1 /* -*- c-basic-offset: 4; indent-tabs-mode: nil -*- */ 2 /* ==================================================================== 3 * Copyright (c) 1999-2004 Carnegie Mellon University. All rights 4 * reserved. 5 * 6 * Redistribution and use in source and binary forms, with or without 7 * modification, are permitted provided that the following conditions 8 * are met: 9 * 10 * 1. Redistributions of source code must retain the above copyright 11 * notice, this list of conditions and the following disclaimer. 12 * 13 * 2. Redistributions in binary form must reproduce the above copyright 14 * notice, this list of conditions and the following disclaimer in 15 * the documentation and/or other materials provided with the 16 * distribution. 17 * 18 * This work was supported in part by funding from the Defense Advanced 19 * Research Projects Agency and the National Science Foundation of the 20 * United States of America, and the CMU Sphinx Speech Consortium. 21 * 22 * THIS SOFTWARE IS PROVIDED BY CARNEGIE MELLON UNIVERSITY ``AS IS'' AND 23 * ANY EXPRESSED OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, 24 * THE IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR 25 * PURPOSE ARE DISCLAIMED. IN NO EVENT SHALL CARNEGIE MELLON UNIVERSITY 26 * NOR ITS EMPLOYEES BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL, 27 * SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT 28 * LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, 29 * DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY 30 * THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT 31 * (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE 32 * OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE. 33 * 34 * ==================================================================== 35 * 36 */ 37 /* 38 * senone.h -- Mixture density weights associated with each tied state. 39 * 40 * ********************************************** 41 * CMU ARPA Speech Project 42 * 43 * Copyright (c) 1996 Carnegie Mellon University. 44 * ALL RIGHTS RESERVED. 45 * ********************************************** 46 * 47 * HISTORY 48 * 49 * $Log$ 50 * Revision 1.1 2006/04/05 20:27:30 dhdfu 51 * A Great Reorganzation of header files and executables 52 * 53 * Revision 1.7 2006/02/22 17:27:39 arthchan2003 54 * Merged from SPHINX3_5_2_RCI_IRII_BRANCH: 1, NOT doing truncation in the multi-stream GMM computation \n. 2, Added .s3cont. to be the alias of the old multi-stream GMM computation routine \n. 3, Added license \n. 4, Fixed dox-doc. \n 55 * 56 * Revision 1.6.4.4 2006/01/16 19:47:05 arthchan2003 57 * Removed the truncation of senone probability code. 58 * 59 * Revision 1.6.4.3 2005/08/03 18:53:43 dhdfu 60 * Add memory deallocation functions. Also move all the initialization 61 * of ms_mgau_model_t into ms_mgau_init (duh!), which entails removing it 62 * from decode_anytopo and friends. 63 * 64 * Revision 1.6.4.2 2005/07/20 19:39:01 arthchan2003 65 * Added licences in ms_* series of code. 66 * 67 * Revision 1.6.4.1 2005/07/05 05:47:59 arthchan2003 68 * Fixed dox-doc. struct level of documentation are included. 69 * 70 * Revision 1.6 2005/06/21 19:00:19 arthchan2003 71 * Add more detail comments to ms_senone.h 72 * 73 * Revision 1.5 2005/06/21 18:57:31 arthchan2003 74 * 1, Fixed doxygen documentation. 2, Added $ keyword. 75 * 76 * Revision 1.2 2005/06/13 04:02:56 archan 77 * Fixed most doxygen-style documentation under libs3decoder. 78 * 79 * Revision 1.1.1.1 2005/03/24 15:24:00 archan 80 * I found Evandro's suggestion is quite right after yelling at him 2 days later. So I decide to check this in again without any binaries. (I have done make distcheck. ) . Again, this is a candidate for s3.6 and I believe I need to work out 4-5 intermediate steps before I can complete the first prototype. That's why I keep local copies. 81 * 82 * Revision 1.4 2004/12/06 10:52:01 arthchan2003 83 * Enable doxygen documentation in libs3decoder 84 * 85 * Revision 1.3 2004/11/13 21:25:19 arthchan2003 86 * commit of 1, absolute CI-GMMS , 2, fast CI senone computation using svq, 3, Decrease the number of static variables, 4, fixing the random generator problem of vector_vqgen, 5, move all unused files to NOTUSED 87 * 88 * Revision 1.2 2004/08/31 08:43:47 arthchan2003 89 * Fixing _cpluscplus directive 90 * 91 * Revision 1.1 2004/08/09 00:17:11 arthchan2003 92 * Incorporating s3.0 align, at this point, there are still some small problems in align but they don't hurt. For example, the score doesn't match with s3.0 and the output will have problem if files are piped to /dev/null/. I think we can go for it. 93 * 94 * Revision 1.1 2003/02/14 14:40:34 cbq 95 * Compiles. Analysis is probably hosed. 