/* -*- c-basic-offset: 4; indent-tabs-mode: nil -*- */ /* ==================================================================== * Copyright (c) 1999-2004 Carnegie Mellon University. All rights * reserved. * * Redistribution and use in source and binary forms, with or without * modification, are permitted provided that the following conditions * are met: * * 1. Redistributions of source code must retain the above copyright * notice, this list of conditions and the following disclaimer. * * 2. Redistributions in binary form must reproduce the above copyright * notice, this list of conditions and the following disclaimer in * the documentation and/or other materials provided with the * distribution. * * This work was supported in part by funding from the Defense Advanced * Research Projects Agency and the National Science Foundation of the * United States of America, and the CMU Sphinx Speech Consortium. * * THIS SOFTWARE IS PROVIDED BY CARNEGIE MELLON UNIVERSITY ``AS IS'' AND * ANY EXPRESSED OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, * THE IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR * PURPOSE ARE DISCLAIMED. IN NO EVENT SHALL CARNEGIE MELLON UNIVERSITY * NOR ITS EMPLOYEES BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL, * SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT * LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, * DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY * THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT * (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE * OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE. * * ==================================================================== * */ /* * kb.h -- Knowledge bases, search parameters, and auxiliary structures for decoding * * ********************************************** * CMU ARPA Speech Project * * Copyright (c) 1999 Carnegie Mellon University. * ALL RIGHTS RESERVED. * ********************************************** * * HISTORY * $Log$ * Revision 1.1 2006/04/05 20:27:30 dhdfu * A Great Reorganzation of header files and executables * * Revision 1.18 2006/02/23 05:44:59 arthchan2003 * Merged from the branch SPHINX3_5_2_RCI_IRII_BRANCH. * 1, Added temp_init_vithistory, this will choose to initialize the correct viterbi history given the mode. * 2, Moved most of the code in kb_setmllr to adaptor.c * * Revision 1.17.4.2 2005/09/18 01:21:18 arthchan2003 * 1, Add a latticehist_t into kb_t, use a temporary method to allow polymorphism of initialization of vithist_t and latticehist_t. 2, remove the logic kb_set_mllr and put it to adapt_set_mllr * * Revision 1.17.4.1 2005/07/03 23:00:58 arthchan2003 * Free stat_t, histprune_t and srch_t correctly. * * Revision 1.17 2005/06/21 23:21:58 arthchan2003 * Log. This is a big refactoring for kb.c and it is worthwhile to give * words on why and how things were done. There were generally a problem * that the kb structure itself is too flat. That makes it has to * maintained many structure that could be maintained by smaller * structures. For example, the count of A and the array of A should * well be put into the same structure to increase readability and * modularity. One can explain why histprune_t, pl_t, stat_t and * adapt_am_t were introduced with that line of reasoning. * * In srch_t, polymorphism of implementation is also one important * element in separting all graph related members from kb_t to srch_t. * One could probably implement the polymorphism as an interface of kb * but it is not trivial from the semantic meaning of kb. That is * probably why srch_t is introduced as the gateway of search interfaces. * * Another phenonemon one could see in the code was bad interaction * between modules. This is quite serious in two areas: logging and * checking. The current policy is unless something required cross * checking two structures, they would be done internally inside a module * initialization. * * Finally, kb_setlm is now removed and is replaced by ld_set_lm (by * users) or srch_set_lm (by developers). I think this is quite * reasonable. * * Revision 1.10 2005/06/19 19:41:23 archan * Sphinx3 to s3.generic: Added multiple regression class for single stream MLLR. Enabled MLLR for livepretend and decode. * * Revision 1.9 2005/05/11 06:10:38 archan * Code for lattice and back track pointer table dumping is now wrapped in reg_result_dump. The function is shared across mode 4 and mode 5. Possibly later for mode 3 and mode 6 as well. * * Revision 1.8 2005/04/25 23:53:35 archan * 1, Some minor modification of vithist_t, vithist_rescore can now support optional LM rescoring, vithist also has its own reporting routine. A new argument -lmrescore is also added in decode and livepretend. This can switch on and off the rescoring procedure. 