摘要
近年來，資訊多媒體裝置的發展迅速，語音處理相關的議題也成為相關研究的焦點，無論是語音辨識或合成都是現今研究的重心。 以語音合成而論，如何讓電腦學習發出類似人類的說話方式，如斷詞判斷、語調節奏及文意表達是相關研究的目標，而在多方的發展下，現今電腦語音合成已能發出清晰語句，但在語調節奏上，仍是機器的平穩音調。 而本論文重點，即是著重於讓電腦學習判斷句子的語調，結合隱藏式馬可夫模型及調整語言參數及韻律參數，讓語音合成更接近人類的自然發音。 並且利用市面上的影音光碟來做比較，雖然無法完全如人類自然發音的結果，但相較其他線上的語音合成系統，是有較佳的韻律感及聲調。 將電腦語音合成結合現今的多媒體裝置，像是平板電腦、電子書等成為閱讀或語言學習的有利工具。
並列摘要
In recent years, the rapid development of information media devices, voice processing-related issues has become the focus of research, either voice recognition or synthesis are the focus of the present study. In terms of voice synthesis, issued a similar learning how to make the computer human way of speaking, such as breaking words judge, tone and rhythm of expression context is relevant objective of the study, and in the development of multi-party, the modern computer speech synthesis has been able to give a clear statementbut in tone on the rhythm, the machine is still smooth tone. The focus of this paper, that is Zhuchong to let the computer learn to judge the tone of the sentence, combined with Hidden Markov Model and adjust the parameters and language of rhythm parameters, so close to humanspeech synthesis more natural pronunciation. CD-ROM and use the market to do more, although not completely as the result of human naturalpronunciation, but compared to other online speech synthesis system, there is a better sense of rhythm andtone. Computer speech synthesis with today's multimedia devices, such as tablet PCs, e-books and other reading or language learning as a powerful tool.