Compare noise before and after a change

1. Decide what you are changing

Choose one comparison: fan low versus high, window open versus closed, or a device before versus after maintenance. Write down the two conditions. Change one thing at a time so the result has a clear explanation.

Use the same microphone, browser, input gain and calibration throughout. A reference instrument is needed if you need absolute levels; a consistent setup is the starting point for a relative comparison.

2. Mark the position

Place the device on a stable surface. Mark the microphone position, orientation and distance from the source, for example 1 m from the same side of a fan. Keep the microphone uncovered and outside direct airflow.

Note background activity before starting. If traffic, conversation or a door interrupts one run, record it and repeat that run under a comparable condition.

3. Run both conditions for the same duration

Let the source settle, then measure condition A for 60 seconds and save the result. Change only the chosen setting, allow it to settle, and measure condition B for the same duration. Repeat the pair three times if practical.

The tool shows captured audio duration. Keep the page visible and the screen on; switching apps or locking the screen can end a run. Use a new measurement for each condition.

4. Compare the averages, then explain the peaks

Place the session averages side by side and look for a consistent difference across the repeated pairs. Use the history and your notes to explain unusual maximum readings. One brief event can increase the maximum without representing the whole run.

5. Keep the comparison honest

In that illustrative pair, the average differs by 6.6 dB but the maxima differ by only 0.5 dB. The maximum alone would hide the sustained difference. Repeat the runs before assuming the pattern persists.

Those figures are invented for explanation, not typical fan levels or calibration targets. If you moved the microphone or changed input settings between runs, repeat the comparison.

• Record the setup with your results
• Resolve calibration and input errors