Hi,
Thanks very much
for your help :)
I am not sure about how to use the trained model and label
file. Does this like what we do in htk, using 'HVite' to do the
recognition. In hts, we use some else tools, 'HMGenS'?
Yes. Besides ,you can choose
different method of parameter generation algorithms.
Could you please give me some examples if you can
or do you know where I can get information about using the trained model and your own label
files?
This model is a English
version. If you want to try other languages, you
should use other question set to train a different model.
BTW, you can turn to the mail
on 8th November, 2009 from jangwon kim. He described his way of generating labal
files.
Many thanks Xu
2009/11/11 那兴宇 <nxy-yzqs@xxxxxxx>
Hi,
Thanks very much for the reply. I used
the SPTK 3.1 instead of SPTK 3.2 and changed the 'sox' option '-w' to '-4' in
the 'Training .pl' file. Then I trained it agian and the resulting 'log' file
seems correct.
I paste the last few lines of the 'log' file below.
Please check it. Also, there are '1mix', '2mix', 'hts_engine', and 'stc'
folders in the 'gen/pst001/ver1' directory.
If the training is
all right, how to use the traning result to do text-to-speech? Thanks a
lot.
Congratulations on your successful training.
You can see that the synthesizing operation has already been done, based on
the given label file.
You can do your own TTS
using the trained model and your own label files.
====================================================================== Start
synthesizing waveforms (2mix) at Tue Nov 10 19:28:39 NZDT
2009 ======================================================================
Processing
directory
/home/xxie024/Desktop/hts/HTS-demo_CMU-ARCTIC-SLT/gen/qst001/ver1/2mix/0: Synthesizing
a speech waveform from alice01.mgc and
alice01.lf0...done Synthesizing a speech waveform from alice02.mgc
and alice02.lf0...done Synthesizing a speech waveform from
alice03.mgc and alice03.lf0...done Synthesizing a speech waveform
from alice04.mgc and alice04.lf0...done Synthesizing a speech waveform from alice05.mgc and
alice05.lf0...done Synthesizing a speech waveform from
alice06.mgc and alice06.lf0...done Synthesizing a speech waveform
from alice07.mgc and alice07.lf0...done Synthesizing a speech
waveform from alice08.mgc and alice08.lf0...done Synthesizing a
speech waveform from alice09.mgc and alice09.lf0...done Synthesizing
a speech waveform from alice10.mgc and
alice10.lf0...done Synthesizing a speech waveform from alice11.mgc
and alice11.lf0...done Synthesizing a speech waveform from
alice12.mgc and alice12.lf0...done Synthesizing a speech waveform
from alice13.mgc and alice13.lf0...done Synthesizing a speech
waveform from alice14.mgc and alice14.lf0...done Synthesizing a
speech waveform from alice15.mgc and alice15.lf0...done Synthesizing
a speech waveform from alice16.mgc and
alice16.lf0...done Synthesizing a speech waveform from alice17.mgc
and alice17.lf0...done Synthesizing a speech waveform from
alice18.mgc and alice18.lf0...done Synthesizing a speech waveform
from alice19.mgc and alice19.lf0...done Synthesizing a speech
waveform from alice20.mgc and alice20.lf0...done Synthesizing a
speech waveform from alice21.mgc and alice21.lf0...done Synthesizing
a speech waveform from alice22.mgc and
alice22.lf0...done Synthesizing a speech waveform from alice23.mgc
and alice23.lf0...done Synthesizing a speech waveform from
alice24.mgc and alice24.lf0...done Synthesizing a speech waveform
from alice25.mgc and alice25.lf0...done Synthesizing a speech
waveform from alice26.mgc and alice26.lf0...done Synthesizing a
speech waveform from alice27.mgc and alice27.lf0...done Synthesizing
a speech waveform from alice28.mgc and
alice28.lf0...done Synthesizing a speech waveform from alice29.mgc
and alice29.lf0...done Synthesizing a speech waveform from
alice30.mgc and alice30.lf0...done Synthesizing a speech waveform
from alice31.mgc and alice31.lf0...done Synthesizing a speech
