Soundproofing vs. Acoustic Treatment – STC, NRC & Why Rooms Echo

Soundproofing and acoustic treatment solve two different problems. Soundproofing keeps sound from passing between rooms through mass, airtight construction and layers held apart, and acoustic treatment controls how sound behaves inside one room, absorbing or scattering it so that speech stays clear. A wall is rated by its sound transmission class, STC, and an absorber by its noise reduction coefficient, NRC. Foam and fabric panels are porous absorbers and count as treatment. Their absorption rating says nothing about how much sound they keep out of the next room. A neighbor’s television coming through a party wall calls for soundproofing. A new open-plan room that rings, or bass that booms in one seat and fades in the next, calls for treatment. Concert halls have taught that second lesson for more than a century. Wallace Clement Sabine was called in to rescue an unusable Harvard lecture room in 1895, and Lincoln Center’s first concert hall, opened in 1962, went through two major reconstructions in the 60 years that followed.

Soundproofing and STC Ratings

Sound transmission class, STC, is the single number ASTM E413 assigns to a wall or floor from laboratory measurements of the airborne sound passing through it. It rises with mass, airtight construction and decoupling. A single layer blocks sound mainly by its weight per unit area, and each doubling of that weight adds about 6 decibels.

A 150-millimeter slab of poured concrete reaches about STC 55, and National Research Council Canada treats that as the practical limit of one layer. Higher ratings come from double construction, two heavy layers with an airspace between them, such as a stud wall with two layers of 5/8-inch gypsum on each side. With no rigid link between the two faces, the air trapped between them acts as a spring. Above the resonance of that spring, the wall’s isolation climbs much faster than its weight alone would give. Insulation in the cavity added about 8 STC points to gypsum-board walls in the council’s tests, and a rigid connection between the layers works like an electrical short circuit.

Sound also travels around a wall, through the floors, ceilings and walls that surround it, and in a poor design those flanking paths carry more energy than the wall itself. A laboratory STC is an upper limit, and the same partition built in a real building can rate lower. For over 50 years North American codes set a minimum STC for the separating assembly alone. Effective noise control needs a requirement that counts flanking as well, such as the apparent sound transmission class, ASTC.

For walls and floors between dwelling units, California’s 2025 building code requires STC 50 in a laboratory test or NNIC 45 on site. Floor-ceilings must also reach IIC 50 for impact sound, or NISR 45 in the field. The requirements sit in Section 1206, numbered the same in the 2024 International Building Code. National Research Council Canada puts the effective rating dwellings need for reasonable privacy higher, at STC 55. The edition that applies depends on what a jurisdiction has adopted, and a house built under the IRC meets a sound provision only where the optional sound appendix is adopted.

What STC Ratings Mean

An STC rating is roughly the decibel reduction a partition provides, though not equally across frequencies, and low-frequency sound is harder to block. STC tracks speech, radio and television noise only in a general way, and ASTM E413 calls it unsuitable for machinery, music and traffic. Facades, windows and doors also carry OITC, a rating against aircraft, road and rail noise, and floors carry IIC for impact noise, measured with a tapping machine that does not simulate a person walking.

Two published scales describe what speech sounds like through a wall at each rating, about five points apart. The National Bureau of Standards’ 1976 homeowner guide has normal speech easily understood through an STC 25 wall, loud speech audible but not intelligible at 40 and shouting barely audible at 50. A scale an acoustical consultant published through the Acoustical Society of America in 2016 has loud speech audible but not intelligible already at 35. It puts STC 60 as the minimum to keep amplified sound from being heard. The 1976 guide put most interior walls inside a typical home or apartment at STC 30 to 35.

In a National Research Council Canada survey of 600 residents and their 300 party walls, published in 1999, the effective ratings measured in the buildings ran from STC 38 to 60. Below about 50, how often residents heard their neighbors did not change with the rating, and above it they heard them less as the rating rose. The paper recommends STC 55 as a realistic goal and STC 60 as the more ideal one.

How to Soundproof an Existing Wall

Seal the Gaps and Outlet Boxes

The first work on an existing wall is sealing its gaps, because any gap that leaks air also leaks sound. The joints where a wall meets the ceiling, the floor and the other walls are the usual flanking paths. Perimeter gaps and penetrations take an acoustical sealant meeting ASTM C919. Holes for wiring and pipes are sealed or caulked, since sound passes through insulation stuffed into a hole on its own. A hollow-core door blocks sound poorly, and where privacy matters it gives way to a solid or insulated-core door gasketed at its edges, as California’s code asks of new dwelling separations.

Outlet boxes leak too. In National Research Council Canada tests on a double-stud wall rated STC 55, untreated metal boxes set back to back cut the rating to 51. Gasketed airtight boxes left it at 55. Offsetting the boxes by at least 400 millimeters and filling the cavity with absorption shrinks the penalty.

