Background noise in buildings comes primarily from mechanical systems — air handling units, fan coil units, diffusers, ductwork, pumps, and chillers. This noise is continuous, broadband, and present in every occupied space served by HVAC. Unlike transient sounds (doors closing, people talking, traffic passing), mechanical noise is always there. If it is too loud, it makes concentration impossible. If it has the wrong spectral shape — too much low-frequency rumble or too much high-frequency hiss — it causes fatigue, annoyance, and complaints even at moderate levels.
Noise Criteria (NC) curves provide a standardised method for rating this background noise. An NC rating distils a complex octave-band noise spectrum into a single number that tells you whether a room is acceptably quiet for its intended use. This guide explains what NC curves are, how to use them, and when to choose NC, NR, or RC instead.
What NC Curves Are
NC curves were developed by Leo Beranek in 1957 as a practical tool for specifying acceptable background noise levels in occupied spaces. The system consists of a family of curves, each identified by a number (NC-15, NC-20, NC-25, NC-30, and so on in increments of 5), plotted on a graph where:
- The x-axis represents octave band centre frequencies: 63 Hz, 125 Hz, 250 Hz, 500 Hz, 1000 Hz, 2000 Hz, 4000 Hz, and 8000 Hz.
- The y-axis represents sound pressure level (SPL) in decibels (dB re 20 micropascals).
The NC Curve Values
The following table shows the octave band SPL values for the most commonly specified NC ratings:
| NC Rating | 63 Hz | 125 Hz | 250 Hz | 500 Hz | 1000 Hz | 2000 Hz | 4000 Hz | 8000 Hz |
|---|---|---|---|---|---|---|---|---|
| NC-15 | 47 | 36 | 29 | 22 | 17 | 14 | 12 | 11 |
| NC-20 | 51 | 40 | 33 | 26 | 22 | 19 | 17 | 16 |
| NC-25 | 54 | 44 | 37 | 31 | 27 | 24 | 22 | 21 |
| NC-30 | 57 | 48 | 41 | 35 | 31 | 29 | 28 | 27 |
| NC-35 | 60 | 52 | 45 | 40 | 36 | 34 | 33 | 32 |
| NC-40 | 64 | 56 | 50 | 45 | 41 | 39 | 38 | 37 |
| NC-45 | 67 | 60 | 54 | 49 | 46 | 44 | 43 | 42 |
| NC-50 | 71 | 64 | 58 | 54 | 51 | 49 | 48 | 47 |
How the Shape Works
The slope of the NC curves reflects the equal loudness contours of human hearing (ISO 226). At 63 Hz, the ear requires approximately 30 dB more SPL to perceive the same loudness as at 1000 Hz. The NC curves account for this by permitting higher absolute SPL values at low frequencies while being progressively stricter at higher frequencies.
This shape also reflects the spectral character of typical HVAC noise. Mechanical systems tend to produce more energy at low frequencies (fan noise, duct rumble) and less at high frequencies (diffuser hiss). The NC curve shape provides a reasonable envelope for typical mechanical noise spectra.
How to Determine the NC Rating of a Room
Determining the NC rating of a measured or predicted noise spectrum is straightforward. The process involves three steps.
Step 1: Measure or Calculate the Octave Band Spectrum
Using a Class 1 or Class 2 sound level meter with octave band analysis capability, measure the unoccupied background noise level in the room with all mechanical systems operating at their normal design condition. Record the SPL at each octave band centre frequency from 63 Hz to 8000 Hz.
If you are predicting noise levels during design (before the building is constructed), calculate the contribution of each noise source at the receiver position using manufacturer sound power data, duct attenuation, room effect correction, and distance attenuation.
Example measurement in a typical open plan office:
| Octave Band | 63 Hz | 125 Hz | 250 Hz | 500 Hz | 1000 Hz | 2000 Hz | 4000 Hz | 8000 Hz |
|---|---|---|---|---|---|---|---|---|
| Measured SPL (dB) | 58 | 49 | 43 | 38 | 34 | 31 | 28 | 25 |
Step 2: Plot Against NC Curves
Plot the measured octave band values on the NC curve chart. The NC rating is determined by the highest NC curve that is touched or exceeded by any single octave band value. It is not the average — it is the worst case.
