Text-to-speech (TTS) applications have been applied in diverse areas all over the world. Considering the fact that Swahili pronunciation is not complicated, and the language spoken by about 45 – 100 million people as their first or second language,, we considered the feasibility, and developed a Swahili Text-to-Speech (TTS) system. This paper gives an account of the Swahili TTS system developed...

A Preliminary Study on Fully Neural Network-based Speech Synthesis System 廖書漢 SHU-HAN Liao ,蔡亞伯 YaBo Chai , 廖元甫 a Yuan-Fu Liao, a 國立台北科技大學電子工程系 [email protected], [email protected], [email protected] 摘要 傳統的語音合成使用先文字分析後語音合成的架構,但是這種兩階段的作法, 通常會有,若前級分析錯誤,就會影響後級合成,且無法挽救的問題。因此,在 本論文中我們希望嘗試把前後級,全部都改成以類神經網路實現,以便將來可以 直接合成一個大的端對端語音合成類神經網路。主要的想法是,直接以字元串為 輸入單位,並盡量用大量未標記語料,進行非監督式類神經網路訓練。我們的系 統包含四個子網路,分...

The Blizzard Challenge 2005 was a large scale international evaluation of various corpus-based speech synthesis systems using common datasets. Six sites from around the world, both academic and industrial, participated in this evaluation, the first ever to compare voices built by different systems using the same data. Here we describe results of the evaluation and many of the observations and l...

To examine the relationship between speech perception and production in second language acquisition, this study investigated whether training in the perception domain transfers to improvement in the production domain. Native speakers of Japanese were trained to identify English /r/-/l/ minimal pairs. Recordings were made of the subjects’ productions of minimal pairs before and after identificat...

A new series of studies on adults' and infants' perception of phonetic "prototypes," exceptionally good instances of phonetic categories, show that prototypes play a unique role in speech perception. Phonetic category prototypes function like "perceptual magnets" for other stimuli in the category. They attract nearby members of the category, rendering nem more perceptually similar to the catego...

During conversations, speakers are assumed to provide information about their emotional state. This study is a systematic investigation of the voice characteristics of the emotion of anxiety. For this purpose speech samples of 63 subjects were collected in a situation which induced mild anxiety. On the basis of the acoustic measurements it appears that anxiety affects vocal parameters and showe...

This paper describes a method of concatenative speech synthesis that makes use of 3-dimensional labelling of speech, and shows how this can be applied to the synthesis of both mono-lingual and foreign-language speech. The dimensions encode phonetic, prosodic, and voicequality information in order to fully describe the acoustic characteristics of each speech segment.