How to Isolate Impact Noise

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

In February 2026, SonusGearFlow was brought into a tenant improvement project at a six-story mixed-use building in Portland, Oregon. The ground floor was being converted into a 110 m² (1,185 ft²) post-production suite and two edit rooms for a small documentary studio. Directly above it sat a boutique fitness studio with heavy footfall: treadmill running, box jumps, and free-weight drops on rubber mats.

The owner’s goal was practical: keep the post rooms bookable during business hours without constant complaints about “thumps,” “booms,” and low-frequency rumble. The fitness studio had a long-term lease and could not be relocated. The building’s existing construction was typical for the neighborhood—wood joists with a gypsum ceiling and a relatively lightweight floor system above—so impact transmission was inevitable without remediation.

The team included a project manager (GC-side), SonusGearFlow acting as acoustic consultant and site verifier, a structural engineer, and the post facility’s audio lead. The “why” was clear: the post studio’s work was dialogue-heavy, and clients expected quiet rooms for editing, ADR pickups, and review sessions. Even if the noise wasn’t loud in an SPL sense, intermittent impact events were disruptive and hard to edit around.

2) Challenges and requirements at the outset

The first site walk revealed two problems: the building’s low-frequency isolation was limited, and the upstairs tenants produced classic impact excitation. We heard footfall as short impulses coupled into the ceiling, followed by a “after-ring” around 50–80 Hz. Weight drops produced longer decay and occasional rattles in recessed light trim.

Constraints shaped the strategy:

Schedule: 7 weeks from notice to proceed to operational handover. The studio had client bookings starting week 8.
Ceiling height: The post rooms had an existing 3.05 m (10 ft) slab-to-finish height; the client wanted to keep at least 2.74 m (9 ft) finished.
Building limitations: The base structure was wood; adding significant dead load was limited by structural review.
Access: Minimal disruption to the fitness studio; work had to occur primarily from below.
Performance requirement: Not a lab-grade spec, but measurable reduction in impact intrusions. The client asked for “no obvious thumps during normal editing.” We translated this into two targets: (1) reduce perceived footfall and drop events by at least 10 dB in the 50–200 Hz range at the listening position, and (2) remove audible rattles/buzzes entirely.

We also discovered HVAC duct hangers and sprinkler lines hard-coupled to the joists. Even if we improved the ceiling, flanking paths could bypass it and reintroduce impact as structureborne vibration.

3) Approach and methodology chosen

The key decision was to treat impact noise as a structureborne problem first, not an airborne one. The upstairs impact energy was exciting the joists; the ceiling was re-radiating it as low-frequency sound, and fixtures were acting as sympathetic buzz points.

Because we couldn’t rebuild the upstairs floor assembly, we chose a “from-below” isolation strategy:

Decoupled ceiling: A spring-isolated ceiling system using isolation hangers to reduce vibration transfer from joists to gypsum.
Added mass + damping: Two layers of 5/8 in Type X gypsum with constrained-layer damping compound between layers to reduce panel resonance and increase loss factor.
Cavity absorption: Mineral wool (45–60 kg/m³) to reduce cavity resonance and mid/high transmission through the assembly.
Flanking control: Isolate or re-hang ducts and pipes, add flexible connections where possible, and address light fixtures and penetrations.
Room detailing: Seal all perimeters, use backer rod and acoustical sealant, and avoid rigid short-circuits between old and new ceilings.

We planned verification around repeatable impact generation and measurement. Before construction, we captured baseline data using a calibrated measurement chain: a class-1 measurement mic (Earthworks M50), a portable interface (RME Babyface Pro FS), and Room EQ Wizard for logging. For impact excitation, we used two methods: (1) a repeatable “heavy heel drop” protocol at marked points upstairs with a 90 kg person (three drops per point), and (2) a controlled 10 kg medicine ball drop from 1 m onto the upstairs rubber floor, which approximated a worst-case impulse the tenant would realistically produce.

4) Step-by-step execution narrative

Week 1: Baseline measurement and path finding

We measured in the future edit room (Room A) and mix room (Room B). In Room B, peak levels from medicine ball drops were reaching 54–58 dB(C) at the listening position with clear energy centered around 63 Hz and 125 Hz. Heel drops were lower but frequent, around 45–49 dB(C), with strong impulsive content. Subjectively, the issue wasn’t continuous loudness; it was interruption and startle.

