Design and Implementation of a Diphone Database for Persian Language for Computational Linguistics Applications
Seyedsaeed Ayat · 2010
Speech databases play important role in researches and implementations of the systems related to computational linguistics. In this paper, after studying different phone units that can be used for this purpose, I will present different steps for implementing a diphone database. Therefore, at first I have prepared a corpus for Persian language. Then I have designed and implemented a software that extracts a special word from the corpus and specifies the diphones that must be extracted from that. In the next step, after studying the conditions for increasing the quality, we recorded the speech signals that contain the diphones. In the last step, the diphones extracted from the recorded signals. We evaluated three different methods for finding the diphone borders: listening, visiting the time signal, and studying the spectrogram of the signal, and we use all of them for the accuracy of our database.