Step 3: Read the NC Rating
In the example above:
- 63 Hz: 58 dB falls between NC-35 (60 dB at 63 Hz) and NC-30 (57 dB at 63 Hz) — exceeds NC-30
- 125 Hz: 49 dB falls between NC-30 (48 dB) and NC-35 (52 dB) — exceeds NC-30
- 250 Hz: 43 dB falls between NC-35 (45 dB) and NC-30 (41 dB) — exceeds NC-30
- 500 Hz: 38 dB falls between NC-30 (35 dB) and NC-35 (40 dB) — exceeds NC-30
- 1000 Hz: 34 dB falls between NC-30 (31 dB) and NC-35 (36 dB) — exceeds NC-30
- 2000 Hz: 31 dB falls between NC-30 (29 dB) and NC-35 (34 dB) — exceeds NC-30
- 4000 Hz: 28 dB equals NC-30 (28 dB) — meets NC-30
- 8000 Hz: 25 dB falls below NC-30 (27 dB) — below NC-30
The convention is to report the NC rating as the lowest NC curve that is not exceeded at any octave band. In this case, the room meets NC-35 but does not meet NC-30. The NC rating is therefore NC-35.
Common NC Targets by Room Type
The following targets are drawn from ASHRAE Handbook — HVAC Applications, Chapter 48 (Noise and Vibration Control), and represent recommended maximum background noise levels for various occupancy types.
Very Quiet Spaces (NC-15 to NC-25)
| Room Type | NC Target | Why |
|---|---|---|
| Recording studios | NC-15 | Any audible noise contaminates recordings |
| Concert halls | NC-15 to NC-20 | Pianissimo passages must be clearly audible |
| Broadcast studios | NC-15 to NC-20 | Microphone sensitivity picks up all background noise |
| Theatres | NC-20 to NC-25 | Actors must be intelligible without amplification |
| Churches and worship spaces | NC-25 | Quiet reflective environment for speech and music |
| Courtrooms | NC-25 | Legal proceedings require complete speech intelligibility |
These spaces require extremely quiet mechanical systems. Achieving NC-15 typically means using displacement ventilation, very low face velocity diffusers (below 1.5 m/s), acoustically lined ductwork, and vibration-isolated equipment. The mechanical system cost premium for achieving NC-15 versus NC-35 can be 30-50% of the total HVAC budget.
Quiet Spaces (NC-25 to NC-35)
| Room Type | NC Target | Why |
|---|---|---|
| Private offices | NC-25 to NC-30 | Speech privacy requires low background noise |
| Conference rooms | NC-25 to NC-30 | Video conferencing microphones are sensitive |
| Classrooms | NC-25 to NC-30 | ANSI S12.60 limits background noise to 35 dBA (~NC-30) |
| Hospital patient rooms | NC-25 to NC-30 | Patient recovery requires quiet environment |
| Libraries | NC-30 | Quiet study environment |
| Hotel guest rooms | NC-25 to NC-30 | Guest comfort and sleep quality |
| Residential living rooms | NC-25 to NC-30 | Relaxation and conversation |
Most commercial buildings target this range. It is achievable with standard HVAC design practices: properly sized ductwork, lined plenums, appropriate diffuser selection, and basic vibration isolation of major equipment.
