Near optimum filtering of quantized signals

The need for representing signals by a finite number of bits implies that quantization noise is present in almost all digital signal processing systems and inherently occurs in the analog-to-digital conversion process. The distortion error, or quantization noise, consists of the difference between the input to the quantizer and the discrete output signal. A near-optimum method for filtering out the quantization noise is presented. Use is made of the result that the spectrum of the quantization noise is related to the probability density function of the signal derivative