invisible layer / FFT spectrum
experiment 03 · microphone · all browsers
what the air
is carrying right now.
The full acoustic spectrum of your environment — from infrasound rumble to near-ultrasonic. Mains hum at 50 or 60Hz. HVAC resonance. The unique acoustic fingerprint of every room. Sounds you cannot hear but your phone already captures.
your room has a unique acoustic signature. Reflection patterns, resonant frequencies, background hum — combined, they can identify a specific room with >90% accuracy. Advertisers have used ultrasonic beacons embedded in TV ads to detect which room you're in across multiple devices simultaneously.
range:
peak freq
peak dB
noise floor
waiting for microphone
0 Hz22 kHz
tap the spectrum to inspect a frequency
waterfall — time → · freq ↑
detected peaks
start microphone to detect peaks
▶ how does this work?

The Web Audio API feeds your microphone through an AnalyserNode running a Fast Fourier Transform (FFT). This converts the time-domain audio signal into frequency bins — each bin shows how much energy exists at a specific frequency.

With an FFT size of 4096 and 44.1kHz sample rate, each bin covers ~10.8Hz. The waterfall scrolls left-to-right: time progresses rightward, frequency increases upward, so you can see how the spectrum evolves. Peaks are detected when a bin exceeds the noise floor by >15 dB.

Known signatures: 50Hz = European mains · 60Hz = North American mains · 100/120Hz = LED flicker harmonic · 18–22kHz = ultrasonic beacon or pest repeller.