How to Calculate Sound Background Noise Level Between Rooms

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

In late February, our team at Sonus Gear Flow was brought into a retrofit project at a mid-size post-production facility in Portland, Oregon. The site had two adjacent rooms on the second floor of a converted warehouse: a dialogue edit suite (Room A) and a small live-record room used for ADR and voiceover (Room B). The facility manager had already upgraded monitors and room treatment but kept losing time to a recurring complaint: “The edit suite is quiet until someone records next door. Then the noise floor changes and the dialogue cleanup gets harder.”

The client asked for a concrete answer to a deceptively simple question: What is the background noise level from Room B as heard in Room A, and how does it compare to our target? They wanted the methodology documented so their project manager could re-run the calculations after future gear changes or construction. The team on this job included:

Project manager: Facilities lead coordinating schedule and contractors
Audio lead: Senior re-recording mixer responsible for room targets
Acoustics consultant: Sonus Gear Flow documentarian/engineer (author) validating measurements and calculations
General contractor: Handling any isolation upgrades if required

The “why” was business-critical: the studio booked ADR in two-hour blocks and delivered TV mixes on tight deadlines. They needed predictable noise behavior so editors could set consistent denoise thresholds and avoid revisiting sessions due to shifting background conditions.

2. Challenges and requirements at the outset

The facility’s baseline targets were aligned with common post-production expectations: Room A needed a low, stable noise floor equivalent to roughly NC-20 to NC-25 during critical work, with no intermittent intrusions from Room B when normal ADR prep was happening (talkback, cue playback at modest level, and occasional footfall).

The constraints were typical of a retrofit:

Shared demising wall: Standard 2x4 stud wall, one layer 5/8" gypsum each side, partially insulated; no resilient channel.
Door coupling: Each room had a solid-core door but with standard perimeter seals; the doors were offset by only ~1.4 m in the corridor, creating a flanking path.
HVAC: A shared duct trunk with branch runs into both rooms; supply diffusers were quiet but the return path was not fully isolated.
Time: Two weeks from assessment to a go/no-go decision on construction work, because bookings were already sold.

The key requirement was not just “measure noise.” It was to calculate the background noise contribution between rooms in a way that the team could repeat. That meant combining measured levels, instrumentation calibration, and a clear distinction between:

Room A’s intrinsic background noise (HVAC, equipment, exterior infiltration)
Transmitted noise from Room B into Room A through structure and flanking paths

3. Approach and methodology chosen

We used a method that’s practical for real facilities and aligns with how standards-based work is often validated in the field:

Establish baseline background noise in Room A with Room B silent (doors in normal operating position).
Create a controlled noise source in Room B using pink noise at stable levels, measured at a reference position.
Measure the resulting level in Room A in 1/3-octave bands (or octave bands when time is limited) to see frequency-dependent transmission.
Calculate the transmitted contribution by energy subtraction: L_trans = 10 log10(10^L_total/10 − 10^L_base/10).
Assess impact against targets using A-weighted levels for operational decisions, and band data for diagnosing pathways (low-frequency issues typically indicate structural/duct paths; mid-high often indicates door/seal leaks).

Equipment choices were aimed at repeatability. We used:

Measurement mic: Earthworks M23 (calibrated)
Interface: RME Babyface Pro FS (stable gain, low noise)
Software: Room EQ Wizard (REW) for RTA logging and averaging; Smaart v9 used for quick cross-check
Acoustic calibrator: NTi Audio Calibrator at 94 dB SPL, 1 kHz
Noise source: Genelec 8030C in Room B plus a sub switched off initially (sub only used later for low-frequency probing)
SPL meter cross-check: NTi XL2 (not strictly necessary, but useful for field confidence)

For timelines: one day for measurements and calculations, two days for interpreting results and proposing fixes, then a one-week window for minor construction if needed.

4. Step-by-step execution narrative

Step 1: Confirm operating conditions and calibrate

Measurements are only as good as the operating assumptions. We first set “normal use” conditions:

HVAC running in standard daytime mode
All studio computers on in Room A (since editors work that way)
Doors closed as they typically are during sessions
No people walking the corridor during capture windows

We calibrated the M23 with the NTi calibrator at 94 dB SPL, 1 kHz and confirmed the software chain level. We then checked that REW’s RTA was reading consistently with the NTi XL2 (within ~0.7 dB A-weighted for steady noise), which was acceptable for this field assessment.

