15, the IBM-ADPAT-HELEN_USEREG-STATE15 experiment is not multiplied by the conversion matrix MMF file, using Hts_engine mgc.pdf and lf0.pdf to replace, get multiplied by the conversion matrix MMF file

2. In the ibm-adpat-helen_usereg-state15/voices/... Mgc.pdf and Lf0.pdf files are generated under the path (all model files multiplied by the transformation matrix)
   1. In the scripts/config.pm file, open the $yu_hhed switch and re-make
   2. in the ibm-adpat-helen_usereg-state15/voice/... Mgc.pdf and Lf0.pdf are generated under the path, the idea is binary
   3. go to the scripts/config.pm file, use another path to hhed, $HHED = '/home/ningys/hts_install/yu-hts-2.2-for-not-binary-hts_engine-modelsfile/bin/HHED ';
   4. Re-make to get non binary mgc.pdf and lf0.pdf
4. Enter Path converthts_engeineformat2mmfformat/cmp/, then copy the Mgc.pdf and lf0.pdf below the voice/path .
6. Modify converthts_engeineformat2mmfformat/cmp/Path below the 1_hangle-mgc.pl, mainly the state number has changed, there are 5 states changed to 15 state. 2 Places to change
   1. Need to increase, $num _mgc_s7 to $NUM_MGC_S16
   2. Modifying the value of an @arrayleafnodes array
8. Execute 1_handle-mgc.pl, generate Notbinarymgc.pdf file
10. from /models/qst001/ver1/cmp/mapmmf/Path under copy re_clusterer_all.mmf to current path
12. Execution 2_handle_mmf-mgc.pl,
    1. There is a problem that cannot match two-digit numbers, such as mgc_s11_1, which cannot be matched.
    2. Modify the regular expression to:
14. Get, AFTERMGC_RE_CLUSTERER_SAT_ALL.MMF file (the MMF file, already on the MGC parameter, multiplied by the adaptive conversion matrix)
16. put Experimnet/ibm-adpat-helen_usereg-state15/converthts_engeineformat2mmfformat/cmp/aftermgc_re_clustered_sat Copy the _all.mmf file to the 2014-04/handlemmffileforstatemapping/can/mgc/path
18. Enter the 2014-04/handlemmffileforstatemapping/can/mgc/path, processing the aftermgc_re_clusterer_sat_all.mmf file, get can_mgc_ S2.txt to Can_mgc_s16.txt file
    1. Modify the handle-mmf.pl file
       1. %hash_state Array
       2. Upper limit for loop
20. Copy the 2015-04/handlemmf.../can/mgc/can_mgc_s2.txt file to the/matlab/adaptcan15engstate-2015-04-16/path
22. Then, execute the ADAPT_CAL_KLD_MGC_S2.M code
    1. Note Modify, ADAPT_CAL_KLD_MGC_S2.M code, read the path of the file
    2. Path to write files
24. Get results:

15, the IBM-ADPAT-HELEN_USEREG-STATE15 experiment is not multiplied by the conversion matrix MMF file, using Hts_engine mgc.pdf and lf0.pdf to replace, get multiplied by the conversion matrix MMF file