Moderate Spaces (NC-35 to NC-45)
| Room Type | NC Target | Why |
|---|---|---|
| Open plan offices | NC-35 to NC-40 | Some masking noise is actually desirable for speech privacy |
| Retail spaces | NC-35 to NC-40 | Ambient noise from customers and music masks HVAC |
| Restaurants | NC-35 to NC-40 | Social noise naturally elevates ambient levels |
| Lobbies and corridors | NC-40 to NC-45 | Transient spaces with higher ambient noise tolerance |
| Gymnasiums | NC-40 to NC-45 | High activity noise masks mechanical systems |
The open plan office case is interesting because background noise in the NC-35 to NC-40 range actually improves speech privacy. If the background noise is too low (below NC-30), conversations are intelligible at greater distances, reducing privacy and increasing distraction. This is why many open plan offices install electronic sound masking systems to raise the background noise to a controlled, uniform NC-35 to NC-40.
Noisy Spaces (NC-45 and Above)
| Room Type | NC Target | Why |
|---|---|---|
| Computer server rooms | NC-45 to NC-55 | Occupied intermittently, equipment noise dominates |
| Mechanical plant rooms | NC-55 to NC-65 | Hearing protection may be required above NC-65 |
| Workshops | NC-45 to NC-55 | Tool and machine noise sets the ambient level |
| Kitchens (commercial) | NC-45 to NC-50 | Extraction systems and equipment generate high noise levels |
These spaces are not designed for sustained quiet work. The NC targets reflect the practical reality that mechanical and process noise will dominate regardless of the HVAC design.
NC vs NR vs RC: Three Rating Systems Compared
NC is the most widely used system in North America. But it is not the only one, and it is not always the best one. Two alternative systems address limitations in the NC methodology.
NR Curves (Noise Rating)
NR curves were developed by the International Organization for Standardization and are defined in ISO R 1996:1971 (later superseded by ISO 1996-1:2003 for environmental noise, though NR curves remain in use for building services noise). They are the preferred system in Europe, the UK, Asia, and Australia.
Key differences from NC:
- NR curves extend from 31.5 Hz to 8000 Hz (NC starts at 63 Hz). This extra low-frequency octave band is important for assessing rumble from large air handling units and cooling towers.
- NR curves have a slightly different shape, particularly at low frequencies, where they are somewhat more lenient than NC curves.
- NR values are approximately 5 dB lower than NC values for the same spectrum at mid-frequencies. An NR-30 room is roughly equivalent to an NC-35 room.
RC Curves (Room Criteria)
Room Criteria (RC) curves were developed by ASHRAE as a successor to NC, first published in 1981 and refined as RC Mark II in 1997. The RC method addresses two significant limitations of the NC system.
Limitation 1: NC ignores spectral balance. Two rooms can have the same NC rating but sound completely different. One might have a low-frequency rumble (boomy) while the other has a high-frequency hiss (sibilant). NC treats both the same because it only reports the worst-case octave band.
Limitation 2: NC does not assess low-frequency vibration. Below 63 Hz, mechanical noise can cause audible rattling and perceptible vibration in lightweight building elements. NC curves do not extend below 63 Hz and therefore cannot assess this risk.
The RC method addresses both limitations:
- The RC rating includes a quality descriptor: (N) for neutral (spectrally balanced), (R) for rumbly (low-frequency dominant), (H) for hissy (high-frequency dominant), and (RV) for rumbly with vibration risk.
- RC extends to 16 Hz and 31.5 Hz, covering the frequency range where vibration-induced rattling occurs.
When to use RC: RC is the recommended method in ASHRAE standards and is particularly valuable for:
- Spaces where spectral quality matters (recording studios, concert halls, conference rooms with sensitive AV equipment)
- Buildings with large mechanical plant that produces significant low-frequency energy
- Projects where vibration-induced noise is a risk (lightweight construction above mechanical plant rooms)
Quick Comparison Table
| Feature | NC | NR | RC |
|---|---|---|---|
| Origin | USA (Beranek, 1957) | ISO (1971) | USA (ASHRAE, 1981) |
| Frequency range | 63 - 8000 Hz | 31.5 - 8000 Hz | 16 - 4000 Hz |
| Spectral quality descriptor | No | No | Yes (N, R, H, RV) |
| Vibration assessment | No | No | Yes |
| Primary usage region | North America | Europe, UK, Asia, Australia | North America (ASHRAE projects) |
| Approximate equivalence | NC-35 | NR-30 | RC-33(N) |
How HVAC Design Affects NC Rating
The NC rating of a room is determined almost entirely by the mechanical system design. In a well-sealed building with no external noise intrusion, the background noise in an unoccupied room comes from three HVAC components.
