Why Post-Construction Testing Matters
A building's acoustic performance on paper and its acoustic performance in practice are not the same. The gap between specification and outcome is created by:
- Workmanship variation: An STC 50 partition requires continuous acoustic seals at head and base tracks, fully sealed service penetrations, and no back-to-back electrical boxes. Miss any of these and field NIC drops by 8–15 dB.
- Flanking transmission: Sound travels around partitions through connected structure — floors, ceilings, beams. Flanking paths are invisible in drawings and unavoidable in some structural configurations.
- Material substitutions: Products substituted during construction may not match the specified acoustic performance.
- Design errors: Even the best specification contains assumptions that may not hold in as-built conditions.
This guide covers what to test, how to test it, what equipment you need, how to interpret results, and what to do when tests fail.
What to Test: The Complete Post-Construction Acoustic Programme
1. Reverberation Time (RT60)
When required: Schools (BB93:2015 mandatory), offices (BREEAM Hea 05), healthcare (HTM 08-01), auditoria, any space with a specified RT60 requirement.
Standard: ISO 3382-2:2008 (Measurement of room acoustic parameters — Reverberation time in ordinary rooms). For performance spaces: ISO 3382-1:2009.
Method overview: A sound source (loudspeaker or pistol shot) produces a broadband noise burst. The room is allowed to fill with reverberant energy. The source is cut off, and the decay of sound level is recorded. The T20 (time for sound to decay by 20 dB, extrapolated to 60 dB) or T30 (30 dB decay extrapolated) is calculated for each octave band.
Measurement configuration:
- Minimum 3 source positions and 3 receiver positions per room
- Source and receiver separation: > 1 m; both > 1 m from any wall
- Minimum room volume for ISO 3382-2 method: V > 50 m³ (for smaller rooms, use decay method with MLS)
- Measurement condition: unoccupied, with HVAC operating in normal mode
2. Airborne Sound Insulation (DnT,w)
When required: Part E Building Regulations (residential dwellings — mandatory), schools (BB93), healthcare (HTM 08-01), any partition with a specified sound insulation requirement.
Standard: BS EN ISO 16283-1:2014+A1:2017 (Field measurement of sound insulation in buildings — Airborne sound insulation).
Method overview: A loudspeaker in the source room generates a continuous broadband noise. The sound level is measured in both the source room and the receiving room in octave or one-third octave bands. The Level Difference (D) is adjusted for the receiving room's reverberation time to give the Standardised Level Difference (DnT) per octave band. The single-number rating DnT,w is calculated by fitting to the reference curve per ISO 717-1.
Measurement configuration:
- Source room: loudspeaker positioned to avoid direct transmission; minimum 2 positions
- Receiving room: minimum 5 microphone positions, spatially separated
- Background noise in receiving room must be at least 6 dB below the measured signal level (10 dB preferred) at each frequency band
- For small rooms (V < 25 m³): low-frequency corrections per ISO 16283-1 Annex A
Part E minimum requirements (England):
- Separating walls: DnT,w + Ctr ≥ 45 dB
- Separating floors (airborne): DnT,w + Ctr ≥ 45 dB
3. Impact Sound Insulation (L'nT,w)
When required: Part E (separating floors between dwellings — mandatory), schools (BB93 for some interfaces), healthcare (HTM 08-01).
Standard: BS EN ISO 16283-2:2015 (Field measurement of sound insulation — Impact sound insulation).
Method overview: A standardised tapping machine (5 hammers, 10 hits per second, 500 g each) is placed on the floor of the source room. The sound level produced in the receiving room (below) is measured per octave band and adjusted for receiving room reverberation. The single-number L'nT,w is calculated per ISO 717-2.
Lower L'nT,w = better impact sound insulation (it is a level, not a reduction).
Part E minimum requirement (England): L'nT,w ≤ 62 dB for separating floors.
4. Background Noise Level
When required: Schools (BB93 — mandatory), healthcare (HTM 08-01), offices (BREEAM, WELL), any space with a maximum background noise requirement.
