A voice activity detection algorithm for wireless communication systems with dynamically varying background noise

Speech can be modeled as short bursts of vocal energy separated by silence gaps. During typical conversation, talkspurts comprise only 40% of each par ty's speech and remaining 60% is silence. Communication systems can achieve spectral gain by disconnecting the users from the spectral resource during silence periods. This letter develops a simple and efficient Voice Activit y Detection (VAD) algorithm to work in a mobile environment exhibiting dyna mically varying background noise. The VAD uses a classification method invo lving the full-band energy, ratio of low-band energy to full-band energy, z ero-crossing rate, and peakiness measure.