--- title: "Decibel Meter Online: Sound Level Meter for Noise in dB(A)" description: "Free online decibel meter: measure sound and noise levels in dB(A), dB(C) or dB(Z) with your phone or laptop mic, with Leq, max, octave bands and calibration." url: "https://theaudiostuff.com/tools/decibel-meter/" type: "website" author: "Jakub Charkiewicz" --- > Markdown rendering of https://theaudiostuff.com/tools/decibel-meter/ > Canonical URL to cite: https://theaudiostuff.com/tools/decibel-meter/ > Generated from the published page. Text and links only; see the HTML for layout. Breadcrumb: [Home](https://theaudiostuff.com/) > [Tools](https://theaudiostuff.com/tools/) > Decibel Meter # Decibel Meter: Online Sound Level Meter Measure how loud it is with a decibel meter that runs in your browser: a sound meter and noise meter for the microphone in your phone, tablet or laptop. It reads the sound level in dB(A) the way a hand-held meter does, with Fast or Slow response, the average (Leq), the maximum and minimum, the octave bands that tell rumble from hiss, and how long NIOSH allows at that level. Nothing is recorded or sent. **Before you start:** a phone or laptop microphone is not a calibrated instrument. Until you calibrate, the readings are an estimate; one reading from a real sound level meter brings this one in line with it for your microphone, and the page remembers it. Press Start and allow the microphone when your browser asks. ## Live sound level {#dbm-level-h} –dB(A) Fast response · estimate, not calibrated - **Average (Leq)**: – - **Maximum**: – - **Minimum**: – - **Measured for**: 0:00 - **dB(A)**: – - **dB(C)**: – - **dB(Z)**: – ### Octave bands {#dbm-bands-h} 1. 31.5 Hz 2. 63 Hz 3. 125 Hz 4. 250 Hz 5. 500 Hz 6. 1k Hz 7. 2k Hz 8. 4k Hz 9. 8k Hz 10. 16k Hz - **You should see**: The figure settle in a quiet room and jump when you clap, speak or turn music up. Fast follows every change; Slow moves less and is the easier one to read off a steady noise. The average (Leq) builds up over the whole run, and the octave bands show where the sound is: the left bars for rumble and bass, the right ones for hiss and voices. - **If it looks wrong**: The figure does not move: pick another microphone in the list, check the mute switch, and the input level in your system's sound settings. Loud sounds read too low or the reading drifts back down on a steady noise: your browser kept its automatic gain on. The line under the Start button says when it did. Every reading is several decibels off a real meter: calibrate below, once, for this microphone. ## Noise log {#dbm-log-h} Write the level down row by row, for a noise complaint, a night of traffic or a neighbor's music, and save it as a CSV for a spreadsheet. Each row holds the average (LAeq), the loudest moment (LAFmax), the levels exceeded 10% and 90% of the time (LA10 and LA90: the noise's peaks and the background under it), and the C-weighted average (LCeq), which keeps the bass that dB(A) discounts. The log stays in this page: nothing is stored or sent, and closing the page clears it. Start logging and keep this page open and in front: a browser stops measuring in a hidden tab, and those rows are marked as not measured. While it logs, the page asks the device to keep its screen on. ## Calibrate the meter {#dbm-cal-h} The browser hands the page numbers, not sound pressure, so every reading is the microphone's digital level plus an offset. Until you set it, the meter assumes a full-scale signal from your microphone is 120 dB SPL: an assumption, so readings can be several decibels off, more on a laptop whose input level has been turned up or down. Put this device next to a real sound level meter, play a steady sound (pink noise from the [tone generator](https://theaudiostuff.com/tools/signal-generator/) works well), set the same weighting on both and type in what the real meter reads. dB dB Not calibrated. The offset is kept in this browser, one for each microphone, and never leaves it. ## How to read a sound level meter {#dbm-guide-h} - **dB(A)**: The level with the A-weighting curve of IEC 61672-1, which discounts low and very high frequencies roughly as hearing does at moderate levels: 26 dB down at 63 Hz, flat at 1 kHz. Noise limits and