What is NC (Noise Criteria)?
NC (Noise Criteria) rates background noise levels from HVAC systems, traffic, and other steady-state sources. Lower NC = quieter space. It considers the frequency spectrum because low-frequency rumble is more intrusive than high-frequency hiss at the same volume.
Target values by space: Recording studios need NC-15 to NC-25. Private offices typically need NC-35 to NC-40. Open offices can tolerate NC-40 to NC-45. Exceeding these targets causes distraction, fatigue, and speech interference.
About NC Curves
The rating is the highest curve the measured spectrum touches or crosses in any octave band โ the tangent method. The band that does it is the one to fix.
The curves allow more level at low frequency than at high because the ear is less sensitive there. Ratings are specified in ANSI/ASA S12.2, Criteria for Evaluating Room Noise, which is paywalled; we have not read it, and the octave-band values used here are the widely published NC table rather than a transcription from the standard.
๐ฏ Where the NC curves came from
Noise Criteria curves are Leo Beranek's. He published them as “Revised criteria for noise in buildings” in Noise Control 3, 19โ27 (1957), after earlier Speech Communication curves in the mid-1950s, and revisited them in Noise Reduction (McGraw-Hill, 1960). The family was built from Speech Interference Levels โ the plain average of the 500, 1000, 2000 and 4000 Hz octave bands โ that people had reported as acceptable in a survey of offices. The number attached to an NC curve is approximately its speech interference level.
That origin explains the shape and the limits. The curves are, in NTi Audio's description, “curves of approximate equal loudness” โ their shapes were set to be monotonic and to sit at loudness levels in phons about 23 units above the corresponding speech interference level. But approximate equal loudness is not the same as a target: the curves were never meant to describe the most pleasant spectrum, only octave-band levels that just permit satisfactory speech communication without becoming annoying. A room can sit on an NC curve exactly and still sound wrong, which is what the later RC and NCB families were built to catch. Ratings are specified today in ANSI/ASA S12.2, Criteria for Evaluating Room Noise. That standard is paywalled and we have not read it; the octave-band values in this calculator are the widely published NC table, not a transcription from the standard.
๐ฌ The tangent method, and the number it actually gives you
An NC rating is not an average. Each octave band of the measured spectrum is compared against the family of curves, and the rating is the highest curve that any single band touches or crosses. One band decides the whole result, which is why the calculator reports a limiting frequency alongside the rating. Fixing anything other than that band changes nothing.
The A-weighted figure beside the rating is that sum: the weighting is applied band by band and the results added on an energy basis. It is the A-weighted level of the spectrum you typed in, not a time-averaged Leq over a measurement period. Two spectra with the same A-weighted level can sit several NC points apart, because A-weighting discards most of the low-frequency energy the NC curves are watching.
The curve family here begins at NC-15. A room quieter than NC-15 in every band is reported as NC-15 rather than as a lower number, so this tool cannot distinguish between very quiet and extremely quiet spaces. That matters mainly for studios and concert halls, where the design target is at or below the bottom of the scale.
๐ Targets, and the one that is a requirement
Almost every NC target in circulation is a design convention. The eight room types in the panel are the defaults this tool ships with; they sit inside the ranges commonly used, but we could not attribute any of them to a published standard, so we do not claim one. Where a client, a specification or a local authority names a figure, that figure wins.
One related criterion is genuinely fixed, and it is A-weighted rather than NC. ANSI/ASA S12.60/Part 1-2010 (R2020), Table 1, limits background noise in an unoccupied, furnished core learning space to 35 dBA and 55 dBC for volumes up to 566 mยณ (20,000 ftยณ), from both interior and exterior sources, evaluated as the greatest one-hour average during teaching hours. Above that volume, and in ancillary spaces, the limits are 40 dBA and 60 dBC.
The A- and C-weighted pair is doing real work. A room can pass an A-weighted limit while a fan is producing enough low-frequency rumble to be thoroughly unpleasant; requiring both weightings catches the spectrum that A-weighting hides. The same logic is why the calculator reports whether your spectrum leans rumbly or hissy against the NC-35 curve shape.
