Multilingual Speech to Text Conversion – A Review

EasyChair Preprint 3064

Abstract

Speech is the first major primary need and the most convenient means of communication among individuals. Automatic Speech Recognition (ASR) introduces natural phenomena for man-machine communication. A great deal of work on various aspects of speech recognition and its implementations has been conducted for more than three decades. Speech recognition systems allow users to use speech as another form of input to communicate easily and efficiently with applications. A detailed study on automatic speech recognition is carried out and this paper offers an overview of the major technological perspective and appreciation of the fundamental progress of multilingual translation of speech-to-text conversion and also provides overview technique developed in each stage of speech-to-text conversion classification. It is possible to build an automated application to resolve the language barrier between countries and states within the world. The specification must include 4 modules of voice recognition, translation and speech synthesis, and the translated language text is provided. The goal of this review paper is to recapitulate and match different speech recognition systems and approaches for the conversion of multilingual speech to text.

Keyphrases: Automatic Speech Recognition, end-to-end (E2E) system, feature extraction tools and techniques, multilingual, speech-to-text conversion system (STT)

@booklet{EasyChair:3064,
  author    = {Saloni and Williamjeet Singh},
  title     = {Multilingual Speech to Text Conversion – A Review},
  howpublished = {EasyChair Preprint 3064},
  year      = {EasyChair, 2020}}