TLDR: What ISO 717 Does and Why You Need to Understand It
ISO 717 is the international standard for converting detailed frequency-by-frequency sound insulation data into single-number ratings that can be specified, compared, and enforced in building codes. Part 1 (ISO 717-1:2020) covers airborne sound insulation and produces the weighted sound reduction index Rw (laboratory) and standardised weighted level difference DnT,w (field). Part 2 (ISO 717-2:2020) covers impact sound insulation and produces the weighted normalised impact sound pressure level Ln,w (laboratory) and standardised weighted impact sound pressure level L'nT,w (field).
These single-number ratings are the language of building acoustic specifications worldwide. When a building code says "party walls shall achieve DnT,w 45 dB," it means the field-measured airborne sound insulation, rated per ISO 717-1, must equal or exceed 45 dB. When a product manufacturer claims "Rw 52 dB," they mean the partition tested per ISO 10140, rated per ISO 717-1, achieves a weighted sound reduction index of 52 dB.
The challenge is that these numbers are more complex than they appear. Laboratory Rw and field DnT,w can differ by 3 to 12 dB for the same partition. The basic Rw rating does not capture low-frequency performance, which matters enormously for bass music and traffic noise. And the reference curve method means two partitions with identical Rw can have very different insulation at specific frequencies. This guide decodes the rating system, explains the gap between lab and field, and shows how to specify correctly.
The Field Story: The 8 dB Flanking Gap
An architecture firm in Birmingham designed a speculative office building with open-plan floors separated by a central core. Two tenants on the same floor required a full-height partition between their demised areas. The acoustic specification called for DnT,w 45 dB — a reasonable requirement for office-to-office separation that would prevent normal speech from being overheard.
The partition manufacturer recommended their proprietary system: double metal stud frame, two layers of 12.5mm plasterboard each side, 100mm mineral wool in the cavity. Laboratory test certificate: Rw 58 dB. The specification seemed to have a comfortable 13 dB margin above the 45 dB target.
Post-installation testing per ISO 16283-1 measured DnT,w 37 dB. Not 45, not 50 — thirty-seven. The tenants could clearly hear telephone conversations through the partition. The 8 dB shortfall represented a perceived loudness increase of approximately 6 times compared to the expected performance.
The investigation identified three flanking paths:
Path 1: Raised access floor (5 dB contribution). The 600mm raised access floor was continuous under the partition. Sound entered the floor void on one side, travelled under the partition through the 600mm air space filled with cables and services, and radiated from the floor panels on the other side. The partition sat on the raised floor, not on the structural slab. This single path reduced the effective insulation by approximately 5 dB.
Path 2: Ceiling void (2 dB contribution). The suspended ceiling was a standard mineral fibre grid system on both sides of the partition. The partition extended to the underside of the suspended ceiling but not to the structural soffit. Sound entered the ceiling void through the mineral fibre tiles (which have sound insulation of approximately Rw 12 dB), crossed the void, and exited through the tiles on the other side.
Path 3: Perimeter curtain wall (1 dB contribution). The exterior curtain wall mullions provided a rigid connection between the two offices. Structural vibration from airborne sound excited the mullion on the source side and re-radiated on the receiving side.
The remediation required extending the partition to the structural slab above and below the raised floor, installing a ceiling barrier above the partition (extending 2 m each side), and detailing the curtain wall connection with resilient isolation pads. Cost: GBP 45,000 and three weeks of disruption to both tenants.
The lesson: Rw 58 dB in the lab means nothing if the building provides flanking paths that bypass the partition entirely. The DnT,w target must be achieved in the field, and field performance depends on every path between the rooms — not just the direct path through the partition.
The Reference Curve Method: How Rw is Calculated
ISO 717-1 uses a reference curve shifting method to convert one-third octave band sound reduction index R(f) data into the single-number Rw rating. The procedure is:
- Measured data: Sound reduction index R in dB at each one-third octave band from 100 Hz to 3150 Hz (16 bands).
- Reference curve: A standard reference curve defined in ISO 717-1 Table 1, starting at 33 dB at 100 Hz and rising to 53 dB at 500 Hz, then remaining at 53 dB up to 1250 Hz, then rising to 55 dB at 2000 Hz and above.
- Shifting: The reference curve is shifted upward in 1 dB increments until the sum of unfavourable deviations (where the measured R is below the shifted reference curve) does not exceed 32.0 dB.
- Rw value: The value of the shifted reference curve at 500 Hz is the Rw rating.