waveform from alice32.mgc and alice32.lf0...done Synthesizing a
speech waveform from alice33.mgc and alice33.lf0...done Synthesizing
a speech waveform from alice34.mgc and
alice34.lf0...done Synthesizing a speech waveform from alice35.mgc
and alice35.lf0...done Synthesizing a speech waveform from
alice36.mgc and alice36.lf0...done Synthesizing a speech waveform
from alice37.mgc and alice37.lf0...done Synthesizing a speech
waveform from alice38.mgc and alice38.lf0...done Synthesizing a
speech waveform from alice39.mgc and alice39.lf0...done Synthesizing
a speech waveform from alice40.mgc and
alice40.lf0...done done
Processing directory
/home/xxie024/Desktop/hts/HTS-demo_CMU-ARCTIC-SLT/gen/qst001/ver1/2mix/1:
Synthesizing a speech waveform from alice01.mgc and
alice01.lf0...done Synthesizing a speech waveform from alice02.mgc
and alice02.lf0...done Synthesizing a speech waveform from
alice03.mgc and alice03.lf0...done Synthesizing a speech waveform
from alice04.mgc and alice04.lf0...done Synthesizing a speech
waveform from alice05.mgc and alice05.lf0...done Synthesizing a
speech waveform from alice06.mgc and alice06.lf0...done Synthesizing
a speech waveform from alice07.mgc and
alice07.lf0...done Synthesizing a speech waveform from alice08.mgc
and alice08.lf0...done Synthesizing a speech waveform from
alice09.mgc and alice09.lf0...done Synthesizing a speech waveform
from alice10.mgc and alice10.lf0...done Synthesizing a speech
waveform from alice11.mgc and alice11.lf0...done Synthesizing a
speech waveform from alice12.mgc and alice12.lf0...done Synthesizing
a speech waveform from alice13.mgc and
alice13.lf0...done Synthesizing a speech waveform from alice14.mgc
and alice14.lf0...done Synthesizing a speech waveform from
alice15.mgc and alice15.lf0...done Synthesizing a speech waveform
from alice16.mgc and alice16.lf0...done Synthesizing a speech
waveform from alice17.mgc and alice17.lf0...done Synthesizing a
speech waveform from alice18.mgc and alice18.lf0...done Synthesizing
a speech waveform from alice19.mgc and
alice19.lf0...done Synthesizing a speech waveform from alice20.mgc
and alice20.lf0...done Synthesizing a speech waveform from
alice21.mgc and alice21.lf0...done Synthesizing a speech waveform
from alice22.mgc and alice22.lf0...done Synthesizing a speech
waveform from alice23.mgc and alice23.lf0...done Synthesizing a
speech waveform from alice24.mgc and alice24.lf0...done Synthesizing
a speech waveform from alice25.mgc and
alice25.lf0...done Synthesizing a speech waveform from alice26.mgc
and alice26.lf0...done Synthesizing a speech waveform from
alice27.mgc and alice27.lf0...done Synthesizing a speech waveform
from alice28.mgc and alice28.lf0...done Synthesizing a speech
waveform from alice29.mgc and alice29.lf0...done Synthesizing a
speech waveform from alice30.mgc and alice30.lf0...done Synthesizing
a speech waveform from alice31.mgc and
alice31.lf0...done Synthesizing a speech waveform from alice32.mgc
and alice32.lf0...done Synthesizing a speech waveform from
alice33.mgc and alice33.lf0...done Synthesizing a speech waveform
from alice34.mgc and alice34.lf0...done Synthesizing a speech
waveform from alice35.mgc and alice35.lf0...done Synthesizing a
speech waveform from alice36.mgc and alice36.lf0...done Synthesizing
a speech waveform from alice37.mgc and
alice37.lf0...done Synthesizing a speech waveform from alice38.mgc
and alice38.lf0...done Synthesizing a speech waveform from
alice39.mgc and alice39.lf0...done Synthesizing a speech waveform
from alice40.mgc and alice40.lf0...done done
Processing
directory
/home/xxie024/Desktop/hts/HTS-demo_CMU-ARCTIC-SLT/gen/qst001/ver1/2mix/2:
Synthesizing a speech waveform from alice01.mgc and
alice01.lf0...done Synthesizing a speech waveform from alice02.mgc
and alice02.lf0...done Synthesizing a speech waveform from
alice03.mgc and alice03.lf0...done Synthesizing a speech waveform
from alice04.mgc and alice04.lf0...done Synthesizing a speech
