Microphone Tester

About Microphone Tester

Microphone Tester uses the Web Audio API to visualize your voice or instrument in real time. Quickly spot low gain, background hum, or clipping. Swap inputs without leaving the page and listen back through your speakers if needed.

1. Choose the microphone you want to evaluate and press Start testing.
2. Speak at your normal volume; watch the meter for healthy green/yellow levels.
3. Enable monitor volume to hear the mic live (use headphones to avoid feedback).
4. Review the activity log to confirm peaks and ensure the signal stays consistent.

How can I tell if my microphone is working?

Click Start, accept the browser microphone permission prompt, then speak normally about thirty centimeters from the mic. You should see the live waveform jump on every syllable, the volume meter rise into the green-yellow band, and the peak hold marker climb to roughly minus twelve to minus six dBFS for conversational speech. If the meter stays flat at minus infinity, the wrong device is selected — use the dropdown to choose another input. If the meter is pinned at zero, the input gain or boost is too high and clipping is destroying audio quality. A working microphone should respond to a finger tap on its grille with a clearly visible spike on the waveform.

Why does the browser ask for permission every time?

The W3C MediaDevices.getUserMedia() specification requires an explicit user grant for each origin and each session by default; this is the same gate that protects your webcam. Chrome and Edge let you mark the permission as Allow on subsequent visits per site, but Safari deliberately re-prompts to make microphone access more visible. You can pin the permission by clicking the lock icon next to the URL, selecting Site settings, and changing Microphone from Ask to Allow. Doing so is safe on this tool because no audio leaves your device — captured frames are analysed directly in the browser via the Web Audio API and discarded as soon as the test ends.

What is dBFS and what level should I aim for?

dBFS stands for decibels relative to full scale, where 0 dBFS is the loudest signal the analog-to-digital converter can encode without clipping. Speech should peak between minus twelve and minus six dBFS — loud enough to bury room noise (which sits below minus sixty on a quiet PC) but with six decibels of headroom for sudden plosives like the P in pop. Singing and acoustic instruments should peak around minus six to minus three dBFS. Anything that touches zero dBFS for more than a few samples is clipping; the waveform flattens at the top, harmonic distortion enters the recording, and the file is permanently degraded. Use a pop filter and back off five centimeters from the capsule to tame plosives.

What is the difference between condenser and dynamic microphones?

Condenser microphones use a thin charged diaphragm above a fixed back plate to capture audio with very high sensitivity and a wide frequency response, typically 20 Hz to 20 kHz. They need 48-volt phantom power and pick up every room reflection, so they shine in treated studios for vocals, voiceover, and acoustic instruments. Dynamic microphones use a moving coil attached to a diaphragm inside a magnetic field — no power required, lower sensitivity, narrower frequency response around 50 Hz to 16 kHz. They reject background noise and handle high sound pressure, so they dominate live stages, podcasts in untreated rooms, and broadcast booths. The Shure SM7B and Electro-Voice RE20 are the classic dynamic broadcast choices.

Why does my recording sound thin, muddy, or distant?

Three issues account for most poor sound. Thin and tinny audio usually means you are too far from the capsule and the room reverb dominates; move within fifteen centimeters of a cardioid mic and aim your mouth slightly off-axis. Muddy or boxy audio is room resonance below 300 Hz bouncing off parallel hard walls; add absorption (a duvet behind the mic helps), engage a high-pass filter at 80 Hz in the recording software, or pick a more directional pattern. Distorted audio is clipping — lower the input gain in your OS sound panel until peaks sit at minus six dBFS. Hiss is preamp self-noise; cheap USB mics with built-in preamps often hiss above 50 dB-A.

What sample rate and bit depth should I record at?