We used a mechanic’s stethoscope and a vibration app with an accelerometer puck (for rough comparative checks, not absolute calibration) on the ceiling grid and ductwork. The loudest buzzes corresponded to a recessed can light housing and a supply duct hanger rigidly tied to a joist. That gave us a prioritized remediation list: isolate the ceiling and remove “hard points.”

Week 2: Design finalization and procurement

The ceiling design settled on spring isolation hangers rated for 25 lb (11 kg) each, spaced at 1.2 m (4 ft) along framing members, with hat channel (25 ga) running perpendicular. We coordinated with the structural engineer to confirm that added dead load would remain within acceptable limits. Total added ceiling weight was estimated at ~6.5–7.5 psf (31–37 kg/m²) including gypsum layers, compound, channel, and insulation—within tolerance for this zone given existing safety margins.

To minimize ceiling height loss, we used low-profile hangers and maintained a 50 mm (2 in) air gap from joists to the top of channel wherever possible. The GC ordered:

Isolation spring hangers (Kinetics or Mason equivalent), quantity ~220 for both rooms and corridor
7/8 in hat channel, ~420 m (1,380 ft)
5/8 in Type X gypsum board, two layers over ~190 m² (2,045 ft²)
Damping compound (Green Glue class product), ~300 tubes to hit ~2 tubes per 4x8 sheet equivalent
Mineral wool batts, 3 in thickness for joist cavity
Acoustical sealant, backer rod, putty pads for electrical penetrations
Flexible duct connectors and neoprene isolation for select duct supports

Weeks 3–4: Demolition, rerouting, and flanking path fixes

The existing gypsum ceiling was removed in both rooms. This exposed several direct-coupled routes: EMT conduit strapped to joists, sprinkler mains hung with rigid threaded rod, and a duct trunk acting like a bridge between the joists and the ceiling plane.

Not everything could be decoupled—sprinkler changes required permits and a fire protection contractor, which the timeline didn’t support. Instead, we prioritized what we could influence:

Duct hangers: Replaced two rigid hangers over Room B with isolation hangers and added a short flexible connector section before the diffuser branch.
Lighting: Eliminated recessed cans. We moved to surface-mounted LED fixtures on the new isolated ceiling to avoid penetrations and buzzing housings.
Conduit and boxes: Used putty pads on all electrical boxes in the ceiling plane and ensured boxes were mounted to the new channel system, not the joists.
Perimeter: Added blocking where necessary but maintained isolation—no rigid bridging between the new ceiling framing and walls.

A practical lesson here: if you only “improve the ceiling” but leave ducts and fixtures hard-coupled, the new ceiling becomes quieter and the remaining flanking paths become more noticeable. In our baseline, the can lights buzzed; after isolation, the buzz would have dominated if we hadn’t removed them.

Weeks 4–5: Installing the decoupled ceiling

The hangers were installed into joists with specified fasteners, and the hat channel grid was leveled. We checked for short-circuits constantly: channel ends were kept clear of walls by 6–10 mm, and any place where the channel could touch framing was corrected. Mineral wool went into the cavity, carefully cut to fit without compressing (compressed batts reduce absorption effectiveness and can create mechanical coupling).

The first layer of 5/8 in gypsum was hung perpendicular to channel, seams staggered. We applied damping compound in a random serpentine pattern and installed the second layer within 15 minutes of application to ensure proper spread. Screws for the second layer were sized to avoid contacting channel in a way that “pulled through” and created rigid points; the crew used depth stops and we spot-checked screw depth.

All perimeters were sealed with backer rod and acoustical sealant. We avoided caulking so heavily that it formed a stiff bridge; the goal was airtightness with flexibility.

Weeks 6–7: Finishing, commissioning, and verification

After mudding and paint, we reinstalled HVAC diffusers with care. Where diffusers attached to duct boots, we ensured gaskets were used and no metal-to-metal rattle points existed. The audio team completed speaker mounts and acoustic treatment (panels and bass trapping), but we asked them to avoid direct mechanical ties into joists for any overhead hardware in these rooms.

We repeated the same impact tests at the same upstairs marked points and recorded levels at the same microphone locations and gain settings. We also did subjective checks during a real fitness class to confirm performance under the actual activity profile.

