Mic Sensitivity Analyzer
A guided 3-step test profiles your microphone: stay silent, then speak normally, then speak loudly. The tool measures your noise floor, typical speech level, loud peak, signal-to-noise ratio and dynamic range (all in dBFS), then rates the mic’s sensitivity and suggests a gain adjustment.
๐ Everything runs locally in your browser. No audio is recorded, uploaded, or stored โ only level numbers are computed.
Microphone
Guided test
Live level
Sensitivity profile
How the Sensitivity Analysis Works
Microphone "sensitivity" describes how much signal a mic produces for a given amount of sound. In a browser we can’t read calibrated sound-pressure levels, but we can measure the relative levels (in dBFS) your mic produces in three conditions and turn that into a practical profile. The tool guides you through silence, normal speech, and loud speech, sampling the input level many times per second in each phase.
From those three captures it derives the numbers that actually matter for setting up a mic: how quiet your background is (noise floor), how strong your voice comes through at a normal level (speech level), how loud it can get before trouble (loud peak), the gap between your voice and the noise (signal-to-noise ratio), the span from quietest to loudest (dynamic range), and how close the loud peak is to clipping (headroom).
What the numbers mean
- Noise floor โ the level with no one speaking. Lower (more negative dBFS) is better; a high floor means a hissy mic, high gain, or a noisy room.
- Speech level โ your normal voice. This drives the sensitivity rating: too hot risks clipping, too quiet means you need more gain.
- Signal-to-noise ratio โ speech minus noise floor. Bigger is better; under ~20 dB and your recordings will sound noisy.
- Dynamic range โ loud peak minus noise floor; how much range your mic + room usefully capture.
- Headroom โ how far your loud peak sits below 0 dBFS. Little headroom means you’ll clip when you get excited; aim for several dB.
Tips for a good measurement
- Run it in your normal setup โ same room, distance, and gain you actually use.
- During silence, genuinely stay quiet so the noise floor is accurate.
- Speak at a consistent, natural distance for the speech step; don’t lean in only for the loud step.
- The tool requests the raw signal (auto-gain off), but some systems force processing that can flatten the results.