Hum

My home network has a mood. Mornings are a slow trickle of phones waking up; evenings are a flood of streams and syncs; at 3am it hums to itself, running backups nobody watches. Hum makes that audible.

A small server on the mini PC reads the live state of the network: throughput, active connections, traffic spikes. The synthesizer runs in the browser with Web Audio, so the knobs respond instantly and no sound card is needed anywhere.

How the network plays it:

• Active connections set the span of an arpeggiated bed; throughput sets brightness and note density
• A traffic spike rings a chime, quantized to the beat so it lands musically
• The root key follows the time of day: bright at dawn, low and dark at night

What keeps it musical:

• A chord-progression engine plays real changes (the axis progression, the dorian vamp, a stretched ii-V-I) phrase by phrase, with occasional twists: neighbor substitutions and sus voicings
• Five moods (Deep, Haze, Glass, Pulse, Aurora), each a full modular patch: a vector of ~20 parameters covering the oscillator mix, two reverb busses, the bass pattern, even which instruments play. A mood knob interpolates between them, and the engine slews so every sweep is a slow crossfade
• The ensemble: a 16-step bass sequencer, FM electric piano, a Karplus-Strong plucked string, a vibrato flute lead, blooming swells, and glitter
• A reference analyzer: feed it a YouTube link and it extracts tempo, key, brightness, and dynamics from the audio, then generates a matching patch

It also taught me a physical lesson about feedback: an early plucked-string voice had a loop gain just over 1, and notes would swell back out of the silence a second after they died. Open source release coming.