![]() |
Sprex![]() |
||||||
|
| |||||||
|
News: Downloadable PDA Demos
Introduction |
LAT2DICT(1) lat2dict 1.0: Man Page Version 1 LAT2DICT(1)
NAME
lat2dict - lattice-to-dictionary dummy-dictionary genera-
tor
SYNOPSIS
lat2dict -i file.lat [ -o file.dct ] [ -h ]
DESCRIPTION
lat2dict generates a dummy dictionary from an HTK SLF word
lattice. lat2dict extracts each word in the the lattice
file and creates a word entry in a dummy HTK-style dictio-
nary which includes the word as both word name and pronun-
ciation. The dictionary file is useful if you happen to
have HTK's HSGen program laying around, for the purpose of
generating random sentences from the lattice (HSGen hap-
pens to require a dictionary for this purpose, but the
dummy dictionary generated by lat2dict is adequate).
You might want to use lat2dict too in a shell script of
some kind, to compare your current pronouncing dictionary
to the word list from the lattice, to make sure there are
no missing words in the pronouncing dictionary.
OPTIONS
-i file.lat The named file is an HTK SLF lattice
file. Each line represents a node or a link, those with
W=WORDNAME have words associated with them; it is those
WORDNAME's that are extracted to generate the dummy dic-
tionary.
-l file.dct Into this output file will be written the
dummy dictionary.
-h Prints usage information.
EXAMPLES
To get a list of options and their explanation, use no
options:
example% lat2dict
Generate a dictionary for use with Entropic's HSGen pro-
gram, then generate 10 sample sentences from the lattice.
example% lat2dict -l file.lat -d file.dct
example% HSGen -n 10 file.lat file.dct
If the lattice was generated by gxc, then there will be a
lot of tags in the generated sentence. To clean that up
and just look at the actual words, follow the rule that
all actual words are in lower case only (including names
and abbreviations), so that then you can use a sed script
to delete all the capital letters, numbers, underscores,
and other funny characters in order to see just the actual
words.
BUGS
None known.
SEE ALSO
gxc, HSGen
AUTHOR
Tom Veatch, Sprex, Inc. All rights reserved. lat2dict is
licensed software, to be used only under a license agree-
ment with Sprex.
Product of Sprex, Inc. 10 February 2003 LAT2DICT(1)
|
||||||
Copyright © 1996-2005
Sprex, Inc.
All rights reserved. Sprex, Speech in the Network, TallyGram and ANSR are trademarks of Sprex, Inc.
|