The Festival Speech Synthesis System

System documentation

Edition 1.4, for Festival Version 1.4.3

27th December 2002

Postscript version of manual festival-1.4.3.ps.gz

• 1 Abstract
• 2 Copying
• 3 Acknowledgements
  • 3.1 SIOD
  • 3.2 editline
  • 3.3 Edinburgh Speech Tools Library
  • 3.4 Others
• 4 What is new
• 5 Overview
  • 5.1 Philosophy
  • 5.2 Future
• 6 Installation
  • 6.1 Requirements
  • 6.2 Configuration
  • 6.3 Site initialization
  • 6.4 Checking an installation
• 7 Quick start
  • 7.1 Basic command line options
  • 7.2 Sample command driven session
  • 7.3 Getting some help
• 8 Scheme
  • 8.1 Scheme references
  • 8.2 Scheme fundamentals
  • 8.3 Scheme Festival specifics
  • 8.4 Scheme I/O
• 9 TTS
  • 9.1 Utterance chunking
  • 9.2 Text modes
  • 9.3 Example text mode
• 10 XML/SGML mark-up
  • 10.1 Sable example
  • 10.2 Supported Sable tags
  • 10.3 Adding Sable tags
  • 10.4 XML/SGML requirements
  • 10.5 Using Sable
• 11 Emacs interface
• 12 Phonesets
• 13 Lexicons
  • 13.1 Lexical entries
  • 13.2 Defining lexicons
  • 13.3 Lookup process
  • 13.4 Letter to sound rules
  • 13.5 Building letter to sound rules
  • 13.6 Lexicon requirements
  • 13.7 Available lexicons
  • 13.8 Post-lexical rules
• 14 Utterances
  • 14.1 Utterance structure
  • 14.2 Utterance types
  • 14.3 Example utterance types
  • 14.4 Utterance modules
  • 14.5 Accessing an utterance
  • 14.6 Features
  • 14.7 Utterance I/O
• 15 Text analysis
  • 15.1 Tokenizing
  • 15.2 Token to word rules
  • 15.3 Homograph disambiguation
    • 15.3.1 Using disambiguators
    • 15.3.2 Building disambiguators
• 16 POS tagging
• 17 Phrase breaks
• 18 Intonation
  • 18.1 Default intonation
  • 18.2 Simple intonation
  • 18.3 Tree intonation
  • 18.4 Tilt intonation
  • 18.5 General intonation
  • 18.6 Using ToBI
• 19 Duration
  • 19.1 Default durations
  • 19.2 Average durations
  • 19.3 Klatt durations
  • 19.4 CART durations
• 20 UniSyn synthesizer
  • 20.1 UniSyn database format
    • 20.1.1 Generating pitchmarks
    • 20.1.2 Generating LPC coefficients
  • 20.2 Generating a diphone index
  • 20.3 Database declaration
  • 20.4 Making groupfiles
  • 20.5 UniSyn module selection
  • 20.6 Diphone selection
• 21 Diphone synthesizer
  • 21.1 Diphone database format
  • 21.2 LPC databases
  • 21.3 Group files
  • 21.4 Diphone_Init
  • 21.5 Access strategies
  • 21.6 Diphone selection
• 22 Other synthesis methods
  • 22.1 LPC diphone synthesizer
  • 22.2 MBROLA
  • 22.3 Synthesizers in development
• 23 Audio output
• 24 Voices
  • 24.1 Current voices
  • 24.2 Building a new voice
    • 24.2.1 Phoneset
    • 24.2.2 Lexicon and LTS
    • 24.2.3 Phrasing
    • 24.2.4 Intonation
    • 24.2.5 Duration
    • 24.2.6 Waveform synthesis
    • 24.2.7 Voice selection function
    • 24.2.8 Last remarks
    • 24.2.9 Resetting globals
  • 24.3 Defining a new voice
• 25 Tools
  • 25.1 Regular expressions
  • 25.2 CART trees
  • 25.3 Ngrams
  • 25.4 Viterbi decoder
  • 25.5 Linear regression
• 26 Building models from databases
  • 26.1 Labelling databases
  • 26.2 Extracting features
  • 26.3 Building models
• 27 Programming
  • 27.1 The source code
  • 27.2 Writing a new module
    • 27.2.1 Example 1: adding new modules
    • 27.2.2 Example 2: accessing the utterance
    • 27.2.3 Example 3: adding new directories
    • 27.2.4 Example 4: adding new LISP objects
• 28 API
  • 28.1 Scheme API
  • 28.2 Shell API
  • 28.3 Server/client API
    • 28.3.1 Server access control
    • 28.3.2 Client control
    • 28.3.3 Server/client protocol
  • 28.4 C/C++ API
  • 28.5 C only API
  • 28.6 Java and JSAPI
• 29 Examples
  • 29.1 POS Example
  • 29.2 Singing Synthesis
• 30 Problems
• 31 References
• 32 Feature functions
• 33 Variable list
• 34 Function list
• Index

This document was generated on 24 January 2003 using the texi2html translator version 1.52.