Files in festival/lib

This is a short description of the scheme files in the directory Festival_1.4/festival/lib. All files are from the English version of Festival as provided by CSTR, University of Edinburgh. A few files have been modified to satisfy our needs (e.g., sable-mode.scm, where German tags have been added).

Sable.v0_2.dtd:

Contains the definition (DTD) of Sable-Tags.

VCLocalRules:

Deals with SIOD (Scheme In One Defun), a Scheme interpreter.

cart_aux.scm:

This file provides some functions for manipulating decision trees.

cslush.scm:

Replaces the C++ - functions by Jacques H. de Villiers, such as ``cslush.getwave'', ``cslush.getphone'', ``cslush''.

cstr.scm:

Provides some SIOD-extensions. CSTR stands for ``Centre for Speech Technology Research'' (University of Edinburgh).

darpa_phones.scm:

This file contains a darpa-phones definition.

display.scm:

``display.scm'' provides a xwaves-display for utterances.

duration.scm:

This file contains modules which should -- based on information in the phrase- and word relation -- predict intonation events and labels.

email-mode.scm:

Provides a TTS text mode for reading email messages, including the interesting headers. It includes a support for dealing with quoted text.

engmorph.scm:

Contains rules for English morphographemics which basically split words into their (potential) morphemes. This is not a Scheme file and can't be loaded and evaluated. It is designed for the use with the wfst tools.

engmorphsyn.scm:

``engmorphsyn.scm'' contains a English morpho-syntax grammar which tries to split unknown words.

f2bdurtreeZ.scm:

This file is the first attempt at a tree to learn durations of segments, trained from F2B.

f2bf0lr.scm:

The first attempt to predict F0 values.

festdoc.scm:

Documentation functions that find all functions and variables with documentation, sorts and dumps the information in ``doc/festfunc.texi'' and ``doc/festvars.texi''.

festival.el:

Provides an emacs-mode for Festival. You now can access various Festival-commands (like Say-Region) from a special menu called ``Say''.

festival.scm:

This file contains the general Scheme specific functions in Festival and the definitions of various standard variables.

festtest.scm:

Contains some basic functions used in tests for Festival.

fringe.scm:

Communication with the fringe program.

gsw_us_diphone.scm:

Contains a voice definition for GSW voice

gswdurtreeZ.scm:

This file contains a tree to predict zscore durations build from gsw 450.

holmes_phones.scm:

Contains a definition of the Holmes phone set used by the Donovan LPC diphone synthesizer.

init.scm:

The initialization file. It is loaded before anything else.

intonation.scm:

Contains modules which should predict intonation events and intonation labels based on information in the phrase and word streams.

java.scm:

Provides functions specific to supporting a java client.

klatt_durs.scm:

This file contains the phone duration for Klatt rules, for mrpa phone set.

languages.scm:

Contains a specification of voices (of different languages) and some major choices of synthesis.

lexicons.scm:

Defines various lexicons.

lts.scm:

Contains functions specific to supporting trained LTS rules.

lts_build.scm:

This file contains functions for building LTS rules sets from lexicons.

mbrola.scm:

Support for MBROLA as an external module.

mettree.scm:

Contains some (experimental) data for investigating metrical trees.

module_description.scm:

This file handles module descriptions.

mrpa_allophones.scm:

Contains a definition of the extended mrpa phone set used for some diphone sets.

mrpa_durs.scm:

Sets mrpa average phoneme durations from gsw 450.

mrpa_phones.scm:

Defines the mrpa phone set.

ogimarkup-mode.scm:

Contains an example tts text mode for reading OGI's CSLU toolkit mark up.

pauses.scm:

Functions that predict pause insertions.

phoneset.scm:

Contains functions to define phone sets.

phrase.scm:

Functions that predict phrase boundaries.

pos.scm:

Contains a POS-tagger

postlex.scm:

This file contains postlexical rules.

radio_phones.scm:

Contains a definition of the radio phone set used in the BU RADIO FM corpus (also: darpa set).

sable-latin.ent:

This file deals with special characters.

sable-mode.scm:

Festival support for SABLE.

scfg.scm:

Contains some functions for manipulating a SCFG parse tree.

scfg_wsj_wp20.gram:

Contains a stochastic context free grammar for the wp20 tag set with 19 nonterminals.

siod.scm:

Contains some basic Scheme functions.

soleml-mode.scm:

Provides a support for an SGML-based mark-up language used in the SOLE project.

speech.properties:

In this file the speech engine is registered.

synthesis.scm:

The utterance types are defined. Here is a list of modules and the relationnames contained in the relationtree after execution of the concerned module:

• (Initialize utt) (nil)
• (Text utt) (Token)
• (Token_POS utt) (Token)
• (Token utt) (Token Word)
• (POS utt) (Token Word)
• (Phrasify utt) (Token Word Phrase)
• (Word utt) (Token Word Phrase Syllable Segment SylStructure)
• (Pauses utt) (Token Word Phrase Syllable Segment SylStructure)
• (Intonation utt) (Token Word Phrase Syllable Segment SylStructure IntEvent Intonation)
• (PostLex utt) (Token Word Phrase Syllable Segment SylStructure IntEvent Intonation)
• (Duration utt) (Token Word Phrase Syllable Segment SylStructure IntEvent Intonation)
• (Int_Targets utt) (Token Word Phrase Syllable Segment SylStructure IntEvent Intonation Target)
• (Wave_Synth utt) (Token Word Phrase Syllable Segment SylStructure IntEvent Intonation Target Wave)

tilt.scm:

This file contains the tilt intonation module to predict accents and boundaries and construct the F0 track.

tilt_ab.scm:

Contains CART trees to predict tilt accents and boundaries, trained from f2b.

tilt_params.scm:

Contains CART trees to predict tilt parameters.

tobi.scm:

Contains CART trees to predict tobi accents (trained from f2b), punctuation and minimal pos.

tobi_rules.scm:

This file serves to generate F0 points from tobi labels (rules from Matthias Jilka, IMS [3])

token.scm:

Contains various tokenizing functions and customization.

tokenpos.scm:

Contains functions used in identifying token types.

tts.scm:

Contains TTS functions.

unitdatabase.scm:

This file contains support routines for dealing with unit databases.

voices.scm:

Prepares to access voices.

web.scm:

Provides functions dealing with the web.