HMM/DNN-based Speech Synthesis System (HTS) -
Reference of hts_engine API
[
Front page
] [
New
|
Page list
|
Search
|
Recent changes
|
Help
|
Log in
]
Start:
//[[HTS_engine_API_english]]
#contents
* Preprocessor symbols for conditional compilation　...
- For embedded device
HTS_EMBEDDED (slightly faster setting is used)
- Audio device setting
AUDIO_PLAY_WIN32 (for Windows 2000/XP/Vista C++ compiler)
AUDIO_PLAY_WINCE (for Windows Mobile C++ compiler)
AUDIO_PLAY_NONE (default)
- For [[Festival speech synthesis system:http://festvox.o...
FESTIVAL
- Endian definition
WORDS_BIGENDIAN (e.g. PowerPC, Cell BE, SPARC, 680x0)
WORDS_LITTLEENDIAN (e.g. x86, Alpha AXP)
- Boolean definition (HTS_Boolean)
TRUE
FALSE
* Engine structures [#z44fc354]
** Model [#le7d813d]
*** HTS_Window [#zb4a3381]
- Window coefficients to calculate dynamic features.
int size - # of windows (static + deltas)
int *l_width - left width of windows
int *r_width - right width of windows
double **coefficient - window coefficients
int max_width - maximum width of windows
*** HTS_Pattern [#k48e4d9f]
- List of patterns in a question and a tree.
char *string - pattern string
HTS_Pattern *next - pointer to the next pattern
*** HTS_Question [#z77715ac]
- List of questions in a tree.
char *name - name of this question
HTS_Pattern *head - pointer to the head of pattern list
HTS_Question *next - pointer to the next question
*** HTS_Node [#c82cbc05]
- List of tree nodes in a tree.
int index - index of this node
int pdf - index of PDF for this node (leaf n...
HTS_Node *yes - pointer to its child node (yes)
HTS_Node *no - pointer to its child node (no)
HTS_Node *next - pointer to the next node
HTS_Question *quest - question applied at this node
*** HTS_Tree [#tccac793]
- List of decision trees in a model.
HTS_Pattern *head - pointer to the head of pattern list ...
HTS_Tree *next - pointer to the next tree
HTS_Node *root - root node of this tree
int state - state index of this tree
*** HTS_Model [#lb0795d7]
- Set of PDFs, decision trees and questions.
int vector_length - vector length (include static a...
int *npdf - # of PDFs at each tree
double ***pdf - PDFs
HTS_Tree *tree - pointer to the list of trees
HTS_Question *question - pointer to the list of questions
*** HTS_Stream [#e8ebe6f1]
- Set of models and a window.
int vector_length - vector_length (include static...
HTS_Model *model - models
HTS_DynamicWindow window - window coefficients
HTS_Boolean msd_flag - flag for MSD
int interpolation_size - # of models for interpolation
*** HTS_ModelSet [#n0021e72]
- Set of duration models, HMMs and GV models.
HTS_Stream duration - duration PDFs and trees
HTS_Stream *stream - parameter PDFs, trees and windows
HTS_Stream *gv - GV PDFs
int nstate - # of HMM states
int nstream - # of stream
** Label [#zdbf97c6]
*** HTS_LabelString [#k2bc37e0]
- Individual label string with time infomation.
HTS_LabelString *next - pointer to the next label string
char *name - label string
HTS_Boolean frame_flag - flag for frame length modificat...
int frame - frame length specified in the g...
*** HTS_Label [#cf3de27c]
- List of label strings.
HTS_LabelString *head - pointer to the head of label str...
int size - # of label strings
double speech_speed - speech speed rate
** State stream [#p6180dbc]
*** HTS_SStream [#g80be9f5]
- Individual state stream.
int vector_length - vector length (include static...
double **mean - mean vector sequence
double **vari - variance vector sequence
double *msd - MSD parameter sequence
int win_size - # of windows (static + deltas)
int *win_l_width - left width of windows
int *win_r_width - right width of windows
double **win_coefficient - window coefficients
int win_max_width - maximum width of windows
double *gv_mean - mean vector of GV
double *gv_vari - variance vector of GV
*** HTS_SStreamSet [#c7e181ed]
- Set of state stream.
HTS_SStream *sstream - state streams
int nstream - # of streams
int nstate - # of states
int *duration - duration sequence
int total_state - total state
int total_frame - total frame
** PDF stream [#fbcf0b4b]
*** HTS_SMatrices [#n16e4df2]
- Matrices/Vectors used in the speech parameter generatio...
double **mean - mean vector sequence
double **ivar - inverse diag variance sequence
double *g - vector used in the forward substitution
double **wuw - W' U^-1 W
double *wum - W' U^-1 mu
*** HTS_PStream [#v2504efe]
- Individual PDF stream.
int vector_length - vector length (include static...
int static_length - static features length
int length - stream length
int width - width of dynamic window
double **par - output parameter vector
HTS_SMatrices sm - matrices for parameter genera...
int win_size - # of windows (static + deltas)
int *win_l_width - left width of windows
int *win_r_width - right width of windows
double **win_coefficient - window coefficients
HTS_Boolean *msd_flag - Boolean sequence for MSD
double *gv_buff - buffer for GV calculation
double *gv_mean - mean vector of GV
double *gv_vari - variance vector of GV
double gv_weight - GV weight
*** HTS_PStreamSet [#fc667ac9]
- Set of PDF streams.
