Title :
Voice activity detection based on complex Laplacian model
Author :
Chang, Joon-Hyuk ; Kim, Nam Soo
Author_Institution :
Sch. of Electr. Eng. & INMC, Seoul Nat. Univ., South Korea
fDate :
4/3/2003 12:00:00 AM
Abstract :
A voice activity detector (VAD) based on the complex Laplacian model is proposed. The likelihood ratio based on the Laplacian model is computed and then applied to the VAD operation. According to experimental results, it is found that the Laplacian statistical model is more efficient for the VAD algorithm compared to the Gaussian model.
Keywords :
discrete Fourier transforms; probability; speech coding; speech enhancement; Laplacian statistical model; VAD algorithm; complex Laplacian model; discrete Fourier transform coefficients; goodness-of-fit test; likelihood ratio; noisy speech distribution; speech enhancement; variable-rate speech coding; voice activity detector;
Journal_Title :
Electronics Letters
DOI :
10.1049/el:20030392
