Title :
VQ codebook design using improved simulated annealing algorithms
Author :
Lu, Ngoc-Ai ; Morrell, Darryl R.
Author_Institution :
Arizona State Univ., Tempe, AZ, USA
Abstract :
The authors investigate the design of codebooks for vector quantization (VQ) using three simulated annealing algorithms: generalized simulated annealing (GSA), simulated annealing with an inner loop criterion (ISA), and simulated annealing combined with a generalized Lloyd algorithm (SA-C). The algorithms´ performance is experimentally evaluated by creating codebooks for speech waveform quantization. It is found that GSA and ISA both provide acceptable performance, although the GSA algorithm typically achieves a higher signal-to-noise ratio
Keywords :
data compression; encoding; speech analysis and processing; SNR; VQ codebook; generalized Lloyd algorithm; generalized simulated annealing; inner loop criterion; signal-to-noise ratio; simulated annealing algorithms; speech waveform quantization; vector quantization; Algorithm design and analysis; Convergence; Costs; Distortion measurement; Instruction sets; Iterative algorithms; Optimization methods; Simulated annealing; Speech analysis; Vector quantization;
Conference_Titel :
Acoustics, Speech, and Signal Processing, 1991. ICASSP-91., 1991 International Conference on
Conference_Location :
Toronto, Ont.
Print_ISBN :
0-7803-0003-3
DOI :
10.1109/ICASSP.1991.150429