[hts-users:00461] HTS-2.0RC3

["Heiga ZEN (Byung Ha CHUN)" <zen@xxxxxxxxxxxxxxxx>]

Dear members of hts-users mailing list,


The HTS working group would like to release the 3rd release candidate of HTS-2.0.
They can be downloaded from

	Patch code:
	http://kt-lab.ics.nitech.ac.jp/~zen/HTS-2.0RC3_for_HTK-3.4-alpha.patch.bz2
	(190670 bytes, MD5 checksum: aa7b6f56fc1dc0f51bfaf46dca9934d7)
	
	Speaker dependent demo:
	http://kt-lab.ics.nitech.ac.jp/~zen/HTS-demo_CMU-ARCTIC-SLT.tar.bz2
	(83386649 bytes, MD5 checksum d5f709b89aea5d4b026d45d6dc4ac6a4)

	Speaker adaptation/adaptive training demo:
	http://kt-lab.ics.nitech.ac.jp/~zen/HTS-demo_CMU-ARCTIC-ADAPT.tar.bz2
	(515973742 bytes, MD5 checksum f795fd33c8f65bf4d3b673fb2f00bfeb)


Differences between RC2 and RC3 are as follows:
- Speaker adaptation/adaptive training of MSD and multi-stream HMMs
- Speech parameter generation with speaker adapted, full covariance, or non left-to-right HMMs.
- Interface of HMGenS is moved from HHEd style to HERest style
- Miscellaneous fixes
- Demonstration for speaker adaptation/adaptive training


This version is released as a patch code to HTK-3.4-alpha (http://htk.eng.cam.ac.uk/).
You can download it from the following addresses:

	HTK 3.4 (alpha):
	http://htk.eng.cam.ac.uk/ftp/software/HTK-3.4-alpha.tar.gz

	HDecode:
	See extensions page on HTK website - 
	http://htk.eng.cam.ac.uk/extensions/index.shtml

After downloading HTK-3.4-alpha, HDecode and HTS-2.0RC2 patch code, please expand them (tar -jxvf) and apply the patch code on htk-3.4 directory as
follows:

	htk-3.4% patch -p1 -d . < HTS-2.0RC3_for_HTK-3.4-alpha.patch

Then please run configure and make.


Festival Speech Synthesis System (http://www.cstr.ed.ac.uk/projects/festival/download.html) and SPTK
(http://kt-lab.ics.nitech.ac.jp/~tokuda/SPTK/index.html) are required to run the demonstration.
Please download, configure, make, and install them.


Many people have asked us about F0 extraction, so from RC2 we decided to include a F0 extraction program.
Fortunately, ESPS get_f0 has been included in the open-source speech toolkit *Snack* (http://www.speech.kth.se/snack/).
In the release we include a Tcl/Tk script to call get_f0 function from Snack.
Therefore, you also have to install Tcl/Tk and Snack on your environment to run the demonstration.
ActiveState (http://www.activestate.com/) provides a Tcl/Tk distribution named ActiveTcl for many platforms (http://www.activestate.com/Products/ActiveTcl/?mp=1).
This distribution includes Snack and it is easy to install and use.
I recommend you to use this for running the demonstration (Actually I'm using it).  


After expanding HTS-demo_CMU-ARCTIC-{SLT,ADAPT}.tar.bz2, please run configure.
If your PATH variable does not include SPTK, Festival, tclsh (with Snack) and HTS binary directory, please run configure as follows:

$ PATH=/usr/local/SPTK/bin:/usr/local/HTS-2.0RC3_for_HTK-3.4/bin:/usr/local/festival/examples:/usr/local/Activetcl/bin:$PATH ./configure

Then run make.
After composing training data, training script will be run in background.


This would be the final release candidate.
We are planning to release the final version at the end of this month.
Any bug reports and comments are highly appreciated, please mail to the hts-users mailing list.
I'm looking forward to hearing from you :-)


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
 Japan

 web: http://kt-lab.ics.nitech.ac.jp/~zen
------------------------------------------------