Title :
Two-stage sigma-delta modulation
Author :
Wong, Ping Wah ; Gray, Robert M.
Author_Institution :
Dept. of Electr. & Comput. Eng., Clarkson Univ., Potsdam, NY, USA
fDate :
11/1/1990 12:00:00 AM
Abstract :
An exact analysis of the nonlinear difference equation describing a discrete time two-stage sigma-delta quantizer with DC and sinusoidal inputs is presented. It is shown that for DC inputs, the binary quantizer error of the second quantizer, which appears at the output of the modulator as a second-order difference, is asymptotically white, uniformly distributed, and uncorrelated with the input. For sinusoidal inputs, this result holds with one exception: the autocorrelation, which is a sum of weighted Bessel functions, does not correspond to that of a white noise process. At large oversampling ratios, the autocorrelation is approximately white and hence the white additive noise model is a good approximation for evaluating the first- and second-order moments. It is found that for both a sinc3 decoder filter and an ideal low-pass filter decoder, the average quantization noise power is inversely proportional to the fifth power of the oversampling ratio R for both types of inputs. This represents an improvement over the single-loop sigma-delta quantizer by a factor of R2 . Simulation results that agree with the theoretical analysis are presented
Keywords :
decoding; delta modulation; low-pass filters; nonlinear differential equations; random noise; signal processing; DC inputs; average quantization noise power; binary quantizer error; discrete time two-stage sigma-delta quantizer; low-pass filter decoder; nonlinear difference equation; sinc decoder filter; sinusoidal inputs; white additive noise; Additive noise; Analytical models; Autocorrelation; Decoding; Delta-sigma modulation; Difference equations; Low pass filters; Quantization; Signal to noise ratio; White noise;
Journal_Title :
Acoustics, Speech and Signal Processing, IEEE Transactions on