Multi-Stud Wall Sound Transmission Class (STC) Estimator

Frequently asked questions

How accurate is this STC estimate?

Treat it as a rough rule-of-thumb only, not a certified rating. It is a widely-used additive estimation rule built from commonly cited point bonuses for decoupling, cavity insulation, and extra gypsum layers, not an ASTM E90-standardized predictive method — actual performance also depends on flanking transmission and workmanship.

Can I use this number in a project specification or acoustic compliance submittal?

No — for any acoustic-critical project, such as a multifamily party wall, specify and verify performance using the manufacturer's actual ASTM E90 lab-tested assembly rating for your exact wall makeup, not this estimate.

Where do the bonus point values come from?

From commonly cited figures in acoustic design guides: roughly +8 for resilient channel/decoupling, +3 for cavity insulation, and +2 to +3 per extra gypsum layer. Real assemblies vary, so confirm against published assembly data for your specific products where possible.

Does it cover sound that gets around the wall rather than through it?

No. Every field on this page describes one panel of wall between two rooms, and the arithmetic has nowhere to put anything outside it. In a finished building the paths that beat a well-built partition are usually the ones going past it: the wall dying at a suspended ceiling with an open plenum carrying over the top into the next room, a floor topping or screed poured continuously under the sole plate, one duct branch serving both rooms so speech travels the ductwork, a bulkhead or soffit run through, a continuous window sill or mullion. Any one of those puts a ceiling on the whole partition, and once it is the weakest path a fourth layer of board changes nothing anybody can hear — which is why the same makeup measured in an occupied building commonly rates below the laboratory figure for the wall alone. The fixes are detailing rather than acoustic product: carry the partition deck to deck and seal it where it lands instead of stopping it above a ceiling grid, interrupt continuous toppings on the line of the wall, and route or attenuate shared duct runs so they are not a shortcut. Worth being clear about what that gap means if a rating has to be met rather than aimed at: a laboratory number describes the wall with these paths deliberately suppressed, while what an occupant hears, and what a field measurement records, includes all of them. It is not a fudge factor you can subtract here — it is designed out at the junctions by somebody reading them.

It shows 44 before I have told it anything about my wall — what is that assuming?

Three things you have not said yet. The 8 points sitting in the decoupling field are the largest single bonus on the page and they assume resilient channel, isolation clips or an equivalent break between board and framing; a plain single-stud partition with board screwed straight to the studs has none of that. Zero that field and the same wall reads 36. The 3 in the cavity insulation field assumes the cavity is insulated — take that out and the default reads 41. Read that one by its label rather than its position: the form sets out in two columns on a wide screen, and the other field showing 3 is the per-extra-layer bonus, which is doing nothing at all while the layer count sits at zero. The base 33 assumes one layer of board on each face. The layer field is the easiest to double by accident. Its label counts layers beyond one per side, so a wall double-boarded on both faces is 1 by that reading, not 2, and the arithmetic will not argue with you either way: it multiplies your bonus by whatever you type, flat, with no taper. One extra layer reads 47, two reads 50, three reads 53, the third worth exactly as much as the first, which is not how board behaves on a leaf that is already heavy. Wind every field to its maximum and the page prints 70 without hesitating. The decoupling field is also carrying more than its name suggests, because it is the only place any form of separation can go — channel, clips and hat channel, staggered studs on a shared plate, a full double-stud wall on separate plates — and those are not worth the same as each other. Build a double-stud wall and you are using that field as a proxy, with its cap of 10 the only thing holding the estimate down. Do not count the same separation twice, either: channel screwed to a pair of leaves that are already separated is not a second +8. What has no field at all is cavity depth, stud spacing, and whether the studs are timber or light-gauge steel — the steel stud is the more flexible of the two and commonly gives the better result at the same makeup, and this page cannot tell them apart. None of the five fields carry units, incidentally, so the metric and imperial switch leaves this one alone; points are points.

What goes wrong on site that this number never sees?

Fasteners first. Wherever the board is decoupled, a screw long enough to reach the stud behind the channel undoes the decoupling at that point, and 8 of the 44 on screen are that decoupling. It is a silent failure: the board goes up, the screw feels right, nobody sees it again, and it happens most often on the second layer, where the longer screws come out of the box. Fix into the channel's board flange and never through it into the framing behind, keep screw length to what that layer actually needs, and stop anyone chasing the stud lines for a firmer bite. Then contact. Board landing tight on the slab, wedged against a return wall or jammed hard under the structure bridges the leaf the same way a screw does; common practice is to hold it clear at floor and abutments — the gap is usually quoted at around 6 mm (1/4 in) — and fill the gap with a non-hardening acoustic sealant rather than let the board bear. Insulation sets the same trap — the cavity wants filling, not stuffing, and batts compressed hard against a decoupled face are another bridge. Then holes. Two electrical boxes back to back in one stud bay is a straight route through both leaves and a classic complaint on a wall that was otherwise built well; put them in separate bays, seal them, and treat plumbing penetrations, recessed fittings and the head track the same way. An unsealed perimeter bead at head or base is a slot the full length of the wall. None of that is visible to this arithmetic, and none of it comes off as points you could subtract. If the wall has a figure to meet, build a tested assembly the way it was tested — board type and thickness, stud gauge and spacing, channel type, screw length, insulation and sealant are all part of what was tested rather than details around it — and treat any substitution as putting you back into estimate territory, which is where this page lives.