[hts-users:00008] Re: japanese demo

[Heiga ZEN <zen@xxxxxxxxxxxxxxxx>]

Hi, 

On Thu, 22 Jan 2004 15:42:01 +0900
Nobuaki MINEMATSU <mine@xxxxxxxxxxxxxxxxxxxx> wrote:

> I found that there is acoustic difference between m001 we made and m001
> you (NIT people) made. The m001 we made sounds more like buzzer.

Because m001 voice on gtalk demo page is post-filtered (formant-emphasized).
In HTS-demo script, post-filtering technique is not used for waveform generation.
However, you can use it by "hts_engine" instead of "HMGenS" when synthesizing speech from HMMs.

"hts_engine" is invoked by typing the command line 

hts_engine -dm mcep_dyn.win   \
           -dm mcep_acc.win   \
           -df lf0_dyn.win    \
           -df lf0_acc.win    \
           -td tree-dur.inf   \
           -tm tree-mcep.inf  \
           -tf tree-lf0.inf   \
           -md duration.pdf   \
           -mm mcep.pdf       \
           -mf lf0.pdf        \
           -b  0.3            \ 
               input.lab

"{mcep,lf0}_{dyn,acc}.win" are included in the distribution.
"tree-{mcep,lf0,dur}.inf" and "{mcep,lf0,duration}.pdf" are generated by HTS-demo script at HTS-demo/voices.

Please try it.

Best regards

Heiga Zen

-- 
 ------------------------------------------------
  Heiga Zen
  Department of Computer Science and Engineering
  Graduate School of Engineering
  Nagoya Institute of Technology
  Japan

  e-mail: zen@xxxxxxxxxxxxxxxx
     web: http://kt-lab.ics.nitech.ac.jp/~zen
 ------------------------------------------------