In this paper, a DPCM (Differential Pulse Code Modulation) system with forward adaptive Lloyd-Max’s quantizer is presented. This quantizer is designed for low bit rate, where the first and the second order linear predictors are used in the proposed DPCM system solution. It is shown how SQNR (Signal to Quantization Noise Ratio) and Gp (Prediction Gain) depend on the correlation coefficients of the predictor of the first and the second order. The obtained experimental values of parameters SQNR and Gp are presented and compared with the corresponding theoretical values for the given system. In this manner the selection of optimal DPCM system correlation coefficients values is performed and the possibilities of this DPCM system application in the speech coding are indicated.

Keywords: Differential pulse code modulation (DPCM), Lloyd-Max’s quantizer, Forward adaptation, Predictor, Correlation coefficient, Prediction gain, Signal to quantization noise ratio