Text processing and acoustic signal in Brazilian Portuguese: a linguistic-engineering interface for science and speech technology

• Grant number: 93/00565-2
• Support Opportunities: Research Projects - Thematic Grants
• Start date: January 01, 1994
• End date: July 31, 1996
• Field of knowledge: Linguistics, Literature and Arts - Linguistics
• Principal Investigator: Eleonora Cavalcante Albano
• Grantee: Eleonora Cavalcante Albano
• Host Institution: Instituto de Estudos da Linguagem (IEL). Universidade Estadual de Campinas (UNICAMP). Campinas , SP, Brazil
• Associated research grant(s): 96/04822-8 - Osamu Fujimura | Ohio State University - Estados Unidos, AV.EXT

Abstract

The construction of interfaces between text processors and speech signal processors is seen as a task to be divided between groups of specialists in Linguistics and Engineering in order to develop suitable models for the Portuguese language and at the same time to implement prototypes of synthesizers and speech recognizers. Schemes of concatenative synthesis and of formants with various levels of linguistic processing are proposed for implementation in Brazilian Portuguese. A recognition scheme will also be developed for a small vocabulary of isolated words. The research will be carried out jointly by a team from the Institute of Language Studies (IEL) and two teams from the complementary areas of the School of Electrical Engineering (FEE) of the State University of Campinas (UNICAMP). The initial objective of the project is to create an interface between Linguistics and Engineering which will tackle the following tasks: 1) phonetic-acoustic description to provide feeds for synthesis schemes; 2) phonetic-acoustic description to provide feeds for recognition schemes; 3) design of joint processing systems for text and signals. Using this shared base it will be possible to tackle the following themes, dividing them into tasks of the linguists' responsibility and/or of the engineers: 1) synthesis; 2) recognition; 3) linguistic processing; 4) signals processing. The integration of themes relating to the text and to the signal within a processing model, in accordance with the proposal for this project will have the additional advantage of allowing a new type of environment to emerge in Brazil, where multidisciplinary scientific and technological production naturally promote the training of professionals in the new and promising speech Science and Technology. (AU)