For voice calls, podcasting, and most YouTube content, 48 kHz at 16-bit is the practical sweet spot — it matches video standards, halves file size versus 24-bit, and exceeds the Nyquist requirement for the human voice range (which tops out near 8 kHz for clear speech). For music production, archival, or post-processing that involves EQ and compression, record at 48 kHz or 96 kHz with 24-bit depth: the extra eight bits of dynamic range give you about 48 dB of additional headroom for processing without introducing quantisation noise. Higher rates like 192 kHz are usually overkill outside specialised mastering and double storage costs. This tester reports the actual sample rate the browser hands the Web Audio API, which on most devices is 44100 or 48000 Hz.

What is audio gain staging and why does it matter?

Gain staging is the discipline of setting an appropriate signal level at every stage of the recording chain: microphone, preamp, interface, operating system, and recording application. The goal is to keep the signal well above each stage's noise floor (commonly minus sixty dBFS) and well below clipping (zero dBFS), with peaks landing near minus twelve to minus six dBFS so subsequent processing has headroom. A common rookie mistake is leaving the OS slider at 100 percent and turning the mic preamp down, which amplifies preamp noise instead of the source. The correct order is preamp first (set the analog gain so the loudest part peaks at minus six dBFS on the interface meter), then leave the OS at 100 percent and the application fader near unity.

Is my voice uploaded anywhere when I use this tool?

No. The microphone test uses MediaDevices.getUserMedia() to receive an audio stream directly from your operating system into a MediaStreamAudioSourceNode inside the browser's Web Audio API. From there an AnalyserNode reads frequency and amplitude data for the live meters and waveform, and the raw stream is never written to disk, never encoded, and never sent over the network. The moment you stop the test or close the tab, the MediaStreamTrack is stopped and the audio buffers are garbage collected. You can verify by opening the developer tools Network tab and confirming zero requests during the test, and by checking the browser's address bar microphone indicator vanishes the instant you press Stop.

Which browsers and devices are supported?

The tester works in any browser that exposes the W3C MediaDevices.getUserMedia() API plus a Web Audio AnalyserNode over a secure (HTTPS) context: Chrome 55+, Edge 79+, Firefox 36+, Safari 14.1+ on macOS, Safari on iOS/iPadOS 14.5+, and Chrome/Samsung Internet on Android. It does not run on Internet Explorer or on insecure HTTP pages — in those cases the tool detects the missing API and shows an unsupported message instead of pretending to listen. iOS Safari additionally requires the page to be served over HTTPS and only grants microphone access in direct response to a tap, so press Start testing yourself rather than expecting it to auto-start. USB, Bluetooth, built-in array, and XLR-via-interface microphones all appear in the device dropdown as long as the operating system already recognises them.

What is the difference between RMS and true peak, and what noise floor passes QA?

The big number on the stats card is RMS dBFS — the root-mean-square level over each analysis frame, which tracks perceived loudness and is the stable reading you certify a unit against. True peak is the loudest single sample held for the whole session; it catches transient clipping that an averaged RMS misses. For QA pass/fail, measure the idle noise floor by staying silent for a few seconds: a healthy consumer mic should sit at minus sixty dBFS RMS or lower, a decent USB/headset mic around minus fifty-five to minus fifty, and anything above minus forty-five indicates audible hiss or room noise that should fail inspection. Signal-to-noise ratio is simply your spoken RMS (around minus eighteen to minus twelve dBFS) minus that idle floor; aim for at least forty dB of SNR. The float-based meter in this tool can read genuine minus sixty dBFS floors, which a coarse 8-bit meter cannot.

How do I export or save a mic test result?

Run a test, then press Stop. The Export CSV and Export JSON buttons in the Session statistics card become active and download the session straight to your device — no upload involved. Each file contains a one-second-interval timeline (timestamp, RMS dBFS, peak dBFS, and a clip flag) plus a summary block holding the device label, the audio sample rate, total duration, min/max/average RMS, the held true-peak, and the clipping-sample count. CSV opens directly in Excel, Google Sheets, or any plotting tool and is ideal for attaching to a repair ticket or batch QA report; JSON is the better choice when you want to feed the result into a script or a test-management system. Start a new test to clear the previous capture.