waveform from alice05.mgc and alice05.lf0...done Synthesizing a
speech waveform from alice06.mgc and alice06.lf0...done Synthesizing
a speech waveform from alice07.mgc and
alice07.lf0...done Synthesizing a speech waveform from alice08.mgc
and alice08.lf0...done Synthesizing a speech waveform from
alice09.mgc and alice09.lf0...done Synthesizing a speech waveform
from alice10.mgc and alice10.lf0...done Synthesizing a speech
waveform from alice11.mgc and alice11.lf0...done Synthesizing a
speech waveform from alice12.mgc and alice12.lf0...done Synthesizing
a speech waveform from alice13.mgc and
alice13.lf0...done Synthesizing a speech waveform from alice14.mgc
and alice14.lf0...done Synthesizing a speech waveform from
alice15.mgc and alice15.lf0...done Synthesizing a speech waveform
from alice16.mgc and alice16.lf0...done Synthesizing a speech
waveform from alice17.mgc and alice17.lf0...done Synthesizing a
speech waveform from alice18.mgc and alice18.lf0...done Synthesizing
a speech waveform from alice19.mgc and
alice19.lf0...done Synthesizing a speech waveform from alice20.mgc
and alice20.lf0...done Synthesizing a speech waveform from
alice21.mgc and alice21.lf0...done Synthesizing a speech waveform
from alice22.mgc and alice22.lf0...done Synthesizing a speech
waveform from alice23.mgc and alice23.lf0...done Synthesizing a
speech waveform from alice24.mgc and alice24.lf0...done Synthesizing
a speech waveform from alice25.mgc and
alice25.lf0...done Synthesizing a speech waveform from alice26.mgc
and alice26.lf0...done Synthesizing a speech waveform from
alice27.mgc and alice27.lf0...done Synthesizing a speech waveform
from alice28.mgc and alice28.lf0...done Synthesizing a speech
waveform from alice29.mgc and alice29.lf0...done Synthesizing a
speech waveform from alice30.mgc and alice30.lf0...done Synthesizing
a speech waveform from alice31.mgc and
alice31.lf0...done Synthesizing a speech waveform from alice32.mgc
and alice32.lf0...done Synthesizing a speech waveform from
alice33.mgc and alice33.lf0...done Synthesizing a speech waveform
from alice34.mgc and alice34.lf0...done Synthesizing a speech
waveform from alice35.mgc and alice35.lf0...done Synthesizing a
speech waveform from alice36.mgc and alice36.lf0...done Synthesizing
a speech waveform from alice37.mgc and
alice37.lf0...done Synthesizing a speech waveform from alice38.mgc
and alice38.lf0...done Synthesizing a speech waveform from
alice39.mgc and alice39.lf0...done
Synthesizing a speech waveform from alice40.mgc and
alice40.lf0...done done
Cheers, Xu
2009/11/9 那兴宇 <nxy-yzqs@xxxxxxx>
Hi,
Thanks very much for your reply. Errors appear again.
The
current version of SPTK is SPTK-3.2. Should I use the SPTK-3.1 version? When
the training is running, can I use the computer to do other things or should
I leave it alone?
Anyone has the similar experience? Please check it.
The following is the details. I highlight the error and invalid parts with
the red color. Thanks a lot.
Last
few lines of Terminal:
# Generating a generation label
script rm -f scp/gen.scp for lab in
/home/xiexu/Desktop/hts/HTS-demo_CMU-ARCTIC-SLT/data/labels/gen/*.lab; do
\ echo ${lab} >> scp/gen.scp;
\ done
make[1]: Leaving directory
`/home/xiexu/Desktop/hts/HTS-demo_CMU-ARCTIC-SLT/data'
echo "Running a training/synthesis perl script (Training.pl) in
background...." Running a training/synthesis perl script (Training.pl) in
background....
/usr/bin/perl scripts/Training.pl scripts/Config.pm > log
2>&1
& xiexu@xiexu-laptop:~/Desktop/hts/HTS-demo_CMU-ARCTIC-SLT$
Last few lines of the "log" file:
====================================================================== Start
synthesizing waveforms (1mix) at Mon Nov 9 18:51:00 NZDT
2009 ======================================================================
Processing
directory
/home/xiexu/Desktop/hts/HTS-demo_CMU-ARCTIC-SLT/gen/qst001/ver1/1mix/0: Synthesizing
a speech waveform from alice01.mgc and alice01 .lf0...mglsadf : Invalid option
'g'!
