dB(A) vs dB(C) vs dB(Z): Frequency Weighting Explained

dB(A) vs dB(C) vs dB(Z) is a comparison of the three standardized frequency-weighting curves that sound level meters apply when converting raw acoustic pressure into a single decibel reading. A-weighting (dBA) suppresses very low and very high frequencies to approximate how the human ear perceives loudness at everyday listening levels; C-weighting (dBC) is nearly flat across the audible range and is reserved for high-level, peak, and low-frequency-rich noise; and Z-weighting (dBZ), or “zero weighting,” applies no frequency correction at all, giving a true unweighted physical reading. All three curves are defined in the international sound-level-meter standard IEC 61672-1:2013 (Electroacoustics — Sound Level Meters).

The letter after “dB” is not decoration — it tells you which filter shaped the number, and comparing readings taken with different weightings is one of the most common sources of confusion in acoustic reporting. A reading of “85 dB” with no letter is scientifically incomplete; “85 dB(A)” and “85 dB(C)” for the same noise event can represent very different real-world sound fields, sometimes with a gap of 20 dB or more when the noise is bass-heavy. Understanding why these curves exist, how they differ mathematically, and which one applies to which situation is essential background for anyone reading a noise report, an equipment datasheet, or a regulatory limit.

Why Do Sound Level Meters Apply Frequency Weighting at All?

Frequency weighting exists because the human ear is not equally sensitive to all frequencies, and a single unweighted decibel number does not, by itself, tell you how loud a sound will actually seem to a listener.

Human hearing sensitivity varies dramatically across the audible spectrum — roughly 20 Hz to 20 kHz for a healthy young ear — with peak sensitivity in the 2,000–5,000 Hz range (the region most important for speech intelligibility) and much lower sensitivity at the low-frequency and high-frequency edges of that range. This behavior was first mapped systematically by Harvey Fletcher and Wilden Munson in 1933 as a family of equal-loudness contours, later refined and standardized as ISO 226 (most recently revised in 2023). Two pure tones at the same physical sound pressure level — say, 100 Hz and 3,000 Hz — do not sound equally loud; the 100 Hz tone will seem noticeably quieter even though a microphone measures identical pressure.

Because raw, unweighted decibel measurements do not reflect this perceptual reality, standards bodies built correction filters directly into sound level meters. A weighting curve is a table of frequency-dependent gain/attenuation values (in dB) that is applied electronically or digitally to the incoming signal before the meter computes a single overall level. The result is a number that better matches either (a) how loud the sound is perceived to be at typical levels (A-weighting), (b) how loud it is perceived at high levels or its true peak pressure (C-weighting), or (c) the sound’s actual, unmodified physical energy (Z-weighting). This entire family of curves — A, B, C, D, and now Z — grew out of that same 1930s research and its later revisions, and is formally codified today in IEC 61672-1 and its U.S. counterpart, ANSI/ASA S1.4.

What Is A-Weighting (dB(A))?

A-weighting is a frequency-correction curve that attenuates low frequencies steeply and high frequencies moderately, while giving a slight boost around 2,000–4,000 Hz, in order to approximate the sensitivity of human hearing at ordinary sound levels. It was originally derived from the 40-phon Fletcher-Munson equal-loudness contour — the curve describing quiet sounds — but decades of field data have shown it correlates well with occupational hearing-damage risk across virtually all everyday sound levels, which is why it remains the mandated weighting for hearing-conservation and environmental-noise regulation today.

A-weighting is the only frequency weighting that IEC 61672-1 requires every compliant sound level meter to include; C- and Z-weighting are optional features (though common on professional instruments). Readings are written as dB(A) or dBA. Nearly every hearing-health regulation in the world is expressed this way: the U.S. National Institute for Occupational Safety and Health (NIOSH) recommends an exposure limit of 85 dB(A) averaged over 8 hours, and the Occupational Safety and Health Administration (OSHA) sets its permissible exposure limit at 90 dB(A) over 8 hours, both measured “on the A scale of a standard sound level meter at slow response,” per 29 CFR 1910.95. Environmental noise ordinances, traffic-noise regulations, and product-noise labeling almost universally use dB(A) as well. Full detail on why this specific number matters for hearing is covered in How Loud Is Too Loud? Safe Decibel Levels.

