[hts-users:01620] Re: GV causes clipping

[tomoki@xxxxxxxxxxx]

Hi,

You can find several samples in the following page:

 http://spalab.naist.jp/~tomoki/IEICE/HTS+GV/

Sorry for only the samples using mel-cep based on
STRAIGHT spectrum. More details are described in
the paper (IEICE Trans. Vol. E90-D, No. 5, pp. 816-824).

Thanks,
Tomoki Toda
   Nara Institute of Science and Technology
   E-mail: tomoki@xxxxxxxxxxx
   TEL: +81-743-72-5282
   FAX: +81-743-72-5289

: Leon Zhang <leonzh@xxxxxxxxxxxxx>
: [hts-users:01618] Re: GV causes clipping

> Thanks for your information. I read the paper of GV and found it 
> is promising for me, could you please tell me where I can download 
> the sample wave pair to demonstrate the effectiveness of GV. Both 
> lsp and mel-cepstral are welcomed.
> 
> Thanks
> Leon
> 
> -----Original Message-----
> From: Heiga ZEN (Byung Ha CHUN) [mailto:zen@xxxxxxxxxxxxxxx]
> Sent: Thursday, July 31, 2008 2:59 PM
> To: hts-users@xxxxxxxxxxxxxxx
> Subject: [hts-users:01603] Re: GV causes clipping
> 
> Hi,
> 
> Leon Zhang wrote (2008/07/31 7:32):
> 
> > I am interesting at the voice quality improvement by GV,
> > who can tell me your observation? I am using lsp to build HTS 
> system,> is GV able to improve voice quality too?
> 
> GV is effective for LSPs, but usually its improvement is 
> relatively smaller than
>  that with mel-cepstral coefficients.  Furthermore, the use of GV 
> with LSPs
> sometimes causes disorders of LSPs, it means that the synthesis 
> filter becomes
> unstable.  Rearranging LSPs by post-processing can avoid it, but 
> it introduces
> artifacts in synthesized speech.
> 
> Best regards,
> 
> Heiga ZEN (Byung Ha CHUN)
> 
> --
> ------------------------------------------------
>  Heiga ZEN     (in Japanese pronunciation)
>  Byung Ha CHUN (in Korean pronunciation)
> ------------------------------------------------
> 
> 
> 
>