5) Technical decisions and trade-offs made

Several trade-offs were explicit and documented:

Decoupling vs. ceiling height: A deeper spring system would have improved low-frequency isolation but cost headroom. We chose low-profile hangers and accepted that 50–80 Hz performance would improve, but not disappear.
Two layers + damping vs. three layers: A third layer of gypsum adds mass but pushes weight and labor. With the structural limit and schedule, two layers with damping compound provided the best performance-per-pound.
From-below only vs. treating the source: The ideal fix is at the source—floating floor or underlayment upstairs. The lease situation made that unrealistic. We designed a robust receiver-side solution and set expectations accordingly.
Sprinkler isolation: Full decoupling wasn’t feasible without major permitting. We chose to isolate around sprinkler penetrations with oversized holes, escutcheons, and flexible sealant to prevent rigid contact with the new ceiling layers.
Lighting choice: Recessed fixtures are common but problematic in isolation assemblies. Switching to surface-mounted LEDs reduced penetrations and saved troubleshooting time later.

6) Results and outcomes with specific details

The post-construction measurements showed consistent improvement:

Medicine ball drop: Peak levels at the mix position in Room B reduced from 54–58 dB(C) to 42–46 dB(C), depending on drop location upstairs. That is a 10–14 dB(C) reduction.
Heel drops: Reduced from 45–49 dB(C) to 34–38 dB(C), roughly 10–12 dB(C) improvement.
Spectral change: The 125 Hz band showed the most improvement (often 12–16 dB reduction). The 63 Hz band improved more modestly (typically 6–9 dB), which matched expectations given structural coupling and the limits of a low-profile ceiling system.
Rattles: The pre-existing buzz from the recessed cans was eliminated entirely (fixtures removed). One diffuser exhibited a faint tick during heavy drops; tightening and adding gasket material resolved it.

Subjectively, the character changed from a sharp “thump” to a dull, lower-level “presence” that was easier to ignore and did not trigger startle responses. The audio lead reported that dialogue editing and nearfield monitoring were no longer interrupted by footfall. During the first client-attended review, a class was active upstairs; the client noticed no disturbance until it was pointed out, and even then described it as “barely there.”

In operational terms, the studio booked full days without noise-related rescheduling. The owner estimated this avoided at least two lost sessions per month at $900/day, quickly justifying the additional ceiling cost.

7) Lessons learned and what could be done differently

The project worked because we treated it as an impact isolation problem with verification, not guesswork. Still, several lessons stood out:

Flanking paths become the new problem: After isolation, small mechanical issues (diffuser ticks, conduit buzz, trim contact) can dominate. Budget time for “noise chasing” at the end.
Low-frequency expectations must be managed: In wood-frame buildings, energy below ~63 Hz is hard to stop from below. If the tenant regularly drops heavy weights, receiver-side treatments can reduce but not fully eliminate it.
Coordination beats heroics: The best improvements came from trades coordinating—HVAC re-hanging, lighting changes, and electrical detailing. If any one trade had ignored isolation constraints, the ceiling could have been short-circuited.
Verification needs repeatable methods: The marked-point protocol and controlled drop weight provided comparable before/after data. Without it, the team would have relied on memory and anecdote.

If we had more time and cooperation from the upstairs tenant, we would have proposed a partial source treatment: adding a high-performance underlayment in the fitness studio’s free-weight zone and enforcing designated drop areas. Even a targeted floating platform under the most abusive equipment often yields outsized gains, especially in the 50–125 Hz region.

8) Takeaways applicable to other projects

Start with a path diagnosis: Identify whether impact is primarily through the ceiling, ductwork, or fixtures. Fix the dominant path first.
Decouple, then add mass, then seal: For impact noise from above, decoupling is usually the largest lever. Mass and damping help, but they work best when the assembly is mechanically isolated.
Avoid recessed penetrations in isolated ceilings: Use surface fixtures, minimize cutouts, and treat any necessary penetrations with putty pads and airtight detailing.
Plan for the low end: If the problem includes 50–80 Hz energy, be honest: full elimination is unlikely without source control or major structural changes. Design for meaningful reduction and removal of secondary noises.
Measure before and after with repeatability: Even simple protocols—consistent drop weight, marked points, fixed mic position, consistent gain—turn a subjective complaint into an engineering decision.
Give the GC a “do not short-circuit” checklist: Isolation assemblies fail most often due to one screw, one rigid bracket, or one tight channel end touching a wall. A short field checklist prevents expensive rework.

Impact noise isolation is rarely about a single magic product. In this case, a decoupled ceiling, disciplined detailing, and targeted flanking fixes delivered a measurable 10–14 dB reduction for the worst events and converted an unusable room into a reliably bookable post suite—within seven weeks, with real-world constraints, and without requiring the upstairs tenant to shut down operations.