40 #ifndef __PS_CMDLN_MACRO_H__ 41 #define __PS_CMDLN_MACRO_H__ 43 #include <sphinxbase/cmd_ln.h> 44 #include <sphinxbase/feat.h> 45 #include <sphinxbase/fe.h> 48 #define POCKETSPHINX_OPTIONS \ 49 waveform_to_cepstral_command_line_macro(), \ 50 cepstral_to_feature_command_line_macro(), \ 51 POCKETSPHINX_ACMOD_OPTIONS, \ 52 POCKETSPHINX_BEAM_OPTIONS, \ 53 POCKETSPHINX_SEARCH_OPTIONS, \ 54 POCKETSPHINX_DICT_OPTIONS, \ 55 POCKETSPHINX_NGRAM_OPTIONS, \ 56 POCKETSPHINX_FSG_OPTIONS, \ 57 POCKETSPHINX_DEBUG_OPTIONS 60 #define POCKETSPHINX_DEBUG_OPTIONS \ 64 "File to write log messages in" \ 69 "Verbosity level for debugging messages" \ 74 "Directory to log feature files to" \ 79 "Directory to log raw audio files to" }, \ 83 "Directory to log senone score files to" \ 87 #define POCKETSPHINX_BEAM_OPTIONS \ 91 "Beam width applied to every frame in Viterbi search (smaller values mean wider beam)" }, \ 95 "Beam width applied to word exits" }, \ 99 "Beam width applied to phone transitions" }, \ 103 "Beam width applied to last phone in words" }, \ 107 "Beam width applied to last phone in single-phone words" }, \ 111 "Beam width applied to every frame in second-pass flat search" }, \ 115 "Beam width applied to word exits in second-pass flat search" }, \ 119 "Phoneme lookahead window size, in frames" }, \ 123 "Beam width applied to phone loop search for lookahead" }, \ 127 "Beam width applied to phone loop transitions for lookahead" } 130 #define POCKETSPHINX_SEARCH_OPTIONS \ 134 "Compute all senone scores in every frame (can be faster when there are many senones)" }, \ 138 "Run forward lexicon-tree search (1st pass)" }, \ 142 "Run forward flat-lexicon search over word lattice (2nd pass)" }, \ 146 "Run bestpath (Dijkstra) search over word lattice (3rd pass)" }, \ 150 "Print results and backtraces to log file." }, \ 154 "Initial backpointer table size" }, \ 158 "Maximum number of distinct word exits at each frame (or -1 for no pruning)" }, \ 162 "Maximum number of active HMMs to maintain at each frame (or -1 for no pruning)" }, \ 166 "Nodes ignored in lattice construction if they persist for fewer than N frames" }, \ 170 "Minimum number of end frames for a word to be searched in fwdflat search" }, \ 174 "Window of frames in lattice to search for successor words in fwdflat search " } 177 #define POCKETSPHINX_FSG_OPTIONS \ 181 "Sphinx format finite state grammar file"}, \ 185 "JSGF grammar file" }, \ 189 "Start rule for JSGF (first public rule is default)" }, \ 190 { "-fsgusealtpron", \ 193 "Add alternate pronunciations to FSG"}, \ 197 "Insert filler words at each state."} 200 #define POCKETSPHINX_NGRAM_OPTIONS \ 204 "Word trigram language model input file" }, \ 208 "Specify a set of language model\n"}, \ 212 "Which language model in -lmctl to use by default"}, \ 216 "Language model probability weight" }, \ 220 "Language model probability weight for flat lexicon (2nd pass) decoding" }, \ 224 "Language model probability weight for bestpath search" }, \ 228 "Inverse of acoustic model scale for confidence score calculation" }, \ 232 "Word insertion penalty" }, \ 236 "New word transition penalty" }, \ 240 "Phone insertion penalty" }, \ 244 "Unigram weight" }, \ 248 "Silence word transition probability" }, \ 252 "Filler word transition probability" }, \ 256 "Bigram-mode: If TRUE only one BP entry/frame; else one per LM state" }, \ 260 "Whether to dump the lextree structure to stderr (for debugging), 1 for Ravi's format, 2 for Dot format, Larger than 2 will be treated as Ravi's format" } 263 #define POCKETSPHINX_DICT_OPTIONS \ 267 "Main pronunciation dictionary (lexicon) input file" }, \ 271 "Noise word pronunciation dictionary input file" }, \ 275 "Dictionary is case sensitive (NOTE: case insensitivity applies to ASCII characters only)" }, \ 279 "Maximum new OOVs that can be added at run time" }, \ 283 "Use within-word phones only" } 286 #define POCKETSPHINX_ACMOD_OPTIONS \ 290 "Directory containing acoustic model files."}, \ 294 "File containing feature extraction parameters."}, \ 298 "Model definition input file" }, \ 302 "Senone to codebook mapping input file (usually not needed)" }, \ 306 "HMM state transition matrix input file" }, \ 310 "HMM state transition probability floor (applied to -tmat file)" }, \ 314 "Mixture gaussian means input file" }, \ 318 "Mixture gaussian variances input file" }, \ 322 "Mixture gaussian variance floor (applied to data from -var file)" }, \ 326 "Senone mixture weights input file (uncompressed)" }, \ 330 "Senone mixture weights floor (applied to data from -mixw file)" }, \ 334 "Inverse weight applied to acoustic scores." }, \ 338 "Senone dump (compressed mixture weights) input file" }, \ 342 "MLLR transformation to apply to means and variances" }, \ 346 "Use memory-mapped I/O (if possible) for model files" }, \ 350 "Frame GMM computation downsampling ratio" }, \ 354 "Maximum number of top Gaussians to use in scoring." }, \ 358 "Beam width used to determine top-N Gaussians (or a list, per-feature)" },\ 362 "kd-Tree file for Gaussian selection" }, \ 366 "Maximum depth of kd-Trees to use" }, \ 370 "Maximum number of Gaussians per leaf node in kd-Trees" }, \ 374 "Base in which all log-likelihoods calculated" } 376 #define CMDLN_EMPTY_OPTION { NULL, 0, NULL, NULL }