Audio Spectrum Analyzer

Press Start analyzer to see what your microphone picks up as a live frequency spectrum and a scrolling spectrogram, and use them to find hum, buzz or whine. Levels are relative dBFS, not sound pressure.

Audio is analyzed in your browser to draw the spectrum. Nothing is recorded, stored or uploaded.

Names appear once you allow microphone access.

Stopped

Peak frequency
Peak level
Bin width

Spectrogram: newest sound at the top, about 10 s shown.

View
Hum markers
Resolution

Browser processing

Echo cancellation
Noise suppression
Auto gain control

Runs entirely in your browser. Test results, audio and video never leave your device.

How to find hum, buzz or whine

  1. Choose your microphone from the Microphone list, or leave Default microphone. Press Start analyzer and allow microphone access when the browser asks.
  2. Stay quiet for a few seconds and watch the spectrum. Steady hum, buzz or whine shows as narrow spikes, and as vertical lines in the spectrogram below it, where the newest sound is at the top. If nothing rises while you speak, check the mic with the Microphone Test.
  3. Under Hum markers, choose 50 Hz or 60 Hz to match the mains power where you are. Spikes on the dashed lines point to electrical hum; spikes on many of them usually sound like buzz. Keep Resolution on Standard or Fine for this: with Fast, bins are 23.4 Hz wide at 48 kHz, too coarse to tell 50 Hz from 60 Hz.
  4. Read Peak frequency for the loudest frequency. Bin width shows how far off it can be: choose Fine under Resolution to tell close frequencies apart, which makes the display react more slowly.
  5. Press Freeze to hold the picture while you look. The Peak hold line keeps the highest level at each frequency until you press Clear peaks.
  6. Press Stop when you are done. The browser releases the microphone.

What is an audio spectrum analyzer?

An audio spectrum analyzer shows how strong each frequency in a sound is, from bass to treble, so a steady tone appears as a peak at its frequency. Mains hum shows as spikes at the local mains frequency, 50 or 60 Hz, and its multiples, and whine as a narrow spike higher up. In a browser it shows relative levels, not how loud a sound is in the room.

Hum, buzz, whine or hiss: what the spectrum shows

These are sounds in what your microphone picks up. If a speaker is what buzzes or rattles, use the Audio Distortion Test instead.

What you hearWhat the spectrum showsCommon causes
Hum: a low, steady toneA spike at the local mains frequency, 50 or 60 Hz, often with a few of its multiplesMains power getting into the audio: a ground loop, a nearby power supply or an unshielded cable
Buzz: a harsh, raspy toneSpikes at many multiples of 50 or 60 Hz, often past the last Hum markerMains interference from dimmers, chargers or a USB port shared with other devices
Whine: a thin, high toneA narrow spike higher up that may move when the computer works harderCoil whine from power parts in a computer or charger, or interference from a USB device
Hiss: a soft rushing soundRaised levels across the high frequencies, with no single spikeThe microphone’s own noise, made louder by a high input volume

How to get rid of microphone hum or buzz

  1. Unplug the laptop charger and watch the spikes again. If they drop, the hum comes through the power supply, often from a ground loop with other plugged-in equipment.
  2. Try another USB port or cable, and connect a USB microphone straight to the computer instead of through a hub.
  3. Move the microphone and its cable away from power bricks, chargers, dimmer switches, monitors and speakers.
  4. For a microphone on an XLR cable, check that both plugs are fully seated, and try another cable.
  5. After each change, record a few seconds with the Mic Test with Playback to hear whether the hum is gone.

What this test can detect

  • Mains hum: spikes at 50 or 60 Hz and their multiples, lined up with the Hum markers.
  • Buzz, whine and hiss: many evenly spaced spikes, a single high spike, or raised levels across the high frequencies.
  • The loudest frequency, shown as Peak frequency with its level and the bin width that sets how far off it can be.
  • Sounds that come and go, kept on screen by the spectrogram and the Peak hold line.
  • Which processing the browser really turned off, listed under Browser processing.

How it works

The page opens the microphone with getUserMedia, asking for echo cancellation, noise suppression and auto gain control to be off. A Web Audio AnalyserNode computes the spectrum of the latest 8192 samples every screen frame (2048 with Fast, 32768 with Fine); bin width is the sample rate divided by that number, 5.9 Hz at 48 kHz.

The display runs from −120 to 0 dBFS, where 0 dBFS is a full-scale sine; a tone between two bins can read up to about 1 dB low. Peak frequency is the center of the loudest bin from 20 Hz to 20 kHz. The spectrogram adds a row every 40 ms (120 ms with reduced motion) and keeps 250 rows.

Limitations

  • Levels are relative dBFS, measured against the loudest signal the input can carry. They are not sound pressure, so the page cannot tell how loud a sound is in the room, and the same sound reads higher when you raise the input volume.
  • It is not a calibrated analyzer or a frequency response measurement. The curve mixes the microphone’s response with the room’s, and small built-in microphones are often weaker in the deep bass and can exaggerate some high frequencies.
  • Peak frequency can be off by up to half the bin width, which runs from 23.4 Hz with Fast to 1.5 Hz with Fine at 48 kHz, and two tones only a few bins apart merge into one peak.
  • Some processing may stay on in some browsers. Safari lets a page control only echo cancellation, and turning it off also turns off Safari’s automatic gain. Some phones and headsets also process sound in hardware that no web page can reach. Browser processing shows what the browser actually applied, not what the hardware does.
  • Nothing above half the sample rate can appear, so the axis ends below 20 kHz at low sample rates. A Bluetooth headset microphone in call mode carries a narrower range of frequencies, so the highest ones may be missing.

Frequently asked questions

Can it tell how loud a sound is in the room?

No. It is not a decibel meter: its levels are in dBFS, relative to the loudest signal your input can carry, and they change with the input volume. Measuring sound pressure needs a calibrated meter, and a browser cannot calibrate a microphone.

Can it measure my microphone’s frequency response?

No. That needs a known test signal, a calibrated reference microphone and a controlled room. Here the curve shows whatever reaches your microphone, shaped by the microphone and the room together, so it is good for comparing setups, not for rating a microphone.

Why does the spectrum change when I choose another resolution?

Because each setting splits the sound into bins of a different width. Fine separates close frequencies but reacts more slowly, and Fast follows quick changes but blurs close frequencies together. Steady noise also reads lower with narrower bins, because each bin collects less of it, while a pure tone keeps its level.

Can I analyze the sound my computer is playing?

Not directly: the page analyzes what your microphone picks up. Play the sound out loud and hold the microphone near the speaker; the result then includes the speaker, the room and the microphone. The Tone Generator plays test tones for this.

Is my audio recorded or uploaded?

No. The sound is analyzed in your browser’s memory only to draw the spectrum and spectrogram, and none of it is saved or sent to a server. Pressing Stop releases the microphone. The privacy policy explains how the site handles data.