Why Open-Office Noise Must Be Measured
In an open-plan office, the trade-off for more collaboration is that speech flows between desks. The real issue is not the volume level (dB) but whether that speech is intelligible from a neighboring workstation.
When a colleague's phone call is audible and its content understood, attention is automatically drawn to it. Workstations where the STI (Speech Transmission Index) exceeds 0.5 are associated with an average cognitive task loss of 6% (Sandy Brown, 2014).
ISO 3382-3 converts this problem into objective numbers: how quickly speech from one workstation decays with distance (spatial decay), how loud it is at a desk 4 m away, and where STI drops below 0.5.
Design decisions follow from there — an absorption design that has never been measured is nothing more than a hypothesis.
Four Single-Number Quantities — What Is Measured and How
ISO 3382-3:2022 defines four Single Number Quantities (SNQs). The first three have been present since the 2012 first edition; comfort distance (rC) was added in the 2022 revision.
The four single-number quantities of ISO 3382-3:2022
The "Good target" column shows the informative annex values from 2012 (retained in 2022). These are guidance, not mandatory thresholds.
| 기준 | 기호 | 단위 | 무엇을 재나 | "좋음" 가이드 |
|---|---|---|---|---|
| D2,S | D2,S | dB | Spatial decay — drop in A-weighted SPL per doubling of distance | ≥ 7 dB |
| Lp,A,S,4m | Lp,A,S,4m | dB | A-weighted speech SPL measured at 4 m from the talker | ≤ 48 dB |
| rD | rD | m | Distraction distance — where STI drops below 0.5 from the talker | ≤ 5 m |
| rC (new in 2022) | rC | m | Comfort distance — where A-weighted speech SPL drops below 45 dB. Integrates D2,S and Lp,A,S,4m into a single metric in meters | No ISO target — smaller is better |
2022 Revision — What Changed from the 2012 Edition
The 2022 revision (2nd edition) incorporated ten years of measurement experience. The core three SNQs (D2,S · Lp,A,S,4m · rD) retain their original definitions and formulas, but a fourth metric has been added and previously ambiguous sections of the measurement procedure have been clarified.
ISO 3382-3 — four changes from 2012 to 2022 edition
| 기준 | 항목 | 2012 (초판) | 2022 (2판) | 실무 영향 |
|---|---|---|---|---|
| Number of SNQs | Number of SNQs | 3 (D2,S, Lp,A,S,4m, rD) + privacy distance | 4 (above three + rC). Privacy distance removed | rC reframes the metric to perceived comfort |
| Source directivity | Source directivity | Mouth-simulator (directional) sources allowed | Omnidirectional source only, all phases | Repeatability up, lab-to-lab agreement improved |
| Measurement procedure | Measurement procedure | Phase definitions partially ambiguous | Setup, measurement, background phases clarified | Measurement uncertainty reduced |
| STI reference signal | STI reference signal | IEC 60268-16 referenced (gender weighting variants) | IEC 60268-16 male α/β weights explicitly specified | STI calculation made consistent |
Same dB(A) — Why Is One Space Noisy and Another Quiet?
Even when a sound level meter reads the same 55 dB(A), one space may be tolerable while another breaks concentration. The difference lies in how long reflected sound lingers in the space and how quickly it decays with distance — precisely what ISO 3382-3's D2,S captures.
"Good" Guideline Values — ISO Informative Annex
The body of ISO 3382-3 specifies only the measurement method. Guidance on "what values are good" is provided as informative-only in an annex (values stated in the 2012 edition; retained in the 2022 edition).
ISO 3382-3 — Good vs insufficient guide values (informative)
Source: Sandy Brown (2014) summarising ISO 3382-3:2012 annex values. Unchanged in 2022.
| 기준 | 지표 | 좋음 (Good) | 부족 (Insufficient) | 체감 |
|---|---|---|---|---|
| D2,S | D2,S | ≥ 7 dB | < 5 dB | Speech attenuates well over distance |
| Lp,A,S,4m | Lp,A,S,4m | ≤ 48 dB | > 50 dB | Low cognitive load at 4 m |
| rD | rD | ≤ 5 m | > 10 m | Distraction zone is small |
Where Do "Grade A–D" Classifications Come From — It Is Not ISO
The "ISO 3382-3 Grade A" notation that frequently appears in open-office acoustic specifications and interior guides is in fact a classification not defined by ISO. ISO 3382-3 provides only the measurement methods for its four SNQs and informative guideline values; the A/B/C/D grading framework is an addition by external standards that have borrowed the ISO metrics.
The prime example is the Finnish Indoor Climate Classification 2018 (FiSIAQ). It takes ISO's D2,S · Lp,A,S,4m · rD and organizes them into S1–S3 or a four-tier scale tied to the national certification framework.