Spectrum Adaptation Terms C and Ctr
Recognising that Rw alone does not capture performance against all noise types, ISO 717-1 Annex A defines two spectrum adaptation terms:
| Term | Noise Spectrum | Typical Sources | Frequency Emphasis |
|---|---|---|---|
| C | A-weighted pink noise | Speech, music, children, medium-speed traffic | Broadband, 100-3150 Hz |
| Ctr | A-weighted urban traffic noise | Road traffic, rail traffic, disco bass, aircraft | Low frequency emphasis, 100-3150 Hz |
These terms are added to Rw to produce adjusted ratings:
- Rw + C: Insulation against broadband noise sources (speech, general activity)
- Rw + Ctr: Insulation against low-frequency-dominant noise (traffic, music bass)
| Construction | Rw | C | Ctr | Rw+Ctr |
|---|---|---|---|---|
| 200mm concrete wall | 56 | -1 | -5 | 51 |
| 100mm concrete block + plaster | 45 | -1 | -6 | 39 |
| Double plasterboard on metal stud | 53 | -2 | -9 | 44 |
| Lightweight timber frame | 48 | -3 | -12 | 36 |
| Single glazing 6mm | 28 | -1 | -3 | 25 |
| Double glazing 6/12/6mm | 33 | -1 | -5 | 28 |
Note the timber frame wall: Rw 48 dB looks adequate, but Rw+Ctr of 36 dB is significantly lower. For a bedroom facade facing a busy road, specifying Rw 48 dB without checking Ctr would result in inadequate traffic noise insulation.
Calculate Now: Use AcousPlan's free calculator to verify your partition and facade designs meet sound insulation targets including spectrum adaptation terms.
Laboratory vs Field: Understanding the Gap
The most important concept in sound insulation is the gap between laboratory and field performance. Different metrics exist for each context:
| Context | Airborne Metric | Impact Metric | Standard |
|---|---|---|---|
| Laboratory (element only) | Rw | Ln,w | ISO 10140 (measurement), ISO 717 (rating) |
| Field (whole building) | DnT,w or R'w | L'nT,w or L'n,w | ISO 16283 (measurement), ISO 717 (rating) |
DnT,w (standardised level difference) is the most commonly specified field metric in European building codes. It normalises for receiving room reverberation time using a reference of 0.5 seconds, allowing comparison across rooms of different sizes. R'w (apparent sound reduction index) normalises for receiving room absorption area and is closer to the laboratory Rw concept but includes flanking.
The gap between Rw and DnT,w (or R'w) is caused by flanking transmission — sound energy reaching the receiving room via paths other than directly through the separating element. The main flanking paths in buildings are:
Flanking Path Contributions
| Path | Description | Typical Contribution |
|---|---|---|
| Floor/ceiling continuity | Structural slab connects source and receiving rooms | 3-8 dB |
| Side wall continuity | Adjacent walls provide rigid vibration path | 1-4 dB |
| Services penetrations | Pipes, ducts, cables through/around partition | 1-5 dB |
| Raised/access floors | Continuous cavity under partition | 3-6 dB |
| Ceiling voids | Continuous void above partition | 2-5 dB |
| Edge sealing gaps | Perimeter gaps at floor, ceiling, wall junctions | 1-10 dB |
The total flanking contribution is not a simple sum — it follows energy addition (logarithmic), and the dominant path sets the limit. If the floor provides a flanking path with an effective Rw of 40 dB, no amount of improvement to the direct partition path will push the overall DnT,w above 40 dB.
Rule of thumb for common construction types:
- Heavy concrete frame, masonry infill: Rw to DnT,w loss = 3-6 dB
- Steel/concrete frame, lightweight partitions: Rw to DnT,w loss = 5-10 dB
- Timber frame: Rw to DnT,w loss = 5-12 dB
ISO 717-2: Impact Sound Rating
Part 2 of ISO 717 rates impact sound insulation using the same reference curve concept but in reverse — lower numbers are better because L'nT,w represents noise reaching the receiving room.
The laboratory metric is Ln,w (weighted normalised impact sound pressure level), measured per ISO 10140-3 using a standard tapping machine on the floor specimen. The field metric is L'nT,w (weighted standardised impact sound pressure level), measured per ISO 16283-2 in situ.
