NAME

genmodel.tcl - Generate mono-phone or tri-phone transcriptions.

AVAILABILITY

script/hmm_1.0

SYNOPSIS

 genmodel.tcl input output [options]

PARAMETERS

input

Input word transcription file (see Master Label File).

output

Output model (mono-phone or tri-phone) transcription file.

OPTIONS

-dictionary [Default = dict.db]

Input database file used to index the pronunciation dictionary (see worddb.tcl)

-cutoff float [Default = 100000]

If timing information is specified in the master label file (btime/etime fields) then only generate model transcriptions for segments which are shorter than the specified cutoff value.

-sillen float [Default = 100ms]

Specify the minimum silence model duration. If a silence is shorter than the specified length, then a short-pause model is used instead.

-shortpause [Default = 1]

By default a short-pause (T model) is inserted between words. The -nosp flag will disable the insertion of a short-pause between each word.

-addsil [Default = 0]

Add a silence model to the sentence begin and end.

-type string [Default = mono]

Generate mono-phone or tri-phone (mono/tri) model transcriptions.

-config string

Read command line options (configuration info) from this file. The configuration file is in essence a Tcl script which sets the required internal variables.

 set config(genmodel,dictionary) dict.db
 set config(genmodel,cutof) 1000000
 set config(genmodel,sillen) 100
 set config(genmodel,shortpause) 1
 set config(genmodel,addsil) 0
 set config(genmodel,type) mono

These values will override the preset default values. Subsequent command line options will override the values specified by the configuration script. Command line parameters are specified using the param variable.

 set param(scribe,input) foo
 set param(scribe,output) foo.mono

Since command line parameters are typically not optional the user needs to specify the command line parameters as a single "-" character for the settings defined in the configuration script to take effect.

-help

Provide immediate command line feedback.

DESCRIPTION

genmodel.tcl converts a word model transcription file to an associated mono-phone or triphone transcription file. Each word is expanded into either a mono-phone or triphone representation using the pronunciations listed in the pronunciation database. The resulting model transcriptions are then used for embedded training.

SEE ALSO

worddb.tcl, hmmscribe.tcl, hmmembed.tcl

AUTHOR

Johan Schalkwyk
Center for Spoken Language Understanding
Oregon Graduate Institute of Science & Technology