Speech Synthesis

Text-To-Speech

  • Text-to-Speech (TTS) system generates speech waveform from the given text
  • The ultimate goal of this system is to synthesize a human-like natural voice
  • It can generate various speech signal by controlling the condition such as speaking speed, emotion
  • This system can be used in many applications such as announcement, Human Machine Interface (HMI)

 

 

Synthesized sample

Speed control (Male)

Speed control (Female)

Multi-Modal Speech Synthesis

  • Brain-to-speech and articulatory-to-speech generate audible speech from human signals (ECoG, EMA)
  • Encoding a speech-related representation can bridge the gap between speech and brain or articulatory signals
  • It can generate various speech signals by controlling the condition such as speaking speed, emotion
  • This system can transform the lives of individuals with speech disorder