Digital microphone diagnostic
Microphone Noise Floor Test
Measure the quiet digital floor of your current microphone input for 30 seconds. Results are real dBFS samples processed locally.
30-second local test
Measure your microphone noise floor
Start the test in the quietest practical setting. The timer stops automatically after 30 seconds. Values are raw dBFS, not environmental dBA.
Current digital level
-- dBFS
Digital full-scale level
- Current
- --
- Minimum
- --
- Energy average
- --
- Maximum
- --
Audio is processed in this browser only. No recording or measurement sample is uploaded.
What the result means
dBFS describes level relative to the maximum digital signal the browser can receive. Values are usually negative; 0 dBFS is digital full scale, not silence. A lower, more negative quiet-room result usually means less digital input noise, but microphone gain and browser processing can change it.
How to run a useful test
- Choose the microphone you actually plan to use and close calling or recording apps.
- Move away from fans, keyboards and handling noise, then leave the device untouched.
- Start the test and let the full 30 seconds complete. Compare only tests made with the same device and settings.
What this test cannot tell you
The result is not dBA or dB SPL and cannot describe environmental loudness without microphone sensitivity data and calibration. Read dBFS vs dB SPL, then see calibration and browser measurement accuracy.
What numbers to expect from a quiet desk
Park a laptop in a silent study, close chat and recorder apps, and let the full 30 seconds run. A modern USB headset might idle near minimum −68 dBFS with Leq near −64 dBFS, while the same room on a built-in laptop microphone beside a spinning fan might sit near minimum −48 dBFS with Leq near −44 dBFS. Both describe digital headroom below 0 dBFS, the ceiling where clipping starts, not loudness in a room. More negative usually means a quieter input path, yet a hotter preamp gain or leftover browser processing can lift the whole result without any change in the air around you.
Compare two inputs without fooling yourself
- Test in the same quiet spot within minutes: same chair position, same browser, same 30-second length.
- Select one input at a time, close calling tabs first, and keep hands off the desk while sampling.
- Write down current, minimum, Leq and maximum plus the input name. Compare only results from the same device and settings, because sensitivity and gain travel with the hardware.
Capture is mono over HTTPS after a once-per-origin permission, with echo cancellation, noise suppression and automatic gain requested off where the browser permits it. When the panel shows idle dashes it means no samples yet, never a placeholder value.
When a high floor is a warning sign
- Maximum sticks only a few decibels above Leq in a dead-quiet room: something steady like a fan, charger whine or keyboard vibration is feeding the capsule.
- Two runs ten minutes apart differ by more than about 6 dB with nothing moved: a call, recorder or second tab likely grabbed the microphone or its processing changed.
- The floor jumps when you plug in a dock or hub: that path has its own gain, so treat it as a different input and retest from scratch.
This diagnostic cannot become a room-loudness figure on its own. Pair it with the background-noise workflow for environmental context, the measurement methodology for the signal path, and microphone troubleshooting when an input refuses to start.