Decibel Meter
A microphone, a browser and an honest caveat: the number is an estimate unless the microphone is calibrated.
What this meter can and cannot tell you
The page reads your microphone, computes the loudness of each 2,048-sample block and converts it to decibels. That part is real measurement of the signal reaching the microphone.
The weak link is the microphone and its gain. Phone and laptop microphones are not calibrated instruments: the same room can read 62 dB on one device and 71 dB on another. Treat the number as a relative level — useful for comparing "quieter" and "louder" — until you set the calibration offset against a known reference.
Why the offset matters
A digital microphone reports its signal relative to full scale: 0 dBFS is the loudest level it can capture without clipping. Turning that into sound pressure needs one number — how loud a full-scale signal actually is on this device. The default here is 94 dB at full scale, a common assumption for consumer microphones, but it is only an assumption.
If you have a reference meter, play a steady tone or a fan, note the reference reading, and move the slider until this page matches. The offset then applies to every reading on that device. Without a reference, keep the default and read the output as approximate.
Reading the number
| Level | Sounds like | Safe daily exposure |
|---|---|---|
| 30 dB | Whisper, quiet bedroom | 8 h+ |
| 50 dB | Quiet office, refrigerator hum | 8 h+ |
| 70 dB | Normal conversation, vacuum cleaner | 8 h+ |
| 85 dB | Heavy traffic, busy restaurant | 8 h |
| 94 dB | Lawn mower, hand drill | 1 h |
| 100 dB | Subway platform, nightclub | 15 min |
| 110 dB | Rock concert front row, chainsaw | 1 min 29 s |
| 120 dB | Ambulance siren, thunderclap | Seconds |
The exposure column follows the NIOSH rule used on the chart page: 85 dBA for eight hours, and the allowed time halves with every 3 dB increase.
Hearing damage is cumulative. A single loud event can cause permanent loss, and so can years of 90 dB without protection. If you need a number for a workplace, a tenancy dispute or a noise complaint, use a calibrated Type 1 or Type 2 sound level meter — this page is not evidence.
Getting a better reading
- Turn off audio processing. The page already requests the raw stream: echo cancellation, noise suppression and automatic gain are disabled, because they change the level.
- Hold still. Wind, handling noise and breathing can add 5–10 dB.
- Measure at a consistent distance and angle. Doubling the distance drops the level by about 6 dB in a free field.
- Watch the peak, not only the live number. A short bang can be missed by the eye but captured by the peak.
Questions
Is this a certified sound level meter?
No. It is a browser tool with an uncalibrated microphone. Certified meters are type-approved instruments with calibration certificates; they cost hundreds and are checked against a reference source.
Why does the reading sit near 0 dB in silence?
Silence is not zero: the microphone still picks up its own electrical noise, and the offset maps the quietest signal to a positive number. A reading that will not go below 25–35 dB usually means the room is quiet and the microphone noise floor is showing.
Does the page record audio?
No. Samples are used once to compute a level, then discarded. Nothing is stored or uploaded. Pressing stop releases the microphone entirely.
Why is dBA different from dB?
dB is a ratio; dBA is that ratio filtered to approximate how the ear responds to different frequencies. A low-frequency rumble can be 90 dB but only 70 dBA. Browser meters do not apply the A-weighting curve, so compare readings cautiously.