Direct Modeling of Frequency Spectra and Waveform Generation Based on Phase Recovery for DNN-Based Speech Synthesis.

Shinji Takaki, Hirokazu Kameoka, Junichi Yamagishi

Browse the full Interspeech paper archive.