libTLK  1.3.1
Functions
Initialization of an acoustic model using linear segmentation
Algorithms

Functions

 tL_initializer_free (tLInitializer *ini)
 Frees memory.
 tL_initializer_new (const tLAModel *amodel, const tLSeqs *seqs, char **err)
 Gets a tLInitializer.
 tL_init (tLInitializer *ini, tLAMiCounts *counts, const tLFea *fea, const size_t seq)
 Initializes an acoustic model from a sample and its transcription.

Function Documentation

tL_init ( tLInitializer *  ini,
tLAMiCounts counts,
const tLFea fea,
const size_t  seq 
)

Initializes an acoustic model from a sample and its transcription.

This function linearly segments the input frames, according to the number of symbols of the transcription. According to this segmentation, acoustic model counts are increased using tL_init_hmm. The transcription is specified as an integer code according to the sequence manager used during the creation of the auxiliary data.

Parameters:
iniInitializer.
countsInteger counts related to amodel.
feaA feature vector manager, which must contain a sequence of feature vectors of an adequate type for the amodel.
seqThe code of the symbol sequence related to fea.
Returns:
0 if the sample has been successfully used to increase the acoustic model counts or the minimum number of required frames.
tL_initializer_free ( tLInitializer *  ini)

Frees memory.

Frees the memory allocated for the initializer.

Parameters:
iniThe initializer.
tL_initializer_new ( const tLAModel amodel,
const tLSeqs seqs,
char **  err 
)

Gets a tLInitializer.

This function creates a new initializer used to initialize an acoustic model from samples and their transcriptions using linear segmentation. Currently supported types are: see tL_hmm_initializer_new.

Parameters:
amodelThe acoustic model.
seqsA symbol sequence manager containing the transcription of the samples.
errPointer to string variable. If not NULL, an error message is allocated in the variable in case of error.
Returns:
Initializer, or NULL in case of error.
 All Data Structures Variables