HTS_PStream *pstream - PDF streams
int nstream - # of PDF streams
int total_frame - total frame
** Generated parameter stream [#cb40623b]
*** HTS_GStream [#r1f0e0dc]
- Generated parameter stream.
int static_length - static features length
double **par - generated parameter
*** HTS_GStreamSet [#qab80b0e]
- Set of generated parameter stream.
int total_nsample - total sample
int total_frame - total frame
int nstream - # of streams
HTS_GStream *gstream - generated parameter streams
short *gspeech - generated speech
** Engine [#j629fb64]
*** HTS_Global [#x0a5815e]
- Global settings.
int stage - Gamma=-1/stage : if stage=0 t...
HTS_Boolean use_log_gain - log gain flag (for LSP)
int sampling_rate - sampling rate
int fperiod - frame period
double alpha - all-pass constant
double beta - postfiltering coefficient
int audio_buff_size - audio buffer size (for audio ...
double *msd_threshold - MSD thresholds
double *duration_iw - weights for duration interpol...
double **parameter_iw - weights for parameter interpo...
double **gv_iw - weights for GV interpolation
double *gv_weight - GV weights
*** HTS_Engine [#wa56bd53]
- Engine itself.
HTS_Global global - global settings
HTS_ModelSet ms - set of duration models, HMMs and GV...
HTS_Label label - label
HTS_SStreamSet sss - set of state streams
HTS_PStreamSet pss - set of PDF streams
HTS_GStreamSet gss - set of generated parameter streams
* Vocoder structures [#gc6a2eb8]
** Audio [#i33a0fbe]
*** HTS_Audio [#se4ce25f]
- For MS Windows (Windows Mobile) audio output device.
HWAVEOUT hwaveout - audio device handle
WAVEFORMATEX waveformatex - wave formatex
short *buff - current buffer
int buff_size - current buffer size
int which_buff - double buffering flag
HTS_Boolean now_buff_1 - double buffering flag
HTS_Boolean now_buff_2 - double buffering flag
WAVEHDR buff_1 - buffer
WAVEHDR buff_2 - buffer
int max_buff_size - buffer size of audio output ...
- For Linux, etc.
int i - make compiler happy
** Vocoder [#lbf2a363]
*** HTS_Vocoder [#p4d69001]
- MLSA/MGLSA filter settings.
int stage - Gamma=-1/stage : if stage=0...
double gamma - Gamma
HTS_Boolean use_log_gain - log gain flag (for LSP)
int fprd - frame shift
int iprd - interpolation period
int seed - seed of random generator
unsigned long next - temporary variable for rand...
HTS_Boolean gauss - flag to use Gaussian noise
double rate - sampling rate
double p1 - used in excitation generation
double pc - used in excitation generation
double p - used in excitation generation
double inc - used in excitation generation
int sw - switch used in random gener...
int x - excitation signal
HTS_Audio *audio - pointer for audio device
double *freqt_buff - used in freqt
int freqt_size - buffer size for freqt
double *spectrum2en_buff - used in spectrum2en
int spectrum2en_size - buffer size for spectrum2en
double r1, r2, s - used in random generator
double *postfilter_buff - used in postfiltering
int postfilter_size - buffer size for postfiltering
double *c, *cc, *cinc, *d1 - used in the MLSA/MGLSA filter
double *pade - used in mlsadf
double *lsp2lpc_buff - used in lsp2lpc
int lsp2lpc_size - buffer size of lsp2lpc
double *gc2gc_buff - used in gc2gc
int gc2gc_size - buffer size for gc2gc
* Engine functions [#l74e4110]
** Initialize engine [#t310270e]
*** HTS_Engine_initialize [#ec43b3cc]
void HTS_Engine_initialize(HTS_Engine *engine, int nstre...
- Use: Initialize engine.
- Arguments:
HTS_Engine *engine - pointer to HTS_Engine structure
- ''Attention!!:'' To start engine, first you must call t...
** Load models [#z7b8be60]
*** HTS_Engine_load_duration_from_fn [#va73cd1a]
void HTS_Engine_load_duration_from_fn(HTS_Engine *engine...
- Use: Load duration PDFs and trees from files using give...
- Arguments:
HTS_Engine *engine - pointer to HTS_Engine structure
char **pdf_fn - duration PDF file names
char **tree_fn - duration tree file names
int interpolation_size - # of duration models to be inte...
- ''Attention!!'': You must initialize variable '''engine...
*** HTS_Engine_load_duration_from_fp [#wc2481e3]
void HTS_Engine_load_duration_from_fp(HTS_Engine *engine...
- Use: load duration PDFs and trees from files using give...
- Arguments:
HTS_Engine *engine - pointer to HTS_Engine structure
FILE **pdf_fp - duration PDF file pointers
FILE **tree_fp - duration tree file pointers
int interpolation_size - # of duration models to be inte...
- ''Attention!!'': You must initialize variable '''engine...
*** HTS_Engine_load_parameter_from_fn [#kd48cfd2]
void HTS_Engine_load_parameter_from_fn(HTS_Engine *engin...
int stream_index,...
- Use: load parameter PDFs, trees and windows from files ...
- Arguments:
HTS_Engine *engine - pointer to HTS_Engine structure
char **pdf_fn - parameter PDF file pointers
char **tree_fn - parameter tree file pointers
char **win_fn - parameter window file pointers
int stream_index - index of stream
HTS_Boolean msd_flag - flag for MSD
int window_size - # of windows
int interpolation_size - # of parameter models to be int...
- ''Attention!!'': You must load duration models using HT...
*** HTS_Engine_load_parameter_from_fp [#j2b19c02]
void HTS_Engine_load_parameter_from_fp(HTS_Engine *engin...
int stream_index,...
- Use: load parameter PDFs, trees and windows from files ...
- Arguments:
HTS_Engine *engine - pointer to HTS_Engine structure
FILE **pdf_fp - parameter PDF file pointers
FILE **tree_fp - parameter tree file pointers
FILE **win_fp - parameter window file pointers
int stream_index - index of stream
HTS_Boolean msd_flag - MSD flag
int window_size - # of windows
int interpolation_size - # of parameter models to be int...
