Decibel Chart: Decibel Levels of Common Sounds

A normal conversation runs 60–70 dBANIDCD. A lawn mower runs 85–90 dBANIOSH. The line worth remembering sits at 85 dBA, which is what NIOSH recommends as an eight-hour average at work, halving the allowed time for every 3 dB above itNIOSH.

How to read this decibel chart

Decibels do not add up the way most numbers do. The scale is logarithmic: each 10 dB step is ten times the sound energy of the one below it, so 90 dB carries ten times the energy of 80 dB and a hundred times the energy of 70 dB. The practical consequence is the one NIOSH builds its limit around — every 3 dB increase doubles the energy reaching your ear, so the time you can safely spend in it halvesNIOSH. Three decibels looks like a rounding error written down. It is not.

The scale describes energy, not how loud something seems. Perceived loudness depends on frequency, on how long the sound lasts, and on the listener. Two sounds at the same decibel figure can feel quite different, which is why the figures below are published as ranges rather than as single points.

Nearly every number on this page is written dBA rather than dB. The A stands for A-weighting: the signal is filtered to discount the very low and very high frequencies that human hearing is least sensitive to, so the figure tracks the risk to hearing rather than the raw energy in the air. Health agencies publish in dBA, exposure limits are written in dBA, and the meter on this site reports dBA. Where you see a decibel figure with no letter after it, you generally cannot tell which of the two it is, and the difference on a bass-heavy sound can run to double figures.

Decibel levels of common sounds

That table is shorter than the ones you will find elsewhere. Charts on competing sites run to fifty or more rows — breathing at 10 dB, a whisper at 30, a library at 40, a refrigerator at 45 — and almost none of them say where those numbers came from. Trace a few back and they lead to each other, or to an acoustics textbook, or nowhere at all.

The rows above are the figures two US health agencies actually publish. NIDCD's list of average decibel ratings for familiar sounds begins at normal conversation. NIOSH's begins at 85 dBA, the level where it starts recommending hearing protectionNIOSH. Neither publishes a figure for a whisper, because a whisper does not put anyone's hearing at risk and a health agency has no reason to put a number on it.

One US figure does exist below the table's quiet end. A 1974 EPA report, summarised by NIOSH, recommended 70 dBA averaged over 24 hours as an environmental exposure limit, and separately named 55 dBA outdoors and 45 dBA indoors as the levels above which noise starts to interfere with speech and become an annoyanceNIOSH. Those are limits rather than descriptions of a sound, which is why they sit here in the prose and not in the table.

Where a range straddles one of the risk thresholds, the bar and the colour follow its lower end. A movie theatre published at 74–104 dBA is shown in the gentler band, because the quiet screening is as real as the loud one and the chart should not imply that every trip to the cinema is hazardous.

How loud is 70 decibels?

Seventy is the ceiling for not having to think about it. NIDCD states that sounds at or below 70 dBA are unlikely to cause hearing loss even after long exposureNIDCD. No duration, no dose arithmetic, no hearing protection. A conversation, a television at a comfortable volume and a busy office all sit at or under it.

That does not make 70 dBA silent or pleasant. The EPA figures above put the threshold for speech interference indoors at 45 dBA, twenty-five decibels lower. A room can be loud enough to wreck a phone call, ruin your concentration and keep you awake while sitting comfortably inside the range that will never damage your hearing. Those are two different questions and the same number cannot answer both. If you came here worried about a neighbour or an office, annoyance is your question, not hearing loss.

How loud is 70 decibels takes that split apart properly: which of the two questions you have, which figures answer each, and why reaching for the hearing literature is why so many noise complaints go nowhere.

How loud is 85 decibels?

Eighty-five is where the published guidance changes character. NIOSH set its recommended exposure limit at 85 dBA averaged over an eight-hour workday, and recommends hearing protection whenever levels exceed it regardless of how long the noise lastsNIOSH. There is a field test that needs no equipment: at 85 dBA you generally have to raise your voice to be heard by someone three feet awayNIOSH. At 95 dBA you have to shoutNIOSH.

Eighty-five is also where American guidance splits in two, and the split catches people out. NIOSH recommends 85 dBA with a 3 dB exchange rate. OSHA enforces a permissible exposure limit of 90 dBA over eight hours with a 5 dB exchange rateOSHA. One is a health recommendation and the other is an enforceable standard that had to account for what was feasible when it was written. They disagree, both are current, and a number that clears one can fail the other.

There is more to that level than a line on a table. How loud is 85 decibels works through what actually sits there, how long it is safe, and why "too loud" is a pair of numbers rather than one.

How loud is 100 decibels?

