How Loud Is This?
Measuring

How to calibrate a phone sound meter

Calibrating means comparing your phone with a trusted reference and entering the difference as an offset. It usually takes a few minutes and is the single biggest improvement you can make to phone readings.

Updated · 5 min read

On this page

To calibrate a phone sound meter, measure the same steady sound with your phone and with a trusted reference, then enter the difference in the meter’s offset field. If the reference reads 72.0 dBA and your phone reads 68.5 dBA, the offset is +3.5 dB. The best reference is a sound level meter that meets IEC 61672; an acoustic calibrator (94 or 114 dB at 1 kHz) also works if you can fit it to the microphone. Recheck the result at a second level, and recalibrate whenever your device or setup changes.

Why phones need calibrating

Every phone microphone, and the electronics behind it, has a slightly different sensitivity. Two phones of the same model can disagree, and different models can be several decibels apart. Browsers and operating systems can also change gain. The online sound meter turns off the browser’s speech processing automatically, but it can’t know how sensitive your particular microphone is. An offset fixes that one big unknown.

Calibration is per device and per setup. If you use the meter on a phone and a laptop, calibrate each one separately, and in each browser you use.

Method 1: compare with a reference sound level meter

This is the most practical method for most people. A friend, a workplace safety officer, a school science department or a hire shop may have a meter you can borrow.

  1. Set the reference meter to A-weighting and, if it has one, Leq mode. Use Fast or Slow; Leq is unaffected by the choice.
  2. Make a steady, broadband sound at a moderate level, around 65 to 80 dBA. A fan on a constant setting, running water, or pink noise played through a decent speaker all work. Avoid pure tones, music and speech; they vary too much and can produce uneven readings across the room.
  3. Put both microphones side by side, as close together as you can, at the same distance from the source (about 1 metre is fine) and away from walls. Point both the same way.
  4. Measure for 30 to 60 seconds with both at the same time, and compare the two Leq values.
  5. Swap positions and repeat once. Average the two differences. This cancels small differences between the two spots.
  6. Enter the offset: reference minus phone. If the phone reads low, the offset is positive; if it reads high, it is negative.
  7. Check at a second level, 10 to 15 dB quieter or louder. If the difference is still within about 1 dB, your offset is good across that range.

Method 2: use an acoustic calibrator

An acoustic calibrator is a small device that produces a precise tone, usually 94 dB (1 pascal) and sometimes 114 dB, at 1 kHz inside a sealed cavity. Professionals use one to check their meter before and after every measurement. Calibrators are specified in IEC 60942.

The catch is that calibrators are built to fit standard measurement microphones, typically half-inch capsules. A phone microphone is a tiny opening in the case, so you can’t seal the calibrator over it properly. If the seal leaks, the level reaching the phone will be lower and less stable than the nominal value, and your offset will be wrong.

  • Use a calibrator only if you have an adapter that seals over the phone’s microphone opening, or an external measurement microphone that fits it.
  • Prefer the 94 dB setting. At 114 dB many phone audio chains are near their limits.
  • Because 1 kHz has no A-weighting correction, the reading should be 94 dB in A, C or Z weighting.
  • If the reading wobbles by more than a decibel as you adjust the fit, the seal isn’t good enough to trust.

Sanity checks without a reference

These won’t calibrate the meter, but they help catch something badly wrong:

  • Quiet room check. In a very quiet room at night, the reading shows your device’s noise floor. A phone typically bottoms out somewhere in the 30s or low 40s dBA. If your quiet bedroom reads 55 dBA with nothing running, something is off. Don’t set the offset from this, because the floor is the device’s self-noise, not the room.
  • Consistency check. Measure the same steady sound, such as a running tap, from the same spot on two days. Results should agree within a decibel or two.
  • Ballpark check. Normal conversation at about 1 metre is usually around 55 to 65 dBA. A reading of 40 or 80 dBA suggests a problem such as a blocked microphone. The decibel chart has more reference points.

Limits of a single offset

An offset shifts every reading by the same amount. It can’t fix:

  • Frequency response. Phones usually lose sensitivity at low bass, so bass-heavy sound may still read low after calibration. See dBA vs dBC vs dBZ for why this matters more for dBC.
  • Noise floor. Very quiet sounds still bottom out at the device’s self-noise.
  • High levels. Many phones distort or limit from about 100 dB, and no offset will recover that.
  • Hidden processing. Some devices apply audio processing in the operating system that a web page can’t disable. It can make the error depend on the type of sound.

So calibrate with the kind of sound you plan to measure, at a similar level. For more on getting good readings, see how to measure noise with your phone.

Phone meters vs Class 1 and Class 2 meters

Professional sound level meters are built and tested to IEC 61672, which defines two performance classes:

Class 1Class 2Calibrated phone
TolerancesTightestWiderNot specified or tested
Frequency rangeWidest, accurate into deep bass and high trebleNarrower at the extremesVaries by model; often weak in deep bass
Typical useEnvironmental noise, legal and planning work, researchWorkplace surveys, general noise checksEstimates, before-and-after comparisons, personal awareness
VerificationPeriodic lab verification and field calibrator checksPeriodic lab verification and field calibrator checksYour own offset against a reference

Class 1 and Class 2 meters can measure accurately at levels well beyond what a phone can handle, and many record peak levels, octave bands and statistical levels such as L90. A phone can’t claim either class, however carefully it is calibrated. If a result will be used for compliance, a legal dispute or a planning decision, use a properly verified meter, and treat your phone as a way to know when that is worth doing.

Frequently asked questions

Can I calibrate a phone sound meter without a reference meter?

Not accurately. You need a known sound level to compare against, which means a sound level meter or an acoustic calibrator. Guessing from a “quiet room” or from a published level for an appliance can easily leave you 5 dB or more out. Without a reference, use the meter for comparisons on the same device rather than absolute numbers.

What is 94 dB in calibration?

94 dB is the sound pressure level of 1 pascal, a standard reference level. Acoustic calibrators produce a 1 kHz tone at 94 dB, and often 114 dB too, inside a small cavity that fits over a measurement microphone.

How often should I recalibrate?

Recheck after changing phone, browser, case or microphone, and after a major operating-system update, since any of these can change the audio path. Otherwise an occasional check against a reference is enough for everyday use.

Will calibration make my phone as good as a professional meter?

No. A single offset corrects the overall sensitivity at the level and type of sound you calibrated with. It doesn’t fix the phone’s frequency response, its noise floor or its distortion at high levels, and it doesn’t make the phone a Class 1 or Class 2 instrument.