Weighted Estimation of Harmonic Components in a Musical Sound Signal

The study of musical sound has become a popular research field. Harmonic regression signal plus noise statistical models have been used to analyze sound signals. However, it is common to give estimates of harmonic parameters without indications of their uncertainties. Least squares estimates for harmonic models have been studied and asymptotic variance expression have been developed. In practice, window-based estimates are used. This paper studies the statistical properties of such estimates; in particular, we use asymptotic variance expressions to develop standard errors and construct confidence intervals. We present applications and examples of the statistical techniques to musical sound signal analysis