Standard: No single standard mandates the measurement procedure, but BS 8233:2014 Annex A and BB93 Appendix A describe the relevant methodology.
Method: Type 1 SLM measuring LAeq (A-weighted equivalent continuous sound level) and octave band levels. All building services (HVAC, lighting, lifts) operating in normal mode. No occupancy. Duration: minimum 5 minutes at each measurement position.
Reporting: Report LAeq and octave band levels (31.5 Hz to 4000 Hz). Compare against specified targets. Also report LCpeak if specified. For critical spaces (music rooms, recording studios), report NC or RC curve values calculated from octave band measurements.
5. Speech Transmission Index (STI)
When required: Auditoria, lecture theatres, places of assembly, PA system verification, classrooms (BB93 encourages STI measurement), any space with a specified speech intelligibility requirement.
Standard: IEC 60268-16:2020 (Sound system equipment — Objective rating of speech intelligibility by speech transmission index).
Method: A calibrated test signal (STIPA, a modulated noise signal) is emitted from a loudspeaker at a defined source level. Modulation transfer is measured at the receiver position. STI is calculated from the modulation transfer function across 14 octave band and modulation frequency combinations.
Reporting: Report STI value (0–1) and the verbal rating (Bad/Poor/Fair/Good/Excellent) per IEC 60268-16 Table 4. For PA systems, report at representative listener positions including the nearest, furthest, and most challenging positions.
Equipment Required for Acoustic Commissioning
Minimum Kit for Compliance Testing
| Equipment | Specification | Approximate Cost |
|---|---|---|
| Sound level meter | Type 1 (Class 1), IEC 61672; integrating function (Leq); octave band filter | £2,500–£5,000 |
| Calibrator | Pistonphone, IEC 60942 Class 1; traceable to national standard | £300–£600 |
| Omnidirectional loudspeaker | Dodecahedron speaker, ≥ 100 dB LW; flat response 100 Hz – 5000 Hz | £1,500–£3,500 |
| Power amplifier | 200–500 W, low THD | £300–£800 |
| Measurement microphones | IEC 61672 Class 1; minimum 2 (simultaneous measurement) | £400–£800 each |
| Tapping machine (impact sound) | ISO 10140-5 standardised machine | £2,000–£4,500 |
| STIPA signal generator | IEC 60268-16 compliant signal; or use 2-channel analyser | £800–£2,000 |
Total investment for a complete acoustic commissioning kit: approximately £8,000–£16,000. Many acoustic consultants use integrated measurement systems (NTi Audio XL2, Brüel & Kjær Type 2250, GRAS 12AA) that combine SLM, microphone, and analysis in a single instrument.
Calibration Requirements
All measurement equipment must be:
- Calibrated by an accredited laboratory (UKAS or equivalent) within 12 months of use
- Field-calibrated at the start and end of each measurement session using the pistonphone calibrator
- Field calibration check recorded in the measurement report
Interpreting Results: Pass, Marginal, Fail
Measurement Uncertainty
All acoustic measurements have inherent uncertainty. ISO 12999-1:2014 provides uncertainty budgets for sound insulation measurements. Typical expanded uncertainties (95% confidence interval):
- DnT,w field measurement: ±1.5–2.5 dB
- L'nT,w field measurement: ±1.5–2.5 dB
- RT60 (T20) field measurement: ±5–10% of value (i.e., ±0.03–0.05 s for a 0.6 s room)
- Background noise LAeq: ±0.5–1.0 dB
Suggested Acceptance Criteria (for specification writers)
Building on ISO 12999-1 uncertainty budgets:
| Parameter | Pass | Marginal | Fail |
|---|---|---|---|
| DnT,w (vs target T) | ≥ T | T − 2 to T | < T − 2 |
| L'nT,w (vs target L) | ≤ L | L to L + 2 | > L + 2 |
| RT60 (vs target τ) | ≤ τ | τ to τ + 0.1 s | > τ + 0.1 s |
| Background noise (vs target N) | ≤ N | N to N + 2 dB | > N + 2 dB |
For regulatory requirements (Part E), there is no marginal zone — the minimum specified value is the pass/fail threshold. For design quality specifications, the marginal zone allows professional judgment before triggering remediation obligations.
