What spectral analysis shows
Any sound is a mix of many frequencies at the same time. A deep bass note has strong low frequencies; a cymbal has energy high up; a human voice has a basic pitch and many overtones above it. Spectral analysis splits a sound into these frequencies, using a method called the Fast Fourier Transform (FFT), and shows how loud each one is. It lets you see what you hear and also what you might miss.
The analyzer runs in your browser. A microphone signal or file is analysed on your device and is not sent anywhere.
The three views
- Spectrum: frequency from left (low, 20 Hz) to right (high, 20 kHz), level from bottom to top in decibels. It shows the sound right now. Good for finding a hum, a whistle or the main frequency of a note.
- Spectrogram: time moves from one side to the other, frequency goes up, and colour shows the level. It shows how the sound changes. Speech, birdsong and music each leave typical patterns, and short noises like clicks are easy to spot as vertical lines.
- Oscilloscope: the waveform itself, the way the air pressure moves. Useful for checking clipping, polarity and simple test tones.
Reading the spectrum
| Range | Frequencies | What you find there |
|---|---|---|
| Sub-bass | 20–60 Hz | rumble, kick drum body, mains hum at 50/60 Hz |
| Bass | 60–250 Hz | bass guitar, warmth of voices |
| Low mids | 250–500 Hz | muddy or boxy sound if too strong |
| Mids | 500 Hz–2 kHz | body of most instruments and speech |
| High mids | 2–6 kHz | clarity of voice, harshness if too strong |
| Highs | 6–20 kHz | air, cymbals, hiss, "s" sounds |
Peak frequency and note name
The analyzer finds the strongest frequency and shows it in hertz together with the nearest musical note, for example 440 Hz = A4. This works well for a single sustained tone like a whistle, a held sung note or one string. For chords it shows the loudest component, which may be an overtone rather than the note you expect.
Practical uses
Keep in mind that every analyzer makes a trade-off: fine frequency detail needs a longer piece of sound, so it reacts more slowly, while a fast display blurs frequencies that are close together. A short click is therefore easier to see in the spectrogram than in the spectrum.
- Find the exact frequency of a hum before removing it with Noise Reduction.
- Check how much bass or treble a mix has before using the equalizer.
- Check whether a file is real high quality: an MP3 converted from a low bitrate often has nothing above 16 kHz, even if it now has a high bitrate.
- Test your microphone and room, together with the microphone test.
- Tune an instrument or check your singing by watching the note name of the peak.
- Learn acoustics: whistle, sing a vowel, or play a single note, and watch how the overtones appear as evenly spaced lines above the main frequency.
When you export a PNG, you get exactly the current view. This is useful for documenting a problem, comparing two microphones, or showing students what a sound looks like. For a fair comparison, analyse both sources with the same view and the same input level.