side of the corridor wall, we added a constrained-layer damping approach: one layer of 16 mm Type X gypsum, viscoelastic compound applied in a consistent bead pattern (approx. 0.5–0.6 tubes per 4×8 sheet), then a second layer. Screws were staggered and kept off studs where possible to avoid creating rigid “shorts.”
Penetration sealing: Back-to-back electrical boxes were reworked by relocating one box offset by a stud bay and installing putty pads rated for acoustic/fire sealing. All conduit penetrations were sealed with non-hardening acoustic sealant.
Door upgrade: The hollow-core corridor door to the studio entry was replaced with a solid-core 45 mm slab, full perimeter compression seals, and an automatic drop seal. Door undercut was reduced to less than 3 mm when closed.

Week 4 (punch list + verification): We did a day of punch-list work focused on small but consequential gaps: unsealed top plate edges, a missed conduit penetration above the ceiling, and a hairline crack at a corner bead. Then we repeated measurements using the same source level and positions.

5. Technical decisions and trade-offs made

Simulation vs constructability: The double-stud “best practice” solution would have delivered the largest improvement on paper, but the floor area loss and cost would have pushed the project beyond the $42k cap. We accepted a smaller modeled improvement in exchange for speed and keeping the studio operational.

Ceiling strategy: We did not replace the ACT ceiling across the entire corridor. Instead, we treated the wall line and immediate adjacent bays as the flanking hotspot. This reduced cost and time, but it relied on careful sealing above the grid—something that is hard to inspect after tiles go back.

Low-frequency expectations: We explicitly did not promise dramatic isolation below 80–100 Hz. The building structure, slab continuity, and shared plenum limited what could be achieved without room-within-room construction. The owner’s complaints were mostly 125–500 Hz; we optimized there.

Door as a system: We treated the door not as a single product choice but as a set: slab mass, seals, threshold behavior, and latch alignment. A premium door without correct seal compression would have wasted money.

6. Results and outcomes with specific details

The post-install measurements showed a clear improvement, but not a perfect match to the optimistic early simulation outputs. The difference—predictable in hindsight—was the residual flanking and the variability in existing construction.

Measured isolation improvement (live room to corridor):

125 Hz: +6 dB improvement (simulation suggested +8 to +10 dB)
250 Hz: +10 dB improvement (simulation suggested +12 dB)
500 Hz: +13 dB improvement (simulation suggested +14 dB)
1 kHz: +15 dB improvement (simulation suggested +15 dB)
2 kHz and above: +12 to +16 dB improvement depending on mic position (door area dominated variance)

In practical terms, corridor audibility changed from “clearly identifiable drum hits and guitar riff content” to “mostly low-frequency thump with limited intelligibility,” which aligned with the owner’s target. The adjacent office reported that afternoon disturbances dropped enough that they stopped filing complaints. The property manager requested a copy of the verification summary for lease records—usually a sign that a retrofit has moved from “argument” to “documented improvement.”

Why the simulation overpredicted at 125 Hz: Two main reasons showed up during diagnostics:

Structural flanking: Vibration transmission through slab edges and shared framing was still present. We could hear and measure energy around the baseboard line even after wall upgrades.
Plenum complexity: Even with barriers, the return plenum created alternate paths. Where the barrier stopped at an inaccessible beam pocket, the intensity probe showed a localized leak.

Timeline and budget performance: The project ran 18 working days from kickoff to final measurement, with 4 days of full studio downtime (within the 5-day constraint). Final cost landed at $39.6k, including a $2.1k allowance for unexpected insulation remediation and electrical relocation.

7. Lessons learned and what could be done differently

Lesson 1: Validate assumptions about existing cavities. Our early model assumed a uniformly insulated stud cavity because that’s what the original spec claimed. Reality was inconsistent. If we had opened a small inspection section during Week 0, we could have adjusted predictions and planned the remediation more efficiently.

Lesson 2: Plenum barriers need inspectable continuity. “Seal it above the ceiling” is easy to say and hard to prove later. In future projects, we would specify photo documentation before closing tiles and require a simple checklist: deck perimeter sealed, stud bays filled, penetrations labeled and sealed, and any inaccessible voids noted explicitly.

Lesson 3: Doors remain the largest source of variance. Even with good seals, door alignment and user behavior matter. During verification we saw a 3–5 dB swing at 1–4 kHz depending on whether the latch fully engaged and how evenly the seals compressed. A door closer with consistent pull force would have reduced variability.

Lesson 4: Simulation should be framed as “assembly performance,” not “project performance.” Our wall TL simulation was fairly accurate for the upgraded assembly, but the building is a system. Flanking paths, weak junctions, and construction tolerances are where field results diverge most.

8. Takeaways applicable to other projects

Start with pathfinding, not products. Before choosing an expensive wall system, identify whether the dominant path is through the wall, over the wall (ceiling plenum), around the wall (doors), or through services (ducts and penetrations). A $200 sealant plan can outperform a $10k mass upgrade if the leak is at the perimeter.
Model what you can, then add a “flanking realism” margin. If simulation suggests +12 dB improvement in the 125–250 Hz band for a retrofit, plan and communicate that field results may be 2–6 dB less unless junction details are rebuilt slab-to-slab.
Specify continuity details like you would for waterproofing. Acoustic isolation fails at seams: top plates, deck intersections, back-to-back boxes, and ceiling grid returns. Treat those junctions as first-class deliverables, not incidental labor.
Keep verification positions consistent. Use repeatable source and mic locations, log source level, and document background noise. Differences in placement can masquerade as “performance changes,” especially in small rooms with modal behavior.
Make trade-offs explicit to the client. We documented that sub-100 Hz isolation would remain limited without structural decoupling. That prevented scope creep and kept success criteria tied to the complaint frequencies that mattered.

In this project, simulation did its job: it guided the selection of a cost-effective assembly and forecast where improvements would land. Field results confirmed the general shape of the prediction but exposed where reality always lives—at flanking paths and imperfect existing construction. For audio engineers and project managers, the practical win is understanding that “TL simulation vs real-world results” is not a contest. It’s a workflow: predict, build, measure, and then iterate based on the system you actually have, not the one you wish you had.