Fan Noise
The air handling unit (AHU) fan is typically the dominant noise source. Fan sound power increases with airflow rate and static pressure. The relationship is approximately:
Lw increases by 5 dB per doubling of airflow and 10 dB per doubling of static pressure.
Reducing fan noise at the source is always more effective than attenuating it downstream. Strategies include:
- Selecting fans that operate at their peak efficiency point (lower specific fan power = lower noise)
- Using variable speed drives (VSDs) to reduce fan speed during part-load operation — a 50% speed reduction yields approximately 15 dB noise reduction
- Choosing backward-curved centrifugal fans over forward-curved fans (lower noise at the same duty point)
- Oversizing AHU casing to reduce air velocity past the fan and coils
Duct-Borne Noise
Sound generated by the fan travels through the ductwork to the occupied space. Attenuation occurs through:
- Natural duct attenuation: Unlined rectangular ductwork provides 0.1-0.6 dB/m attenuation depending on duct size and frequency. Smaller ducts attenuate more per metre.
- Duct lining: 25 mm or 50 mm acoustic lining (mineral wool or fibrous glass) inside rectangular ductwork provides 3-12 dB/m attenuation at mid and high frequencies but has limited effect below 250 Hz.
- Silencers (attenuators): Splitter silencers or cylindrical silencers provide 10-30 dB insertion loss across a defined frequency range. They are the primary tool for achieving low NC ratings in spaces served by long duct runs.
- Duct bends: Each 90-degree bend provides 1-7 dB attenuation depending on frequency and the presence of turning vanes.
Breakout and Diffuser Noise
The final component of room noise comes from sound breaking out through duct walls into the room and from turbulent airflow at the supply diffuser.
Duct breakout is significant in lightweight rectangular ductwork passing through or adjacent to the occupied space. Round ductwork has much higher transmission loss than rectangular ductwork (approximately 15 dB better at mid-frequencies) and should be used wherever breakout is a concern.
Diffuser noise is generated by turbulent airflow at the air supply terminal. It is controlled by limiting the face velocity. The following face velocities are recommended for different NC targets:
| NC Target | Maximum Face Velocity |
|---|---|
| NC-25 | 1.5 m/s |
| NC-30 | 2.0 m/s |
| NC-35 | 2.5 m/s |
| NC-40 | 3.0 m/s |
| NC-45 | 3.5 m/s |
Step-by-Step NC Assessment Worked Example
Consider a 30-person conference room requiring NC-30 compliance. The mechanical engineer has provided the following predicted octave band noise levels from the HVAC system:
| Source | 63 Hz | 125 Hz | 250 Hz | 500 Hz | 1k Hz | 2k Hz | 4k Hz | 8k Hz |
|---|---|---|---|---|---|---|---|---|
| AHU fan (at diffuser) | 45 | 38 | 32 | 27 | 23 | 20 | 17 | 14 |
| Duct breakout | 40 | 33 | 26 | 20 | 16 | 13 | 10 | 8 |
| Diffuser self-noise | 30 | 28 | 30 | 29 | 27 | 24 | 20 | 16 |
| Total (log sum) | 46 | 39 | 35 | 31 | 28 | 25 | 22 | 18 |
To combine multiple sources, use logarithmic addition: L_total = 10 log10(10^(L1/10) + 10^(L2/10) + 10^(L3/10)).
