Title :
Optimal transformation of LSP parameters using neural network
Author :
Vu, Hai Le ; Lois, Laszlo
Author_Institution :
Dept. of Telecommun., Tech. Univ. Budapest, Hungary
Abstract :
The intraframe correlation properties of line spectrum pair (LSP) are used to develop an efficient encoding algorithm using the Karhunen-Loeve (KL) transformation. The important nonuniform statistical characteristics of LSP frequencies are investigated. Based upon this nonuniform property the neural network based techniques for generating the transform vectors via system training are studied. Using the principal component analysis (PCA) network to decorrelate LSP coefficients, we show that these new approaches lead to as good or better distortion as compared to other methods for speech analysis-synthesis
Keywords :
correlation methods; neural nets; spectral analysis; speech coding; speech processing; speech synthesis; transform coding; transforms; Karhunen-Loeve transformation; LSP parameters; decorrelation; distortion; efficient encoding algorithm; intraframe correlation properties; line spectrum pair; low bit rate; neural network; nonuniform statistical characteristics; optimal transformation; principal component analysis network; speech analysis-synthesis; speech coding; system training; transform vectors; Encoding; Filters; Frequency; Karhunen-Loeve transforms; Linear predictive coding; Neural networks; Principal component analysis; Quantization; Speech analysis; Speech synthesis;
Conference_Titel :
Acoustics, Speech, and Signal Processing, 1997. ICASSP-97., 1997 IEEE International Conference on
Conference_Location :
Munich
Print_ISBN :
0-8186-7919-0
DOI :
10.1109/ICASSP.1997.596194
