Title :
Low-complexity predictive trellis-coded quantization of speech line spectral frequencies
Author :
Kang, Sangwon ; Shin, Yongwon ; Fischer, Thomas R.
Author_Institution :
Sch. of Electr. Eng. & Comput. Sci., Hanyang Univ., Kyunggi-Do, South Korea
fDate :
7/1/2004 12:00:00 AM
Abstract :
Low-complexity block-constrained trellis coded quantization (BC-TCQ) structures are studied, and a predictive BC-TCQ encoding method is developed for quantization of line spectrum frequency (LSF) parameters for speech coding applications. The performance is compared with the linear predictive coding (LPC) vector quantizers used in the AMR-WB (ITU-G.722.2) and IS-641 speech coding standards, demonstrating reduction in spectral distortion (SD) and significant reduction in encoding complexity.
Keywords :
linear predictive coding; spectral analysis; speech coding; trellis codes; vector quantisation; AMR-WB standard; BC-TCQ encoding; IS-641 standard; ITU-G.722.2 standard; LPC; LSF parameters; line spectrum frequency parameters; linear predictive coding vector quantizers; low-complexity block-constrained trellis coded quantization; narrowband speech; predictive coding; speech coding; tail-biting approach; wideband speech; Computer science; Convolutional codes; Costs; Encoding; Frequency; Lattices; Linear predictive coding; Quantization; Rate-distortion; Speech coding;
Journal_Title :
Signal Processing, IEEE Transactions on
DOI :
10.1109/TSP.2004.828916