2, I am reaching the final difficulty of mode 5 implementation. That is, to implement an algorithm which dynamically decide which tree copies should be entered. However, stuffs like score propagation in the leave nodes and non-leaves nodes are already done. 3, As briefly mentioned in 2, implementation of rescoring , which used to happened at leave nodes are now separated. The current implementation is not the most clever one. Wish I have time to change it before check-in to the canonical. * * Revision 1.7 2005/04/20 03:36:18 archan * Remove setlm from kb entirely, refactor it to search implementations, do the corresponding change for the changes in ascr and pl * * Revision 1.6 2005/03/30 01:22:47 archan * Fixed mistakes in last updates. Add * * * 14-Jun-2004 Yitao Sun (yitao@cs.cmu.edu) at Carnegie Mellon University * Modified struct kb_t to save the last hypothesis. * * 07-Jul-1999 M K Ravishankar (rkm@cs.cmu.edu) at Carnegie Mellon University * Added kb_t.ci_active. * * 02-Jun-1999 M K Ravishankar (rkm@cs.cmu.edu) at Carnegie Mellon University * Started. */ #ifndef _S3_KB_H_ #define _S3_KB_H_ #include #include #include "kbcore.h" #include "lextree.h" #include "vithist.h" #include "ascr.h" #include "fast_algo_struct.h" #include "mllr.h" #include "cmn.h" #include "stat.h" #include "adaptor.h" #include "cb2mllr_io.h" /** \file kb.h * \brief The global wrapper structure for all variables in 3.X * search. We may want to use it for sphinx 3.0 as well. */ #ifdef __cplusplus extern "C" { #endif #if 0 } /* Fool Emacs into not indenting things. */ #endif /* * ARCHAN :20040229. Starting from s3.4, we accept users to specify mutliple LMs * using -lmctlfile. To avoid potential overhead caused by rebuilding the trees * at every sentence. We allocate all lexical trees in the intialization. If we * assume the number of words to be less than 64k and the number of LMs is smaller * than 100. This memory loss should be fine. Later: We should allow users to specify * the memory address mode by their own discretion. */ /* The wrapper structure of all operation in the Sphinx 3.X decoder. * */ typedef struct { /** Core models, defined as acoustic and language models, dictionary (pronounciation models_, front-ends, filler-penalties approximate acoustic models such as sub vector quantization map and Gaussian selector */ kbcore_t *kbcore; /**< Core model structures */ /** Feature generation related variables*/ fe_t *fe; /**< Front-end processing object */ mfcc_t **mfcc; /**< Input feature frames (may be NULL) */ float32 ***feat; /**< Dynamic feature frames */ /** Structures of storing parameters for different techniques. */ ascr_t *ascr; /**< Senone and composite senone scores for one frame. */ beam_t *beam; /**< Structure that wraps up parameters related to beam pruning. */ fast_gmm_t *fastgmm; /**< Structure that wraps up parameters for fast GMM computation. */ pl_t *pl; /**< Structure that wraps up parameters for phoneme look-ahead. */ /** Structure that wraps up adaptation variables. such as regression matrices in MLLR */ adapt_am_t * adapt_am; /**< Structure that wraps up parameters for adaptation such as MLLR. */ stat_t *stat; /**< Structure of statistics including timers and counters. */ /** FILE handle that handles output. */ FILE *matchfp; /**< File handle for the match file */ FILE *matchsegfp; /**< File handle for the match segmentation file */ FILE *hmmdumpfp; /**< File handle for dumping hmms for debugging */ /* The only variable I intend to make it be alone in the whole structure. It has its own uniqueness. (dhuggins: I have no idea what that means) */ int32 op_mode; /** A mode for specifying operation */ char *uttid; /**< Utterance ID. The one thing that should move to somewhere like srch */ char *uttfile; /**< Utterance Filename. The other thing that should move somewhere else. */ void *srch; /**< The search structure */ } kb_t; /** Initialize the kb */ S3DECODER_EXPORT void kb_init (kb_t *kb /**< In/Out: A empty kb_t */ ); /** Initialize the kb (re-entrant version) */ S3DECODER_EXPORT void kb_init_r(kb_t *kb, /**< In/Out: A empty kb_t */ cmd_ln_t *config /**< In: The command-line object */ ); /** Deallocate the kb structure */ S3DECODER_EXPORT void kb_free (kb_t *kb ); /* RAH 4.16.01 */ /** Set MLLR */ S3DECODER_EXPORT void kb_setmllr(char* mllrname, /**< In: The name of the mllr model */ char* cb2mllrname, /**< In: The filename of the MLLR class map */ kb_t *kb ); /* ARCHAN 20040724 */ /** Set Utterance ID to uttid */ S3DECODER_EXPORT void kb_set_uttid(char *uttid, /**< In: the new utt id */ char *uttfile, /**< In: the new utt filename */ kb_t *kb ); /* ARCHAN 20041111 */ #if 0 { /* Stop indent from complaining */ #endif #ifdef __cplusplus } #endif #endif