You can check the mglsadf in the current version3.2 SPTK
mannual. There is no '-g' option, so you should use SPTK3.1
instead.
mglsadf - MGLSA digital filter for
speech synthesis
usage: mglsadf [ options ] mgcfile [
infile ] > stdout
options: -m m : order of
mel-generalized cepstrum [25] -a
a :
alpha
[0.35] -c c : gamma = -1 /
(int)
c
[1] -p p : frame
period
[100] -i i : interpolation
period
[1] -t : transpose
filter
[FALSE] -k :
filtering without
gain
[FALSE] -P P : order of Pade
approximation
[4] -h : print
this message infile:
filter input
(float)
[stdin] stdout: filter
output (float) mgcfile:
mel-generalized cepstrum (float)
notice: if c==0, MLSA filter is
used, P should be 4 or 5
SPTK: version 3.2 CVS Info:
$Id: mglsadf.c,v 1.23 2008/11/06 15:40:51 tatsuyaito Exp $ /usr/bin/sox: invalid option -- w Failed:
invalid option
SoX doesn't support a single '-w'
option.
It refers to the reverberation effect or the window selection
options.
So I guess '-w' means the data length of the waveform is a 'word'. You
can change it to '-2'(means 2 Byte) or '-4'(means 4 Byte).
/usr/bin/sox: SoX v14.2.0
Usage summary:
[gopts] [[fopts] infile]... [fopts] outfile [effect
[effopts]]...
SPECIAL FILENAMES (infile,
outfile): -
Pipe/redirect input/output (stdin/stdout); use with -t -d,
--default-device Use the default audio device (where
available) -n,
--null
Use the `null' file handler; e.g. with synth effect -p,
--sox-pipe Alias
for `-t sox -'
SPECIAL FILENAMES (infile only): "|program
[options] ..." Pipe input from external program (where
supported) http://server/file Use the given URL as
input file (where supported)
GLOBAL OPTIONS (gopts) (can be specified
at any point before the first effect): --buffer
BYTES Set the
size of all processing buffers (default 8192) --combine
concatenate Concatenate multiple input files (default for
sox, rec) --combine sequence Sequence
multiple input files (default for play) --effects-file FILENAME
File containing effects and options -h,
--help
Display version number and usage information --help-effect
NAME Show usage of effect NAME, or
NAME=all for all --help-format NAME
Show info on format NAME, or NAME=all for all --input-buffer
BYTES Override the input buffer size (default: as
--buffer) --interactive
Prompt to overwrite output file -m, --combine
mix Mix multiple input files
(instead of concatenating) -M, --combine
merge Merge multiple input files (instead of
concatenating) --plot gnuplot|octave Generate script to
plot response of filter effect -q, --no-show-progress Run in
quiet mode; opposite of -S --replay-gain track|album|off Default:
off (sox, rec), track
(play) -R
Use default random numbers (same on each run of SoX) -S,
--show-progress Display progress while
processing audio
data --version
Display version number of SoX and
exit -V[LEVEL]
Increment or set verbosity level (default 2);
levels:
1: failure
messages
2:
warnings
3: details of
processing
4-6: increasing levels of debug messages FORMAT OPTIONS (fopts): Input
file format options need only be supplied for files that are
headerless. Output files will have the same format as the input file
where possible and not overriden by any of various means including
providing output format options.
