6, before Zhijun 40 sentences Corpus synthesis experiment, obtains the time-length HMM model, replaces 40 sentence adaptive time-length model

Experimental Purpose:

1. Miss Wu thought Hw2000-adapt-zhijun40-state10, the experimental results of the speed of the sound is faster than the Zhijun 40 sentences, so like with the time-length model of the Zhijun 40 sentences, and then the adaptive spectrum, the basic frequency parameters, to obtain a result, How about a look?

Experimental process:

2. Create a new path:Experimnet/zhijun40-syn
4. Labels (all copied from the hw2000-adapt-zhijun40-state10/data/labels/path)　　　　　　
   1. Full
   2. Mono
   3. Gen
6. Questions
8. Raw
10. Data/makefile

1. Prjdir =/home/yuqj/experimnet/zhijun40-syn

2. Mgcorder = 34, changed to 35 order, consistent with Hw2000-adapt-zhijun40-state

12. scripts/config.pm

    1. %ordr = (' MGC ' = ' + ')

    2. %band = (' MGC ' = ' + ')

    3. $prjdir = '/home/yuqj/experimnet/zhijun40-syn ';

    4. M

Experimental results:

6, before the Zhijun 40 sentence Corpus synthesis experiment, obtains the time-length HMM model, replaces 40 sentence adaptive time-length model