ACTIVE SPEAKER DETECTION

Active Speaker Detection
API with Anti-Flip-Flop Logic

An active speaker detection API is a programmable interface that identifies which participant in a video conference is currently speaking and emits real-time events for UI updates, auto-zoom, and simulcast layer priority. V100's implementation runs server-side in Rust with exponential smoothing, hysteresis (1-second minimum switch time, 500ms hold time), audio level beacons, auto-zoom integration, and a ranked speakers list -- eliminating the distracting flip-flopping that plagues naive client-side implementations. Backed by 10 passing tests.

Rust server-side detection
10 tests passing
Auto-zoom integration