Phoneme-based spectral voice conversion using temporal decomposition and Gaussian mixture model | IEEE Conference Publication | IEEE Xplore