Add Mass, Decouple and Damp

A sealed wall gains isolation from added mass, from decoupling the board on resilient channel, or from a damping layer between two sheets of gypsum. In the National Bureau of Standards’ table of approximate ratings, a 2×4 wall with 5/8-inch board nailed to each side rates about STC 35. The same wall with two layers on each side rates about 40. In a National Research Council Canada study of 70 gypsum walls, doubling the layers on one side of a decoupled, insulated wall typically added 4 to 5 points.

Resilient channel holds the board off the studs. The council’s 1985 digest puts drywall fixed directly to both sides of wood studs at about STC 33. Walls with resilient channels on one side and absorber in the cavity rate 48 to 56 in the same digest, depending on the number of board layers. The channel works only while the board hangs on it alone. A screw long enough to reach the framing, a channel mounted upside down or a cabinet bolted through to the studs short-circuits it. In the council’s later tests, spacing the channels 610 millimeters apart instead of 406 took one wall from STC 50 to 55, as fewer channels bridged less sound.

Damping is the third route. A 2011 National Research Council Canada test used a wood-stud wall with two layers of 16-millimeter board on each side, and a viscoelastic layer between the sheets on both faces raised it from STC 44 to 52. The paper does not name the product, and the wall it tested had no resilient channel.

Mass loaded vinyl is a limp, heavy sheet sold by weight, 1 pound per square foot at ⅛ inch in one supplier’s range. Two seller-commissioned tests give figures for different specimens. A bare sheet of 1.18 pounds per square foot rated STC 27 on its own. A steel-stud wall with insulation, one layer of 5/8-inch Type X board on each face and a lighter vinyl layer on one side rated STC 55. No matching wall was tested without the vinyl, so neither result shows how much the vinyl adds.

Soundproof Windows

A window isolates better with a wider air space between sealed panes and worse when it opens, in National Research Council Canada’s 1988 laboratory digest on wood-framed glazing. Sealed double glazing gains about 3 STC points each time the air space doubles, and an openable window with good weatherstripping usually rates 3 to 5 points below the same window sealed. Lightweight metal frames gave lower results than wood, and single glazing gained very little as the glass thickened.

Triple glazing performed essentially like double in those tests unless the panes were spread very far apart. Laminated glass can give much higher transmission loss above the glass’s coincidence frequency through the damping in its plastic interlayers, though the gain may be drastically reduced at typical Canadian winter temperatures. In a worked example for a bedroom on a busy road, New Zealand’s building research body BRANZ uses a double window with 6- and 4-millimeter panes 110 millimeters apart. Closed and sealed, it brings about 75 decibels outside down to 31 inside.

Acoustic Treatment and NRC Ratings

Acoustic treatment absorbs or scatters sound inside a room, and absorbers are rated by NRC, the average share of sound they remove at 250, 500, 1,000 and 2,000 hertz. ASTM C423 measures that share in a reverberation room from how quickly sound decays with the sample inside. Its current single number is the sound absorption average, SAA, across twelve bands from 200 to 2,500 hertz, and laboratories still report both.

An NRC of 0.8 means a material removes, on average across those four bands, about four fifths of the sound that strikes it, and a higher NRC means more absorption across them. A coefficient above 1.0 is a test artifact, caused by diffraction at the sample’s edges. NRC and STC measure different things in different tests, absorption inside a room and isolation between rooms, so one cannot be read from the other.

Foam is an easily processed, low-cost absorber, but its performance between 100 and 800 hertz is often limited, and the European rating of ISO 11654 says nothing at all below 250 hertz. Isolation still belongs to the wall or floor assembly, whatever soundproofing materials it is built from.

Diffusion, the other treatment, scatters sound instead of absorbing it. Manfred Schroeder, a physicist at Bell Telephone Laboratories, proposed in 1975 that hard walls be shaped to number sequences so that they scatter reflections widely. Four years later he proposed ceilings that send early sound sideways to listeners, as tall, narrow old halls do. On a rear wall, absorption or diffusion keeps a speaker’s voice from bouncing back to the front.

Why Rooms Echo and How to Reduce It

Rooms echo when their surfaces send back most of the sound that reaches them, and in an open-plan room finished in drywall, glass and poured concrete nearly every surface does. Half-inch drywall on studs absorbs less than a tenth of the sound above 500 hertz, and concrete and brick absorb almost none between 125 hertz and 4 kilohertz. Soft materials work mostly in the treble. Heavy carpet takes little of the bass. Velour drapery absorbs half or more of the sound above 500 hertz.