- ''Attention!!'': You must load duration models using HT...
*** HTS_Engine_load_gv_from_fn [#id008a1a]
void HTS_Engine_load_gv_from_fn(HTS_Engine *engine, char...
- Arguments:
HTS_Engine *engine - pointer to HTS_Engine structure
char **pdf_fn - GV PDF file names
int stream_index - index of streams
int interpolation_size - # of GV models to be interpolated
- ''Attention!!'': You must load parameter models using H...
*** HTS_Engine_load_gv_from_fp [#g6d9d662]
void HTS_Engine_load_gv_from_fp(HTS_Engine *engine, FILE...
- Arguments:
HTS_Engine *engine - pointer to HTS_Engine structure
FILE **pdf_fn - GV PDF file pointers
int stream_index - index of streams
int interpolation_size - # of GV models to be interpolated
- ''Attention!!'': You must load parameter models using H...
** Set parameter [#b8dd5d39]
*** HTS_Engine_set_sampling_rate [#q41e65a1]
void HTS_Engine_set_sampling_rate(HTS_Engine *engine,int...
- Use: set sampling frequency.
- Arguments:
HTS_Engine *engine - pointer to HTS_Engine structure
int i - sampling frequency (Hz), 1= < i <= ...
- ''Attention!!:'' Default value is 16000.
*** HTS_Engine_set_fperiod [#c5c89881]
void HTS_Engine_set_fperiod(HTS_Engine *engine,int i)
- Use: set frame shift.
- Arguments:
HTS_Engine *engine - pointer to HTS_Engine structure
int i - frame shift (point), 1= < i <= 2000
- ''Attention!!:'' Default value is 80.
*** HTS_Engine_set_alpha [#tbe7cee5]
void HTS_Engine_set_alpha(HTS_Engine *engine,double f)
- Use: set frequency warping parameter alpha.
- Arguments:
HTS_Engine *engine - pointer to HTS_Engine structure
double f - alpha, 0.0 <= f <= 1.0
- ''Attention!!:'' Default value is 0.42.
*** HTS_Engine_set_gamma [#wdc30080]
void HTS_engine_set_gamma(HTS_Engine *engine,int i)
- Use: set Gamma.
- Arguments:
HTS_Engine *engine - pointer to HTS_Engine structure
int i - Gamma=-1/i : if i=0 then Gamma=0, 0...
- ''Attention!!:'' Default value is 0.
*** HTS_Engine_use_log_gain [#lfbc8c63]
void HTS_Engine_set_log_gain(HTS_Engine *engine,HTS_Bool...
- Use: set log gain flag.
- Arguments:
HTS_Engine *engine - pointer to HTS_Engine structure
HTS_Boolean i - log gain flag.
- ''Attention!!:'' Default value is FALSE.
*** HTS_Engine_set_beta [#m44c517c]
void HTS_SetBeta(HTS_Engine *engine,double f)
- Use: set postfiltering coefficient parameter beta.
- Arguments:
HTS_Engine *engine - pointer to HTS_Engine structure
double f - beta, -0.8 <= f <= 0.8
- ''Attention!!:'' Default value is 0.0. If you set beta...
*** HTS_Engine_set_audio_buff_size [#ua56725a]
void HTS_Engine_set_audio_buff_size(HTS_Engine *engine, ...
- Use: set buffer size for direct audio output.
- Arguments:
HTS_Engine *engine - pointer to HTS_Engine structure
int i - buffer size (sample), 0 <= i <= 48000
- ''Attention!!:'' Default value is 0. If i=0, direct aud...
*** HTS_Engine_set_msd_threshold [#qd019ba4]
void HTS_Engine_set_msd_threshold(HTS_Engine *engine, in...
- Use: set MSD threshold.
- Arguments:
HTS_Engine *engine - pointer to HTS_Engine structure
int stream_index - index of streams
double f - threshold
*** HTS_Engine_set_duration_interpolation_weight [#rdd5ff...
void HTS_Engine_set_duration_interpolation_weight(HTS_En...
- Use: set weight for duration interpolation.
- Arguments:
HTS_Engine *engine - pointer to HTS_Engine structure
int interpolation_index - index of duration models
double f - interpolation weight
*** HTS_Engine_set_parameter_interpolation_weight [#k8bc0...
void HTS_Engine_set_parameter_interpolation_weight(HTS_E...
- Use: set weight for parameter interpolation.
- Arguments:
HTS_Engine *engine - pointer to HTS_Engine structure
int stream_index - index of streams
int interpolation_index - index of parameter models
double f - interpolation weight
*** HTS_Engine_set_gv_interpolation_weight [#lc3e00bc]
void HTS_Engine_set_gv_interpolation_weight(HTS_Engine *...
- Use: set weight for GV interpolation.
- Arguments:
HTS_Engine *engine - pointer to HTS_Engine structure
int stream_index - index of streams
int interpolation_index - index of GV models
double f - interpolation weight
*** HTS_Engine_set_gv_weight [#eaa46840]
void HTS_Engine_set_gv_weight(HTS_Engine *engine, int st...
- Use: set GV weight.
- Arguments:
HTS_Engine *engine - pointer to HTS_Engine structure
int stream_index - index of streams
double f - GV weight
** Synthesize speech [#sf7321ed]
*** HTS_Engine_load_label_from_fn [#j87fa620]
void HTS_Engine_load_label_from_fn(HTS_Engine *engine, c...
- Use: load label from file using given file name.
- Arguments:
HTS_Engine *engine - pointer to HTS_Engine structure
char *fn - label file name
*** HTS_Engine_load_label_from_fp [#i892f523]
void HTS_Engine_load_label_from_fp(HTS_Engine *engine, F...
