How to Reduce HVAC Noise in Broadcast Studios

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

In early 2025, Sonus Gear Flow was asked to document and support a noise-mitigation retrofit for a mid-market broadcast facility in Columbus, Ohio. The client operated two on-air rooms (Studio A: news/talk, Studio B: sports) and a shared control room that also served as a voiceover/edit suite during off-hours. The facility was built in the late 2000s in a leased commercial building and had accumulated “patch fixes” over time—mostly around furniture, IT, and mic upgrades. The HVAC system, however, had never been tuned for modern broadcast noise expectations.

The project team included the station’s engineering manager (client-side lead), a mechanical contractor familiar with the building’s rooftop units (RTUs), an acoustical consultant, and our audio engineering lead responsible for translating “what we hear” into specifications the trades could execute. The “why” was straightforward: complaints from on-air talent about constant low-frequency rumble and intermittent rushing noise were increasing, and the station had begun producing more long-form spoken-word content where noise floors become painfully obvious.

The immediate business driver was a new syndication partner requiring measurable noise performance. While the contract didn’t mandate a specific room criterion, the partner’s QC process routinely flagged HVAC rumble and broadband hiss. The station set a practical target: reduce HVAC contribution to an equivalent of NC-20 to NC-25 in the studios, with no “events” (clunks, pops, pressure bursts) audible during takes.

2) Challenges and requirements at the outset

Two issues showed up in the first walk-through. First, the supply air velocity at the studio diffusers was high enough that you could hear turbulent noise from the grille face, especially in Studio B. Second, there was clear low-frequency energy around 40–80 Hz that varied with compressor staging and fan speed changes. The studios were otherwise quiet: the building was not in a high-traffic location, and the isolation construction was decent for a leased space (double layer gypsum on resilient channel in most partitions, solid core doors, moderate sealing).

We established baseline measurements over two days to catch different operating conditions:

Studio A (on-air position): HVAC ON Leq (A-weighted) 31–33 dBA; HVAC OFF 23–24 dBA. Distinct 63 Hz bump of 8–10 dB above adjacent bands.
Studio B (talent mic position): HVAC ON Leq 34–36 dBA; HVAC OFF 24–25 dBA. Noticeable midband “air hiss” between 500 Hz and 2 kHz from diffuser turbulence.
Control room: HVAC ON 32–34 dBA; HVAC OFF 25–26 dBA. Less turbulence, more low-frequency structure-borne feel.

The station’s technical requirements were not only acoustical. The studios ran hot due to lights, people, and equipment. Shutting air off during live segments wasn’t acceptable; even a 15–20 minute HVAC pause led to temperature drift and client complaints in summer. Any solution had to maintain airflow and comfort while reducing noise. Additionally, changes had to fit into a narrow downtime window: two weekends and three late-night shifts, coordinated around live programming.

3) Approach and methodology chosen

We treated the HVAC noise as a system problem rather than a “throw a silencer at it” problem. The methodology had four tracks running in parallel:

Measure and attribute: Separate diffuser-generated turbulence, duct-borne fan noise, and structure-borne vibration.
Reduce at the source: Fan speed control behavior, duct velocities, and compressor/fan staging where feasible.
Interrupt transmission paths: Add attenuation and decouple vibration in targeted locations.
Validate with repeatable tests: Same mic positions, same HVAC modes, and consistent measurement settings to confirm real improvement.

Instrumentation included a Class 1 measurement microphone (Earthworks M50 with a calibrator), a handheld analyzer for spot checks, and DAW-based recordings for subjective review. We also logged RTU operating states from the building management system (BMS) to correlate noise spikes to mechanical events (fan ramping, economizer changes, compressor staging).

4) Step-by-step execution narrative

Week 1: Baseline and localization

The first week was observation and attribution. We performed nearfield measurements at:

Each diffuser face (to quantify turbulence and grille noise)
The supply trunk near the studio branch takeoffs (to identify duct-borne fan noise)
Ceiling grid and stud reference points (to check for structure-borne vibration)

Studio B’s primary issue was diffuser turbulence. The diffuser served a relatively high CFM through a 2x2 lay-in with a restrictive neck. Air velocity at the face was measured around 650–750 fpm in typical cooling mode—high for a critical listening environment. Studio A had lower face velocity but more low-frequency content that did not change much when diffuser dampers were adjusted, suggesting upstream fan/duct-borne noise or vibration rather than local turbulence.

