[hts-users:00137] demo script problems & using festival voices in hts
- Subject: [hts-users:00137] demo script problems & using festival voices in hts
- From: Tomas Sykora <tomx@xxxxxx>
- Date: Wed, 05 Oct 2005 00:40:45 +0200
- User-agent: Mozilla Thunderbird 1.0.6 (X11/20050716)
problems with the demo-script.
I've installed hts following the readme file which can be found in
http://hts.ics.nitech.ac.jp/README.html.
The configuration was done step by step using this readme, with exactly
the same versions of software.
Even though the demo script Training_cmu_us_arctic_awb.pl after hours
finished successfully (also with gdb and strace) there are *no final
synthetized speech samples*.
I've set datadir to HTS-demo_CMU-ARCTIC-AWB and projectdir to HTS-demo.
In the datadir can be found trained hidden markov models.
Pit and lf0 files are missing.
In the gen_wave function is one command commented out by default:
#."sox -t raw -r $samprate - -t wav -r $samprate $gendir/${base}.wav";
No idea what is wrong. If I have the same versions of software and was
strictly following the readme file, the only problem could be in
settings in makefiles which were not listed in the readme file
(projectdir vs datadir?).
Festival is up and running. HTK test script finished successfully.
I'm running Gentoo Linux with 2.6.9 kernel, gcc version 3.3.5,
glibc-2.3.5, perl-5.8.6.
To speed up the troubleshooting process I would appreciate sample
Makefiles of HTK, HTS-demo_CMU-ARTIC-AWB and HTS-demo.
Another question is - *will HTS work with another festival voice?* And
in another language?
Or there are any dependencies?
Thanks and greetings
Tomas Sykora
- Follow-Ups
-
- [hts-users:00138] Re: demo script problems & using festival voices in hts, Heiga ZEN (Byung Ha CHUN)