Observation-Model Error Compensation For Enhanced Spectral Envelope Transformation In Voice Conversion
Fernando Marquez Villavicencio, Jordi Sanjaume Bonada

A strategy to enhance the signal quality and naturalness was designed for performing probabilistic spectral envelope trans- formation in voice conversion. The existing modeling error of the probabilistic mixture to represent the observed enve- lope features is translated generally as an averaging of the in- formation in the spectral domain, resulting in over-smoothed spectra. Moreover, a transformation based on poorly mod- eled features might not be considered reliable. Our strategy consists of a novel definition of the spectral transformation to compensate the effect of both over-smoothing and poor mod- eling. The results of an experimental evaluation show that the perceived naturalness of converted speech was enhanced.