On the Generation of Gaussian Noise

The paper by Chapeau-Blondeau and Monir [IEEE Trans. Signal Process., vol.50, no.9, p.2160-5, 2002 September] addresses the problem of synthesizing a generalized Gaussian noise with exponent 1/2 by means of a nonlinear memoryless transformation applied to a uniform noise. The transformation involves the Lambert W function and the computation of this function required development of a rational approximation as well as a C program. Here, we show that same could be performed more easily for a generalized Gaussian noise with any exponent