What is STC?
Sound Transmission Class (STC) measures how well a wall, floor, or ceiling blocks airborne sound like voices, TV, and music. It's the primary rating for speech privacy between rooms. Unlike IIC which measures footsteps, STC focuses on sounds that travel through the air.
Higher STC = better blocking. STC 25 means loud speech is clearly understood; STC 50 means loud speech is inaudible; STC 60+ provides studio-grade isolation where most sounds are completely blocked.
Only the first row is a code requirement. The other three are the targets this tool ships with; we could not attribute them to a published standard, so they are labelled as targets rather than dressed up as requirements.
🎯 What STC is, and what it is not
Sound Transmission Class is a single number condensed from sixteen transmission-loss measurements. The wall is built into an opening between two reverberation chambers, noise is played on one side, and the difference in level is measured in one-third octave bands from 125 Hz to 4000 Hz. ASTM E90 governs how that measurement is made; ASTM E413 governs how the sixteen numbers become one. Both are paywalled and we have not read them; they are named here to identify what governs each step, and the numbers on this page come from a laboratory report we could read.
Three things follow that matter on site. STC is a laboratory rating of one partition with no flanking paths. It says almost nothing about bass, because the rating stops at 125 Hz — below which a home cinema does most of its damage. And it belongs to the assembly, not to a product: a sheet of drywall does not have an STC.
🔬 How E413 turns sixteen numbers into one
A fixed reference contour is slid up against the measured curve until it can go no higher without breaking two rules:
- the sum of the shortfalls below the contour, across all sixteen bands, must not exceed 32 dB; and
- no single band may fall more than 8 dB below the contour.
The STC is the value of the contour where it crosses 500 Hz. The contour itself rises 3 dB per one-third octave from 125 to 400 Hz, then 1 dB per one-third octave from 400 to 1250 Hz, then runs flat to 4000 Hz:
The deficiency counters in the results panel are those two rules made visible. Because a single band may only be 8 dB down, one deep dip — a coincidence notch, a resonance, a leak — holds the whole rating back while every other band is comfortable. That is why two walls with the same STC can sound completely different, and why the curve is worth more than the number.
One detail in the procedure does real work: the measured levels are rounded to the nearest decibel before the contour is fitted, not after. Running the 350 walls in IRC-IR-761 through this page's routine with their unrounded values reproduces the published rating for 198 of them; rounding first reproduces it for 333 of 350. If you fit a rating to your own third-octave data, round it first.
The 17 walls this page does not reproduce all miss the same way, and they all miss high. Every one of them contains a transmission-loss value reported at exactly half a decibel, and every one comes out one point above the published STC — never below. The disagreement is entirely a tie-breaking convention: rounding halves up matches 333 of 350 but errs only upward, rounding halves down matches 331 and errs only downward, and rounding halves to even matches 330 and errs in both directions. This page rounds halves up, so where it is wrong it is optimistic by one point. Read a result that sits exactly on a requirement as a result that may not meet it.
🧱 Where the assembly data comes from
The measured assemblies in this calculator are specimens from IRC-IR-761, Gypsum Board Walls: Transmission Loss Data — 350 walls tested by the National Research Council of Canada and published in full, one-third octave band by band, with the STC for each. Tests were run to ASTM E90-1990 in NRC's reverberation chamber suite and classified to ASTM E413-1987. Ratings across the series run from STC 32 to STC 69, and each preset carries its NRC specimen number so you can look the wall up.
One result in that report is worth more than any single rating. The same wall — one layer of 15.9 mm type X board each side, 38×89 mm studs at 400 mm o.c., 90 mm glass fibre, resilient channels at 600 mm — was completely rebuilt six times over about a year using new materials from the same lot each time. It measured STC 45 to 46. A second wall, identical but with two layers of board each side, was rebuilt the same way and measured STC 56 to 59. Same drawing, same laboratory, same crew: a three-point spread.
That spread is the honest error bar on any STC you did not measure yourself. If a design sits one point above a requirement, it does not meet the requirement — it is inside the noise.
📊 What the code actually asks for
IBC 2021 §1206.2 requires walls, partitions and floor-ceiling assemblies separating dwelling units and sleeping units from each other, and from public and service areas, to achieve a sound transmission class of not less than 50 when tested to ASTM E90. Where the assembly is field-tested instead, §1206.2 accepts a Normalized Noise Isolation Class of not less than 45, measured to ASTM E336. §1206.3 sets the matching impact requirement, covered by the IIC calculator.
The five-point gap is deliberate: field performance is expected to be lower. Flanking — sound going around the partition through the floor, the ceiling plenum, a shared duct, back-to-back outlet boxes, an unsealed perimeter — is absent in a laboratory and almost always present on site. Field-measured ASTC or NNIC and laboratory STC are different quantities; a lab STC 50 wall does not guarantee a field result of 50.