- Use: load label from file using given file pointer.
- Arguments:
HTS_Engine *engine - pointer to HTS_Engine structure
FILE *fp - label file pointer
*** HTS_Engine_load_label_from_string [#q36e8c4c]
void HTS_Engine_load_label_from_string(HTS_Engine *engin...
- Use: load label from string.
- Arguments:
HTS_Engine *engine - pointer to HTS_Engine structure
char *data - label string
*** HTS_Engine_load_label_from_string_list [#i9e769ac]
void HTS_Engine_load_label_from_string_list(HTS_Engine *...
- Use: load label from string list.
- Arguments:
HTS_Engine *engine - pointer to HTS_Engine structure
char **data - label string list
int size - size of label string list
*** HTS_Engine_create_sstream [#c2e0aaed]
void HTS_Engine_create_sstream(HTS_Engine *engine)
- Use: parse label, determine state duration and create s...
- Arguments:
HTS_Engine *engine - pointer to HTS_Engine structure
*** HTS_Engine_create_pstream [#b1fda60c]
void HTS_Engine_create_pstream(HTS_Engine *engine)
- Use: create PDF stream using state stream.
- Arguments:
HTS_Engine *engine - pointer to HTS_Engine structure
*** HTS_Engine_create_gstream [#fab90e37]
void HTS_Engine_create_gstream(HTS_Engine *engine)
- Use: synthesis speech and store generated parameter usi...
- Arguments:
HTS_Engine *engine - pointer to HTS_Engine structure
- ''Attention!!:'' To synthesize speech, you must set str...
*** HTS_Engine_save_infomation [#e7c9463d]
void HTS_Engine_save_infomation(HTS_Engine *engine, FILE...
- Use: output trace infomation.
- Arguments:
HTS_Engine *engine - pointer to HTS_Engine structure
FILE *fp - output file pointer
*** HTS_Engine_save_label [#sb2d1f59]
void HTS_Engine_save_label(HTS_Engine *engine, FILE *fp)
- Use: output label with time.
- Arguments:
HTS_Engine *engine - pointer to HTS_Engine structure
FILE *fp - output file pointer
*** HTS_Engine_save_generated_parameter [#ya641edd]
void HTS_Engine_save_generated_parameter(HTS_Engine *eng...
- Use: output generated parameter.
- Arguments:
HTS_Engine *engine - pointer to HTS_Engine structure
FILE *fp - output file pointer
*** HTS_Engine_save_generated_speech [#k8327b8d]
void HTS_Engine_save_generated_speech(HTS_Engine *engine...
- Use: output generated speech.
- Arguments:
HTS_Engine *engine - pointer to HTS_Engine structure
FILE *fp - output file pointer
*** HTS_Engine_save_riff [#wb9640f7]
void HTS_Engine_save_riff(HTS_Engine *engine, FILE *fp)
- Use: output riff format file.
- Arguments:
HTS_Engine *engine - pointer to HTS_Engine structure
FILE *fp - output file pointer
*** HTS_Engine_refresh [#a8a1a225]
void HTS_Engine_refresh(HTS_Engine *engine)
- Use: free label, state streams, PDF streams and generat...
- Arguments:
HTS_Engine *engine - pointer to HTS_Engine structure
** Free engine [#mef9b660]
*** HTS_Engine_clear [#gd323729]
void HTS_Engine_clear(HTS_Engine *engine)
- Use: free engine.
- Arguments:
HTS_Engine *engine - pointer to HTS_Engine structure
* Vocoder functions [#tffbab44]
** Initialize vocoder [#v5bc81e4]
*** HTS_Vocoder_initialize [#ff9170fc]
void HTS_Vocoder_initialize(HTS_Vocoder *v, const int m,...
- Use: initialize the vocoder.
- Arguments:
HTS_Vocoder *v - pointer to HTS_Vocoder struct...
int m - order of mel-cepstral coeffic...
int stage - Gamma=-1/stage : if stage=0 t...
HTS_Boolean use_log_gain - log gain flag
int rate - sampling frequency
int fperiod - frame shift
int buff_size - buffer size for direct audio ...
** Synthesize speech [#d6645d7f]
*** HTS_Vocoder_synthesize [#h737ba80]
void HTS_Vocoder_synthesize(HTS_Vocoder *v, const int m,...
- Use: run the vocoder and synthesize waveform.
- Arguments:
HTS_Vocoder *v - HTS_Vocoder structure pointer
int m - order of spectrum coefficients
double lf0 - log F0 value
double *spectrum - spectrum coefficients
double alpha - frequency warping parameter alpha
double beta - postfiltering parameter beta
short *rawdata - short pointer to store synthesized wa...
*** HTS_Vocoder_postfilter_mcp [#r92930b7]
void HTS_Vocoder_postfilter_mcp(HTS_Vocoder *v, double *...
- Use: postfilter for mel-cepstrum.
- Arguments:
HTS_Vocoder *v - HTS_Vocoder structure pointer
double *mcp - spectrum coefficients
int m - order of mel-cepstral coefficients
double alpha - frequency warping parameter alpha
double beta - postfiltering parameter beta
** Free vocoder [#rb2ef784]
*** HTS_Vocoder_clear [#wb420be0]
void HTS_Vocoder_clear(HTS_Vocoder *v)
- Use: free the vocoder.
- Arguments:
HTS_Vocoder *v - pointer to HTS_Vocoder structure
* Other functions [#q32f7c6d]
** For copyright [#e99f219b]
*** void HTS_show_copyright(FILE *fp) [#y30d6252]
- Use: show hts_engine_API copyright.
- Arguments:
FILE *fp - output file pointer
*** void HTS_get_copyright(char *str) [#d24078ea]
- Use: copy hts_engine_API copyright to string.
