ALGEBRAIC VECTOR QUANTIZATION OF LSF PARAMETERS WITH LOW STORAGE AND COMPUTATIONAL-COMPLEXITY

This correspondence presents quantization scheme for encoding line spe ctral parameters used in linear predictive coding (LPC) of speech. The scheme is based on low-dimensionality regular-point lattices. The alg ebraic codebook need not be stored, and the optimum codevector is foun d through simple rounding of the input vector. Thus, the scheme result s in significant savings of memory and reduced computational complexit y when compared to traditional vector-quantizer solutions. The quantiz er achieves an average spectral distortion of about 1 dB at 28 b/frame for the telephone bandwidth.