Are Phone Decibel Meter Apps Accurate?
The short answer: close enough for a rough check, not close enough to rely on. In NIOSH testing, 4 of 10 shortlisted apps stayed within ±2 dB of a professional meter, and an external calibrated microphone got the best of them to about ±1 dB. Most apps, uncalibrated, are further off than that.
What the testing actually found
NIOSH researchers Chucri Kardous and Peter Shaw screened the app stores and tested the shortlist against a professional sound level meter in a controlled setup. Ten apps made it through screening. Of those, four stayed within ±2 dB of the reference meter, and adding a calibrated external microphone to the phone tightened the best results to about ±1 dB.
2 dB sounds small, but it is the difference between a reading that says 83 and one that says 85 — which is the line where NIOSH says hearing protection and time limits start to matter. And that is the good group. Apps that were never calibrated, or that were built for a different phone, can be off by much more.
Where the error comes from
A phone is not a measuring instrument. It has a microphone designed for speech in a call, and its output is shaped by software to make voices sound clear. Four things push decibel readings off:
- No calibration. Every microphone has its own sensitivity. A meter is calibrated against a known reference so the number means something; an app generally is not.
- Automatic gain control. Many phones turn quiet input up and loud input down. That flattens exactly the thing you are trying to measure.
- Uneven frequency response. Phone mics over- or under-weight some frequencies. Bass-heavy noise is where readings drift most, because low frequencies are handled worst.
- Physical setup. A case over the mic port, a thumb across it, wind, or the phone lying on a soft surface all change the reading by several decibels.
A browser meter adds one more limit on top: it can read the raw microphone stream, but it cannot apply a factory correction for that specific phone model the way a tuned app can. That is why ours has a calibration offset you can set by hand.
What a phone reading is good for
Plenty, as long as you use it for the right question. Phones are excellent at telling you that something is louder than it seemed, and they are the only meter most people have on them at 1 a.m.
| Use it for | Why it works |
|---|---|
| Spotting that a room is louder than it feels | Relative readings in the same spot, same device |
| Recording when and how often noise happens | Timestamps, duration and peaks |
| Comparing two rooms, or one room before and after a fix | The difference is what matters, not the absolute number |
| Checking whether a sound crosses a rule-of-thumb limit | Use it as a warning, then confirm properly |
What it is not good for
| Not for | Why not |
|---|---|
| Proving a noise nuisance to a landlord or council | They want a calibrated meter or their own measurement |
| Workplace noise compliance | OSHA-style readings need a meter that meets the standard |
| Any number you would put in a contract or a letter | Uncalibrated readings invite a challenge |
| Measuring very low or very high frequencies | Phone mics are weakest at the ends |
The distinction that matters in a dispute: a phone shows the pattern, a meter shows the level. "Loud music from 23:40 to 01:00, 15 dB above my quiet room, on nine nights" is a strong complaint even from a phone. "It was 57.3 dBA" from an uncalibrated phone is a weak one, because the other side can reasonably ask what it was measured with.
How to get a better reading
- Take the case off, or at least clear the microphone port.
- Put the phone at the height your ears are, not on a sofa or a table that muffles it.
- Watch the live level for a minute before trusting it. Peaks from a door slam are not the room's level.
- Note a quiet baseline in the same room, so you can report the difference as well as the number.
- Calibrate the offset against a real meter if you can borrow one. If you cannot, say the reading is uncalibrated whenever you share it.
The free decibel meter shows live level, average and peak, and has a calibration offset field for exactly this. It never records or uploads audio.
When to stop using a phone
The moment someone else has to trust the number. That means a landlord, a council, a lawyer, an employer, or your own hearing protection decisions at work. A Class 2 meter (IEC 61672) covers home and general checks; Class 1 is for work that has to hold up to scrutiny. How to choose a sound level meter goes through the differences, and when a phone is genuinely enough.
Questions
- Are phone decibel meter apps accurate?
- Some are close, most are not. In NIOSH testing, 10 apps made the shortlist after screening and 4 of those stayed within ±2 dB of a professional meter. Adding a calibrated external microphone brought the best ones to about ±1 dB. Any app without calibration is a rough estimate, not a measurement.
- Is a browser decibel meter as good as a phone app?
- It runs on the same microphone, so the ceiling is the same. A browser meter has one extra weak point: the browser cannot correct for the microphone's frequency response the way a tuned app can, and the numbers depend on the device. Treat both as estimates and use the calibration offset to line them up with a real meter.
- Can I use a phone app as proof of a noise problem?
- As a record of the pattern, yes: it shows when noise happened, how often, and how much louder it was than your quiet baseline. As proof of an exact level, no. Most landlords, councils and courts want readings from a calibrated meter, or they send someone to measure it themselves.
- Why does my phone read 10 dB higher or lower than a real meter?
- Three usual causes: the microphone is not calibrated, automatic gain control is boosting quiet sounds, or the phone's mic responds unevenly across frequencies so bass or hiss gets over-weighted. A case, a covered mic port and wind add more error on top.
- How do I calibrate a phone decibel meter?
- Compare it with a known reference — a calibrated meter, or a calibrator — in the same steady sound, and note the difference. Then apply that number as an offset. It only holds for that device, that microphone position and that kind of sound, so recalibrate when any of those change.