Carpet added to a classroom floor does not significantly shorten reverberation, especially at low frequencies, though it quiets chairs and desks scraping. A sharp hand clap finds flutter echo, a ringing between two parallel hard walls. The cure is absorbent panels on one or both walls, staggered so each faces an untreated surface, or walls splayed at least eight degrees out of parallel. Drapery and panels handle the treble. The bass needs absorbers with depth behind them, or membrane absorbers tuned to it.

Where to Put Acoustic Panels, and Bass Traps

Acoustic panels go on one of each pair of parallel walls, staggered so each faces an untreated surface, and on the rear wall, where absorption or diffusion keeps a speaker’s voice from bouncing back to the front. They work best spread around the room rather than gathered on one wall, the floor or the ceiling. Bass traps are inherently large and must cover a significant share of the room’s surface to work.

Porous panels need depth to reach the bass. A thin layer fixed flat to a hard wall absorbs poorly, because the air barely moves right at the surface. Spaced off the wall, it absorbs well, best where the gap is about a quarter wavelength. Glass-fiber board on a solid backing absorbs 0.22 at 250 hertz at 1 inch thick, and 0.78 at 2 inches.

Below the Schroeder frequency, named after Manfred Schroeder, a room’s resonant modes are pronounced and audible. Small home listening rooms typically suffer bass booms that differ from seat to seat. In a critical listening room, early reflections are those reaching the listener within 15 milliseconds of the direct sound, and the European Broadcasting Union asks that they stay at least 10 decibels below it.

One layer of drywall on studs is a useful bass absorber, the same mechanism that let the airspace behind Lincoln Center’s old side walls drain the bass. Membrane absorbers work in the low range where porous panels are weak. Multiple subwoofers can cancel room modes only for the sound the system itself plays, so in a room for live music electronic correction does not replace acoustic treatment.

Reverberation Time, RT60, and Sabine’s Lecture Room

Reverberation time, written RT60, is how long a sound takes to fade by 60 decibels, and it is measured band by band under ISO 3382-2, because it changes with frequency as absorption does. Wallace Clement Sabine worked out how to predict it. In 1895 Harvard asked him to correct the lecture room of the new Fogg Art Museum. The room had been found impracticable and abandoned as unusable. In Sabine’s account, the room copied Sanders Theatre, the best auditorium in Cambridge, and was the worst.

Sabine timed how long an organ note stayed audible after it stopped, about 5.6 seconds in the empty room. He then carried seat cushions over from Sanders Theatre and laid them out, and every added length of cushion shortened the decay. Across rooms whose sizes differed by a factor of nearly 150, he found that reverberation time multiplied by total absorption is proportional to volume, the relation now called the Sabine equation.

Boston’s Symphony Hall, opened on 15 October 1900, was designed with the equation, and the Boston Symphony Orchestra calls it the first hall built on scientifically derived acoustical principles. The orchestra’s founder Henry Lee Higginson and his architects, McKim, Mead & White, brought Sabine in, and he refused payment. The orchestra’s exhibit gives the target as 1.9 to 2.1 seconds. Sabine’s own calculation, on his measure of how long a sound stays audible, came to 2.31 seconds, and the occupied hall runs 1.9 seconds as reverberation time in Leo Beranek’s figures.

Rooms for speech need far less. The reverberation-time standard of ANSI/ASA S12.60-2010 for classrooms of up to 283 cubic meters is 0.6 seconds, readily adaptable to 0.3 for children who are deaf or hard of hearing. In many American classrooms, listeners with normal hearing catch only three words in four. Lecture halls work at about 1 second and a broadcaster’s reference listening room at 0.2 to 0.4.

Rooms That Failed Through the Ear

Philharmonic Hall, the first venue to open at Lincoln Center, was found particularly lacking in bass warmth when it opened in September 1962. Yet Beranek’s measurements gave the bass the longest reverberation, 3.0 seconds at 125 hertz. Leo Beranek had designed it as a 2,400-seat shoebox modeled on Boston’s Symphony Hall. Lincoln Center had the architect Max Abramovitz add at least 200 seats over Beranek’s objections, widening the hall toward the rear.

Manfred Schroeder led the measurements Bell Laboratories made in the hall, with Bishnu Atal, Gerhard Sessler and James West, and blamed Beranek’s ceiling clouds for filtering out the missing frequencies. An advisory committee that included Schroeder drew up a plan of its own. When it prevailed, Beranek formally disavowed the work, in September 1963. Heinrich Keilholz replaced the clouds with a stepped wood ceiling in 1969, and the critic Harold Schonberg called the result a complete success, though he still found the bass weaker than desired.

Avery Fisher’s gift renamed the building in 1975 and paid for a rebuild by the acoustician Cyril Harris, with Philip Johnson as architect. Audiences returned in the fall of 1976 to far shorter reverberation than the hall of 1962 had, 1.6 seconds at 125 hertz against 3.0, the largest drop of any band. Harold Schonberg greeted the new hall as opening on an acoustical high note, yet critics went on complaining that the sound was not up to standard.

