HMM-Based Distributed Text-to-Speech Synthesis Incorporating Speaker-Adaptive Training

摘要: In this paper, a hidden Markov model (HMM) based distributed text-to-speech (TTS) system is proposed to synthesize the voices of various speakers in client-server framework. The on speaker-adaptive training for constructing HMMs corresponding target speaker, and its computational complexity balanced by distributing processing modules TTS at both client server achieve real-time operation. other words, fewer complex operations, such as text inputs HMM-based speech synthesis, are conducted client, while training, which very operation, assigned server. It shown from performance evaluation that operates real time provides good synthesized quality terms intelligibility similarity.