Typical values:
| Floor Construction | Ln,w (lab) | L'nT,w (field) |
|---|---|---|
| 150mm concrete slab, bare | 78 | 80-85 |
| 150mm concrete + 50mm screed | 73 | 75-80 |
| 150mm concrete + floating floor (25mm resilient layer + 65mm screed) | 52 | 55-60 |
| 150mm concrete + floating floor (50mm mineral wool + 65mm screed) | 45 | 48-55 |
| Timber joist floor, bare | 82 | 85-90 |
| Timber joist floor + resilient bars + 2x plasterboard ceiling | 58 | 62-68 |
Impact sound is more sensitive to workmanship than airborne. A single rigid bridge through a floating floor — a screw penetrating the resilient layer, a pipe sleeve touching the screed — can increase L'nT,w by 5-10 dB. This is why field impact results are often worse than laboratory predictions.
Comparing Rw, STC, and DnT,w
For those working across international markets, the relationship between rating systems matters:
| Feature | Rw (ISO 717-1) | STC (ASTM E413) | DnT,w (ISO 717-1) |
|---|---|---|---|
| Context | Laboratory | Laboratory | Field |
| Frequency range | 100-3150 Hz | 125-4000 Hz | 100-3150 Hz |
| Reference curve | ISO 717-1 Table 1 | ASTM E413 contour | Same as Rw |
| Max single deficiency | No limit | 8 dB | No limit |
| Total unfavourable deviation | <= 32.0 dB | Implicit in method | <= 32.0 dB |
| Low-frequency adjustment | C, Ctr terms | None standard | C, Ctr terms |
| Includes flanking | No | No | Yes |
STC has a maximum 8 dB single-frequency deficiency rule that ISO 717 lacks. This means STC penalises severe dips at single frequencies more than Rw does. In practice, Rw and STC are within 1-2 points for most constructions, but lightweight constructions with resonance dips can show differences of 3-5 points.
Common Mistakes
Mistake 1: Specifying Rw when you mean DnT,w. Building codes specify field performance (DnT,w). Product data sheets show laboratory performance (Rw). Specifying "Rw 45 dB" for a party wall means you need a product with Rw of approximately 50-55 dB to account for flanking losses. Always clarify whether the target is laboratory or field, and budget for the flanking gap.
Mistake 2: Ignoring Ctr for facades. Specifying Rw 35 dB for a bedroom facade sounds adequate. But if the site faces a motorway, the relevant metric is Rw+Ctr. A lightweight facade with Rw 35 dB and Ctr -10 dB provides only 25 dB against traffic noise. Many complaints about road noise stem from facades that meet Rw targets but fail Rw+Ctr.
Mistake 3: Comparing products tested to different standards. Rw (ISO 717/ISO 10140) and STC (ASTM E413/ASTM E90) use different test standards, different frequency ranges, and different reference curves. They are not interchangeable. A product with STC 52 does not necessarily achieve Rw 52. Always convert using actual test data, not approximate equivalences.
Mistake 4: Treating sound insulation as additive. Two Rw 30 dB walls in series do not give Rw 60 dB. The combined insulation depends on the air gap between them, the coupling through the structure, and resonance effects. In some cases, two walls with an inadequate cavity can perform worse than predicted due to mass-spring-mass resonance.
Mistake 5: Forgetting impact spectrum adaptation (CI and CI,50-2500). Just as airborne Rw has C and Ctr, impact Ln,w has CI (for standard impact sources like footfall) and CI,50-2500 (extended to 50 Hz for heavy/soft impacts like children jumping). A floor that meets Ln,w requirements can still generate complaints from bass-frequency impacts if CI is not considered.
Summary
ISO 717 translates complex frequency-dependent sound insulation data into the single-number ratings that building codes, specifications, and product data sheets use worldwide. Understanding this standard means understanding:
- The difference between laboratory (Rw, Ln,w) and field (DnT,w, L'nT,w) ratings
- Why flanking transmission creates a 3-12 dB gap between lab and field
- How spectrum adaptation terms (C, Ctr) reveal performance against specific noise types
- Why comparing products by single-number rating alone can mislead
For every sound insulation specification, start with the field target (DnT,w or L'nT,w), add a flanking margin appropriate to the construction type (5-10 dB), select products with laboratory Rw exceeding that adjusted target, and detail every flanking path — floor, ceiling void, side walls, services, junctions. AcousPlan's acoustic calculator helps you model these interactions and verify compliance before construction.
Verify your insulation design: Use AcousPlan's free sound insulation calculator to check Rw, DnT,w, and flanking transmission for your building design.