- Arguments:
char *str - output string
End:
//[[HTS_engine_API_english]]
#contents
* Preprocessor symbols for conditional compilation　...
- For embedded device
HTS_EMBEDDED (slightly faster setting is used)
- Audio device setting
AUDIO_PLAY_WIN32 (for Windows 2000/XP/Vista C++ compiler)
AUDIO_PLAY_WINCE (for Windows Mobile C++ compiler)
AUDIO_PLAY_NONE (default)
- For [[Festival speech synthesis system:http://festvox.o...
FESTIVAL
- Endian definition
WORDS_BIGENDIAN (e.g. PowerPC, Cell BE, SPARC, 680x0)
WORDS_LITTLEENDIAN (e.g. x86, Alpha AXP)
- Boolean definition (HTS_Boolean)
TRUE
FALSE
* Engine structures [#z44fc354]
** Model [#le7d813d]
*** HTS_Window [#zb4a3381]
- Window coefficients to calculate dynamic features.
int size - # of windows (static + deltas)
int *l_width - left width of windows
int *r_width - right width of windows
double **coefficient - window coefficients
int max_width - maximum width of windows
*** HTS_Pattern [#k48e4d9f]
- List of patterns in a question and a tree.
char *string - pattern string
HTS_Pattern *next - pointer to the next pattern
*** HTS_Question [#z77715ac]
- List of questions in a tree.
char *name - name of this question
HTS_Pattern *head - pointer to the head of pattern list
HTS_Question *next - pointer to the next question
*** HTS_Node [#c82cbc05]
- List of tree nodes in a tree.
int index - index of this node
int pdf - index of PDF for this node (leaf n...
HTS_Node *yes - pointer to its child node (yes)
HTS_Node *no - pointer to its child node (no)
HTS_Node *next - pointer to the next node
HTS_Question *quest - question applied at this node
*** HTS_Tree [#tccac793]
- List of decision trees in a model.
HTS_Pattern *head - pointer to the head of pattern list ...
HTS_Tree *next - pointer to the next tree
HTS_Node *root - root node of this tree
int state - state index of this tree
*** HTS_Model [#lb0795d7]
- Set of PDFs, decision trees and questions.
int vector_length - vector length (include static a...
int *npdf - # of PDFs at each tree
double ***pdf - PDFs
HTS_Tree *tree - pointer to the list of trees
HTS_Question *question - pointer to the list of questions
*** HTS_Stream [#e8ebe6f1]
- Set of models and a window.
int vector_length - vector_length (include static...
HTS_Model *model - models
HTS_DynamicWindow window - window coefficients
HTS_Boolean msd_flag - flag for MSD
int interpolation_size - # of models for interpolation
*** HTS_ModelSet [#n0021e72]
- Set of duration models, HMMs and GV models.
HTS_Stream duration - duration PDFs and trees
HTS_Stream *stream - parameter PDFs, trees and windows
HTS_Stream *gv - GV PDFs
int nstate - # of HMM states
int nstream - # of stream
** Label [#zdbf97c6]
*** HTS_LabelString [#k2bc37e0]
- Individual label string with time infomation.
HTS_LabelString *next - pointer to the next label string
char *name - label string
HTS_Boolean frame_flag - flag for frame length modificat...
int frame - frame length specified in the g...
*** HTS_Label [#cf3de27c]
- List of label strings.
HTS_LabelString *head - pointer to the head of label str...
int size - # of label strings
double speech_speed - speech speed rate
** State stream [#p6180dbc]
*** HTS_SStream [#g80be9f5]
- Individual state stream.
int vector_length - vector length (include static...
double **mean - mean vector sequence
double **vari - variance vector sequence
double *msd - MSD parameter sequence
int win_size - # of windows (static + deltas)
int *win_l_width - left width of windows
int *win_r_width - right width of windows
double **win_coefficient - window coefficients
int win_max_width - maximum width of windows
double *gv_mean - mean vector of GV
double *gv_vari - variance vector of GV
*** HTS_SStreamSet [#c7e181ed]
- Set of state stream.
HTS_SStream *sstream - state streams
int nstream - # of streams
int nstate - # of states
int *duration - duration sequence
int total_state - total state
int total_frame - total frame
** PDF stream [#fbcf0b4b]
*** HTS_SMatrices [#n16e4df2]
- Matrices/Vectors used in the speech parameter generatio...
double **mean - mean vector sequence
double **ivar - inverse diag variance sequence
double *g - vector used in the forward substitution
double **wuw - W' U^-1 W
double *wum - W' U^-1 mu
*** HTS_PStream [#v2504efe]
- Individual PDF stream.
int vector_length - vector length (include static...
int static_length - static features length
int length - stream length
int width - width of dynamic window
double **par - output parameter vector
HTS_SMatrices sm - matrices for parameter genera...
int win_size - # of windows (static + deltas)
int *win_l_width - left width of windows
int *win_r_width - right width of windows
double **win_coefficient - window coefficients
HTS_Boolean *msd_flag - Boolean sequence for MSD
double *gv_buff - buffer for GV calculation
double *gv_mean - mean vector of GV
double *gv_vari - variance vector of GV
double gv_weight - GV weight
*** HTS_PStreamSet [#fc667ac9]
- Set of PDF streams.
HTS_PStream *pstream - PDF streams
int nstream - # of PDF streams
int total_frame - total frame
** Generated parameter stream [#cb40623b]
*** HTS_GStream [#r1f0e0dc]
- Generated parameter stream.
int static_length - static features length
double **par - generated parameter
*** HTS_GStreamSet [#qab80b0e]
- Set of generated parameter stream.
int total_nsample - total sample
int total_frame - total frame
int nstream - # of streams
HTS_GStream *gstream - generated parameter streams
short *gspeech - generated speech
** Engine [#j629fb64]
*** HTS_Global [#x0a5815e]
- Global settings.
int stage - Gamma=-1/stage : if stage=0 t...
