← 🎧 Education

🎧 Hive Acoustics

Deeper probe = stronger, cleaner hive signal. Wind and traffic raise the noise floor and lower the SNR.
Highlight 300–500 Hz piping band
Dominant frequency: —
Signal-to-noise: —
Colony read: —
FPS: —
The science

Each state is modelled as a sum of Gaussian spectral lines, S(f) = base + sum A_k exp(-((f - f_k)/w_k)^2), scaled by a probe attenuation 0.15 + 0.85 d (d = mic depth 0-1) and added to a uniform noise floor. Workers' wingbeat hum sits near 200-300 Hz; queen piping/tooting and pre-swarm quacking lie roughly at 300-500 Hz.

Signal-to-noise: SNR = 10 log10( sum S^2 / sum N^2 ) over the 1 kHz band shown. Pre-swarm piping is pulsed at about 1.4 Hz.

Honest limits: the spectra are synthetic and stylised, tuned to published ranges, not recordings; the attenuation law is linear, not a real comb acoustics model; no FFT of a real waveform is computed. Real classifiers use many more features and hive-specific calibration.

Drag — rotate · Scroll — zoom

🎧 Listening to the Hive

A probe microphone inside a 3D hive cutaway feeds a live spectrogram and circular frequency analyzer, showing how colony state, microphone placement and background noise combine to produce the acoustic signal researchers use to read a colony without opening it.

🔬 What It Demonstrates

Each colony state — queenright, queenless, pre-swarm piping, or winter cluster — has a distinct frequency signature. The simulation synthesises that signature live and renders it as a scrolling spectrogram and a spectrum-analyzer ring, the same kind of view a bioacoustic monitoring rig would produce from real hive audio.

🎮 How to Use

Pick a colony state, then slide the microphone deeper toward the broodnest or shallower toward the hive wall to see signal strength change, and raise ambient noise to see the signal-to-noise ratio degrade. Toggle the 300–500 Hz band to see where queen piping and swarm tooting/quacking typically sit.

💡 Did You Know?

Continuous acoustic monitoring only became practical for ordinary apiaries in the last decade or so — cheap digital recorders, in-hive sensors and machine-learning classifiers turned a research curiosity into a tool hobbyists can run year-round.