Calibrating a sound meter app means finding how far off your phone reads and entering that difference as an offset. The reliable way: put your phone next to a calibrated sound level meter in steady noise, measure both for 30 to 60 seconds, subtract, and enter the difference in the app’s calibration setting.
What you can’t do is calibrate a phone the way a lab calibrates an instrument. An offset shifts the whole scale up or down by a fixed number of decibels. It doesn’t fix the microphone’s frequency response, its behavior at very loud or very quiet levels, or any processing the phone applies on its way to the app.
What calibration actually fixes #
A phone microphone has a sensitivity nobody publishes, and it varies between models and sometimes between units. The app converts the digital level it receives into a dB SPL figure using an assumed sensitivity, so if that assumption is 4 dB off for your phone, every reading is 4 dB off in the same direction. That constant error is what an offset removes.
| Error source | Fixed by an offset? |
|---|---|
| Constant level error from mic sensitivity | Yes |
| Uneven frequency response (weak bass, peaky treble) | No |
| Automatic gain or noise suppression still active | No |
| Microphone noise floor in very quiet rooms | No |
| Overload and clipping at very high levels | No |
| Case, finger or wind blocking the mic | No |
Because of that, calibration helps most in the middle of the scale, roughly where you calibrated it, with broadband sound. A phone calibrated at 85 dB with pink noise can still be off at 35 dB in a quiet bedroom or at 110 dB at a concert.
Method 1: match a reference sound level meter #
This is the practical option for most people, and the one professional guidance points toward. You need access to a calibrated meter for ten minutes: a workplace safety officer, a local maker space, an audio engineer, or a rental from a test-equipment company usually has one.
- Find steady noise. A running fan, an HVAC unit, a shop vac, or pink noise from a speaker. Avoid speech, music with dynamics, and anything intermittent.
- Set both instruments the same way. A-weighting on both, the same time weighting (Slow is easier to read), and the same range if the meter has one.
- Put the microphones side by side, a few centimeters apart, same height, both facing the source. Hold them away from your body and any hard surface.
- Measure for 30 to 60 seconds and compare the energy average, not the instantaneous reading. Comparing two jumping numbers by eye is how people end up calibrating to noise.
- Subtract. Reference meter 78.4, phone 74.6, so your phone reads 3.8 dB low and your offset is +4 (most apps take whole decibels).
- Verify at a different level. Repeat at something quieter and something louder. If the gap is roughly constant, the offset is good. If it swings by more than a few dB, the mismatch is in the frequency response and no single offset will fix it.
- Write down what you did: the phone model, the date and the offset. Offsets belong to one phone, not to the app.
Method 2: an acoustic calibrator and an external microphone #
This is how real meters are checked. An acoustic calibrator is a small barrel that fits over a microphone and produces a known tone. The Extech 407766, for example, produces 1,000 Hz at 94 dB or 114 dB, accurate to ±0.5 dB at 94 dB, and meets IEC 60942.
The catch is fit. Calibrators are made for standard half-inch and quarter-inch measurement microphones, not for a hole in the side of a phone. To use one properly you need an external measurement microphone that plugs into the phone, which is also the setup NIOSH recommends for serious work: its Sound Level Meter app is accurate within ±2 dBA on its own, and meets the Type 2 requirements of IEC 61672-3 with a calibrated external microphone. The NIOSH app can calibrate either the internal or an external microphone.
If you’re buying an external mic and a calibrator anyway, you’re close to the price of a decent Class 2 meter. Sound meter app vs SPL meter compares the two paths.
Method 3: sanity checks with no equipment #
Without a reference, you can’t calibrate. You can still catch a phone that’s badly wrong.
- Check a quiet room and a conversation. A quiet bedroom at night should read well below a conversation at a meter away, which usually lands in the 60s. If your phone shows 55 dB in a silent room or 45 dB during a conversation, something is off.
- Compare two phones and two apps. Run them side by side on the same steady source. Agreement doesn’t prove accuracy, but a phone that disagrees with everything else is the suspect.
- Don’t calibrate to a YouTube tone. The level reaching the mic depends on the speaker, the volume setting and the room, so the “94 dB test tone” on a laptop tells you nothing about absolute level.
- Check your habits before blaming the phone. A case over the mic, a finger on the bottom edge, wind, or a phone resting on a hard surface all move readings by several decibels. Our guide to measuring noise with a phone covers technique.
Setting the offset in Sound Meter dB #
In Sound Meter dB, calibration lives in Settings. The offset adjusts in steps of 1 or 5 decibels, up to ±30 dB in either direction, with a reset that puts it back to zero. Once set, it’s applied to every reading, and the offset in force is stored alongside each saved session along with the weighting and response you used, so an old measurement stays interpretable after you change the setting.
Two other settings matter while you calibrate: frequency weighting (A-weighted dBA, the standard for noise rules, or flat) and response (Fast at 125 ms or Slow at 1 second). Match whatever the reference meter is using. The app is free, works offline and doesn’t record audio, and its own documentation is clear that even a calibrated phone is an estimate rather than a certified measurement.
How often should you re-check? #
Phone microphones don’t drift the way a precision capsule does, but a few events are worth a re-check:
- After a phone replacement or repair, since the offset belongs to the hardware.
- After a major OS update that changes audio behavior.
- Before a series of measurements you plan to keep, such as a noise log for a neighbor complaint.
- If a reading looks implausible against a familiar sound.
For day-to-day comparisons, an uncalibrated phone is often good enough, because the error is roughly constant and cancels out when you subtract two readings from the same phone.
Frequently asked questions #
Can I calibrate a sound meter app without a sound level meter? #
Not properly. Without a known reference you have nothing to compare against, so the best you can do is sanity-check against familiar sounds and against other devices. If you need real accuracy, borrow or rent a calibrated meter for ten minutes.
Does calibrating a phone app make it accurate enough for OSHA compliance? #
No. Workplace noise measurements need a calibrated Class 1 or Class 2 sound level meter, or a dosimeter, with documented calibration. A phone with an offset is useful for awareness and for deciding when to get real measurements taken. See class 1 vs class 2 sound level meters.
Why do two phones show different decibel readings in the same room? #
Different microphone capsules, different placement in the case, different audio processing, and different assumed sensitivities in the app. Differences of several decibels between models are normal, which is exactly what an offset is for.
Should I calibrate with A-weighting or flat? #
Whichever you plan to use, and the same on both instruments. Since almost all noise guidance is in dBA, calibrate in A-weighting unless you have a specific reason to work with flat readings. dB vs dBA explains the difference.
My phone reads lower than the meter at low levels but matches at high levels. What now? #
That pattern points to the microphone’s noise floor and its frequency response, not to a constant error, so an offset can’t fix both ends. Calibrate for the range you care about most and treat readings far from it as rough.