96 * 97 * Revision 1.1 2000/04/24 09:39:41 lenzo 98 * s3 import. 99 * 100 * 101 * 13-Dec-95 M K Ravishankar (rkm@cs.cmu.edu) at Carnegie Mellon University. 102 * Added senone_eval_all(). 103 * 104 * 12-Nov-95 M K Ravishankar (rkm@cs.cmu.edu) at Carnegie Mellon University. 105 * Created. 106 */ 107 108 109 #ifndef _LIBFBS_SENONE_H_ 110 #define _LIBFBS_SENONE_H_ 111 112 113 /* SphinxBase headers. */ 114 #include <sphinxbase/err.h> 115 #include <sphinxbase/ckd_alloc.h> 116 #include <sphinxbase/cmd_ln.h> 117 #include <sphinxbase/logmath.h> 118 119 /* Local headers. */ 120 #include "ms_gauden.h" 121 #include "bin_mdef.h" 122 123 /** \file ms_senone.h 124 * \brief (Sphinx 3.0 specific) multiple streams senones. used with ms_gauden.h 125 * In Sphinx 3.0 family of tools, ms_senone is used to combine the Gaussian scores. 126 * Its existence is crucial in Sphinx 3.0 because 3.0 supports both SCHMM and CDHMM. 127 * There are optimization scheme for SCHMM (e.g. compute the top-N Gaussian) that is 128 * applicable to SCHMM than CDHMM. This is wrapped in senone_eval_all. 129 */ 130 131 #ifdef __cplusplus 132 extern "C" { 133 #endif 134 135 typedef uint8 senprob_t; /**< Senone logs3-probs, truncated to 8 bits */ 136 137 /** 138 * \struct senone_t 139 * \brief 8-bit senone PDF structure. 140 * 141 * 8-bit senone PDF structure. Senone pdf values are normalized, floored, converted to 142 * logs3 domain, and finally truncated to 8 bits precision to conserve memory space. 143 */ 144 typedef struct { 145 senprob_t ***pdf; /**< gaussian density mixture weights, organized two possible 146 ways depending on n_gauden: 147 if (n_gauden > 1): pdf[sen][feat][codeword]. Not an 148 efficient representation--memory access-wise--but 149 evaluating the many codebooks will be more costly. 150 if (n_gauden == 1): pdf[feat][codeword][sen]. Optimized 151 for the shared-distribution semi-continuous case. */ 152 logmath_t *lmath; /**< log math computation */ 153 uint32 n_sen; /**< Number senones in this set */ 154 uint32 n_feat; /**< Number feature streams */ 155 uint32 n_cw; /**< Number codewords per codebook,stream */ 156 uint32 n_gauden; /**< Number gaussian density codebooks referred to by senones */ 157 float32 mixwfloor; /**< floor applied to each PDF entry */ 158 uint32 *mgau; /**< senone-id -> mgau-id mapping for senones in this set */ 159 int32 *featscr; /**< The feature score for every senone, will be initialized inside senone_eval_all */ 160 int32 aw; /**< Inverse acoustic weight */ 161 } senone_t; 162 163 164 /** 165 * Load a set of senones (mixing weights and mixture gaussian codebook mappings) from 166 * the given files. Normalize weights for each codebook, apply the given floor, convert 167 * PDF values to logs3 domain and quantize to 8-bits. 168 * @return pointer to senone structure created. Caller MUST NOT change its contents. 169 */ 170 senone_t *senone_init (gauden_t *g, /**< In: codebooks */ 171 char const *mixwfile, /**< In: mixing weights file */ 172 char const *mgau_mapfile,/**< In: file specifying mapping from each 173 senone to mixture gaussian codebook. 174 If NULL all senones map to codebook 0 */ 175 float32 mixwfloor, /**< In: Floor value for senone weights */ 176 logmath_t *lmath, /**< In: log math computation */ 177 bin_mdef_t *mdef /**< In: model definition */ 178 ); 179 180 /** Release memory allocated by senone_init. */ 181 void senone_free(senone_t *s); /**< In: The senone_t to free */ 182 183 /** 184 * Evaluate the score for the given senone wrt to the given top N gaussian codewords. 185 * @return senone score (in logs3 domain). 186 */ 187 int32 senone_eval (senone_t *s, int id, /**< In: senone for which score desired */ 188 gauden_dist_t **dist, /**< In: top N codewords and densities for 189 all features, to be combined into 190 senone score. IE, dist[f][i] = i-th 191 best <codeword,density> for feaure f */ 192 int n_top /**< In: Length of dist[f], for each f */ 193 ); 194 195 #ifdef __cplusplus 196 } 197 #endif 198 199 #endif 200