NC vs RC vs NCY — Which Noise Criterion? Most Engineers Pick Wrong | AcousPlan

An office completes its post-construction acoustic commissioning. The HVAC noise level is measured at 38 dBA, well under the NC 40 limit specified in the mechanical engineer's scope of works. The space is signed off. Six months later, the facilities manager is receiving complaints about "rumbling" in the open-plan area near the main air handling unit. The tenants say the space "feels noisy" despite being within the specified limits.

The engineer is correct. The space is NC 38. The tenants are also correct. The space is unacceptable. Both can be true simultaneously — and the reason is that NC 38 is the wrong metric for characterising what the tenants are experiencing.

This article explains when NC, RC Mark II, and NCY are each the right tool, what each one misses, and how to select the appropriate noise criterion for a given space before design rather than after complaint.

Background: Three Criteria, Three Different Questions

NC, RC, and NCY are all single-number ratings derived from octave-band sound pressure level (SPL) measurements, typically from 63 Hz to 8000 Hz (NC) or 31.5 Hz to 8000 Hz (RC). All three are used to characterise HVAC background noise. They differ in what aspect of the noise they prioritise:

NC (Noise Criteria): Answers "Is the overall level acceptable for speech privacy and general comfort?" Developed by Beranek (1957), NC was designed for telephone communication environments. It evaluates the level of each octave band against reference curves and returns the highest exceedance. NC is widely used in the USA and is the most common criterion in architectural mechanical specifications.

RC Mark II (Room Criteria): Answers "Is the spectral balance of HVAC noise acceptable — and specifically, does it have a problematic character (rumbly or hissy)?" Developed by Blazier (1981, revised 1997), RC Mark II evaluates the same octave bands but uses a smoothed average of the mid-frequency bands (250–2000 Hz) as the rating index, then applies a Quality Assessment Index (QAI) to identify spectral imbalance. RC Mark II is the preferred method in ASHRAE Handbook — HVAC Applications for quality assessment.

NCY (Noise Criteria with Yen Modification): Answers "Does the NC curve underrate low-frequency noise?" NCY, proposed by Yen and adopted by some ASHRAE practitioners, applies a penalty factor to the 63 Hz and 125 Hz bands to account for their greater perceptual impact in enclosed spaces compared to what the NC contour weighting implies. NCY identifies spaces where NC appears acceptable but occupants find the low-frequency energy intrusive.

The NC Contour: What It Measures and What It Doesn't

The NC contour system works as follows. Plot the measured octave-band SPL on a graph with the NC reference curves. The NC rating is the integer value of the highest NC curve tangent to the measured spectrum — the point where one octave band just touches or exceeds its reference curve.

NC reference values for selected octave bands:

Recommended NC values by space type (ASHRAE):

The NC curve system has one significant structural weakness: the NC contours are relatively tolerant of low-frequency energy. At 63 Hz, NC 40 permits 53 dB SPL. At 500 Hz, NC 40 permits 31 dB SPL. This 22 dB difference between the NC 40 limit at 63 Hz versus 500 Hz means a spectrum with heavy 63 Hz content can technically pass NC 40 while producing an audible low-frequency rumble character.

This is exactly what happened in the office example at the top of this article. The space measured NC 38 — the 500 Hz band was the tangent, touching the NC 38 contour. But the 63 Hz band was at 51 dB — just under the NC 38 limit of 53 dB — generating a perceptible rumble that the NC system rated as acceptable. By NC, the space is fine. In the room, the rumble is real.

RC Mark II: The Better Tool for Quality Assessment

RC Mark II addresses the NC low-frequency tolerance problem by introducing the Quality Assessment Index (QAI). The procedure:

Step 1: Calculate the RC rating. Take the arithmetic average of the measured SPL at 500 Hz, 1000 Hz, and 2000 Hz. This is the RC(M) value. (The (M) suffix indicates the mark II procedure.)

RC(M) = (SPL₅₀₀ + SPL₁ₖ + SPL₂ₖ) / 3

Step 2: Calculate the QAI. Compare the measured spectrum to the "neutral" RC spectrum shape. The neutral spectrum (N) is a −5 dB/octave slope. A spectrum that closely follows this shape has QAI near zero. Departures from the neutral shape — whether toward excess low-frequency energy (rumbly) or excess high-frequency energy (hissy) — increase the QAI.

The QAI is computed as:

QAI = sqrt(Σ(ΔLᵢ²) / n)

Step 3: Classify the character.