What Is C-Weighting (dB(C))?

C-weighting is a frequency-correction curve that is nearly flat across the audible spectrum — with only mild roll-off below about 31.5 Hz and above about 8 kHz — designed to approximate perceived loudness at high sound levels, where the ear’s low-frequency sensitivity increases compared to quiet conditions. Because it barely attenuates bass frequencies, C-weighting captures far more low-frequency energy than A-weighting does, which makes it the standard choice whenever the low-frequency content of a noise, or its true instantaneous peak pressure, matters more than its perceived loudness at moderate volume.

Readings are written as dB(C) or dBC. C-weighting is specified whenever regulations or engineering practice need to capture peak sound pressure levels rather than time-averaged loudness: the European Union’s Physical Agents (Noise) Directive sets a maximum permitted instantaneous peak sound level of 140 dB(C) (equivalent to a peak pressure of about 200 Pa) to protect workers against sudden, damaging pressure spikes such as impact noise or gunfire, and municipal noise ordinances sometimes use dB(C) specifically to regulate bass-heavy or impulsive sound — Atlanta’s noise code, for example, caps impulsive sound at 100 dB(C) at property lines. C-weighting is also common in audio and home-theater calibration, where engineers need a measurement that does not artificially suppress the deep bass content of film soundtracks and music, and in occupational hygiene, where comparing a location’s dB(C) reading against its dB(A) reading is a quick diagnostic for whether a noise source has significant low-frequency character that A-weighting alone would understate.

What Is Z-Weighting (dB(Z))?

Z-weighting, short for “zero weighting,” is a flat frequency response — nominally 0 dB gain from 10 Hz to 20 kHz, within a tolerance of about ±1.5 dB — that applies no perceptual correction whatsoever, providing an unmodified, purely physical measurement of sound pressure across the full audible range. It was formally introduced into IEC 61672 in 2003 specifically to replace the inconsistent “Linear” or “Flat” settings that different manufacturers had previously implemented with varying, non-standardized bandwidths — meaning two “flat” readings from different older meters were not always directly comparable, a problem Z-weighting was designed to eliminate.

Readings are written as dB(Z) or dBZ. Z-weighting is used whenever an analyst needs the true, unweighted acoustic energy at every frequency — for detailed octave-band or one-third-octave-band spectral analysis, environmental impact assessments that must characterize the full noise spectrum (rather than just its perceived loudness), architectural acoustics testing, and any situation where downstream analysts (rather than the meter itself) will apply their own frequency corrections, such as calculating Noise Criteria (NC) or Room Criteria (RC) curves from raw octave-band data.

dB(A) vs dB(C): What’s the Real Difference?

The core practical difference between dB(A) and dB(C) comes down to how each treats low-frequency sound, and the numbers make the contrast concrete. At 63 Hz, A-weighting subtracts roughly 26 dB from the raw signal, while C-weighting subtracts less than 1 dB; at 31.5 Hz the gap is even larger — A-weighting cuts about 39 dB, C-weighting only about 3 dB. In practical terms, this means A-weighting treats deep bass as almost inaudible for measurement purposes (matching how quiet, low-frequency tones are genuinely perceived by the ear), while C-weighting treats deep bass as nearly as significant as midrange sound — which matters for peak pressure, hearing damage from very loud low-frequency events, and detecting bass-heavy noise sources like traffic, HVAC rumble, or subwoofers that a dB(A) reading alone would substantially undercount.

Acoustic engineers and industrial hygienists sometimes exploit this gap directly, comparing the dB(C) and dB(A) readings of the same noise event (informally, the “C-minus-A” method): when the two readings are close together, the noise has a fairly flat, broadband spectrum well captured by A-weighting; when the dB(C) reading is many decibels higher than the dB(A) reading of the same event, it signals that a large share of the acoustic energy sits at low frequencies that A-weighting is deliberately designed to suppress — a useful flag for choosing appropriate hearing protection or further octave-band analysis. See How to Measure Sound: Sound Level Meters for how modern meters let you log both weightings simultaneously.

dB(C) vs dB(Z): When Do They Actually Diverge?