If a specification lists "Grade A," you must first identify which external classification — not ISO itself — defines that grade before measurement results can be compared against it.
Design Prescription — What to Address First in Each Space
D2,S and rD are functions of absorptive area. On the same floor plan, how ceiling and wall absorption are distributed determines the measured values. Here are five representative scenarios and their primary prescriptions.
Five scenarios — first-priority remedies for ISO 3382-3 metrics
New build — exposed or polished concrete ceiling
"The ceiling is bare concrete and reverberates"
Tenant fit-out — ceiling untouchable
"We cannot touch the ceiling. We can change desks"
Call centre / heavy video-call use
"All-day calls and video meetings — someone is always talking"
Existing space with hard floor (tile/concrete)
"Footfalls and chair drags are constantly distracting"
Glass partitions + compact floor plan
"Glass walls bounce sound into adjacent rooms"
Who Is Responsible for What — Phase-by-Phase Accountability
What ISO 3382-3 Cannot Capture — Limitations and Complements
ISO 3382-3 measures the acoustic response to "one person speaking." That is, the cumulative effect of multiple people speaking simultaneously (speech masking by speech) is not captured by the four SNQs.
Absorb first. Block next. Mask last.
In Summary — Measure → Design → Verify: One Loop
ISO 3382-3:2022 answers the user-experience question "how intelligible is my speech from the next desk?" with four single numbers. For new construction, embed D2,S and rD targets in the specification at the design stage; for renovation, treat ceiling absorption as the single largest lever. Measurements should follow the 2022 procedure using an omnidirectional source and an array of six or more receiver positions.
When material selection for open-office acoustic design is proving difficult — Tornex melamine foam (ceiling NRC 0.9+) and ANKER carpet tile (low-frequency absorption and impact sound attenuation) are the materials that translate this article's prescriptions into actual installation.
ISO 3382-3:2022 FAQ
Q1Can a space be "certified" to ISO 3382-3?
Q2Can 2012 measurements be reused under the 2022 edition?
Q3How does the new rC differ from rD?
Q4Does sound masking alone improve ISO 3382-3 numbers?
Q5The spec says "ISO 3382-3 Grade A". How should we respond?
Glossary
- SNQ (Single Number Quantity) — a metric that compresses spatial acoustic performance into a single number in an acoustic measurement standard. ISO 3382-3 defines four SNQs: D2,S · Lp,A,S,4m · rD · rC.
- D2,S — Spatial Decay Rate. The attenuation (dB) in A-weighted speech sound pressure level when the distance from the speaker is doubled. Higher values indicate absorption-dominant spaces.
- Lp,A,S,4m — A-weighted speech sound pressure level measured 4 m from the speaker (dB).
- rD — Distraction Distance. The distance from the speaker (m) at which STI falls to 0.5 or below.
- rC (added 2022) — Comfort Distance. The distance (m) at which A-weighted speech SPL falls to 45 dB or below. Consolidates D2,S and Lp,A,S,4m into a single distance metric.
- STI (Speech Transmission Index) — a single number on a scale of 0–1. Below 0.5: distracting; 0.5–0.7: audible; 0.7+: highly intelligible. Calculated per IEC 60268-16.
- NRC (Noise Reduction Coefficient) — arithmetic mean of absorption coefficients across the 250 / 500 / 1,000 / 2,000 Hz octave bands. Range: 0–1.
- RT60 — Reverberation Time. The time (s) required for sound pressure to decay 60 dB after the source stops. WELL v2 conference room requirement: ≤ 0.6 s.
- ABW (Activity-Based Working) — a working style in which individuals move between different workstations based on the type of task. Acoustic requirements vary by seat type, making a single ISO 3382-3 measurement insufficient on its own.
References
- ISO 3382-3:2022 — Acoustics — Measurement of room acoustic parameters — Part 3: Open plan offices (2nd edition).
- ISO 3382-3:2012 — 1st edition. The primary source for the "good" vs. "insufficient" guideline values cited in this article.
- Sandy Brown (2014) — BS EN ISO 3382-3:2012, is it worth the effort? Source for the 6% cognitive loss figure at STI 0.5.
- ASTM E1130-16 (2021) — Standard Test Method for Objective Measurement of Speech Privacy in Open Plan Spaces Using Articulation Index.
- WELL v2 — Reverberation Time (source for the RT60 ≤ 0.6 s conference room recommendation).
- WELL v2 — Sound Reducing Surfaces (source for the ceiling NRC ≥ 0.9 and wall coverage ≥ 25% at NRC ≥ 0.7 recommendations).
- Treble Technologies docs — Parameter definitions for open-plan office acoustics.
- FiSIAQ (2019, CLIMA conference) — Revised Finnish classification of indoor climate 2018. An example of an external four-tier classification system built on ISO metrics.