Step 2: Measure Room A baseline background noise

In Room A, we placed the mic at the primary listening position (mix chair), 1.2 m above the floor, pointing upward to reduce directional bias. We logged 90 seconds of A-weighted and 1/3-octave data and captured the Leq (equivalent continuous level).

Room A baseline (Room B silent):

L_A,eq: 23.8 dBA
Octave-band highlights (dB SPL): 63 Hz: 33 dB, 125 Hz: 28 dB, 250 Hz: 22 dB, 500 Hz: 19 dB, 1 kHz: 17 dB, 2 kHz: 16 dB, 4 kHz: 15 dB

That baseline was actually respectable for a retrofit. The issue wasn’t “Room A is loud,” it was “Room A changes when Room B is active.”

Step 3: Generate a controlled noise source in Room B

In Room B, we placed a Genelec 8030C approximately where the talkback/cue monitor usually sat: 1.3 m off the floor, 0.6 m from the demising wall. We played pink noise and set a reference level of 78 dBA at 1 m from the speaker (measured with the M23 in REW, cross-checked with the XL2).

This level was chosen for two reasons: it was above the Room A baseline enough to make energy subtraction stable, and it represented a “moderate operational noise” scenario (cue playback at a level that doesn’t endanger hearing but reflects real session behavior).

Step 4: Measure Room A total noise with Room B active

With Room B playing pink noise continuously, we repeated the Room A measurement at the same position, logging for 90 seconds.

Room A total with Room B noise active:

L_A,eq,total: 28.6 dBA
Octave-band highlights (dB SPL): 63 Hz: 38 dB, 125 Hz: 32 dB, 250 Hz: 25 dB, 500 Hz: 21 dB, 1 kHz: 19 dB, 2 kHz: 18 dB, 4 kHz: 16 dB

The “total” number by itself isn’t yet actionable. The calculation step is where the transmitted component becomes clear.

Step 5: Calculate transmitted background noise contribution

We used energy subtraction to estimate the noise attributable to Room B transmission:

Formula: L_trans = 10 log10(10^L_total/10 − 10^L_base/10)

Plugging in the A-weighted levels:

L_total = 28.6 dBA
L_base = 23.8 dBA

Converting to linear energy:

10^28.6/10 ≈ 724
10^23.8/10 ≈ 240
Difference ≈ 484
L_trans ≈ 10 log10(484) ≈ 26.8 dBA

This result often surprises teams: the transmitted component (26.8 dBA) is not the same as the delta between total and baseline (4.8 dB). Because decibels are logarithmic, a seemingly small increase can indicate a substantial additional energy contribution.

We repeated the same subtraction in octave bands to identify pathways. The largest transmitted contributions were in the 63–125 Hz bands, with secondary contributions around 250 Hz.

5. Technical decisions and trade-offs made

Several practical trade-offs came up during the day:

A-weighted vs band-based reporting: Project managers like a single number (dBA), but isolation failures show up in bands. We reported both: A-weighted for operational impact, octave/1/3-octave for diagnosis.
Pink noise vs real program: Pink noise made the calculation repeatable. We validated with two “real” sources afterward (spoken word at ~70 dBA in Room B and cue playback of music at ~76 dBA) to ensure the practical outcome matched the controlled test.
Measurement position choices: One position is not enough for a full certification, but the client needed a quick, repeatable method. We used the mix position as primary and did two spot checks: near the door in Room A and near the shared wall.
Subwoofer usage: We initially avoided a sub to keep the stimulus realistic. Later we ran a low-frequency sweep (40–200 Hz) to confirm that the worst coupling aligned with mechanical/structural paths rather than air leaks alone.

The trade-off was clear: we were not producing a formal lab-grade STC or ISO 16283 report. We were producing a field-calculation method that answered the operational question and pointed to fixes.