๐ ๏ธ NC, NCB, RC and RNC
NC is the oldest and most widely quoted of several competing families, and its main weakness is that it says nothing about the quality of the noise โ two spectra that touch NC-35 in different bands can sound completely different. Later schemes address that directly. Room Criterion curves are W.E. Blazier's, from “Revised Noise Criteria for Application in the Acoustical Design and Rating of HVAC Systems”, Noise Control Engineering Journal 16(2), 64โ73 (1981); the RC reference curves are straight lines sloping at โ5 dB per octave through the stated value at 1000 Hz, and the rating carries a letter for rumble, hiss or neutral rather than a number alone. Beranek later published balanced NC (NCB), and Schomer and Bradley's test of proposed revisions for the National Research Council of Canada compares NCB and RC against measured annoyance and argues for a third family, RNC.
This calculator implements NC only. If a specification calls for RC, RNC or NR, an NC number is not a substitute and the two are not interchangeable.
โ Common questions
What are the noise criteria curves?
They are a family of octave-band curves, one per NC value, that a measured background-noise spectrum is compared against. Leo Beranek published them in 1957 in Noise Control, having derived them from Speech Interference Levels that people had reported as acceptable in a survey of offices, and ratings are specified today in ANSI/ASA S12.2. Each curve allows more sound at low frequencies than at high, because the ear is less sensitive there, and the number attached to a curve is approximately its speech interference level.
How to calculate NC rating?
Measure the octave-band sound pressure levels from 63 Hz to 8000 Hz with an octave-band analyser, at the listening position, with the building services running. Then compare each band against the NC curve family and take the highest curve that any single band touches or crosses. That is the tangent method: the rating is set by one band, not by an average, so the band that reached it is the one worth spending money on. An A-weighted single number will not give you an NC rating, because it discards most of the low-frequency information the curves depend on.
What is NC 30 noise criteria?
NC-30 means no octave band of the measured spectrum exceeds the NC-30 curve, which allows 57 dB at 63 Hz falling to 27 dB at 8000 Hz. It is the sort of level at which normal speech is comfortable at several metres and a quiet mechanical system is barely noticeable. It is a rating, not a requirement: the target for any particular room is a design decision, and we could not attribute the common NC-30 recommendation for private offices to any published standard.
What is NC 40 sound?
NC-40 allows 64 dB at 63 Hz falling to 37 dB at 8000 Hz. In practice it is an audible but unobtrusive background โ comfortable for open-plan work, where a little steady noise actually helps mask conversations, and too loud for a room where people need to listen carefully. Whether it is acceptable depends entirely on the use of the room; NC-40 is a description of a spectrum, not a pass or fail.
What is the recommended NC level for a conference room?
There is no code figure for a conference room, and we could not attribute the commonly quoted NC-30 to NC-35 range to a published standard, so this tool treats NC-35 as a default rather than a requirement. The nearest thing to a fixed criterion for a room where speech has to be understood is ANSI/ASA S12.60 Part 1, which limits background noise in an unoccupied furnished classroom to 35 dBA and 55 dBC. That is an A- and C-weighted limit rather than an NC rating, and the two are not interchangeable.
The NC curve family is attributed to L.L. Beranek, “Revised criteria for noise in buildings”, Noise Control 3, 19โ27 (1957) and Noise Reduction (McGraw-Hill, 1960), as cited in NTi Audio's Noise Curves application note, which also records the survey of office Speech Interference Levels the curves were built from. RC curves are W.E. Blazier, Noise Control Engineering Journal 16(2), 64โ73 (1981); the โ5 dB per octave description and the NCB/RC/RNC comparison are from Schomer & Bradley, A test of proposed revisions to room noise criteria curves (NRC Canada / NCEJ 48, 2000). Background-noise limits for learning spaces are from ANSI/ASA S12.60/Part 1-2010 (R2020), Table 1. A-weighting values follow the weighting function of IEC 61672. ANSI/ASA S12.2 and IEC 61672 are named to identify what governs each rating; both are paywalled and we have not read them, so the tabulated NC levels here are the widely published values rather than a transcription from the standard.