Tone Question of Tree Based Context Clustering for Hidden Markov Model Based Thai Speech Synthesis

Abstract

Problem statement:In HMM-based Thai speech synthesis, tone is an important issue that brings about the intelligibility of the synthesized speech. Tone distortion resulted from imbalance of the training data should be appropriately treated. Approach: This study described an HMM-based speech synthesis system for Thai language. In the system, spectrum, pitch and state duration are modeled simultaneously in a unified framework of HMM, their parameter distributions are clustered independently by using a decision-tree based context clustering technique. The contextual factors which affect spectrum, pitch and duration, i.e., part of speech, position and number of phones in a syllable, position and number of syllables in a word, position and number of words in a sentence, phone type and tone type, are taken into account for constructing the questions of the decision tree. Since Thai is a tonal language, tone questions play an important role in the context clustering process. Results: An experimental result compared F0 contours between those of synthesized speech with and without tone questions; furthermore the size of Thai speech corpus is varied to investigate the synthesized speech quality. Conclusion: By using the tone questions in the tree-based context clustering process, the tone distortion is relieved significantly.

Fujisaki’s Model of Thai’s Fundamental Frequency Contours with Environmental Noises

“…Both male and female speech has been constructed in the speech database. The extraction tools are applied in this study (Mixdorff and Fujisaki, 1997;Chomphan and Kobayashi, 2007a;2007b). For each parameter, the frequency distribution over its range is constructed, subsequently the distributions of standard Thai are plot in a graph.…”

Improvement of Tone Intelligibility for Average-Voice-Based Thai Speech Synthesis

“…As for speaker dependent HMM-based Thai speech synthesis research, a speech synthesizer has been implemented (Chomphan and Kobayashi, 2007a). In the system, a group of contextual factors which affect spectrum, fundamental frequency (F0) and state duration, such as tone type and part of speech are taken into account especially for the purpose of producing natural sounding prosody of the tonal speech.…”

“…It has been found that it can provide speech with the better reproduction of prosody over the unit-selection-based Vaja TTS system from NECTEC (National Electronics and Computers Technology Center) (Chomphan and Kobayashi, 2007b). A decision tree with a toneseparated structure (Chomphan and Kobayashi, 2007a) presents the considerable improvement of tone correctness of the synthesized speech. However, some distortion of syllable duration is noticeable when the system is trained with a small amount of data.…”

“…However, some distortion of syllable duration is noticeable when the system is trained with a small amount of data. The other structures of the decision tree are designed for not only the purpose of maximal correctness of tone but also the purpose of elimination of the syllable duration distortion (Chomphan and Kobayashi, 2007a).…”

Evaluation of Hidden Markov Model Based Adaptive Provisioning of Optical Burst Switching Networks Amenable for Upgradation to Green Flexigrid Networks

“…Therefore appropriate predictive schemes for path-dependent connection-holding times will be required. Hidden Markov Model (HMM), provides the solution for wide variety of applications such speech tagging and noun-phrase chunking (Chomphan, 2012), is also flexible enough to predict any type of internet traffic (Dainotti et al, 2008) and has been demonstrated for wavelength assignment in WDM networks by Johnson et al (2001). In this study, we use thresholding of HMM predicted path-dependent connection holding times and show that this connection-time impairment aware provisioning of Burst Sizes and Associated Wavelength Allocation (ABS-WA) can reduce delays and improve throughput.…”