Energy-Based Voice Activity Detection (VAD) identifies speech by measuring the audio signal's energy level. When the signal exceeds a predefined threshold, the system classifies it as speech rather than silence or background noise.
Voice AI platforms use Energy-Based VAD as a lightweight speech detection method for telephony, embedded devices, and low-latency applications. More advanced systems often combine it with machine learning-based detection.