This calculator does not produce a compliance result. It classifies transmission-loss data to the E413 procedure, which is useful for comparing options and for reading a test report you have been handed. Code compliance requires a tested assembly, and a submittal requires the test report. Do not put a number from this page on a drawing as the assembly's rating.
🛠️ Mass law, and why it disagrees with real walls
The Mass Law tab estimates transmission loss for a single homogeneous leaf from its surface density alone: TL ≈ 20 log₁₀(m × f) − 47 dB for random incidence, with m in kg/m² and f in Hz. Tom Irvine's Acoustic Transmission Loss gives the same relationship in imperial units as TL = 20 log W + 20 log f − 33 with W in lb/ft², citing George Diehl's Machinery Acoustics (1973), and derives the constant as a normal-incidence value less 5 dB to account for random incidence. The two forms are the same equation: 20 log 4.882 is 13.8 dB, so −33 with pounds per square foot becomes −46.8, which is the −47 the calculator uses.
Doubling the mass, or doubling the frequency, adds 6 dB. That is the whole of the mass law, and it explains why mass alone runs out of road: to gain the next 6 dB you have to double again. What it does not capture is that mass and decoupling multiply. Two of the measured pairs above isolate the second layer of board cleanly, with the same board type and the same cavity fill on both sides of the comparison: on a staggered frame it was worth eight points, STC 47 to 55, and on a double frame another eight, 58 to 66. The apparently equivalent pair on shared studs — STC 33 to 38 — is not a clean comparison, because those two specimens also differ in board type and in cavity fill, so the five points cannot be attributed to the extra layer alone. The comparison that matters most needs no controls at all: a single layer each side on a double frame measured STC 58, twenty points above two layers each side on shared studs. Decoupling beat mass by a margin no amount of board would close.
Mass law has no term for a cavity, for stud coupling, for damping or for the coincidence dip, so it will disagree with any real stud wall. Use it to sanity-check a solid leaf — a slab, a sheet of glass, a masonry wall — and use test data for everything else.
❓ Common questions
How do you calculate STC rating?
You measure transmission loss in the sixteen one-third octave bands from 125 Hz to 4000 Hz, then fit the standard reference contour to that curve. The contour is raised until either the shortfalls below it add up to 32 dB or a single band falls 8 dB short, whichever happens first, and the STC is the contour value at 500 Hz. The measurement is made to ASTM E90 in a laboratory and the classification to ASTM E413. There is no way to calculate STC from a wall description alone: without transmission-loss data you are estimating, not rating.
What is a good STC rating for soundproofing?
The only figure that is fixed rather than fashionable is the code minimum: IBC 2021 section 1206.2 requires STC 50 between dwelling units and sleeping units when the assembly is laboratory tested to ASTM E90, or a Normalized Noise Isolation Class of 45 when it is field tested. For a home studio or a room where bass is the problem, STC is a poor guide at all, because the rating stops at 125 Hz. Aim at the specific noise you are trying to stop rather than at a number.
What does an STC rating of 50 mean?
It means a laboratory measured the assembly to ASTM E90 and the E413 contour fit landed on 50 at 500 Hz. In practice loud speech through such a wall is generally not intelligible, and often not audible over normal background noise. It does not mean you will get 50 on site: the laboratory has no flanking paths, and sound that travels through the floor, the ceiling plenum, a shared duct or an unsealed perimeter is not counted. That is why the code accepts a field-tested figure five points lower.
Is an STC rating of 27 good?
No. STC 27 is well below the STC 50 that IBC 2021 section 1206.2 requires between dwelling units and sleeping units, and it is below every one of the 350 gypsum-board walls in the National Research Council of Canada test series, the weakest of which measured STC 32. Ratings in the high twenties usually belong to a door or a window rather than to a wall, and in a partition the weakest element sets the result: a door at 27 in a wall at 55 gives you neither.
What type of wall is most effective at reducing sound transmission?
A double frame with a gap, insulated, is the most effective of the common constructions, because it removes the structural path between the two faces instead of just adding weight. In the National Research Council of Canada test series IRC-IR-761, a single layer of 13 mm gypsum board each side of a double 2×4 frame with a 25 mm gap and glass fibre in both cavities measured STC 58, while two layers each side on ordinary shared studs measured STC 38. The catch is thickness and cost, and a double wall is still only as good as its weakest penetration.
Transmission-loss curves, STC ratings and the rebuild-to-rebuild spread quoted on this page are from the National Research Council of Canada, IRC-IR-761, Gypsum Board Walls: Transmission Loss Data (Halliwell, Nightingale, Warnock & Birta, 1998), which reports tests made to ASTM E90-1990 and classified to ASTM E413-1987. The mass-law relationship and its random-incidence constant are from Tom Irvine, Acoustic Transmission Loss (Vibrationdata, 2012), citing George Diehl, Machinery Acoustics (Wiley, 1973). The code requirement is IBC 2021 §1206.2. ASTM E90, ASTM E413 and ASTM E336 are named to identify what governs each measurement and classification; all are paywalled and we have not read them.