Text to Speech

High quality conversion of text into clear multi language speech signal without a large word content database which is especially suited for devices with a lower battery life such as toys, phones, etc

• Best in class among embedded systems
• Natural speech quality
• Low footprint (e.g. ARM7)
• Support for 11 languages
• American English, British English, German, Spanish, Italian, French, Dutch, Turkish, Mandarin Chinese, Japanese, Brazilian / Portuguese

Key Features of text-to-speech

• Voice personalization
  • speed, pitch, spectral shape, formant sharpening, ...
• Voice conversion
  • male, female, child, dwarf, ...
• Emotion control
  • speak friendly, angry, weepy, exited, whisper, …
• Cross language speaker support
  • E.g. Italian person speaking German
• External interfaces
  • “viseme” information for “talking face”
  • event outputs: marker, token, sentence
  • speech Synthesis Markup Language  (SSML) support

Status: Ready to be implemented