Banco de dados orais: uma nova perspectiva aos estudos do português brasileiro

Abstract

This work represents an attempt to facilitate the automatic access to the spontaneous Portuguese data for studies on experimental phonetic area and speech synthesis, maintaining qualitative and quantitative information of the data. This corpus constitutes more than 30 hours of recording with samples of two Portuguese spoken dialects in Amazonia Region: the Amazon Portuguese (AM) and the Afro-Brazilian Portuguese (ABP). As a baseline, samples of the Standard Brazilian Portuguese (PB) were also used. The corpus inventory, the digitalization of the speech data and the phonetic segmentation using the PRAAT software preceded the construction of the oral database. Using language HTML and other resources of computer science a documentation of the database in a Internet site becomes possible.