A simple method to estimate the amplitude and standard deviation of sinusoidal signals corrupted with additive Gaussian noise is proposed. For this, a two-parameter model is developed by sorting the samples of the signal. This reduced parametric model allows robust parameter estimation, even if the phase function of the sinusoid is nonlinear, discontinuous, and unknown. The functionality and performance of the proposed method are analyzed by several computer simulations; the used GNU Octave program is provided. The proposed method can be useful for unbiased envelope estimation in fringe pattern normalization among other potential applications.