Common Commissioning Failures and Their Causes
Failure 1: Low DnT,w — Partition Flanking
Symptom: Measured DnT,w is 8–12 dB below specified Rw. The difference is too large to be measurement uncertainty.
Cause: Flanking transmission through a shared floor slab, ceiling void, or structural element bypasses the partition. Sound travels around the partition rather than through it.
Diagnosis: Measure sound level in the source room and the receiving room at multiple octave bands. Compare the frequency pattern against the partition's TL curve. If the measured DnT,w is relatively flat across frequencies (rather than following the mass-law increase with frequency), flanking from a lightweight path is dominant.
Remedy: Identify and seal the flanking path. Common paths: ceiling void above partition (add acoustic barrier above the partition to the slab), steel beam continuing through partition (add acoustic isolation layer), floor screed continuing through (saw-cut and seal with acoustic mastic).
Failure 2: High RT60 — Insufficient Absorption
Symptom: Measured RT60 is 0.2–0.4 s above target. Room sounds excessively reverberant.
Cause: Either (a) not enough absorption area installed, (b) acoustic products were substituted with lower-NRC products, or (c) installed products are not achieving their tested NRC in this installation (wrong edge detail, wrong mounting).
Diagnosis: Calculate the total absorption in the room using the Sabine formula backwards: A = 0.161 × V / RT60_measured. Compare against the specified absorption. If A is significantly less than specified, check product installation.
Remedy: Add absorptive surface area. If ceiling is already fully treated, add wall panels. Check that tiles are properly seated in the grid (gapped tiles reduce absorption by 10–15%) and that baffles have adequate clearance above and below.
Failure 3: High Background Noise — HVAC Noise
Symptom: Background noise level is 5–10 dB above specified maximum. NC or RC curve shows dominant low-frequency content.
Cause: Air handling unit or duct work is generating excessive noise. Either (a) the HVAC specification was insufficient, (b) duct liners were omitted, (c) air velocities are too high causing turbulence noise, or (d) AHU is vibrating and transmitting through structure.
Diagnosis: Measure octave band levels in the room. Identify dominant frequencies. Low-frequency (63–250 Hz) dominance suggests AHU blade passing frequency or fan noise. High-frequency (1000–4000 Hz) dominance suggests duct turbulence (high velocity) or unlined terminal units.
Remedy: Reduce supply duct velocity (target < 3.5 m/s at diffuser), add duct silencer on supply and return branches serving the affected room, add vibration isolation to AHU mounting, add acoustic lining to first 1–2 m of duct downstream of AHU.
Failure 4: Low STI — Excessive Reverberation
Symptom: STI below target. Typically STI < 0.70 in a classroom or lecture theatre.
Cause: RT60 is too high, reducing modulation contrast at mid frequencies. STI is very sensitive to RT60: reducing RT60 from 1.0 to 0.5 s at 1000 Hz typically increases STI by 0.10–0.15.
Relationship (Houtgast-Steeneken estimate): STI ≈ 1 − log10(C/10) where C is a function of SNR and RT60. Practically: every 0.2 s reduction in RT60 (mid-frequency) adds approximately 0.05 STI.
Remedy: Increase acoustic absorption to reduce RT60 (same diagnosis and remedy as Failure 2). Also check that background noise is not excessive (high background noise also reduces STI independently).
The Commissioning Report
A complete acoustic commissioning report should contain:
- Executive summary: One page. Pass/marginal/fail status for each tested parameter. List of any remediation required.
- Introduction: Project description, parties involved, testing dates.
- Equipment list: All instruments with calibration certificate references.
- Measurement conditions: Occupancy status, HVAC operating mode, weather conditions.
- Test results by parameter: Tabulated results vs targets, with graphs of octave band data.
- Assessment against specification: Explicit pass/fail determination for each measurement.
- Recommendations: For any failures, specific remediation proposals with predicted outcome.
- Appendices: Raw measurement data, calibration certificates, room sketches showing source/receiver positions.