The script used in the TUTORIAL 
compute_mfcc.csh - a thin wrapper of wave2feat, current converter actually support batch processing. 
   -> wave2feat  

01.ci-hmm/slave_convg.csh -Can also be used in training semi-continous model.
   -> flat_initialize.csh
	->mk_mdef_gen
	->mk_flat
        ->maketopology.csh
  	->init_gau
	->norm.mop64
	->cp_parm
   -> baum_welch.csh
	->bw.mop64

02.cd_untied/slave_convg.sh 
   -> initialized_untied.csh
      -> init_mixw
   -> baum_welch.csh
      -> 
   -> norm_and_launchbw.csh 

03.buildtree/slave_buildtrees.sh
   -> make_questions.csh
	->make_quest
   -> build_tree.csh
	->bldtree

04.tiestate/slave_tiestate.csh
   -> prunetree.csh
	->prunetree
   -> tiestate.csh
	->mk_mdef_gen
	->tiestate

05.cd-chmm/slave_convg.sh
   ->initialize.cd.csh
   ->baum_welch.csh
   ->norm_and_launchbw.csh
	-> split_gaussians.csh


wave2feat - a feature converter that seems to work on only MFCC_D_DD_POW
(can be found in wave2feat/new_fe_wrapper.c)
->main
  -> fe_parse_options
  -> fe_convert_files 
     -> fe_build_filenames
     -> fe_openfiles
     -> fe_readblock_spch
     -> fe_process_utt
     -> fe_writeblock_feat

mk_mdef_gen 
  -> main (mk_mdef_gen/main.c)
    -> parse_cmd_ln
    -> make_ci_list_cd_hash_frm_mdef
	->yank_top
	->free_heap_element
	->acmod_set_new
	->acmod_set_set_n_ci_hint
	->acmod_set_set_n_tri_hint
	->acmod_set_add_ci
	->acmod_set_name2id
	->posnstr2wordpos
	->acmod_set_add_tri
	->output_model_def (in the same file)
    -> make_ci_list_Cd_frm_phnlist
    -> make_ci_list_frm_mdef
    -> make_mdef_from_list
    -> read_dict
    -> make_dict_tripone_list
    -> make_CD_heap
    -> make_mdef_from_list
    -> count_triphones
    -> print_Counts

init_gau
->main
  ->initialize
  ->init_gau (in init_gau.c)
    -> feat_vecsize
    -> gauden_alloc_param
    -> s3gau_read
    -> corpus_next_utt
    -> corpus_get_sent
    -> corpus_get_seg
    -> mk_wordlist
    -> mk_phonelist
    -> ck_seg
    -> mk_sseq
    -> cvt2tripohne
    -> mk_sseq
    -> corpus_get_mfcc
    -> corpus_get_generic_featurevec
    -> corpus_utt_brief_name
    -> feature_compute
    -> corpus_get_sildel
    -> accum_state_mean
    -> accum_state_var
    -> s3gaucnt_write

norm.mpp64?
    -> initialize
    -> normalize
    ->
	
mk_flat
->main
  -> topo_read
  -> s3tmat_write
  -> s3mixw_write

cp_parm
->main
  ->feat_set
  ->rd_param
      ->rd_mixw
      ->rd_gau
      ->rd_tmat
  ->cp_param
      ->cp_mixw
      ->cp_gau
      ->cp_tmat
  ->wr_param
      ->wr_mixw
      ->wr_gau
	->s3gau_write
      ->wr_tmat

bw
main
->prefetch_init
->main_initialize
  -> train_cmd_ln_parse
  -> mod_inv_set_n_feat
  -> model_def_read
->main_reestimate
  -> feat_set_in_veclen
  -> corpus_get_begin
  -> corpus_next_utt
  -> corpus_get_mfcc
  -> corpus_get_generic_featurevec
  -> corpus_get_sildel
  -> silcomp_set_del_seg
  -> feat_compute
  -> corpus_get_sent
  -> next_utt_states
  -> baum_welch_update
  -> corpus_get_seg
  -> mk_wordlist
  -> mk_phonelist
  -> cvt2triphone
  -> mk_sseq
  -> mk_trans_seq
  -> viterbi_update
  -> accum_dump
 
make_quest

bld_tree

prune_tree

tie_state