HTS_Boolean use_log_gain - log gain flag (for LSP)
int sampling_rate - sampling rate
int fperiod - frame period
double alpha - all-pass constant
double beta - postfiltering coefficient
int audio_buff_size - audio buffer size (for audio ...
double *msd_threshold - MSD thresholds
double *duration_iw - weights for duration interpol...
double **parameter_iw - weights for parameter interpo...
double **gv_iw - weights for GV interpolation
double *gv_weight - GV weights
*** HTS_Engine [#wa56bd53]
- Engine itself.
HTS_Global global - global settings
HTS_ModelSet ms - set of duration models, HMMs and GV...
HTS_Label label - label
HTS_SStreamSet sss - set of state streams
HTS_PStreamSet pss - set of PDF streams
HTS_GStreamSet gss - set of generated parameter streams
* Vocoder structures [#gc6a2eb8]
** Audio [#i33a0fbe]
*** HTS_Audio [#se4ce25f]
- For MS Windows (Windows Mobile) audio output device.
HWAVEOUT hwaveout - audio device handle
WAVEFORMATEX waveformatex - wave formatex
short *buff - current buffer
int buff_size - current buffer size
int which_buff - double buffering flag
HTS_Boolean now_buff_1 - double buffering flag
HTS_Boolean now_buff_2 - double buffering flag
WAVEHDR buff_1 - buffer
WAVEHDR buff_2 - buffer
int max_buff_size - buffer size of audio output ...
- For Linux, etc.
int i - make compiler happy
** Vocoder [#lbf2a363]
*** HTS_Vocoder [#p4d69001]
- MLSA/MGLSA filter settings.
int stage - Gamma=-1/stage : if stage=0...
double gamma - Gamma
HTS_Boolean use_log_gain - log gain flag (for LSP)
int fprd - frame shift
int iprd - interpolation period
int seed - seed of random generator
unsigned long next - temporary variable for rand...
HTS_Boolean gauss - flag to use Gaussian noise
double rate - sampling rate
double p1 - used in excitation generation
double pc - used in excitation generation
double p - used in excitation generation
double inc - used in excitation generation
int sw - switch used in random gener...
int x - excitation signal
HTS_Audio *audio - pointer for audio device
double *freqt_buff - used in freqt
int freqt_size - buffer size for freqt
double *spectrum2en_buff - used in spectrum2en
int spectrum2en_size - buffer size for spectrum2en
double r1, r2, s - used in random generator
double *postfilter_buff - used in postfiltering
int postfilter_size - buffer size for postfiltering
double *c, *cc, *cinc, *d1 - used in the MLSA/MGLSA filter
double *pade - used in mlsadf
double *lsp2lpc_buff - used in lsp2lpc
int lsp2lpc_size - buffer size of lsp2lpc
double *gc2gc_buff - used in gc2gc
int gc2gc_size - buffer size for gc2gc
* Engine functions [#l74e4110]
** Initialize engine [#t310270e]
*** HTS_Engine_initialize [#ec43b3cc]
void HTS_Engine_initialize(HTS_Engine *engine, int nstre...
- Use: Initialize engine.
- Arguments:
HTS_Engine *engine - pointer to HTS_Engine structure
- ''Attention!!:'' To start engine, first you must call t...
** Load models [#z7b8be60]
*** HTS_Engine_load_duration_from_fn [#va73cd1a]
void HTS_Engine_load_duration_from_fn(HTS_Engine *engine...
- Use: Load duration PDFs and trees from files using give...
- Arguments:
HTS_Engine *engine - pointer to HTS_Engine structure
char **pdf_fn - duration PDF file names
char **tree_fn - duration tree file names
int interpolation_size - # of duration models to be inte...
- ''Attention!!'': You must initialize variable '''engine...
*** HTS_Engine_load_duration_from_fp [#wc2481e3]
void HTS_Engine_load_duration_from_fp(HTS_Engine *engine...
- Use: load duration PDFs and trees from files using give...
- Arguments:
HTS_Engine *engine - pointer to HTS_Engine structure
FILE **pdf_fp - duration PDF file pointers
FILE **tree_fp - duration tree file pointers
int interpolation_size - # of duration models to be inte...
- ''Attention!!'': You must initialize variable '''engine...
*** HTS_Engine_load_parameter_from_fn [#kd48cfd2]
void HTS_Engine_load_parameter_from_fn(HTS_Engine *engin...
int stream_index,...
- Use: load parameter PDFs, trees and windows from files ...
- Arguments:
HTS_Engine *engine - pointer to HTS_Engine structure
char **pdf_fn - parameter PDF file pointers
char **tree_fn - parameter tree file pointers
char **win_fn - parameter window file pointers
int stream_index - index of stream
HTS_Boolean msd_flag - flag for MSD
int window_size - # of windows
int interpolation_size - # of parameter models to be int...
- ''Attention!!'': You must load duration models using HT...
*** HTS_Engine_load_parameter_from_fp [#j2b19c02]
void HTS_Engine_load_parameter_from_fp(HTS_Engine *engin...
int stream_index,...
- Use: load parameter PDFs, trees and windows from files ...
- Arguments:
HTS_Engine *engine - pointer to HTS_Engine structure
FILE **pdf_fp - parameter PDF file pointers
FILE **tree_fp - parameter tree file pointers
FILE **win_fp - parameter window file pointers
int stream_index - index of stream
HTS_Boolean msd_flag - MSD flag
int window_size - # of windows
int interpolation_size - # of parameter models to be int...
- ''Attention!!'': You must load duration models using HT...