Now compare the total against the NC-30 curve (57, 48, 41, 35, 31, 29, 28, 27):
- 63 Hz: 46 dB vs NC-30 limit 57 dB — passes (11 dB margin)
- 125 Hz: 39 dB vs NC-30 limit 48 dB — passes (9 dB margin)
- 250 Hz: 35 dB vs NC-30 limit 41 dB — passes (6 dB margin)
- 500 Hz: 31 dB vs NC-30 limit 35 dB — passes (4 dB margin)
- 1000 Hz: 28 dB vs NC-30 limit 31 dB — passes (3 dB margin)
- 2000 Hz: 25 dB vs NC-30 limit 29 dB — passes (4 dB margin)
- 4000 Hz: 22 dB vs NC-30 limit 28 dB — passes (6 dB margin)
- 8000 Hz: 18 dB vs NC-30 limit 27 dB — passes (9 dB margin)
If the 1000 Hz band had shown 33 dB instead of 28 dB, it would exceed NC-30 (limit 31 dB) and the room would rate as NC-35. The corrective action would be to add duct silencer attenuation at 1000 Hz, increase the ceiling tile NRC (which primarily affects mid-frequency reverberant field contribution), or reduce the supply air velocity to lower diffuser noise.
Common Mistakes in NC Assessments
Mistake 1: Using dBA Instead of Octave Bands
A-weighted sound level (dBA) is a single-number metric that applies a frequency weighting curve to approximate human hearing sensitivity. It is useful for screening but insufficient for NC assessment because it hides the spectral detail. A room can be 35 dBA and still fail NC-30 if the spectrum has a peak in a single octave band. Always assess NC using octave band data, never dBA alone.
Mistake 2: Measuring with Occupants Present
NC ratings represent the background noise from building services, not from occupants. Measurements must be taken in the unoccupied room with all HVAC systems running at design capacity. Occupant noise (conversation, footfall, equipment) is excluded from the NC assessment.
Mistake 3: Ignoring the 63 Hz Band
Many sound level meters default to displaying 125 Hz as the lowest octave band. The 63 Hz band is critical for identifying low-frequency rumble from large fans, cooling towers, and transformers. Always ensure the measurement covers 63 Hz to 8000 Hz.
Mistake 4: Specifying NC Without Stating the Standard
NC-35 according to ASHRAE, NC-35 according to the original Beranek 1957 curves, and NC-35 according to various proprietary software tools can differ by 1-3 dB at individual octave bands due to rounding and interpolation differences. Always state the reference: "NC-30 per ASHRAE Handbook 2019, Chapter 48."
NC in the Context of Green Building Certification
WELL v2 Feature S07 (Sound)
WELL v2 specifies maximum background noise levels in terms of dBA (not NC directly), but the underlying performance is equivalent. WELL S07 requires:
- Private offices: 40 dBA maximum (approximately NC-35)
- Open plan offices: 45 dBA maximum (approximately NC-40)
- Conference rooms: 35 dBA maximum (approximately NC-30)
BREEAM Hea 05
BREEAM references BS 8233:2014, which specifies indoor ambient noise levels using NR curves (not NC). The target values depend on room type and are generally equivalent to NC values minus 5 (NR-30 is roughly NC-35).
LEED
LEED v4.1 EQ Credit: Acoustic Performance references ASHRAE standards and specifies NC targets directly. The credit requires that HVAC background noise does not exceed the ASHRAE recommended NC level for each space type.
Summary
NC curves are the foundational tool for specifying and assessing background noise in buildings. They translate a complex octave-band noise spectrum into a single number that architects, mechanical engineers, and building owners can understand and act on. The key takeaways:
- NC is a worst-case metric: The rating is determined by the single octave band that comes closest to exceeding the curve, not by the average across all bands.
- NC targets must match room function: NC-25 for conference rooms, NC-35 for open offices, NC-40 for retail. Specifying a single NC target for an entire building is a common error.
- Choose the right system for your region: NC for North America, NR for Europe and Asia, RC when spectral quality matters.
- HVAC design is 90% of the solution: Fan selection, duct sizing, silencer specification, and diffuser velocity control determine whether the NC target is met.
- Measure with octave band data in unoccupied rooms: dBA is insufficient, and occupant noise must be excluded.