Because C-weighting is nearly flat across most of the audible spectrum, dB(C) and dB(Z) readings are often close for typical broadband noise. They diverge at the extremes of the spectrum: C-weighting rolls off below its −3 dB points at roughly 31.5 Hz and above roughly 8 kHz, while Z-weighting stays essentially flat all the way from 10 Hz to 20 kHz. This means for noise with substantial energy in the deep infrasound range (below ~20 Hz, such as wind turbine or HVAC rumble) or in the extreme high-frequency range near ultrasound, only Z-weighting captures that content without any built-in roll-off — which is exactly why Z-weighting is the standard choice for full-spectrum diagnostic and research measurements, while C-weighting remains the practical choice for peak-level regulatory compliance where a defined, standardized bandwidth is required.

Comparison Table: A-, C-, and Z-Weighting Correction Values by Frequency

The table below gives the standardized correction values (in dB) that A-weighting and C-weighting apply at representative octave-band center frequencies, per IEC 61672-1. A positive value means the raw signal is boosted; a negative value means it is attenuated. Z-weighting applies 0 dB at every frequency shown (by definition), which is why it is omitted from the numeric columns.

Notice the symmetry: both curves are normalized to 0 dB at 1,000 Hz (the reference frequency where all weightings agree), but A-weighting falls away steeply on both sides of the spectrum while C-weighting stays close to flat until the very edges. This single table explains nearly every practical difference discussed above — why a bass-heavy noise source reads much higher on dB(C) than dB(A), and why a mid-frequency tone (near 1 kHz, the range of a typical dial tone or timer beep) reads almost identically on both scales.

When Should You Use dB(A), dB(C), or dB(Z)? Application Table

Whatever Happened to B-Weighting and D-Weighting?

Two additional weighting curves — B-weighting and D-weighting — were part of earlier sound level meter standards but have since fallen out of active use. B-weighting was introduced alongside A- and C-weighting in the 1936 American standard as an intermediate curve intended for medium sound levels, sitting between A and C; in practice it saw almost no adoption and was dropped from the current IEC 61672 standard, though its definition survives in the withdrawn IEC 60651. D-weighting was designed specifically for measuring high-level aircraft noise, with a pronounced peak around 6 kHz reflecting how the ear integrates broadband noise (rather than pure tones) differently across the cochlea’s frequency-selective hair cells; it has since been superseded by loudness-corrected metrics such as EPNdB (Effective Perceived Noise in decibels) for certifying modern civil aircraft, and A-weighting is now used for light civilian propeller aircraft measurements instead. Neither B- nor D-weighting appears on modern sound level meters as a standard option — today’s instruments offer A, C, and Z, exactly the three curves compared in this article.

Myth vs Fact: Is “dB(A)” Just a More Modern Version of “dB”?

Myth: “dB(A) is simply the updated, more accurate version of plain dB — newer meters use dB(A) instead of dB because it’s a better unit.”

Fact: This is false. dB(A) is not a replacement or an improvement on the decibel itself — it is the same physical decibel scale with a specific, standardized frequency filter applied on top of it for a specific purpose (approximating perceived loudness at ordinary levels). Plain, unweighted dB (equivalent to dB(Z) on a modern meter) is not obsolete or “less accurate”; it is simply a different, complementary measurement that captures raw physical sound energy without any perceptual adjustment, which is exactly what is needed for spectral analysis, full-bandwidth peak measurements, and scientific reporting. Neither dB(A) nor dB(Z) is “more correct” in an absolute sense — each is the right tool for a different question: dB(A) answers “how loud does this sound to a typical human ear at typical levels?”, while dB(Z) answers “how much acoustic energy is physically present at every frequency, unmodified?” Using the wrong one for a given regulation or engineering task — for example, relying on dB(A) alone to characterize a bass-heavy industrial noise complaint — can significantly understate the low-frequency energy that is actually present and potentially damaging.