The last reconstruction gutted the room to the walls of the 1962 hall, pushed the stage almost 25 feet into the auditorium and cut the seats to about 2,200. Its side walls became beech panels on a solid backing, replacing an airspace that had weakened the bass, and its balconies were faceted to scatter sound. David Geffen Hall reopened in October 2022, designed by Diamond Schmitt Architects and Tod Williams Billie Tsien Architects with the acoustic consultancy Akustiks.

Paul Scarbrough, whose firm did the acoustics, calls the critics’ response overwhelmingly positive. He writes that the hall needed 60 years, multiple minor renovations and two major reconstructions to reach acoustics worthy of its orchestra. The critics’ published verdicts were more guarded. Zachary Woolfe found the hall “mightily improved” but feared an almost clinical feeling, and Justin Davidson called the first results “provisionally encouraging and firmly inconclusive”.

In London, the Royal Albert Hall’s notorious echo came from the shape of its dome. Glass-fiber diffusers twelve feet across, designed on Kenneth Shearer’s investigations, were hung beneath it at the beginning of 1969. They nearly eliminated the echo and sent grazing reflections toward the gallery, and glass wool laid on top of them brought the reverberation down to an acceptable value.

Hearing a Room Before It Is Built

Designers can now hear a hall before it is built through auralization, the technique that makes a computer model of a room audible. Computer models of room sound date from the late 1950s, and Schroeder and colleagues set out the principles of auralization in 1962, decades before the technique had a name. In 1967 Bishnu Atal and Schroeder traced rays through two-dimensional rooms. A 1968 paper by A. Krokstad, S. Strøm and S. Sørsdal introduced the first method that computed room responses by ray tracing for different listening positions.

Schroeder’s 1973 review predicted that computers would change concert hall design. Six years later J. B. Allen and D. A. Berkley simulated a small rectangular room by the image method and published the code, a method exact only for rectangular rooms with rigid walls. By the early 1990s Kleiner, Dalenbäck and Svensson were describing auralization by its present name.

Research now tries to predict a room’s acoustics from a single image, the sound that an AI-assisted render leaves out. Image2Reverb, presented in 2021, generates a room’s acoustic response from one picture, and its authors warn that many images portray size, shape and materials inaccurately. A 2025 classroom study stored absorption coefficients in a BIM model’s material properties and let a generative-design optimizer choose materials and room height to balance reverberation against cost.

In the Grand Hall of the Elbphilharmonie in Hamburg, the walls were generated panel by panel. Herzog & de Meuron worked with the acoustician Yasuhisa Toyota on its white skin of some 10,000 individually milled gypsum fiber panels. Their flat points reflect sound, and their deep indentations scatter it. Benjamin Koren’s firm One to One programmed the panels, and by Koren’s account no two absorb or scatter sound alike. Once the parameters were set, Koren said, “I hit play and it creates a million cells, all different and all based on these parameters.” The architects also required the skin to look consistent throughout the room.

Acoustic Slat Panels and Perforated Wood

Acoustic slat panels absorb mainly through the absorbent behind the slats. In measurements at Aalto University, slats with 18 to 50 percent open area changed almost nothing below 250 hertz, where the absorbent behind them set the absorption. Above 250 hertz absorption rose with open area, and slats of varying height absorbed more than uniform ones. A slat panel over felt, laid flat on the test surface, rated NRC 0.55.

A perforated panel in front of a wall works as an array of Helmholtz resonators, each hole the neck of a small resonator over the air behind. Its tuning follows the open area, the cavity depth and the length of the holes. In measurements of finely perforated wooden panels, absorption peaked between 200 and 500 hertz. Thicker panels or wider spacing moved the peak lower, and larger holes moved it higher.

A wooden plate with 7 percent perforation peaked at about 0.35 bare, and at about 0.7 to 0.85 with 5 centimeters of porous material behind it. Above roughly 20 percent open area the liner controls absorption entirely, a rule that does not hold for slats. Micro-perforated panels shrink the holes below a millimeter so that the holes do the absorbing, an idea D.-Y. Maa published in 1975, and transparent micro-perforated absorbers were hung in the German Parliament hall in Bonn in 1993.

A test result also depends on the mounting, since a finish absorbs differently depending on how it is mounted, and ASTM E795 defines the standard mountings that absorption tests specify. Wall and ceiling finishes need the fire class their occupancy requires as well. Under the 2024 IBC, finishes tested to ASTM E84 are Class A with a flame spread index of 0 to 25, Class B at 26 to 75 and Class C at 76 to 200. Table 803.13 sets the class for each occupancy where that edition is adopted. Mounting, pattern, open area, tested absorption and fire class are compared together, in a tool for perforated finishes or in a manufacturer’s test report.