Logo multitel

Speech processing

Two main groups in speech technology

Multitel has worked successfully for several years on research in voice technology. More specifically, the Centre is involved in Automatic Speech Recognition, Text-to-Speech Synthesis and Speech Coding.

The Speech Processing department includes two units: the TTS unit (Text-To-Speech) for the voice synthesis and the ASR unit (Automatic Speech Recognition) for the voice recognition.

Text-to-Speech Synthesis

The Text-to-Speech Synthesis unit (TTS) works on the voice synthesis of text. The voice synthesis includes 2 phases :

  1. a natural language processing phase (phonetic transcription of text with automatic generation of prosody)
  2. a signal processing phase (result conversion of the previous phase into a sound signal)

Research at Multitel in Text-to-Speech synthesis started a few years ago and has led to the development of eLite. This software is a complete Text-to-Speech synthesis system and provides a useful platform for the development and the assessment of new algorithms used in synthesis.

More about Text-to-Speech Synthesis

Automatic Speech Recognition

The Automatic Speech Recognition unit (ASR) works on voice recognition for which Multitel has allowed rapid development of applications according to various constraints, such as :

  • Words format to be recognised : isolated words, strings of words, continuous speech
  • Dependence degreeof the speaker : mono and multi-speaker
  • Vocabulary size
  • Practical conditions : noisy environments, transmission channels
More about Automatic Speech Recognition

Research Projects

Phone : +32 65 34 27 79
Fax : +32 65 34 27 98