At a hundred, the arithmetic stops being abstract. Fifteen minutes at 100 dBA reaches the whole daily noise dose NIOSH allows for an eight-hour shiftNIOSH. Not fifteen minutes of discomfort — fifteen minutes and the day's budget is spent.

Plenty of ordinary places reach it. NIDCD publishes concerts and headphones at maximum volume at 94–110 dBA, and motorcycles at 80–110 dBANIDCD. NIOSH puts bars, nightclubs, chain saws and large sporting events at 95 dBA and upNIOSH.

How loud is 100 decibels goes further into what that level costs, and into why the places that reach it are the hardest ones to manage: nobody at a venue is counting your exposure or handing you protection.

Safe listening time under the NIOSH limit

Read that table for what it is. NIOSH is explicit that the recommended exposure limit is an occupational limit, set to protect workers over a forty-year career, and that it is not meant to be applied to environmental or recreational noise such as power tools, earbuds or concertsNIOSH. This chart therefore prints the durations NIOSH published and does not put a safe time next to a lawn mower. Doing that arithmetic for you would be applying a workplace standard to a Saturday afternoon, which is the mistake the publishing agency spends a page warning against.

The same page adds that a level at one moment can exceed the limit without creating risk, provided the rest of the day is quiet enough to balance it outNIOSH. A single number tells you about a moment. Risk is about a day.

Three rows on the chart have wide enough ranges that the range alone is not much help, and each will get a page of its own explaining what sits at that level and how long it takes to matter: how loud is 70 decibels, how loud is 85 decibels, and how loud is 100 decibels.

What moves a number in a real room

A chart row is a category. The room you are standing in is a specific case, and several things separate the two.

Distance is the first. NIDCD names your distance from the source alongside duration as a factor in whether a sound puts your hearing at riskNIDCD, and it is the reason NIOSH's voice test is specified at three feetNIOSH rather than at an unstated distance. A mower is one number at the handle and another across the garden. Published figures are usually taken at the operator's position, which is the loudest place a person is likely to stand, so treat them as the top of what you will encounter rather than the middle.

What the source is doing matters as much as what it is. NIOSH publishes lawn mowers, vacuums, power tools and printing presses together at 85–90 dBA — one band covering four machines, because within any of them the load, the age and the setting move the figure more than the category doesNIOSH. A mower cutting long wet grass is not the mower idling.

The room itself is the third. Hard floors, bare walls and glass reflect sound back at you; carpet, curtains and soft furniture absorb it. The same machine in a tiled kitchen and a carpeted lounge does not give the same figure, and neither of those is the figure a chart quotes.

None of this makes the chart wrong. It makes it a starting point, which is what a reference table is for.

Measure your own room

A published range tells you what a category of sound usually does. It cannot tell you what your kitchen, your commute or the flat above you is doing right now. For that you need a number from where you are standing, and the online decibel meter on this site will give you one from your device microphone without an app or an account.

Hold the device at arm's length, roughly where your head would be, and keep your hand off the microphone — covering it is the single most common way to get a number that is tens of decibels wrong. Give it thirty to sixty seconds rather than reading the first figure that appears, because most environments swing by several decibels from second to second and one instant is not representative.

Then treat the result as an estimate. Of ten iOS sound apps NIOSH compared against a reference instrument, only three came within ±2 dBA and the two furthest sat about 13 dBA underKardous & Shaw. A browser has less to work with than a native app — no device-specific calibration profile at all — so this site publishes its figure as an uncalibrated estimate within roughly ±10 dB of what a hardware sound level meter would show. That is enough to tell an 85 dBA room from a 60 dBA one, which is the question most people actually have. It is not enough for an OSHA claim, an insurance report, a legal complaint or a medical decision, and nothing on this site should be used for any of them.

How this chart was compiled

Every row was taken from a page that was opened and read, not from another chart. The sources are listed in full below, and each figure on this page carries the name of the publication it came from directly after it, so any single claim can be checked without reading the rest.

Three rules decided what the table contains. A figure appears only if a named health agency published it, which is why the quiet end is missing rather than filled in. Ranges are printed as ranges: where a source gives 85–90 dBA for a lawn mowerNIOSH the table says 85–90, never 87.5, because averaging two published bounds invents a precision nobody claimed. And where an agency groups several sounds on one line, the grouping is kept rather than split into separate rows with separate-looking figures.

That leaves gaps a longer chart would fill. It also means nothing here is inherited from a source that cannot be checked, which is the trade this page makes deliberately. If you find a figure that has moved, or an agency publication this chart should be reading and is not, the contact form reaches someone who will fix it.