hearing-safety rules are written in dB(A), so it is the default. - **dB(C) and dB(Z)**: dB(C) is almost flat, only falling off below about 31 Hz and above 8 kHz, and dB(Z) is not weighted at all. When dB(C) reads well above dB(A), the difference is low-frequency sound, which A discounts and C keeps: traffic, an engine, a subwoofer, the bass through a wall. - **Fast and Slow**: How quickly the reading follows the sound. Fast averages over 125 ms and shows every word and clap; Slow averages over 1 second and gives a steadier figure to read off a fan, a road or music. Both are the standard's exponential averages. - **Average (Leq)**: The equivalent continuous level: the steady level that would carry the same sound energy as everything measured since the start. It averages energy, not decibels, so a short loud moment counts for a lot. It is the figure exposure limits are written for. - **Maximum and minimum**: The highest and lowest readings with the chosen weighting and response since the start (the first half second is left out while the microphone settles). Reset readings starts them, the average and the chart again. - **Hearing safety**: NIOSH recommends no more than 85 dB(A) averaged over 8 hours, and half the time for every 3 dB more: 4 hours at 88, 1 hour at 94, 15 minutes at 100. The line under the readings applies that to the average measured here, which is only as right as the calibration. ## What a phone or laptop decibel meter can and cannot tell you {#tprimer-title} A sound level meter turns air pressure into a number on the decibel scale, with frequency and time weightings that make the number mean the same thing on every meter. A phone or laptop runs the same arithmetic on what its microphone delivers: this page applies the IEC 61672-1 A and C weighting curves as digital filters, averages with the standard Fast and Slow time constants, and keeps the energy average, so every figure is computed the way a hand-held meter computes it. What separates it from a hand-held meter is the microphone in front of the arithmetic. A meter's microphone is calibrated, so a known pressure gives a known reading. A phone's is made for voices: the browser receives its digital level, and how that level relates to sound pressure depends on the microphone, its gain and the operating system. One reading from a real meter settles that relation for one microphone. Some uses need no calibration at all. Comparing two sounds on the same microphone, such as the level of each speaker in a room, or the room before and after you close a window, is a difference, and the unknown offset cancels out of a difference. ### The ideas behind the controls {#tprimer-ideas} - **dB SPL**: Sound pressure level: 20 times the base-10 log of the pressure over 20 micropascals, about the quietest sound a young ear hears at 1 kHz. Every 6 dB doubles the pressure; every 10 dB sounds roughly twice as loud. - **Frequency weighting**: A filter applied before the level is read. A follows hearing at moderate levels and discounts the bass, C is nearly flat, Z is not weighted at all. - **Time weighting**: How fast the reading follows the sound: Fast averages over 125 ms, Slow over 1 second. - **Leq**: The steady level carrying the same energy as the whole measurement. Exposure limits are written for it. ### Common mistakes {#tprimer-mistakes} - Reading an uncalibrated phone figure as a measurement. It is an estimate until one reading from a real meter has set the offset for that microphone. - Averaging decibel figures by eye. Decibels are logarithmic: 60 dB for half the time and 90 dB for the other half averages to an Leq of 87 dB, because the loud half carries almost all the energy. - Measuring with the phone in a pocket, a hand over the microphone or the case rubbing against something. Each of those changes what reaches the microphone more than the room does. - Comparing a dB(A) reading with a dB(C) one. On a sound with a lot of bass they can be more than 10 dB apart. - Leaving the browser's voice processing on. Automatic gain turns loud sound down and noise suppression removes noise, which is what a meter is there to measure; the page asks for it off and says when the browser kept it. ### What this tool cannot tell you {#tprimer-limits} - A phone or laptop microphone is not flat across the audio band and can overload on very loud sound, a concert or a power tool, where its reading flattens out. Calibration fixes the overall offset, not those. - The weighting filters follow the standard's curves within 0.1 dB to 2 kHz and within about 1 dB to 10 kHz at 48 kHz; at the very top they fall short of it, as every digital filter at that rate does. That is far smaller than what the microphone itself adds. - It measures while the page is open and in front: a browser stops the page's drawing in a background tab, and the average leaves those gaps out. It is not a dosimeter for a working day. - For a reading you can rely on without a reference meter, the free [NIOSH Sound Level Meter App](https://www.cdc.gov/niosh/noise/about/app.html) for iPhone and iPad is rated accurate within 2 dB(A), tested in a reverberant chamber at the NIOSH acoustics laboratory. ### Where the math comes from {#tprimer-sources} - [IEC 61672-1: Sound level meters, Part 1: Specifications](https://webstore.iec.ch/en/publication/5708) Standard International Electrotechnical Commission Paid full text Defines the A and C frequency weightings and the Fast and Slow time weightings the meter applies. - [NIOSH 98-126: Occupational noise exposure](https://www.cdc.gov/niosh/publications/numbered/98-126.html) Standard National Institute for Occupational Safety and Health The 85 dB(A) eight-hour limit and the 3 dB exchange rate behind the hearing line under the readings. - [W3C Media Capture and Streams: Constrainable properties](https://www.w3.org/TR/mediacapture-streams/#constrainable-properties) Standard World Wide Web Consortium The autoGainControl, noiseSuppression and echoCancellation settings the meter asks to switch off and reports back. - [W3C Web Audio API: FFT windowing and smoothing over time](https://www.w3.org/TR/2021/REC-webaudio-20210617/#fft-windowing-and-smoothing-over-time) Standard World Wide Web Consortium The AnalyserNode FFT and its window, from which the octave bands are summed. ## Decibel meter FAQ {#dbm-faq-title} 1. ### How do I measure decibels with my phone or laptop? {#dbm-faq-how-do-i-measure-decibels-with-my-phone} Open this page, press Start decibel meter and allow the microphone. The large figure is the sound level in dB(A), updated as the sound changes, with the average, maximum and minimum since you started under it. Hold the phone where you want the reading, microphone uncovered, and keep it still: a hand rubbing the case or a breath on the microphone reads as loud sound. 2. ### How accurate is an online decibel meter? {#dbm-faq-how-accurate-is-an-online-decibel-meter} Uncalibrated, it is an estimate. A browser receives the microphone's digital level, and how that relates to sound pressure depends on the microphone, its gain and the operating system, so the page assumes a typical relation and says so beside every reading. Calibrated once against a real sound level meter, it follows that meter for the same microphone within the limits of the microphone itself: a phone microphone is not made to be flat across the whole range, and very loud sounds can overload it. 3. ### What is the difference between dB(A), dB(C) and dB(Z)? {#dbm-faq-what-is-the-difference-between-db-a-db} They are three frequency weightings defined in IEC 61672-1. dB(A) discounts low and very high frequencies roughly as hearing does at moderate levels, and is the weighting noise limits and hearing-safety rules use. dB(C) is nearly flat and keeps the bass, so it is used for loud sound and low-frequency noise. dB(Z) applies no weighting at all. When dB(C) reads well above dB(A), most of the difference is low-frequency sound. 4. ### How loud is too loud? {#dbm-faq-how-loud-is-too-loud} NIOSH recommends a limit of 85 dB(A) averaged over an 8-hour day, halved for every 3 dB more: 4 hours at 88 dB(A), 1 hour at 94 and 15 minutes at 100. The meter applies that rule to the average (Leq) it has measured and says how long it allows. The figure is only as right as the calibration, and for headphones the safe listening calculator works the level out from the headphone's own specifications. 