00000ctm a22000004a 4500 UP-99796217608094582 Buklod 20230503092542.0 m |o d | ta 080122s2007 xx r |||| u|eng d (iLib)UPD-00000523637 DENGII eng LG 993.5 2007 E64 M46 Mendigorin, Alfredo Jr. S. Automatic English text to Filipino speech translator using dictionary-based translation and concatenative synthesizer with prosody development for Filipino Speech Alfredo S. Mendigorin Jr. 2007. ix, 51 leaves + 1 computer laser optical disc (4 3/4 in.) Thesis (B.S. Computer Engineering)--University of the Philippines, Diliman. Nowadays, automatic-speech translation is of great interest and importance. Furthermore, this offers unparalleled technological support in communicating with people of different languages. The objective of this project is to implement an English text to Filipino speech translation: Machine Translation and Text-to-Speech Synthesis, and determine if the text-to-speech translator is feasible to be implemented in a cellular phone or a personal digital assistant. Machine Translation translates texts from English to Filipino. This was implemented using a translation by directly mapping the English words or phrases to their corresponding Filipino translation. Text-to-speech synthesis was done for the conversion from Filipino text to Filipino speech. A closely-related project which focused on text-to-speech synthesis of the Filipino speech was done in the UP DSP laboratory. It was entitled Prosody Development for Filipino Text-to-Speech Systems created by Melvin O. Co. MSEE. The research outputs a synthesized Filipino speech with prosody. Although the resulting synthesized speech has been found to be prosodically correct, the text-to-speech system can still be further enhanced by choosing the most suitable prosody model for Filipino speech. Two models were compared in this research. The first is the one implemented by Melvin Co, and the other is using MOMELINTSINT. The latter was successful in automatically modeling the intonation of speech while the previous has been shown that it performs well even when implemented on a different programming language. The Mean Opinion Score Test compared the synthesized speech and the concatenated speech. The synthesized speech with prosody got a score of 4.00 out of five in the listening session for overall impression, while the concatenated speech with no prosody got a score of 3.64. The accuracy in determining the kind of utterance was determined using the Declarative/Interrogative Test. It was determined that the listeners identified the correct utterance 97.25% of the time. Machine translating. Speech synthesis. Speech processing systems. Filipino language Versification Automation. English language Translating into Filipino. Automatic speech translation system. Text-to-speech Synthesis. Speech translator. Thesis FI UP UPD DENG-II LG 993.5 2007 E64 M46 Thesis