How Do You Know Which Weighting a Meter Reading Uses?

Any properly reported sound level should always state its weighting explicitly, because the same physical noise event can produce meaningfully different numbers depending on which curve was applied — the gap between dB(A) and dB(C) readings of the same bass-heavy source, as shown in the comparison table above, can easily exceed 15–20 dB. A reading of “85 dB” with no letter attached is scientifically incomplete: it is comparable to reporting a temperature without specifying Celsius or Fahrenheit. Professional-grade sound level meters compliant with IEC 61672-1 display the weighting used directly on-screen (A, C, or Z) and, in Class 1 precision instruments, typically allow simultaneous logging of more than one weighting at once — useful for exactly the C-minus-A diagnostic described earlier. See How to Measure Sound: Sound Level Meters for a full walkthrough of meter classes, calibration, and correct measurement technique.

Frequently Asked Questions

What does the “A” in dB(A) actually stand for?
“A” simply labels which of the standardized weighting curves (A, B, C, D, or Z) was applied — it does not stand for a word like “amplitude.” A-weighting is the specific curve, defined in IEC 61672-1, that attenuates low and high frequencies to approximate the ear’s sensitivity at everyday sound levels.

Is dB(A) always a lower number than dB(C) for the same sound?
Usually yes, but not always — it depends on the noise’s spectrum. Because A-weighting removes far more low-frequency energy than C-weighting does, any noise with meaningful bass content will read lower in dB(A) than in dB(C). For a noise concentrated purely around 1,000 Hz, however, the two readings are nearly identical, since both curves are normalized to 0 dB at that reference frequency.

Why do occupational noise regulations use dB(A) instead of dB(C)?
Because dB(A) has been shown, across decades of field data, to correlate well with the risk of noise-induced hearing loss and with how workers actually perceive workplace loudness at the sound levels typically encountered in industry. dB(C) remains reserved for specific cases — peak/impulse limits and diagnosing low-frequency noise — where A-weighting would understate the risk.

Can a sound be quiet in dB(A) but still loud or damaging in dB(C)?
Yes. A noise dominated by deep bass — such as heavy machinery rumble, a large HVAC system, or certain industrial impact noise — can produce a comparatively modest dB(A) reading while still registering a much higher dB(C) reading, because A-weighting is specifically designed to discount that low-frequency energy. This is precisely why impulse and peak noise limits are specified in dB(C) rather than dB(A).

Is dB(Z) the same thing as unweighted “linear” dB from older meters?
Conceptually yes, but not always numerically identical. Z-weighting was introduced by IEC 61672 specifically to standardize what older “Linear” or “Flat” settings did inconsistently across manufacturers, defining an exact flat response (0 dB ±1.5 dB) from 10 Hz to 20 kHz. An old meter’s “Linear” setting may have used a different, non-standardized bandwidth, so results are not guaranteed to match a modern Z-weighted reading exactly.

Do consumer smartphone sound-level apps use dB(A), dB(C), or dB(Z)?
Most default to dB(A), since it is the standard for everyday loudness and hearing-health reporting, though some professional-grade apps (including NIOSH’s own Sound Level Meter app) allow switching between A, C, and Z weighting, similar to a dedicated instrument. Always check the app’s settings and any accompanying documentation before comparing its readings to a regulatory limit.

Related Reading

Frequency weighting only makes sense once you understand the underlying decibel scale: start with What Is a Decibel (dB)? The Logarithmic Scale Explained and What Is Sound Pressure Level (SPL)? for the physics and formulas beneath every weighted reading. From there, see how these curves are actually applied in the field in How to Measure Sound: Sound Level Meters, how the weighted numbers translate into hearing-health guidance in How Loud Is Too Loud? Safe Decibel Levels, and how they connect back to the broader physics of hearing in The Human Hearing Range. For the full physics foundation, return to the pillar guide, Sound & Acoustics Fundamentals: The Complete Guide.

Understanding which weighting curve applies to a given number is a small habit that prevents a surprisingly large share of misread acoustic data — whether you are comparing a product noise spec, an environmental noise complaint, or a workplace exposure report.