5. ### How do I calibrate the decibel meter? {#dbm-faq-how-do-i-calibrate-the-decibel-meter} Put this device next to a real sound level meter, play a steady sound such as pink noise from the tone generator, and set both to the same weighting and Slow response. Type the reference meter's reading into the calibration box and press Match it. The page stores the offset in this browser for that microphone, so it applies the next time you open the meter with it. 6. ### Can I use it to set my speaker levels? {#dbm-faq-can-i-use-it-to-set-my-speaker} Yes, and for that it does not need calibrating. Play pink noise through one speaker at a time from the tone generator's noise tab, hold the device at your listening position with dB(C) and Slow selected, and adjust each speaker's level until they all read the same. Matching is a comparison on one microphone, so an uncalibrated offset cancels out. 7. ### What is Leq? {#dbm-faq-what-is-leq} The equivalent continuous level: the steady level that would carry the same sound energy as everything measured over the run. It averages energy, so a few loud seconds raise it far more than an average of the decibel figures would. Exposure limits are written for it, which is why the hearing line under the readings uses the A-weighted Leq. 8. ### Does the decibel meter record or upload anything? {#dbm-faq-does-the-decibel-meter-record-or-upload-anything} No. The sound is measured inside your browser and nothing is recorded, stored or sent. The only thing the page keeps is your calibration offset and the weighting and response you chose, in this browser's storage. A noise log stays in the page until you download it as a CSV yourself. Turn microphone off releases the microphone at once. 9. ### Can I record a noise log for a noise complaint? {#dbm-faq-can-i-record-a-noise-log-for-a} Yes. Start logging under the readings and the meter writes a row every 10 seconds, minute, 5 minutes or 15 minutes, on the clock, with the A-weighted average (LAeq), the loudest moment (LAFmax), the levels exceeded 10% and 90% of the time (LA10 and LA90) and the C-weighted average, which shows bass through a wall. Download CSV saves it for a spreadsheet with the date, time and time zone of every row. Keep the page open and in front, with the device on its charger: a browser stops measuring in a hidden tab, and those rows say they were not measured. The page asks the device to keep its screen on while it logs. Calibrate first if the levels matter, and keep in mind that whoever handles the complaint decides what counts as evidence. ## Related reviews and guides {#rt-editorial-title} - [Best Speakers](https://theaudiostuff.com/guides/best-speakers/) - [Speaker reviews](https://theaudiostuff.com/reviews/speakers/) Reviews - [Diora Acoustics Chors 5 Scored 8.5 out of 10](https://theaudiostuff.com/reviews/diora-acoustics-chors-5/) - [Taga Harmony DIAMOND B-60 v.3 Scored 8.2 out of 10](https://theaudiostuff.com/reviews/taga-harmony-diamond-b60/) - [Triangle Australe EZ Scored 8.1 out of 10](https://theaudiostuff.com/reviews/triangle-australe-ez/) Head-to-heads - [Diora Acoustics Chors 5 vs Triangle Australe EZ](https://theaudiostuff.com/compare/diora-acoustics-chors-5-vs-triangle-australe-ez/) - [Diora Acoustics Chors 5 vs Triangle Borea BR09](https://theaudiostuff.com/compare/diora-acoustics-chors-5-vs-triangle-borea-br09/) ## You might also use these tools {#rt-title} - Test **Can I actually hear the difference?** [ABX Blind Test](https://theaudiostuff.com/tools/abx-test/) Eight built-in synthetic test pairs, no upload needed - or drag in two files of your own, DSD included, checked for what actually differs. Loudness-matched and sample-accurate, with a blind preference mode and a shareable certificate. - Training **How good are my ears, really?** [EQ Ear Training Game](https://theaudiostuff.com/tools/ear-training/) Train your ears to identify frequencies from 20 Hz to 20 kHz, or only the range you can hear. Five training modes, score tracking, and a shareable Golden Ear Rating. - Test **How high can I hear, in each ear?** [Hearing Frequency Test](https://theaudiostuff.com/tools/hearing-frequency-test/) Find the highest tone you can hear in each ear and the lowest you hear through your headphones, with a slow sweep you stop when the tone fades out.