• QAI < 5 dB: Neutral character — no perceptual concern
• QAI 5–10 dB: Marginal — may be perceptible
• QAI > 10 dB: Clearly perceptible character

Example: The Rumbly Office

NC calculation:

• 63 Hz: 55 dB, NC 40 limit = 53 dB → exceeds NC 40 by 2 dB → NC rating driven by 63 Hz → NC 42
• 500 Hz: 28 dB, below NC 40 limit of 31 dB ✓
• Final NC rating: NC 42

RC(M) = (28 + 22 + 18) / 3 = 22.7 → RC(M) ≈ RC 23

QAI calculation (simplified): The neutral contour at RC 23 predicts:

• 63 Hz: ~57 dB (neutral)
• 125 Hz: ~52 dB (neutral)
• 250 Hz: ~47 dB (neutral)
• 500 Hz: ~42 dB (neutral)

• 63 Hz: 55 dB measured vs 57 dB neutral → ΔL = −2 (below neutral)
• 125 Hz: 48 dB measured vs 52 dB neutral → ΔL = −4
• 250 Hz: 38 dB measured vs 47 dB neutral → ΔL = −9 (below neutral but less so)
• 500 Hz: 28 dB vs 42 dB neutral → ΔL = −14 (below neutral — spectrum is rumbly)

By NC, this room is NC 42 — marginal but possibly acceptable for an open office. By RC Mark II, it is RC 23 with a Rumbly character — which explains exactly the complaints about "low-frequency rumble" that NC never caught.

The NCY Modification: A Simpler Alternative

RC Mark II is the rigorous method but it is more complex to apply in practice. NCY provides a simpler approximation that addresses the same low-frequency problem.

The Yen (Y) modification applies a correction factor to the NC calculation at the 63 Hz and 125 Hz octave bands:

• If SPL₆₃ > NC reference at 63 Hz − 8 dB: Apply Yen correction
• Yen correction at 63 Hz: Y₆₃ = 0.5 × (SPL₆₃ − NC_ref₆₃ + 8) (when positive)
• Similarly for 125 Hz: Y₁₂₅ = 0.5 × (SPL₁₂₅ − NC_ref₁₂₅ + 5) (when positive)

Applying to the example above (NC 42 room):

• NC 42 limit at 63 Hz = 53 dB. Measured = 55 dB.
• Y₆₃ = 0.5 × (55 − 53 + 8) = 0.5 × 10 = 5 dB penalty
• NCY = NC 42 + 5 = NCY 47

When to Use Which Criterion

The critical design decision is this: if your mechanical specification uses NC only, you have specified a necessary condition for acoustic acceptability but not a sufficient one. A space can be NC 35 and rumbly. A space can be NC 30 and hissy. NC measures the peak octave band exceedance; it does not evaluate whether the spectral balance is perceptually acceptable.

For new construction and fit-out projects with quality requirements above "base building commercial", specify RC Mark II criteria in addition to NC. The ASHRAE Handbook — HVAC Applications (Chapter 49, 2019 edition) provides complete guidance on RC Mark II implementation and target values by space type.

Applying Criteria to Common Space Types

Open-Plan Offices

The appropriate criterion for an open-plan office depends on the speech privacy strategy. If sound masking is installed:

• NC 35–40 is the appropriate criterion for the HVAC background (the masking system provides the privacy-relevant background noise)
• RC Mark II is not needed for privacy purposes but is useful for quality check
• dBAeq 35–40: measured before masking activation

• NC 35 is the target
• RC Mark II should confirm the character is neutral (QAI < 5)
• Note: NC 35 alone without masking will not achieve Normal speech privacy per ASTM E1130 — see the speech privacy article

Conference Rooms and Boardrooms

Conference rooms require strict background noise control because speech intelligibility depends on the SNR between the speaker and the background. A boardroom with NC 40 background has a signal-to-noise ratio deficit at low speaking levels that increases listener fatigue and reduces comprehension.

Appropriate criteria:

• NC 25–30 (target)
• RC Mark II: neutral character required, QAI < 5 dB
• 31.5 Hz octave band: ≤ 65 dB (ASHRAE vibration criterion for speech use spaces)

Recording Studios and Broadcast Rooms

NC 15–20 is the specified criterion. At these levels, NC and RC Mark II align well because any exceedance at any frequency is significant. The 31.5 Hz criterion is most important here: studios near mechanical plant or lift shafts frequently fail at 31.5 Hz (structure-borne vibration) while passing all other octave bands.