*** HTS_Engine_load_gv_from_fn [#id008a1a]
void HTS_Engine_load_gv_from_fn(HTS_Engine *engine, char...
- Arguments:
HTS_Engine *engine - pointer to HTS_Engine structure
char **pdf_fn - GV PDF file names
int stream_index - index of streams
int interpolation_size - # of GV models to be interpolated
- ''Attention!!'': You must load parameter models using H...
*** HTS_Engine_load_gv_from_fp [#g6d9d662]
void HTS_Engine_load_gv_from_fp(HTS_Engine *engine, FILE...
- Arguments:
HTS_Engine *engine - pointer to HTS_Engine structure
FILE **pdf_fn - GV PDF file pointers
int stream_index - index of streams
int interpolation_size - # of GV models to be interpolated
- ''Attention!!'': You must load parameter models using H...
** Set parameter [#b8dd5d39]
*** HTS_Engine_set_sampling_rate [#q41e65a1]
void HTS_Engine_set_sampling_rate(HTS_Engine *engine,int...
- Use: set sampling frequency.
- Arguments:
HTS_Engine *engine - pointer to HTS_Engine structure
int i - sampling frequency (Hz), 1= < i <= ...
- ''Attention!!:'' Default value is 16000.
*** HTS_Engine_set_fperiod [#c5c89881]
void HTS_Engine_set_fperiod(HTS_Engine *engine,int i)
- Use: set frame shift.
- Arguments:
HTS_Engine *engine - pointer to HTS_Engine structure
int i - frame shift (point), 1= < i <= 2000
- ''Attention!!:'' Default value is 80.
*** HTS_Engine_set_alpha [#tbe7cee5]
void HTS_Engine_set_alpha(HTS_Engine *engine,double f)
- Use: set frequency warping parameter alpha.
- Arguments:
HTS_Engine *engine - pointer to HTS_Engine structure
double f - alpha, 0.0 <= f <= 1.0
- ''Attention!!:'' Default value is 0.42.
*** HTS_Engine_set_gamma [#wdc30080]
void HTS_engine_set_gamma(HTS_Engine *engine,int i)
- Use: set Gamma.
- Arguments:
HTS_Engine *engine - pointer to HTS_Engine structure
int i - Gamma=-1/i : if i=0 then Gamma=0, 0...
- ''Attention!!:'' Default value is 0.
*** HTS_Engine_use_log_gain [#lfbc8c63]
void HTS_Engine_set_log_gain(HTS_Engine *engine,HTS_Bool...
- Use: set log gain flag.
- Arguments:
HTS_Engine *engine - pointer to HTS_Engine structure
HTS_Boolean i - log gain flag.
- ''Attention!!:'' Default value is FALSE.
*** HTS_Engine_set_beta [#m44c517c]
void HTS_SetBeta(HTS_Engine *engine,double f)
- Use: set postfiltering coefficient parameter beta.
- Arguments:
HTS_Engine *engine - pointer to HTS_Engine structure
double f - beta, -0.8 <= f <= 0.8
- ''Attention!!:'' Default value is 0.0. If you set beta...
*** HTS_Engine_set_audio_buff_size [#ua56725a]
void HTS_Engine_set_audio_buff_size(HTS_Engine *engine, ...
- Use: set buffer size for direct audio output.
- Arguments:
HTS_Engine *engine - pointer to HTS_Engine structure
int i - buffer size (sample), 0 <= i <= 48000
- ''Attention!!:'' Default value is 0. If i=0, direct aud...
*** HTS_Engine_set_msd_threshold [#qd019ba4]
void HTS_Engine_set_msd_threshold(HTS_Engine *engine, in...
- Use: set MSD threshold.
- Arguments:
HTS_Engine *engine - pointer to HTS_Engine structure
int stream_index - index of streams
double f - threshold
*** HTS_Engine_set_duration_interpolation_weight [#rdd5ff...
void HTS_Engine_set_duration_interpolation_weight(HTS_En...
- Use: set weight for duration interpolation.
- Arguments:
HTS_Engine *engine - pointer to HTS_Engine structure
int interpolation_index - index of duration models
double f - interpolation weight
*** HTS_Engine_set_parameter_interpolation_weight [#k8bc0...
void HTS_Engine_set_parameter_interpolation_weight(HTS_E...
- Use: set weight for parameter interpolation.
- Arguments:
HTS_Engine *engine - pointer to HTS_Engine structure
int stream_index - index of streams
int interpolation_index - index of parameter models
double f - interpolation weight
*** HTS_Engine_set_gv_interpolation_weight [#lc3e00bc]
void HTS_Engine_set_gv_interpolation_weight(HTS_Engine *...
- Use: set weight for GV interpolation.
- Arguments:
HTS_Engine *engine - pointer to HTS_Engine structure
int stream_index - index of streams
int interpolation_index - index of GV models
double f - interpolation weight
*** HTS_Engine_set_gv_weight [#eaa46840]
void HTS_Engine_set_gv_weight(HTS_Engine *engine, int st...
- Use: set GV weight.
- Arguments:
HTS_Engine *engine - pointer to HTS_Engine structure
int stream_index - index of streams
double f - GV weight
** Synthesize speech [#sf7321ed]
*** HTS_Engine_load_label_from_fn [#j87fa620]
void HTS_Engine_load_label_from_fn(HTS_Engine *engine, c...
- Use: load label from file using given file name.
- Arguments:
HTS_Engine *engine - pointer to HTS_Engine structure
char *fn - label file name
*** HTS_Engine_load_label_from_fp [#i892f523]
void HTS_Engine_load_label_from_fp(HTS_Engine *engine, F...
- Use: load label from file using given file pointer.
