[hts-users:01071] Re: about duration model streams

["Heiga ZEN (Byung Ha CHUN)" <zen@xxxxxxxxxxxxxxx>]

Hi,

gues wrote (2007/12/21 18:23):

> We can see the following code in the demo script.
>    for ($i=2;$i<=$nstate{$t2s{$type}}+1;$i++){
>       print EDFILE "TB $thr{$type} ${type}_s${i}_ 
> {*.state[${i}].stream[$strb{$type}-$stre{$type}]}\n";
>    }
>  
> but when $type = "dur", $strb{$type}==1 and $stre{$type}==5, as in the 
> 'Config.pm.in <http://Config.pm.in>' file,
> %stre = ('mgc' => '1',  # stream end
>          'lf0' => '@F0STREN@',
>          'dur' => '@NSTATE@');
>  
> So I think this is a bug, for duration models stream is 1.

No, it's not a bug.

In the previous versions we used a multi-variate Gaussian PDF to represent state duration PDFs.
However, since version 2.0.1 we have changed it to multi-stream structure.
You can find this notation in README file and on the HTS website.

Therefore, # of streams should be equal to the number of states for state duration models.

Regards,

Heiga ZEN (Byung Ha CHUN)

--
------------------------------------------------
Heiga ZEN     (in Japanese pronunciation)
Byung Ha CHUN (in Korean pronunciation)

Department of Computer Science and Engineering
Nagoya Institute of Technology
Gokiso-cho, Showa-ku, Nagoya 466-8555 Japan

http://www.sp.nitech.ac.jp/~zen
------------------------------------------------