How to measure appliance noise at home, and whether to trust a phone app

Yes, you can measure how loud your fridge, heat pump or dishwasher is at home, and a phone is a reasonable place to start. But treat the number as a reading of your room, not a verdict on the product. A good app used carefully will tell you whether one setting is louder than another, whether a machine has got louder since you first measured it, and roughly where it sits. It will not reliably reproduce the figure on the spec sheet, and it was never designed to.

This page covers what the US National Institute for Occupational Safety and Health, NIOSH, found when it tested phone apps, why your reading and the maker's figure will differ, and a simple method that gives you readings you can compare with each other.

What NIOSH found when it tested phone apps

NIOSH studies workplace noise, and it was asked often enough whether phone apps could be relied on that it tested them. Its researchers described the work on the NIOSH Science Blog.

Most apps did not qualify. In the 2014 pilot study, a total of 130 iOS apps were examined and 10 met the selection criteria. Of 62 Android apps, only 4 partially met the criteria. NIOSH says Android apps lacked the features and functionalities found in iOS apps, which it put down to iOS audio capabilities, the open Android ecosystem and the many Android manufacturers using different components.

A few were close to a professional meter. Measured in a reverberant chamber against a reference system, three apps had mean differences within ± 2dBA of the reference for A-weighted readings. NIOSH noted that type 2 sound measurement instruments are specified to an accuracy of ± 2dBA, so some apps could potentially be used in the occupational setting, especially with a type 2 external microphone.

An external microphone helped a lot. A follow-up study using external microphones found sound level measurements within ±1 dB of the reference system, using the same test setup and apps. NIOSH concluded that external, calibrated microphones greatly improve the overall accuracy and precision of smartphone sound measurements, and remove much of the variability of built-in microphones.

The test range was loud. NIOSH generated pink noise at levels from 65 dB to 95 dB. That is a workplace range. Fridges, dishwashers and the indoor units of heat pumps are usually rated well below it, so NIOSH's results tell you nothing either way about how an app behaves at those quieter levels. That is a limit of the evidence, not a finding that apps fail there.

The NIOSH Sound Level Meter app

After that study, NIOSH worked with an app developer, EA LAB, on its own app. NIOSH's page lists what it offers:

• Free to download, and only available on iOS devices.
• "Accurate within ± 2 dBA", and tested according to standards in a reverberant chamber at the NIOSH acoustics laboratory.
• "Meets Type 2 requirements of IEC 61672:3 SLM standard with calibrated external microphone."
• All three major weighting networks, A, C and Z, with averages such as LAeq and TWA, maximum and peak levels, and noise dose.
• The capability to calibrate either the internal or an external microphone.

NIOSH is careful about what the app is for. Its 2017 announcement says the app "is not meant to replace a professional sound level meter or a noise dosimeter or be used for compliance purposes," and recommends that anyone interested in proper noise measurements use an external microphone that can be calibrated with an acoustical calibrator. With such a microphone, NIOSH says the app measured within ± 1 dB of the reference over its 65 to 95 dB SPL testing range. The same post adds that the app was not designed to calculate exposure metrics based on environmental or non-occupational noise limits.

For a home reading, the parts that matter are the A weighting, the maximum, and LAeq, which NIOSH lists among the averages: it gives you one number for a period of running rather than a jumpy live figure. Most of the dose and exposure features are for workers and can be ignored.

Why your reading will not match the maker's figure

Expect a gap, and do not read a fault into it. Our other explainers cover each reason in depth, so here is the short version.

