Title :
Speech enhancement based on second-order architecture and maximum entropy algorithm
Author :
Xiao, Yu ; Guangrui, Hu
Author_Institution :
Dept. of Electr. Eng., Shanghai Jiaotong Univ., China
Abstract :
Based on the idea of adaptive noise cancellation (ANC), a second-order architecture is proposed for speech enhancement. According to the information maximization theory, the gradient descent algorithm is proposed. By using real speech signals, the simulation results demonstrate that the new algorithm has good performance in speech enhancement. The main advantage of the proposed architecture is that a clean speech signal can be obtained with less filter distortion
Keywords :
adaptive filters; filtering theory; gradient methods; interference suppression; maximum entropy methods; optimisation; speech enhancement; adaptive noise cancellation; filter distortion; gradient descent algorithm; information maximization theory; maximum entropy algorithm; performance; second-order architecture; simulation results; speech enhancement; Computer architecture; Distortion; Entropy; Feedback; Filters; Noise cancellation; Signal processing; Source separation; Speech enhancement; Speech processing;
Conference_Titel :
Signal Processing Proceedings, 1998. ICSP '98. 1998 Fourth International Conference on
Conference_Location :
Beijing
Print_ISBN :
0-7803-4325-5
DOI :
10.1109/ICOSP.1998.770276
