Camera, Mic & Media
Spectrum Analyzer - Real-Time Audio Frequency Analyzer
Visualise the frequency spectrum of your microphone or system audio in real time. See which frequency ranges are most prominent using an FFT bar chart.
Reading the spectrum analyzer
The X-axis runs from 20 Hz (deep bass, left) to ~22 kHz (high treble, right) - the full range of human hearing. The Y-axis shows amplitude in decibels (dB): taller bars mean more energy at that frequency. The colour transitions from blue (quiet) to red (loud) as a quick visual cue for intensity.
Frequency bands
| Band | Range | What you hear |
|---|---|---|
| Sub-bass | 20–60 Hz | Deep rumble, kick drum thump felt more than heard |
| Bass | 60–250 Hz | Bass guitar, lower male vocals, warmth |
| Low-mid | 250–500 Hz | Body of most instruments |
| Mid | 500 Hz–2 kHz | Vocal clarity, snare, piano attack |
| Upper-mid | 2–4 kHz | Presence, intelligibility of speech |
| Treble | 4–8 kHz | Consonants, cymbals, sibilance ("s" sounds) |
| Air | 8–20 kHz | Sparkle, breath, ultra-high harmonics |
Peaks vs. noise floor
Sharp tall peaks indicate dominant frequencies - a pure tone or strong harmonic. The noise floor is the baseline level where no signal is present. A quiet recording has a low noise floor with peaks standing clearly above it. Broadband hiss or background noise raises the noise floor evenly across all frequencies.
Privacy
All audio processing happens locally in your browser using the Web Audio API. No audio data is transmitted to any server.
How to interpret your mix
When analyzing a music mix or recording, common problem areas to look for:
- Muddy bass (200–400 Hz): a buildup in this range makes the mix sound thick and unclear. Common causes are too many instruments occupying the same frequency range without EQ separation.
- Harsh vocals (2–4 kHz): excessive energy in the upper-mid range causes listening fatigue. This is often addressed with a narrow cut EQ.
- Sibilance (6–10 kHz): "s" and "sh" sounds that are too prominent or distorted. De-essing is a compression technique that targets this specific range.
dBFS vs. dBSPL
This analyzer shows levels in dBFS (decibels relative to full scale) - 0 dBFS is the maximum digital level before clipping. Levels below 0 dBFS are negative numbers; a well-mastered piece typically peaks at −1 to −0.5 dBFS.
dBSPL (decibels sound pressure level) is a physical measurement of actual air pressure - the loudness you hear in a room. 0 dBSPL is the threshold of human hearing. Normal speech is ~60 dBSPL; a rock concert is ~110 dBSPL. Digital audio level (dBFS) and physical loudness (dBSPL) are related by your system's output volume - this analyzer measures only the digital signal.
FFT window size
The analyzer uses a Fast Fourier Transform (FFT) to decompose audio into frequency components. A larger FFT window (e.g., 4096 or 8192 points) provides higher frequency resolution (more precise frequency bins) but lower time resolution (slower to respond to transients). A smaller window (512 or 1024 points) reacts quickly to changes but merges neighboring frequencies together. The default window size balances visual clarity and responsiveness for real-time analysis.