Does anyone know of any literature or research concerning skewness in the sampled values of speech or audio? Thanks, Matt