COMPUTER SCIENCE. CALCULATION EQUIPMENT. MANAGEMENT

Creating a database for modeling system for speech synthesis based on a deep convolution neural networks

E. S. Izrailova

Academy of Sciences of the Chechen Republic, Grozny State Technical Oil University, 364051, Chechen Republic, Grozny, Ordzhonikidze square, 100

Abstract: The stages of creating a phonetic-acoustic database, adapted for learning the system of automatic speech synthesis based on deep convolutional neural networks, are described. A phonetic transcriptor designed with consideration of the grapheme-phonemic transformations of the Chechen language is presented.

Keywords: speech database, automatic synthesis of Chechen speech, end-to-end speech synthesis system, machine learning, phonetic transcriptor.

UDC: 004.934.5

Received: 23.11.2018

Bibliographic databases:

•