[hts-users:03375] CMU ARCTIC voices (other than SLT) for flite_hts_engine

[Jason Livingston <jettoblack@xxxxxxxxx>]

Hello,
I'm using flite_hts_engine 1.03 and it works well with the pre-built
SLT voice files downloaded from:
http://downloads.sourceforge.net/hts-engine/hts_voice_cmu_us_arctic_slt-1.04.tar.gz

I couldn't find the other CMU ARCTIC voices like BDL, RMS, etc.
pre-built for flite_hts_engine.  If someone has these available,
please let me know.

I tried to build them using HTS-demo_CMU-ARCTIC-ADAPT.  I installed
the required versions of each software package as listed in the
INSTALL file.

Here is the config I used for RMS:
./configure --with-tcl-search-path=/opt/ActiveTcl-8.4/bin
--with-hts-search-path=/usr/local/HTS-2.2beta/bin
--with-fest-search-path=/usr/share/festival/examples
--with-sptk-search-path=/usr/local/SPTK/bin
--with-hts-engine-search-path=/usr/local/bin ADAPTSPKR=rms
TRAINSPKR='bdl clb jmk slt' F0_RANGES='bdl 40 210 clb 130 260 jmk 50
180 slt 110 280 rms 40 200'

After running ./configure, I had to edit the scripts/Config.pm file and change:
$SOXOPTION = 'w';
to:
$SOXOPTION = '2';

This is because recent versions of sox deprecated the -w option and
changed it to '-2' or '-b 16' for 16-bit audio samples.  However I
don't know if the resulting output from the newer sox is different
from the old sox with the -w option.

After running 2-3 days, the voice files were built successfully.
However when I used them in flite_hts_engine, the resulting speech is
very high pitch and fast.

You can download the resulting model here:
https://dl.dropbox.com/u/31668283/rms.zip

The flite_hts_engine command I used is:
echo "Hello this is a test." | ./flite_hts_engine  -s 48000 -p 240 -a
0.55 -td rms/tree-dur.inf -tm rms/tree-mgc.inf -tf rms/tree-lf0.inf
-md rms/dur.pdf -mm rms/mgc.pdf -mf rms/lf0.pdf -dm rms/mgc.win1 -dm
rms/mgc.win2 -dm rms/mgc.win3 -df rms/lf0.win1 -df rms/lf0.win2 -df
rms/lf0.win3 -em rms/tree-gv-mgc.inf -ef rms/tree-gv-lf0.inf -cm
rms/gv-mgc.pdf -cf rms/gv-lf0.pdf -k rms/gv-switch.inf -o "test.wav"

This is the same command that works fine with the pre-built SLT voice.

Did I make a mistake in my configure or flite_hts_engine options?
Does anyone have any suggestions?

Thanks,
Jason