Practical measurement: record at least 3 × 30-second measurement periods at 30-minute intervals to capture variable operating conditions (air handling unit cycling, variable speed drives, occupancy changes).

Hospitals and Healthcare

Healthcare acoustic requirements are governed by HTM 08-01 (UK), FGI Guidelines (USA), and AS/NZS 2107 (Australia). HTM 08-01 specifies NR 35 for wards and NR 30 for single bedrooms — using the NR criterion (similar to but not identical to NC). NR and NC differ primarily at low frequencies: NR curves are slightly more permissive at 63 Hz than NC curves.

For US hospitals following FGI guidelines: NC 35 maximum for patient rooms, NC 30 for ICU. RC Mark II quality check is appropriate to verify low-frequency character — patients are in sustained occupation and low-level rumble that an office worker would eventually tune out becomes chronically fatiguing in a healthcare environment.

Use AcousPlan's calculator to check compliance against multiple criteria simultaneously — enter octave-band measurements and the platform computes NC, NR, and RC(M) values alongside the ASHRAE target for the selected room type.

Practical Measurement Protocol

To correctly evaluate background noise against all three criteria, the measurement must capture the full octave-band spectrum at the design condition:

1. Measurement positions: Minimum 3 positions per space — one at each quarter-point of the room length, at seated head height (1.2 m above floor for offices, 0.6 m for bedrooms). Average across positions.

2. HVAC operating condition: Measure at full cooling load with all air handling units, fan coil units, and variable air volume boxes at their design operating condition. Also measure at minimum load — the noise character often shifts between maximum and minimum flow, and the minimum-load condition may generate more problematic low-frequency tones as variable speed drives hunt.

3. Background exclusion: Verify that the measurement is not contaminated by external noise (traffic, HVAC outdoor units, nearby plant) by measuring with all building systems off. If the "all off" condition is above 35 dBA, external noise is a significant contributor and must be addressed separately.

4. Frequency range: Measure at octave bands from 31.5 Hz to 8000 Hz, not 63 Hz to 4000 Hz. The 31.5 Hz band is relevant for vibration assessment; the 8000 Hz band is relevant for hissy character detection.

5. Duration: Minimum Leq 30 seconds per measurement position to average out transient HVAC cycling.

Common Specification Errors

Error 1: Specifying NC only without a maximum octave band level at 63 Hz. Fix: Add "63 Hz octave band SPL ≤ 50 dB" to the NC criterion specification. This catches rumbly HVAC systems regardless of their NC rating.

Error 2: Using dBAeq alone as the mechanical acoustic criterion. Fix: Supplement dBAeq with NC or RC(M) criteria. dBAeq is an energy average that de-emphasises low frequencies by 26 dB at 63 Hz (A-weighting curve). A space with 60 dB at 63 Hz contributes only approximately 34 dBA to the A-weighted level — a heavy bass condition that sounds unacceptable reads as modest on the A-weighted meter.

Error 3: Measuring at single occupancy condition. Fix: Measure at multiple HVAC operating states: maximum cooling load, minimum load, and heating mode if applicable. Fan noise character changes substantially with fan speed.

Error 4: Not checking for tonal components. NC, RC, and NCY all evaluate broadband octave-band levels. A discrete tone (e.g., fan blade pass frequency, compressor hum, transformer buzz) at 47 Hz that falls within the 31.5 Hz or 63 Hz octave band may be the dominant annoyance source while contributing only modestly to the overall octave-band level. A narrow-band (FFT) measurement is needed to identify tonal components, and tonal adjustments (typically +5 dB penalty) must be applied before the NC/RC assessment.

Summary

NC is the most widely used criterion for HVAC background noise in buildings. It is also the most likely to produce a specification that passes measurement and still generates occupant complaints — because it is relatively tolerant of low-frequency energy and provides no information about spectral character.

RC Mark II is the correct tool for quality assessment of completed HVAC systems in quality-sensitive spaces. It identifies the rumbly/hissy character that explains complaints in NC-compliant rooms. NCY provides a simpler low-frequency check that catches the most common failure mode — heavy 63–125 Hz energy — without the full RC Mark II procedure.

The decision framework is straightforward:

• Design phase: specify NC for budget and contract purposes, plus RC(M) ≤ [value] N (neutral) for quality
• Post-installation: measure both NC and RC Mark II; if QAI > 5, investigate source before handover
• Complaint investigation: always compute RC Mark II and NCY; NC alone will frequently mislead you