[hts-users:00283] Re: about dynamic f0
- Subject: [hts-users:00283] Re: about dynamic f0
- From: "Heiga ZEN (Byung Ha CHUN)" <zen@xxxxxxxxxxxxxxxx>
- Date: Wed, 26 Apr 2006 02:35:10 +0900
Hi,
Nicholas Volk wrote:
$ egrep -R "svec" hts-engine/ SPTK-3.0/
through my hts_engine and SPTK-3.0 files and didn't get
any hits, so I'm a bit puzzled about the origin of
following piece of code:
for (s=1;s<=f0pst.vSize;s++) {
f0pst.sm.mseq[pt][s] =
hmm->svec[j].info->pdf[s+1].info->spdf.cpdf[1].mpdf->mean[1];
f0pst.sm.ivseq[pt][s] = (bound && s!=1) ? 0.0 :
hmm->svec[j].info->pdf[s+1].info->spdf.cpdf[1].mpdf->cov.var[1];
}
Oops, sorry, it's my mistake.
They are from HMGenS (HTS-1.1.2).
In hts_engine, the following part corresponds to the above statements:
/* copy pdfs */
if (voiced[mcepframe]) {
lf0pst->sm.mseq[lf0frame][k+1] = m->lf0mean[state][k+1];
if (nobound || k==0)
lf0pst->sm.ivseq[lf0frame][k+1] = finv(m->lf0variance[state][k+1]);
else /* the variances for dynamic feature are set to inf on v/uv boundary */
lf0pst->sm.ivseq[lf0frame][k+1] = 0.0;
}
The above statements are included in mlpg.c.
Heiga: You did not publish a new hts engine (nor new SPTK version)
with 2.0 release candidates
and I'm sort of guessing that this code fragment might be from it.
(I'm just randomly guessing here.)
For some reasons we did not include hts_engine in the HTS-2.0RC1.
In the next or final release hts_engine will be included.
Best 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://kt-lab.ics.nitech.ac.jp/~zen
------------------------------------------------
- References
-
- [hts-users:00279] Re: about dynamic f0, 刘 磊
- [hts-users:00280] Re: about dynamic f0, Heiga ZEN (Byung Ha CHUN)
- [hts-users:00282] Re: about dynamic f0, Nicholas Volk