CSLU Toolkit Formats


Parts Format
Version: 1.0
Created:
25 May 2003
Modified:
25 May 2003

Overview
The Parts format specifies how phonemes should be expanded into multi-state biphone categories.


Synopsis
FILE-BASED GRAMMAR
Tcl-LIST BASED GRAMMAR
clusterName := token1  token2  ...  tokenN ;
clusterName := token1  token2  ...  tokenN
metaClusterName == token1  token2  ...  tokenN ;
metaClusterName == token1  token2  ...  tokenN 
token -> state ;
token -> state
token -> state1 ... stateN :: leftMetaCluster __ rightMetaCluster ;
token -> state1 ... stateN :: leftMetaCluster __ rightMetaCluster 
--- blank lines ARE allowed --- --- blank lines are NOT allowed ---

Variables
Variable
Description
token
An item name, where an item typically represents a phoneme.
clusterName
A cluster name identifes a group of tokens.  This group of tokens can then be represented by clusterName in a context description.
metaClusterName A cluster name identifes a group of tokens and/or clusters.  This meta-cluster name is then used to define all the ways of expanding a token in a left or right context.
state
A state name associated with a token.  A state name may include a clusterName as a left or right context, as specified in the state specification format.

Description
The Parts format specifies how phonemes should be expanded into multi-state biphone categories.  The parts format is read by the statenet writeCDRules command.

Example

$sil   := .pau .garbage tc kc /BOU /EOU;
$den_l := s z th;
$den_r := s z th;

$LC  == f v T n w I & ^ 9r E i: u ei aI oU kh $den_l $sil ;
$RC  == f v T n w I & ^ 9r E i: u ei aI oU kh $den_r $sil ;

.pau      -> <.pau> ;
tc        -> <tc> ;
kc        -> <kc> ;
.garbage  -> <.garbage> ;

f        -> $LC<f  f>$RC                :: $LC __ $RC ;
v        -> $LC<v  v>$RC                :: $LC __ $RC ;
T        -> $LC<T  T>$RC                :: $LC __ $RC ;
s        -> $LC<s  s>$RC                :: $LC __ $RC ;
z        -> $LC<z  z>$RC                :: $LC __ $RC ;
n        -> $LC<n  n>$RC                :: $LC __ $RC ;
w        -> $LC<w  w>$RC                :: $LC __ $RC ;
I        -> $LC<I  I>$RC                :: $LC __ $RC ;
&        -> $LC<&  &>$RC                :: $LC __ $RC ;
^        -> $LC<^  ^>$RC                :: $LC __ $RC ;
9r       -> $LC<9r 9r>$RC               :: $LC __ $RC ;
E        -> $LC<E  E>$RC                :: $LC __ $RC ;
i:       -> $LC<i: <i:> i:>$RC          :: $LC __ $RC ;
u        -> $LC<u <u> u>$RC             :: $LC __ $RC ;
ei       -> $LC<ei <ei> ei>$RC          :: $LC __ $RC ;
aI       -> $LC<aI <aI> aI>$RC          :: $LC __ $RC ;
oU       -> $LC<oU <oU> oU>$RC          :: $LC __ $RC ;
th       -> th>$RC                      :: __ $RC ;
kh       -> kh>$RC                      :: __ $RC ;


See Also
The Statenet package

Author
John-Paul Hosom, hosom@{cslu, bme, cse}.ogi.edu