Week 2: Mechanical coordination and design decisions

We held a coordination meeting with the mechanical contractor and the building’s facilities rep. The constraints were clear: the RTUs were shared across multiple tenant zones; we could not swap the entire unit or re-engineer the building system. The most realistic improvements were within the studio branch ducts and the immediate mechanical room above the control room ceiling.

We agreed on three physical modifications and two control changes:

Replace two diffusers in Studio B with larger, low-velocity perforated-face diffusers and increase the neck size from 8" to 10" where possible.
Add lined duct sections (2" fiberglass liner, 1.5 lb/ft³) for at least 10 feet upstream of each studio branch where access allowed.
Install rectangular splitter silencers on Studio A and Studio B supply branches (targeted for 125 Hz–2 kHz attenuation, moderate pressure drop).
Add spring isolation for a ceiling-hung fan-powered VAV box feeding the studios (the box was previously hard-hung and transmitting vibration).
Retune fan speed control to reduce aggressive ramping events at the start of live blocks.

Week 3–4: Procurement and prep

Lead time for silencers and diffusers was the schedule risk. We selected commercially available components with predictable delivery: Nailor-style rectangular silencers (packaged to size, 24" length class) and Titus perforated diffusers with a reputation for low self-noise at modest pressure. The mechanical contractor verified ceiling plenum clearances: in Studio A, we had only 14" above the grid near the branch takeoff, requiring a shorter silencer than ideal. We accepted reduced low-frequency attenuation there and planned to rely more on control changes and lining.

During prep, we also addressed non-HVAC noise that often gets blamed on HVAC: one studio had a light fixture with a faint buzz at 120 Hz. That was corrected with a driver replacement so it wouldn’t contaminate post-work measurements.

Weekend 1: Physical installation in Studio B

The first weekend focused on Studio B because it was easiest to improve quickly and would build confidence. The contractor removed the existing restrictive diffuser and short neck. They installed a 10" neck feeding a larger perforated diffuser designed for lower throw velocity. We also added an 8-foot lined flex section between the hard duct and the diffuser neck. Flex duct is not a universal solution—improper installation can increase turbulence—but with a gentle radius and no compression, it provided both decoupling and some broadband attenuation.

After installation, we rebalanced the airflow to maintain the original cooling performance. The goal was not to starve the room; it was to deliver the same CFM at lower face velocity and with less turbulence.

Weekend 2: Silencers, lining, and isolation for Studio A and common feeds

The second weekend was more invasive. We installed a rectangular splitter silencer on Studio A’s supply branch just downstream of the main trunk, along with 12 feet of lined duct where the ceiling path allowed. The VAV/fan-powered box feeding the studios was re-hung using spring isolators and flexible connectors on inlet/outlet to break vibration paths. We also added neoprene isolation pads under a small pump serving an adjacent air handler coil that was contributing intermittent structure-borne pulses.

Late-night shifts: Controls tuning and verification

With hardware in place, we addressed the “events.” The BMS trend logs showed the supply fan performing fast ramps at schedule transitions and economizer changes. Those ramps correlated with audible “whoosh” moments and occasional duct oil-canning (a brief pop as pressure changed). The mechanical contractor adjusted fan ramp rates and minimum airflow setpoints to reduce sudden pressure changes. We also verified that the VAV minimum positions weren’t forcing high velocity through small openings during low-load periods.

5) Technical decisions and trade-offs made

A broadcast studio retrofit lives in trade-offs. Three decisions were debated internally:

Silencer length vs. available plenum space: Longer silencers provide better low-frequency attenuation, but Studio A’s ceiling height limited us. We chose a shorter silencer and compensated with lining and control adjustments. This was a pragmatic compromise for a leased facility.
Lined duct vs. external wrap: Internal lining offers good broadband attenuation but raises concerns about fiber shedding and cleanliness. We specified coated liner rated for air handling and required clean-cut edges and proper sealing. External wrap would have been cleaner but less effective for mid/high noise inside the duct.
Flex duct usage: Flex can solve structure-borne transmission and add some attenuation, but it can also become a noise source if kinked or compressed. We limited flex to short, gently curved sections and documented installation photos for QA.

