Title :
Adaptive bit allocation for LSP parameter quantization
Author :
Hui, Fang ; McLoughlin, I.V.
Author_Institution :
Sch. of Comput. Eng., Nanyang Technol. Inst., Singapore
Abstract :
Line spectral pairs (LSPs) convey vocal tract information for reconstruction of speech in an analysis-synthesis speech coding system. Quantization of LSP parameters is necessary to reduce bit-rate, but causes unavoidable degradation in quality. This paper proposes an adaptive quantization method to reduce bit-rate whilst maintaining reasonable speech quality. The method exploits the observed relationship between LSP locations and underlying speech characteristics
Keywords :
adaptive signal processing; quantisation (signal); speech coding; LSP locations; LSP parameter quantization; adaptive bit allocation; adaptive quantization method; analysis-synthesis speech coding system; bit-rate reduction; line spectral pairs; quality degradation; speech characteristics; speech quality; speech reconstruction; vocal tract information; Bandwidth; Bars; Bit rate; Degradation; Information analysis; Linear predictive coding; Quantization; Spectral analysis; Speech analysis; Speech coding;
Conference_Titel :
Circuits and Systems, 2000. IEEE APCCAS 2000. The 2000 IEEE Asia-Pacific Conference on
Conference_Location :
Tianjin
Print_ISBN :
0-7803-6253-5
DOI :
10.1109/APCCAS.2000.913525
