[hts-users:02881] Re: stream weight and aperiodicity

[Sébastien Le Maguer <Sebastien.Le_maguer@xxxxxxxx>]

Hello,

thank you for the answer.

Kind regards,
Sébastien

On 07/06/2011 17:54, Keiichiro Oura wrote:
> Hi,
>
> I guess that the forward-backward alignments should depend only on the
> spectral and F0 features.
> Therefore, the stream weight for the aperiodic measure was set to zero
> during HMM training in demo scripts.
>
> Regards,
> Keiichiro Oura
>
>
> 2011/6/7 Sébastien Le Maguer<Sebastien.Le_maguer@xxxxxxxx>:
> > Hello,
> >
> > while I was checking my config file i've found that the stream weight of the
> > aperiodicity was 0. Is there a specific reason?
> >
> > Regards,
> > Sébastien