Two numbers and nothing else
Sabine decay time is a constant times room volume divided by total absorption. Absorption is every surface area multiplied by its coefficient and summed, plus whatever people and furniture bring. Nothing else enters — not the shape, not the finish, not the layout — except through those two quantities.
The constant depends on the unit system and this is where the silent error lives. Cubic feet with sabins uses 0.049; cubic metres with metric sabins uses 0.161. Converted properly they agree to a fraction of a percent, which the page demonstrates by working the same room both ways. Use 0.161 with cubic feet and the answer is out by a factor of about three and a third, in the direction that makes a bad room look acceptable.
Run the default room: 7200 cubic feet, painted block walls, a concrete floor and a suspended tile ceiling. The schedule comes to 536 sabins, the average coefficient is 0.213, and Sabine gives 0.66 seconds. The ceiling is doing 87 percent of the work on its own, which is the usual shape in a commercial room and is why the ceiling specification decides more than the wall finishes do.
Sabine and Eyring
Sabine assumes a sound ray goes on reflecting indefinitely, losing a fixed fraction each time. Eyring accounts properly for the energy each reflection removes, using the logarithm of one minus the average coefficient. The two agree closely in a live room and diverge as the room gets more absorbent, with Sabine always giving the longer time. The default room shows 0.66 against 0.59, an eleven percent gap at an average coefficient just over 0.2.
In a heavily treated room the gap matters. Sabine will tell you there is more reverberation left than there really is, which sounds harmless until you are using it to justify a further round of panels. The practical rule is that below about 0.2 average absorption either will do, and above it Eyring is the one to quote.
You only gain the difference
The commonest mistake in a treatment schedule is counting the whole coefficient of the new material instead of the improvement over what it covers. Absorption is a difference. A panel reading 0.85 mounted over painted block at 0.05 gains 0.80 per square foot; the same panel over a soft surface at 0.60 gains 0.25, so it takes more than three times as much of it to do the same job.
Taking the default room from 0.66 seconds to 0.50 needs 706 sabins against 536, a shortfall of 170. At a net gain of 0.80 that is 212 square feet, or eight percent of every surface in the room. That is a manageable job. Set the target at 0.3 instead and the arithmetic wants 1176 sabins, which is a different project entirely.
When the required area starts approaching the total surface of the room, surfaces are the wrong route. Free hanging baffles and banners absorb on both faces and are counted by area of material rather than by area of wall, which is how large hard spaces — gyms, workshops, atria — are normally dealt with.
Where the single number stops being useful
Coefficients move enormously with frequency. A 25 mm panel that reads 0.85 at 500 Hz can read 0.15 at 125, so a schedule assembled at one frequency describes the room only at that frequency. Rooms treated to a mid range target and nothing else come out boomy, because everything above 250 Hz got shorter and the bottom two octaves did not move. If the room matters, run at least 125, 500 and 2000 separately and look at the shape rather than the single figure.
Both formulas also assume a diffuse field. A small room at low frequency is a handful of modes rather than a diffuse field; a long corridor or a very flat room is not diffuse at any frequency; and a room with every absorber on one wall behaves worse than the arithmetic promises because the untreated axis keeps ringing. The estimate improves as the room gets bigger and the treatment gets spread more evenly, and it is an order of magnitude rather than a measurement in any case.
Questions people ask
What is the difference between 0.049 and 0.161?
Units, and nothing else. Volume in cubic feet with absorption in sabins uses 0.049; volume in cubic metres with metric sabins uses 0.161. Converted properly the two give the same answer to a fraction of a percent, which the page shows by working the room both ways. Using the metric constant with imperial volume is an error of roughly three and a third times and it produces no warning of any kind, which is why it is worth checking every time.
Sabine or Eyring?
Sabine always returns the longer time and the gap grows with how absorbent the room is. Below about 0.2 average absorption the two are close enough that it does not matter. Above that, Eyring is the better description, because Sabine assumes reflections continue indefinitely while Eyring accounts for what each one removes. The default room shows 0.66 against 0.59.
How much panel does it take to halve the decay time?
Halving the time means doubling the absorption, and the material only earns you the difference between its coefficient and whatever it covers. The default room holds 536 sabins; halving 0.66 seconds needs about 1070, so 534 more, and at a net gain of 0.80 per square foot that is 668 square feet — over a quarter of every surface in the room. Halving a decay time is always a larger job than it sounds.
Do I count people and furniture?
Yes, and in a room that fills up they are frequently the largest entry in the schedule. Seating and occupants are quoted per seat or per person in absorption units rather than as a coefficient, so they go in the separate field. It is also why an empty restaurant and a full one are two different rooms, and why a measurement taken in an empty room describes a condition nobody experiences.
Will this make the room quieter?
It shortens the decay, which usually makes speech clearer and the room feel calmer, and in a room where the noise is mostly reverberant it lowers the level too. It does nothing whatever about sound arriving from another room or from outside. Absorption is porous and light; blocking transmission needs mass, decoupling and airtight sealing, which is work on the wall assembly rather than something hung on the surface.