8th International Conference on Spoken Language Processing

Jeju Island, Korea
October 4-8, 2004

Hidden Semi-Markov Model Based Speech Synthesis

Heiga Zen (1), Keiichi Tokuda (1), Takashi Masuko (2), Takao Kobayashi (2), Tadashi Kitamura (1)

(1) Nagoya Institute of Technology, Japan
(2) Tokyo Institute of Technology, Japan

In the present paper, a hidden-semi Markov model (HSMM) based speech synthesis system is proposed. In a hidden Markov model (HMM) based speech synthesis system which we have proposed, rhythm and tempo are controlled by state duration probability distributions modeled by single Gaussian distributions. To synthesis speech, it constructs a sentence HMM corresponding to an arbitrarily given text and determine state durations maximizing their probabilities, then a speech parameter vector sequence is generated for the given state sequence. However, there is an inconsistency: although the speech is synthesized from HMMs with explicit state duration probability distributions, HMMs are trained without them. In the present paper, we introduce an HSMM, which is an HMM with explicit state duration probability distributions, into the HMM-based speech synthesis system. Experimental results show that the use of HSMM training improves the naturalness of the synthesized speech.

Full Paper

Bibliographic reference.  Zen, Heiga / Tokuda, Keiichi / Masuko, Takashi / Kobayashi, Takao / Kitamura, Tadashi (2004): "Hidden semi-Markov model based speech synthesis", In INTERSPEECH-2004, 1393-1396.