A Speech Retention Algorithm Based on Blind Source Separation
Wearing noise-canceling earmuffs blocks both noise and speech signals.To directly transmit environmental speech signals into the earmuffs,this paper proposes a speech Retention algorithm based on blind source separation.The algorithm uses blind source separation to decompose the mixed environmental signals into noise and speech signals.Then,based on the short-term energy characteristics of the signals,conditions for identifying the speech signal are defined and used to preserve the speech signal.Through simulation experiments,the identification conditions are determined.Finally,simulations are conducted under various noise types and signal-to-noise ratio environments,successfully preserving the speech signal and verifying the effectiveness of the speech-preservation algorithm.
speech retentionblind source separationshort-time energy of signal
NETL
NSTL
万方数据
