Estimation From Quantized Gaussian Measurements: When and How to Use Dither

• Published in: IEEE transactions on signal processing 2019-07, Vol.67 (13), p.3424-3438
• Main Authors: Rapp, Joshua, Dawson, Robin M. A., Goyal, Vivek K
• Format: Article
• Language: English
• Subjects: alpha-trimmed mean; Dependence; Economic models; Gaussian distribution; generalized Gaussian distribution; Iterative methods; L-estimator; Maximum likelihood estimation; Maximum likelihood estimators; Measurement; midrange; Noise; Noise measurement; Normal distribution; order statistics; Quantization; Quantization (signal); Random variables; Solvers; Statistical methods; subtractive dither
• ISSN: 1053-587X; 1941-0476
• DOI: 10.1109/TSP.2019.2916046

Subtractive dither is a powerful method for removing the signal dependence of quantization noise for coarsely quantized signals. However, estimation from dithered measurements often naively applies the sample mean or midrange, even when the total noise is not well described with a Gaussian or uniform distribution. We show that the generalized Gaussian distribution approximately describes subtractively dithered, quantized samples of a Gaussian signal. Furthermore, a generalized Gaussian fit leads to simple estimators based on order statistics that match the performance of more complicated maximum likelihood estimators requiring iterative solvers. The order statistics-based estimators outperform both the sample mean and midrange for nontrivial sums of Gaussian and uniform noise. Additional analysis of the generalized Gaussian approximation yields rules of thumb for determining when and how to apply dither to quantized measurements. Specifically, we find subtractive dither to be beneficial when the ratio between the Gaussian standard deviation and quantization interval length is roughly less than one-third. When that ratio is also greater than 0.822/K^{{\text{0.930}}} for the number of measurements K>\text{20}, estimators we present are more efficient than the midrange.