[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

[hts-users:04221] [spam] Re: Re: Help require to generate speech using forced-alignment label of natural speech



If this email is not spam, click here to submit the signatures to FortiGuard - AntiSpam Service.
  
Hello,

if I understand clearly, you want to synthesize speech aligned with a natural (original) signal. Am I right ?

If it's the case, Dr. Keiichiro Oura proposed to you to use HSMMAlign (with option -f) and HMGenS (with the options -s). The first command is to get duration of each state into a label file. Then you will use this file as the input of HMGenS and specified the option "-s" to indicate the use of the state duration defined in the label file.

The use of the commands HSMMAlign is illustrated in the demo file (look stage FALGN, around line 570 in 2.3alpha demo (sorry I don't have the beta demo yet)) and for HMGens (look stage PGEN1, around line  650).

Kind regards,
Sébastien

Nirmesh Shah writes:

> Hi,
>
> It was not failed. I tried ./hts_engine command for that purpose. which was
> mentioned in previous mail.
> In one of the earlier thread it was mentioned that you can do it with HSMMAlign
> -f option but I don't know how to run this command as I m new in HTS.
> So, it would be nice if someone can help us in detail for that command.
>
> http://hts.sp.nitech.ac.jp/hts-users/spool/2013/msg00108.html
>
> Thanks and Regards,
> Nirmesh
>
>
> On Mon, Feb 2, 2015 at 10:55 PM, Sébastien Le Maguer <
> slemaguer@coli.uni-saarland.de> wrote:
>
>     Hello,
>
>     if even HSMMAlign failed, it means that you're models are corrupted.
>     Therefore, you should try to solve this problem first.
>
>     Kind regards,
>     Sébastien
>
>     Nirmesh Shah writes:
>
>     > Hi,
>     >
>     > We have train HTS system.
>     > Now in order to generate wav file with given label file, fololwing
>     command has
>     > been used but at the end we get output wav file contains only silence.
>     >
>     >
>     > hts_engine -s 48000 -p 240 -a 0.55 -td tree-dur.inf \
>     > -tm tree-mgc.inf -tf tree-lf0.inf -tl tree-lpf.inf \
>     > -md dur.pdf -mm mgc.pdf -mf lf0.pdf -ml lpf.pdf \
>     > -dm mgc.win1 -dm mgc.win2 -dm mgc.win3 -df lf0.win1 \
>     > -df lf0.win2 -df lf0.win3 -dl lpf.win1 -em tree-gv-mgc.inf \
>     > -ef tree-gv-lf0.inf -cm gv-mgc.pdf -cf gv-lf0.pdf \
>     > -k gv-switch.inf -ow output.wav fsew0_001.lab
>     >
>     > Can anyone please suggest some help?
>     >
>     > PS: In earlier thread, we have found HSMMAlign and HMGenS can be used but
>     we
>     > are unable to run in our end.
>
>     --
>     Dr. Sébastien Le Maguer
>     Postdoctorate researcher
>
>     Saarland University
>     Campus C7.4 - room 2.03
>     D-66123 Saarbrücken
>     Germany
>
>     phone : +49-681-302-70030
>     Mail: slemaguer@coli.uni-saarland.de
>     website : http://sites.google.com/site/seblemaguer
>
>
>
>
>
> --

-- 
Dr. Sébastien Le Maguer
Postdoctorate researcher

Saarland University
Campus C7.4 - room 2.03
D-66123 Saarbrücken
Germany

phone : +49-681-302-70030
Mail: slemaguer@coli.uni-saarland.de
website : http://sites.google.com/site/seblemaguer