/* -*- 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. * * ==================================================================== * */ /* gmm_wrap.h * HISTORY * $Log$ * Revision 1.1 2006/04/05 20:27:30 dhdfu * A Great Reorganzation of header files and executables * * Revision 1.3 2006/02/23 05:38:39 arthchan2003 * Merged from the branch SPHINX3_5_2_RCI_IRII_BRANCH: Added multi-stream GMM computation routine. * * * Revision 1.1.4.3 2005/08/02 21:31:35 arthchan2003 * Added interface for 1, doing multi stream gmm computation with/without composite senone. 2, doing gmm computation (ms or ss optimized) with/wihout composite senone. Haven't tested on the SCHMM on s3.x yet. I think it will work though. * * Revision 1.1.4.2 2005/07/24 01:35:41 arthchan2003 * Add a wrapper for computing senone score without computing composite senone score. Mainly used in mode FSG now * * Revision 1.1.4.1 2005/06/27 05:30:25 arthchan2003 * Merge from the tip of the trunk * * Revision 1.2 2005/06/22 08:00:09 arthchan2003 * Completed all doxygen documentation on file description for libs3decoder/libutil/libs3audio and programs. * * Revision 1.1 2005/06/21 22:48:14 arthchan2003 * A wrapper that provide the function pointer interface of approx_cont_mgau_ci_eval and approx_cont_mgau_frame_eval. They are used in srch_gmm_compute_lv1 and srch_gmm_compute_lv2 respectively. This will also be the home of other gmm computation routine. (Say the s3.0 version of GMM computation) * * Revision 1.2 2005/06/13 04:02:59 archan * Fixed most doxygen-style documentation under libs3decoder. * * Revision 1.1 2005/04/22 04:22:37 archan * Add gmm_wrap, this will share code across op_mode 4 and op_mode 5. Also it also separate active senone selection into a different process. I hope this is the final step before making the WST search works. At the current stage, the code of mode-5 looks very much alike mode-4. This is intended because in Prototype 4, tail sharing will be used to reduce memory. * */ /** \file gmm_wrap.h * \brief Function pointer wrapper of GMM computation. */ #include #include "approx_cont_mgau.h" #ifdef __cplusplus extern "C" { #endif #if 0 /* Fool Emacs. */ } #endif int32 approx_ci_gmm_compute(void *srch, /**< a pointer to a srch_t */ float32 *feat, /**< feature vector */ int32 cache_idx, /**< cache index */ int32 wav_idx /**< frame index */ ); /** This wrapper calls the approximate GMM computation routine which compute the senone score. Then the composite senone will also be computed. */ int32 approx_cd_gmm_compute_sen_comp( void *srch, /**< a pointer to a srch_t */ float32 **feat, /**< feature vector */ int32 wav_idx /**< frame index */ ); /** This wrapper that calls the approximate GMM computation routine which compute only normal senone. */ int32 approx_cd_gmm_compute_sen( void *srch, /**< a pointer to a srch_t */ float32 **feat, /**< feature vector #stream x #coeff*/ int32 wav_idx /**< frame index */ ); /** This wrapper calls the multi-stream exact GMM computation routine which compute the senone score. Then the composite senone will also be computed. */ int32 ms_cd_gmm_compute_sen_comp( void *srch, /**< a pointer to a srch_t */ float32 **feat, /**< feature vector #stream x #coeff*/ int32 wav_idx /**< frame index */ ); /** This wrapper that calls the approximate GMM computation routine which compute only normal senone. */ int32 ms_cd_gmm_compute_sen( void *srch, /**< a pointer to a srch_t */ float32 **feat, /**< feature vector #stream x #coeff*/ int32 wav_idx /**< frame index */ ); /** Depends on which data structure has been initialized, s3_cd_gmm_compute_sen_comp calls ms_cd_gmm_compute_sen_comp (if ms_mgau is initialized) or approx_cd_gmm_coupute_sen_compu (if _mgau is initialized) Then composite triphone will also be computed. */ int32 s3_cd_gmm_compute_sen_comp( void *srch, /**< a pointer to a srch_t */ float32 **feat, /**< feature vector #stream x #coeff*/ int32 wav_idx /**< frame index */ ); /** Depends on which data structure has been initialized, s3_cd_gmm_compute_sen calls ms_cd_gmm_compute_sen (if ms_mgau is initialized) or approx_cd_gmm_coupute_sen (if _mgau is initialized) Only normal senone will be computed. */ int32 s3_cd_gmm_compute_sen( void *srch, /**< a pointer to a srch_t */ float32 **feat, /**< feature vector #stream x #coeff*/ int32 wav_idx /**< frame index */ ); #ifdef __cplusplus } #endif