Skip to content
rajbot edited this page Feb 24, 2011 · 8 revisions

The Internet Archive BookReader now has a real-time Text-To-Speech (TTS) feature to better support our print-disabled users.

Tools

These tools are used to implement the TTS feature:

  • Festival and its text2wave utility for speech synthesis
  • The cmu_us_slt_arctic_hts Festival voice
  • oggenc, which is part of vorbis-tools, for converting wav files to ogg vorbis format for Firefox
  • ffmpeg to convert wav files to mp3 format for webkit and flash-enabled browsers
  • SoundManager2, which provides a consistent audio API across HTML5 and and flash-enabled browsers

Code

  • BookReader.js contains the BookReader TTS code. ** In particular, see the ttsStart(), ttsPlay(), and ttsNextChunk() functions.
  • BookReaderGetText.py returns a chunk of OCRed text to feed to the TTS engine.
  • BookReaderGetTTS.php takes a text string and produces mp3 or ogg audio data.
Clone this wiki locally