• Distance. Sound fades as you move away, and almost no appliance maker says how far away its figure was taken. OSHA gives the open-space rule that sound pressure falls 6 dB each time the distance from a point source is doubled.
• The room. Your kitchen is not a test chamber. OSHA describes a reverberant field wherever sound waves reflect off hard surfaces such as walls and ceilings and combine with sound coming straight from the source. Tiled kitchens and bare floors add to what the phone hears.
• Background noise. The phone records everything in the room: traffic, a television, the fridge while you are trying to measure the dishwasher.
• The setting. Makers usually quote the quietest setting. If you measure the dishwasher's drain pump or the heat pump on boost, you are measuring a different moment. Our page on what dBA means goes through how makers choose the flattering figure.
• A different quantity. Some makers publish sound power, the total sound a machine gives off, rather than sound pressure at a spot, which is what any microphone reads. The two do not convert into each other by moving a decimal point. See sound power vs sound pressure.
• The test standard. A figure measured to a named standard fixes the setting, distance and room. Your kitchen measurement follows none of them. Our page on why dB ratings rarely line up across brands shows what that does to the numbers.

The useful conclusion is this: compare your readings with your own earlier readings, taken the same way, rather than with the box.

How to measure, step by step

This method is built for consistency. A careful reading you can repeat is worth more than a single precise one.

1. Pick a quiet time. Late evening or early morning, with windows shut and other appliances off where you safely can. Do not switch off a fridge or freezer for long, and do not disable anything that heats or protects the home.
2. Measure the background first. With the appliance off, or between cycles, take a reading and write it down. If the reading with the appliance running is barely above it, the phone is mostly hearing the room, and the result says more about your house than the machine.
3. Choose a fixed distance and spot. One metre in front of the appliance is easy to repeat. Mark the spot on the floor with tape, or pick the place you actually stand or sleep. For a heat pump, measure the indoor unit from your bed or sofa, and the outdoor unit from the patio or the property line if neighbors are the concern.
4. Hold the phone the same way each time. OSHA notes that your body can act as a barrier between the noise source and the microphone, and says to hold the meter away from your body. A phone stand or a pile of books at a set height works well. Keep the microphone end pointed toward the appliance.
5. Set A weighting and leave it there. In the NIOSH app the weighting is selectable in Settings. Do not mix A and C or Z readings in one comparison.
6. Use the same setting on the appliance. Same fan speed, same mode, same cycle and the same stage of the cycle. Write down which it was. A dishwasher washing and a dishwasher draining are two different readings.
7. Take several readings. Run the meter for a minute or more and note the LAeq and the maximum. Repeat three times, moving away and back between runs. If they disagree by a lot, something in the room is changing.
8. Record everything. Date, time, distance, weighting, the appliance setting and the background level. That record is what makes a later reading worth comparing.

If you want a better tool, NIOSH's advice is an external microphone that can be calibrated with an acoustical calibrator. OSHA's technical manual also requires pre and post use calibration for the instruments its own inspectors use, which tells you how seriously professionals treat calibration.

What a change in your readings means

Two rules of thumb come from official sources.

• 10 dB sounds about twice as loud. The National Park Service says an increase of 10 dB causes a doubling of perceived loudness and represents a ten-fold increase in sound level.
• 3 dB is a doubling of sound power. OSHA states as a general rule that doubling the sound power increases the noise level by 3 dB.

Set those against the accuracy figures above. NIOSH's criterion for apps was a mean difference of ± 2 dB, so a difference of a decibel or two between two sessions may be nothing more than the measurement. A difference that shows up every time, on the same setting, from the same spot, is real for your room. To see where everyday sounds sit on the scale, see our decibel chart.

When to suspect a fault

A reading that climbs over weeks, or a new sound that was not there before, is worth checking. Before calling anyone, run through the routine causes the makers themselves list: a unit that is not level, a clogged filter, loose hardware or a louder setting than you thought. We collected them in why your appliance got louder. Your written readings help here too: telling a technician "it was this level at one metre on normal cycle in spring, and it is clearly higher now" is more useful than "it seems louder."

If you are choosing a replacement, our guides to the quietest refrigerators, quietest dishwashers and quietest mini-splits quote each maker's figure with whatever setting and conditions it states.

We did not measure anything for this page, and we do not measure the products in our guides. The accuracy figures and advice above come from NIOSH, OSHA and the National Park Service, quoted as they publish them, and the method is a way to make your own readings consistent, not a substitute for a professional assessment.