- Arguments:
HTS_Engine *engine - pointer to HTS_Engine structure
FILE *fp - label file pointer
*** HTS_Engine_load_label_from_string [#q36e8c4c]
void HTS_Engine_load_label_from_string(HTS_Engine *engin...
- Use: load label from string.
- Arguments:
HTS_Engine *engine - pointer to HTS_Engine structure
char *data - label string
*** HTS_Engine_load_label_from_string_list [#i9e769ac]
void HTS_Engine_load_label_from_string_list(HTS_Engine *...
- Use: load label from string list.
- Arguments:
HTS_Engine *engine - pointer to HTS_Engine structure
char **data - label string list
int size - size of label string list
*** HTS_Engine_create_sstream [#c2e0aaed]
void HTS_Engine_create_sstream(HTS_Engine *engine)
- Use: parse label, determine state duration and create s...
- Arguments:
HTS_Engine *engine - pointer to HTS_Engine structure
*** HTS_Engine_create_pstream [#b1fda60c]
void HTS_Engine_create_pstream(HTS_Engine *engine)
- Use: create PDF stream using state stream.
- Arguments:
HTS_Engine *engine - pointer to HTS_Engine structure
*** HTS_Engine_create_gstream [#fab90e37]
void HTS_Engine_create_gstream(HTS_Engine *engine)
- Use: synthesis speech and store generated parameter usi...
- Arguments:
HTS_Engine *engine - pointer to HTS_Engine structure
- ''Attention!!:'' To synthesize speech, you must set str...
*** HTS_Engine_save_infomation [#e7c9463d]
void HTS_Engine_save_infomation(HTS_Engine *engine, FILE...
- Use: output trace infomation.
- Arguments:
HTS_Engine *engine - pointer to HTS_Engine structure
FILE *fp - output file pointer
*** HTS_Engine_save_label [#sb2d1f59]
void HTS_Engine_save_label(HTS_Engine *engine, FILE *fp)
- Use: output label with time.
- Arguments:
HTS_Engine *engine - pointer to HTS_Engine structure
FILE *fp - output file pointer
*** HTS_Engine_save_generated_parameter [#ya641edd]
void HTS_Engine_save_generated_parameter(HTS_Engine *eng...
- Use: output generated parameter.
- Arguments:
HTS_Engine *engine - pointer to HTS_Engine structure
FILE *fp - output file pointer
*** HTS_Engine_save_generated_speech [#k8327b8d]
void HTS_Engine_save_generated_speech(HTS_Engine *engine...
- Use: output generated speech.
- Arguments:
HTS_Engine *engine - pointer to HTS_Engine structure
FILE *fp - output file pointer
*** HTS_Engine_save_riff [#wb9640f7]
void HTS_Engine_save_riff(HTS_Engine *engine, FILE *fp)
- Use: output riff format file.
- Arguments:
HTS_Engine *engine - pointer to HTS_Engine structure
FILE *fp - output file pointer
*** HTS_Engine_refresh [#a8a1a225]
void HTS_Engine_refresh(HTS_Engine *engine)
- Use: free label, state streams, PDF streams and generat...
- Arguments:
HTS_Engine *engine - pointer to HTS_Engine structure
** Free engine [#mef9b660]
*** HTS_Engine_clear [#gd323729]
void HTS_Engine_clear(HTS_Engine *engine)
- Use: free engine.
- Arguments:
HTS_Engine *engine - pointer to HTS_Engine structure
* Vocoder functions [#tffbab44]
** Initialize vocoder [#v5bc81e4]
*** HTS_Vocoder_initialize [#ff9170fc]
void HTS_Vocoder_initialize(HTS_Vocoder *v, const int m,...
- Use: initialize the vocoder.
- Arguments:
HTS_Vocoder *v - pointer to HTS_Vocoder struct...
int m - order of mel-cepstral coeffic...
int stage - Gamma=-1/stage : if stage=0 t...
HTS_Boolean use_log_gain - log gain flag
int rate - sampling frequency
int fperiod - frame shift
int buff_size - buffer size for direct audio ...
** Synthesize speech [#d6645d7f]
*** HTS_Vocoder_synthesize [#h737ba80]
void HTS_Vocoder_synthesize(HTS_Vocoder *v, const int m,...
- Use: run the vocoder and synthesize waveform.
- Arguments:
HTS_Vocoder *v - HTS_Vocoder structure pointer
int m - order of spectrum coefficients
double lf0 - log F0 value
double *spectrum - spectrum coefficients
double alpha - frequency warping parameter alpha
double beta - postfiltering parameter beta
short *rawdata - short pointer to store synthesized wa...
*** HTS_Vocoder_postfilter_mcp [#r92930b7]
void HTS_Vocoder_postfilter_mcp(HTS_Vocoder *v, double *...
- Use: postfilter for mel-cepstrum.
- Arguments:
HTS_Vocoder *v - HTS_Vocoder structure pointer
double *mcp - spectrum coefficients
int m - order of mel-cepstral coefficients
double alpha - frequency warping parameter alpha
double beta - postfiltering parameter beta
** Free vocoder [#rb2ef784]
*** HTS_Vocoder_clear [#wb420be0]
void HTS_Vocoder_clear(HTS_Vocoder *v)
- Use: free the vocoder.
- Arguments:
HTS_Vocoder *v - pointer to HTS_Vocoder structure
* Other functions [#q32f7c6d]
** For copyright [#e99f219b]
*** void HTS_show_copyright(FILE *fp) [#y30d6252]
- Use: show hts_engine_API copyright.
- Arguments:
FILE *fp - output file pointer
*** void HTS_get_copyright(char *str) [#d24078ea]
- Use: copy hts_engine_API copyright to string.
- Arguments:
char *str - output string
Page:
![[PukiWiki]](skin/logo.png "[PukiWiki]"){kind=logo}