-v|--volume
FACTOR Input file volume adjustment
factor (real number) -t|--type
FILETYPE File type of
audio -s/-u/-f/-U/-A/-i/-a/-g Encoding
type=signed-integer/unsigned-integer/floating-
point/mu-law/a-law/ima-adpcm/ms-adpcm/gsm-full-rate -e|--encoding
ENCODING Set encoding (ENCODING in above list) -b|--bits
BITS Encoded
sample size in
bits -1/-2/-3/-4/-8
Encoded sample size in bytes -N|--reverse-nibbles
Encoded
nibble-order -X|--reverse-bits
Encoded bit-order --endian little|big|swap Encoded byte-order; swap means
opposite to
default -L/-B/-x
Short options for the above -c|--channels CHANNELS Number of
channels of audio data; e.g. 2 = stereo -r|--rate
RATE Sample rate
of audio -C|--compression FACTOR Compression factor for output
format --add-comment TEXT Append
output file comment --comment
TEXT Specify
comment text for the output file --comment-file FILENAME File
containing comment text for the output file
AUDIO FILE FORMATS: 8svx
aif aifc aiff aiffc al amb au avr caf cdda cdr cvs cvsd dat dvms f4 f8 fap
flac fssd gsm hcom htk ima ircam la lpc lpc10 lu mat mat4 mat5 maud nist ogg
paf prc pvf raw s1 s2 s3 s4 sb sd2 sds sf sl smp snd sndfile sndr sndt sou
sox sph sw txw u1 u2 u3 u4 ub ul uw vms voc vorbis vox w64 wav wavpcm wv wve
xa xi PLAYLIST FORMATS: m3u pls AUDIO DEVICE DRIVERS:
alsa
EFFECTS: allpass band bandpass bandreject bass bend chorus
compand contrast dcshift deemph delay dither earwax echo echos equalizer
fade filter flanger gain highpass ladspa loudness lowpass mcompand mixer
noiseprof noisered norm oops pad phaser pitch rate remix repeat reverb
reverse riaa silence spectrogram speed splice stat stretch swap synth tempo
treble tremolo trim vol
EFFECT OPTIONS (effopts): effect dependent;
see --help-effect Error in
/usr/local/SPTK-3.2/bin/excite -p 80
/home/xiexu/Desktop/hts/HTS-demo_CMU-ARCTIC-SLT/gen/qst001/ver1/1mix/0/alice01.pit
| /usr/local/SPTK-3.2/bin/mglsadf -m 24 -p 80 -a 0.42 -g 0
/home/xiexu/Desktop/hts/HTS-demo_CMU-ARCTIC-SLT/gen/qst001/ver1/1mix/0/alice01.mgc
| /usr/local/SPTK-3.2/bin/x2x +fs | /usr/bin/sox -c 1 -s -w -t raw -r 16000
- -c 1 -s -w -t wav -r 16000
/home/xiexu/Desktop/hts/HTS-demo_CMU-ARCTIC-SLT/gen/qst001/ver1/1mix/0/alice01.wav This
occurs because of the two errors above.
Many thanks Xu
2009/11/9 那兴宇 <nxy-yzqs@xxxxxxx>
Hi,
Yes, the
training process is running background so you cannot shutdown your
computer and cygwin.
You can check if
it is over by typing "ps -ef". It is like the task manager in
windows.
发件人: 谢旭
发送时间: 2009-11-09 06:32:28
收件人: hts-users
抄送:
主题: [hts-users:02279] Re: Questions
aboutHTS-demo_CMU-ARCTIC-SLT
Hi,
Thanks very much for 那兴宇 and Sébastien's replys. I talk about the the
training part. I have checked the 'gen' folder, and there is no waveform.
I am doing the training again now. The following is the last few
lines of the Terminal:
echo "Running a training/synthesis perl script (Training.pl) in
background...."
Running a training/synthesis perl script (Training.pl) in
background....
/usr/bin/perl scripts/Training.pl scripts/ config.pm > log 2>&1 &
xiexu@xiexu-laptop:~/Desktop/hts/HTS-demo_CMU-ARCTIC-SLF$
Does this mean that the training is still in process but in
background and that I can not shut down the computer until the
training finishes? If it is as I just said, how to check whether the
training is over?
Thanks a lot
Xu
2009/11/9 那兴宇 <nxy-yzqs@xxxxxxx>
Have you checked the background running
of the demo?
If so, check the generated waveforms. I don't
think there is any waveform in your 'gen' directory.
Then you can see the log file to locate
where error occured.
发件人: 谢旭
发送时间: 2009-11-08 18:29:13
收件人: hts-users
抄送:
主题: [hts-users:02276] Questions about
HTS-demo_CMU-ARCTIC-SLT
Hi all,
I have installed the demo and no errors
has occurred. But the installation took about 30 minutes, not like the
INSTALL file says "After composing training data, HMMs are estimated and
speech waveforms are synthesized. It takes about 6 to 12 hours
:-)".
If the installation is right, what should I do next? Is
there a way like festival SayText?
Many thanks
Xu
|