Text to Speech

 

Introduction

Text-to-Speech (TTS) service converts text into audio data. It accomplishes this by dividing the inputted text into sentences, words, and phonemes. It then converts the divided segments into speech signals and synthesizes those signals into audio data.

 

 

TTS service by ThinQ.AI supports the following features:

Feature	Description
Comprehensive analysis	Analyzes text inputted by the user, examining pronunciation, syllables, phrases, context, parts of speech, etc.
Advanced speech synthesis	Learns and restores audio signals based on artificial neural network technology, and generates voice data using Vocoder. Speech synthesis is performed using long short-term memory (LSTM) and recurrent neural network (RNN) architecture.
Fast response time	Generates audio data as soon as the user inputs text.
Human-like speech	Employs natural, human-like speech using the latest deep learning technology.
Multi-language support	Supports Korean and English languages.
Voice selection	Allows the selection of voice gender according to user preferences.
Speech speed control	Offers five-speed levels to deliver speech output.
Pitch tuning	Customizes voice pitch according to user preferences.
Volume control	Offers ten levels of volume control.
Sound effect	Adds sound effects such as robotization and echo to voices according to preferences.

 

Structure

All functions of TTS service are run on a server. TTS service receives text and JSON data as input, and generates audio data output.

 

 

Examples of Use

TTS service can be used in a variety of situations that require voice-based services. TTS service can also be used to assist the visually impaired, narrate audiobooks, or elevate the functions of any application that plays voice audio.