Use of generation process model for synthesizing fundamental frequency contours in HMM-based speech synthesis | IEEE Conference Publication | IEEE Xplore