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.
Honey bee colonies produce a constant acoustic background: wingbeat buzz from thousands of workers, the low steady hum of a broodnest, and distinctive tonal events like queen piping or the sudden roar that follows queen loss. Researchers and advanced hobbyists insert a small microphone or accelerometer into the hive, record continuously, and turn that sound into a spectrogram — a picture of frequency versus time — to non-invasively track queen presence, swarm preparation and overall colony health without opening the box.
Studies on queen "tooting" and worker "quacking" sounds date back to the 1950s, but only with cheap digital recorders, machine-learning classifiers and long-duration in-hive sensors has continuous acoustic monitoring become practical for everyday apiaries rather than just research colonies.
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.
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.
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.
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.