The rating is not the drop
Noise reduction between two rooms is the partition rating plus ten times the log of the receiving room absorption divided by the partition area. The rating describes the wall; the second term describes what happens to the sound once it is through. A soft receiving room soaks it up and the difference is larger. A hard empty one lets it build back up and the difference is smaller. The term is often worth three or four decibels either way, and it is the reason two identical walls in two different rooms measure differently.
Work the defaults. A 45 dB partition of 120 square feet into a 1600 cubic foot room with a measured decay of 0.6 seconds holds 131 sabins, which is a correction of plus 0.4, so the level difference is 45.4. Eighty five decibels in the source room lands at 39.6 through the wall, and with a 35 dB background in the room the meter reads 40.9.
Absorption from a decay time
Sabine ties three things together: decay time, volume and absorption. Any two give the third, so a measured RT60 and a tape measure give the absorption without touching a single material coefficient. In feet the constant is 0.049 and volume is in cubic feet; in metres it is 0.161 and volume is in cubic metres. They describe the same physics — convert one into the other and they agree to a fraction of a percent — but swapping them without converting the volume is a silent error of about three and a third times, in whichever direction hurts most.
The measurement is worth doing. Clap or burst a balloon, record it, and read the decay: even a phone gets you close enough to see whether the correction is helping you or hurting you. The alternative is building the absorption up surface by surface from coefficients, which is a longer job and depends on datasheet figures for materials you may not be able to identify.
The background sets the ceiling
Nothing takes a room below its own background. If the receiving room reads 35 with everything off, then a source arriving at 39.6 gives a total of 40.9, and an infinitely good partition still leaves 35. That is the shape of the second comparison on the page: taking the rating from 45 to 50 moves the level through the wall from 39.6 to 34.6, but the meter only falls from 40.9 to 37.8, because the background is now most of what is left. Five decibels of wall bought three decibels of reading.
The corollary is uncomfortable for anyone selling wall upgrades. Once the source is within a few decibels of the background, further work on the partition is close to wasted, and the useful question becomes whether the background itself can be raised or lowered — which is a ventilation and a site question rather than a construction one.
What broadband levels hide
A single number for a level, like a single number for a rating, conceals what is happening band by band. A source arriving below the background can still be perfectly audible if it is a steady tone, a bass line or an intelligible voice, because hearing picks structure out of noise far below the point where a broadband meter registers anything. Complaints about music through a wall almost always concern energy two or three octaves below where the rating and the A weighting are paying attention.
Everything on this page also assumes the wall is the only route. It is not, usually. A continuous floor slab, a shared ceiling plenum and a duct serving both rooms all arrive on top of this figure, and if any of them is worse than the wall it decides the result instead.
None of this is a measurement, an adequacy determination or a privacy judgement. The target belongs to your brief, any limit belongs to your local ordinance, field performance is measured with an instrument, and anything the code touches is decided by the authority having jurisdiction.
Questions people ask
Why is my measured level different from the rating suggested?
Two reasons, usually. The first is the room correction: a rating is a property of the wall, and what lands in the far room also depends on how absorbent that room is relative to the partition area — worth several decibels either way. The second is that the wall is rarely the only route. A floor slab, a shared plenum or a duct serving both rooms arrives on top of the figure and is not in it.
How do I get the absorption of the receiving room?
The quick way is a measured decay time and the room volume, because Sabine ties the three together: absorption is the constant times volume divided by RT60. Burst a balloon, record it, read the decay. The long way is to sum every surface area times its coefficient from the material datasheets, which is more work and depends on identifying what the surfaces actually are.
Does a better wall always help?
Only until the background in the receiving room takes over. With the defaults, moving the rating from 45 to 50 drops the level through the wall by 5 dB but drops the meter reading by 3.1, because the 35 dB background is most of what remains. Past that point the wall is no longer the thing deciding the reading, and no partition takes a room below its own background.
What is the difference between the two Sabine constants?
Nothing but units. 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 agree to a fraction of a percent. Using one constant with the other unit system is a silent error of roughly three and a third times, which is why the form asks which system you are in before it asks for a volume.
Can this tell me whether a room is private enough?
No. Privacy depends on what is being said, what else is going on in the receiving room, and who is listening, and it is judged with speech privacy measurements rather than with a level difference. Nothing here is an adequacy or privacy determination. The target on the form is yours, any limit is whatever your local ordinance sets, and where the outcome matters the answer is an acoustical consultant with an instrument.