Compare room noise before changing your setup

What do you want to change?

Pick the decision you actually need to make. For the fan, that could be whether the lower speed reduces the sound reaching your chair enough to be worth using. Leave the fan where it is and measure from the desk. Moving it at the same time would make the result harder to explain.

Call the usual setting A and the lower setting B. For a window test, A could be open and B closed. Write down the two conditions and keep the door and other appliances as they were. If your question is about a combined change, such as closing both a window and a door, say so in the notes.

Where the phone sits matters

Use the same device and browser for every run. Put the phone on a stable support with its microphone opening clear, and keep its orientation unchanged. A small tape mark makes it easier to return to the same spot if you need to move it.

A phone lying on fabric can pick up a different signal from one standing on the desk. Leaning over it can also block sound. Start the run, then step back to the same place each time.

If you are comparing two seats, move only the measuring device between the marked positions. Keep the fan setting fixed. That test describes those two listening positions; it does not describe every part of the room.

Steady fans and passing traffic

Thirty to sixty seconds is a manageable starting point for a steady fan. Let it settle before each run and use a similar duration for A and B. A refrigerator that starts and stops every few minutes needs a longer observation that includes the cycle you care about.

Jot down the time and anything that could change between runs. Traffic outside an open window may become quieter while you are closing it. Someone talking in the next room may stop.

Keep the page visible and the device awake. If the browser pauses or a sampling gap stops the session, that is missing observation, not a quiet stretch to include in the comparison.

Why take one more reading?

1. Set up A, let the sound settle, then measure. Stop and save the report.
2. Change to B. Wait for sounds from handling the fan or window to pass, then measure for a similar duration.
3. Return to A and take another measurement.

The second A run checks whether the background changed while you were testing B. If the two A runs differ noticeably, try again at a steadier time. There is no universal acceptable difference for this browser setup. Look for a result that repeats across runs, rather than choosing the pair with the largest change.

Start a fresh session for each condition and leave the calibration offset alone. An offset change moves the displayed number even when the microphone signal stays the same.

Which number should you compare?

Compare Average first. It accounts for the energy and duration of the valid sampled windows, so a loud interruption can have more influence than its brief appearance suggests. Check Min, Max, and the history to see what happened during the run.

If you bumped the phone, repeat that run and note why. If you are testing traffic through a window, passing vehicles belong in the observation. Removing every noisy vehicle would answer a different question.

The live chart shows the latest 30 seconds. A saved report gives an overview of the full session and its sampled statistics. In a long report, neighboring windows are combined in the energy domain to keep the curve readable; summary statistics retain all valid samples. The overview cannot show every brief change separately.

What the result supports

If B repeatedly has a lower estimated Average and the A runs are reasonably consistent, you have a useful comparison for that setup. Keep the duration and phone position with the report.

The size of the difference can still be affected by microphone processing and the frequencies of the sound. Our accuracy limits explain those constraints. This test does not certify a product's noise reduction, establish a legal limit, or show that a room is safe.

For an absolute level or workplace exposure decision, use an appropriate calibrated instrument and procedure. NIOSH's app information shows how sound measurement validation depends on defined hardware and test conditions; it does not validate this browser tool.