We also made an audio-facing decision: do not rely on gating or aggressive noise reduction in the broadcast chain to “solve” HVAC. The station used standard voice processing (including mild expansion), but the goal was to improve the room so that microphone choice, mic technique, and processing could remain consistent rather than constantly compensating.

6) Results and outcomes with specific details

After completion, we repeated measurements at the same positions, same HVAC modes, and similar outdoor conditions (within a 6°F range, comparable humidity). Results were reviewed both objectively and with listening tests using raw mic recordings (Shure SM7B and Neumann TLM 103, preamps at typical gain settings).

Studio B: HVAC ON Leq reduced from 34–36 dBA to 27–29 dBA at the talent mic position. The 500 Hz–2 kHz turbulence band dropped by roughly 6–8 dB, and subjective “hiss” was no longer obvious on pauses. Airflow remained within 5% of pre-work CFM after rebalance.
Studio A: HVAC ON Leq reduced from 31–33 dBA to 26–28 dBA. The 63 Hz bump decreased by about 4–6 dB depending on RTU stage. The remaining low-frequency energy was still detectable on ultra-close listening, but no longer triggered QC flags on program exports during the following month.
Control room: HVAC ON Leq improved modestly (about 3 dB), primarily due to reduced vibration transmission from the re-hung VAV box. The room’s noise floor became more consistent—fewer momentary pressure-related events.

Timeline-wise, the project ran six weeks from first site walk to final sign-off: two weeks for assessment and design, roughly two weeks for procurement and scheduling, and two weekends plus three late-night sessions for execution and tuning. The station reported a clear operational benefit: talent stopped requesting “air off” during long-form segments, and producers noticed fewer edits required to manage room tone shifts between takes.

7) Lessons learned and what could be done differently

The most instructive lesson was that HVAC noise rarely has a single cause. Studio B sounded bad because of diffuser velocity and turbulence; Studio A sounded bad because of low-frequency duct-borne energy and vibration paths. Treating both rooms with the same solution would have wasted money.

Three items we would change on a similar project:

Bring mechanical into the conversation earlier: We involved the contractor in week two, but earlier access to RTU performance data would have sped up attribution and reduced the number of measurement visits.
Reserve space for silencers during original build: The biggest limitation was plenum height. In new construction, we would allocate straight duct runs specifically for silencers and lined sections rather than “finding space later.”
Document control sequences before and after: The control tweaks solved the “events,” but documenting the exact setpoint and ramp changes more rigorously would make future troubleshooting easier when building staff changes.

8) Takeaways applicable to other projects

For audio engineers and project managers tackling HVAC noise in broadcast environments, several practices from this case translate well:

Measure like you mix: Use repeatable mic positions and capture both numbers (Leq, spectra) and recordings. The recording is often what convinces stakeholders, while the spectrum helps trades act.
Separate turbulence from transmission: If noise changes dramatically when you partially close a diffuser damper (even temporarily), you’re likely dealing with velocity/turbulence. If it doesn’t, look upstream for fan noise or vibration paths.
Fix “events,” not just averages: A room can measure acceptably on average and still fail in production because of pressure pops, fan ramps, or economizer transitions. Trend logs from the BMS are as valuable as SPL data.
Choose silencers and lining with pressure drop in mind: Every dB of attenuation costs something in static pressure. Coordinate early so the mechanical system can still deliver required CFM without pushing fan speeds higher (which can erase your gains).
Installation quality matters: A well-specified silencer won’t help if duct transitions create new turbulence. Similarly, flex duct only helps when installed gently and kept fully extended.
Don’t depend on processing to mask a mechanical problem: Gates and expanders can hide noise in pauses, but they don’t fix rumble under speech and they make room tone inconsistent. Quiet rooms are easier to mix and easier to QC.

The retrofit in Columbus didn’t turn a leased commercial suite into a purpose-built scoring stage, but it met the operational goal: consistent, broadcast-appropriate noise floors without sacrificing comfort or forcing awkward workarounds. The key was treating HVAC noise as a measurable, multi-source engineering problem and executing changes that were small enough to schedule, but specific enough to matter.