IMPROVED HARMONIC+NOISE MODEL FOR VOCAL AND MUSICAL INSTRUMENT SOUNDS

摘要

The paper presents an improvement to Harmonic+Noise Model (HNM) for quality analysis/synthesis of vocal and mono-phonic music instrument sounds. So far these models were considered not to offer sufficient quality for modeling of singing voice or music instruments. The problems with HMN model are presented and it is shown that it can not handle certain specific types of music instruments as well as highly reverberant recordings. In this paper we present a new analysis method that significantly improves the HNM performance. The main difficulty in the estimation of the stochastic part of the signal is overcome using non-linear (polyspectral) methods.