[hts-users:00633] Re: HMGenS
Hi,
Mohammed Samir wrote:
The command was:
HMGenS -A -C syn.cnf -D -T 1 -f 0.005 -p hmmlist1
dmlist label.lab
the output was:
PdfStream[1]:
#streams: 1(vSize=20)
#order: 20
file ext: mcp
0-th window: /home/work....../mcp.win1
Pruning-off
Generating Label lable.lab
ERROR[+9935] Generator: Cannot find duration model
da in current list
You *must* specify HMM and duration model mmfs using -H and -N options or directories including HMMs and duration models using -d and -n options.
USAGE: HMGenS [options] hmmList dmList labFiles...
Option Default
-a Use an input linear transform off
-b f Mixture pruning threshold 10.0
-c n type of parameter generation algorithm 0
0: both mix and state sequences are given
1: mix sequence is given,
but state sequence is hidden
2: both state and mix sequences are hidden
-d s dir to find hmm definitions current
-e use model alignment from label for pruning off
-f f frame shift in 100 ns 50000
-h s [s] set speaker name pattern to s, *.%%%
optionally set parent patterns
-m use model alignment for duration off
-n s dir to find duration model definitions current
-p output pdf sequences off
-r f speaking rate factor (f<1: fast f>1: slow) 1.0
-s use state alignment for duration off
-t f [i l] set pruning to f [inc limit] inf
-v f threshold for switching spaces for MSD 0.5
-x s extension for hmm files none
-y s extension for duration model files none
-N mmf Load duration macro file mmf
-A Print command line arguments off
-B Save HMMs/transforms as binary off
-C cf Set config file to cf default
-D Display configuration variables off
-E s [s] set dir for parent xform to s off
and optional extension
-G fmt Set source label format to fmt as config
-H mmf Load HMM macro file mmf
-I mlf Load master label file mlf
-J s [s] set dir for input xform to s none
and optional extension
-L dir Set input label (or net) dir current
-M dir Dir to write HMM macro files current
-S f Set script file to f none
-T N Set trace flags to N 0
-V Print version information off
-X ext Set input label (or net) file ext lab
-----------------------------
the content of the file hmmlist:
da
----------------------------------
where da is the name of the hmm file generated by
HRest
the content of the file dmlist:
duration
--------------------------------
where duration is the name of the duration model
file
generated by HRest (-g option)
the content of the file label.lab:
da duration
---------------------------------
In your case, hmmlist, dmlist, and label should be
% cat hmmlist
da
% cat dmlist
da
% cat label.lab
da
where da is the name of phoneme, not name of hmm file.
by the way, I work on a single phoneme so I don't need
labeling or clustering I just want the parameters and
pdf sequence to input it to the mlsadf filter
manually.
Write code by yourself.
Did you refer to HTKBook before asking any questions?
Formats of model list files and label files are written on it.
I'm extremely tired to answer your question :-(
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://www.sp.nitech.ac.jp/~zen
------------------------------------------------
- References
-
- [hts-users:00632] Re: HMGenS, Mohammed Samir