6. Results and outcomes with specific details

The main finding was that Room A’s baseline was fine, but the transmitted background from Room B during moderate activity was ~26.8 dBA at the mix position, with disproportionate low-frequency energy (63–125 Hz).

Diagnostics revealed two dominant transmission paths:

Door/corridor flanking: At the Room A door location, levels were 2–3 dB higher than at the mix position in the 500 Hz–2 kHz range, suggesting leakage and flanking via the corridor. Door seals were incomplete at the threshold; there was a visible 3–5 mm gap in places.
HVAC return coupling: When we temporarily blocked the Room B return grille with a dense temporary panel (not a permanent solution, strictly a diagnostic), Room A’s 63–125 Hz transmission dropped by ~2 dB. That indicated shared duct/return paths were contributing to the low-frequency change.

The client opted for a “fast intervention” package rather than opening the demising wall:

Door upgrades: Added full perimeter compression seals and an automatic drop seal on both room doors (materials ~$900; installed in one day).
HVAC mitigation: Installed an additional lined flex duct section and a small return silencer for Room B (materials and labor ~$2,400; installed over two nights).
Operational tweak: Moved the Room B cue monitor 1.5 m away from the demising wall and placed it on an isolation pad (minor but measurable benefit in 125–250 Hz).

After the upgrades, we repeated the same test protocol one week later.

Post-fix measurements (same conditions and levels):

Room A baseline: 23.6 dBA (essentially unchanged)
Room A total with Room B noise active: 26.4 dBA
Calculated transmitted component:
- 10^26.4/10 ≈ 437
- 10^23.6/10 ≈ 229
- Difference ≈ 208
- L_trans ≈ 10 log10(208) ≈ 23.2 dBA

In practical terms, the edit suite’s noise floor still rose slightly when Room B was active, but the shift was smaller and less low-frequency heavy. Editors reported they no longer had to change denoise thresholds mid-session, and the re-recording mixer stopped hearing “the room change” during quiet dialogue passages.

The schedule impact was minimal: measurement day (Friday), approvals (Monday), door seals (Tuesday), HVAC work (Wednesday/Thursday nights), re-test (next Friday). Total downtime in either room was kept under six hours.

7. Lessons learned and what could be done differently

The project reinforced a few realities that don’t show up in simplified isolation discussions:

Small dB changes can be meaningful: A 3–5 dB rise in total background level can mean a large transmitted component once you do energy subtraction. Without the calculation, teams underreact.
Low-frequency problems are rarely “just the wall”: The worst bands (63–125 Hz) pointed us toward HVAC/structural paths quickly. If we had focused only on door seals, the client would have been disappointed.
Repeatability beats perfection for operations: Using the same mic position, same stimulus level, and the same averaging window made the calculation useful for future checks. A more formal multi-position survey would be appropriate for new builds, but this was a retrofit under time pressure.

What we would do differently with more time: run a fuller set of positions (four corners plus mix position) and log longer averages (3–5 minutes) to capture HVAC cycling. We would also measure the corridor as a separate receiving space to quantify flanking more explicitly.

8. Takeaways applicable to other projects

If you need to calculate the background noise level transmitted between rooms in a way that supports decisions, the workflow from this case study is portable:

Measure a baseline in the receiving room under real operating conditions (HVAC on, gear on).
Introduce a stable, repeatable noise source in the source room (pink noise is fine) and document the reference level at a fixed distance.
Measure the new total in the receiving room at the same mic position and with the same averaging time.
Calculate the transmitted component with energy subtraction (A-weighted for impact, banded data for diagnosis).
Use the frequency signature to choose fixes:
- Mid/high leakage often points to doors, seals, penetrations, or weak flanking through corridors.
- Low-frequency dominance often points to HVAC paths, structure-borne transmission, or insufficient decoupling.
Re-test using the identical protocol after any change. If your method isn’t repeatable, your conclusions won’t be either.

The most valuable deliverable on this job wasn’t a single number; it was a documented, field-friendly method that connected subjective complaints (“the room changes”) to a measurable transmitted noise component and a targeted mitigation plan. That combination is what keeps post rooms predictable—and keeps schedules from slipping.