open-jtalk-users — For users.

[open-jtalk-users] Questions About Setting Accent Phrase

[<ope...@li...>]

Dear Developers,

I hope this email finds you well. First, I would like to express my thanks
to all developers of OpenJTalk.

I'm Francis Hu, a Chinese college student interested in Japanese language
processing and text-to-speech. One major challenge in Japanese frontend
processing is determining the accent phrase boundary, without which correct
pitch accent can't be assigned.

As far as I know, 18 rules have been implemented in njd_set_accent_phrase.
They are some general rules predicting the boundary of accent phrases. But
in my experiment, the accuracy is not very high, thus affecting the quality
of synthesized speech.

Therefore, I have two questions. One is where do you get these rules? Is
there any papers? And the other is whether there are more possible rules to
be applied?

I really appreciate it if you could give me some guidance. Thanks
in advance.

Best regards
Francis Hu

[open-jtalk-users] (no subject)

[<ope...@li...>]

こんにちは。
この度OpenJTalkを利用中にバグを発見しましたが、SourceForgeのバグ報告チケットが作成できなかったためこちらで報告させていただきます。

- 概要
漢数字(正確にはMecabの形要素解析で漢数字と判定されたもの)と数字が連続したとき、一つの数字として読んでしまう
例: 「いち1」→「じゅういち」

- 想定される挙動
漢数字と数字は別のものとして扱われ、繋げて読まれない

- 環境
OpenJTalk version 1.11
Dictionary for OpenJTalk version 1.11
HTS voice version 1.05

よろしくお願いいたします。

Re: [open-jtalk-users] Does Not Sound Like Demo Page

[<ope...@li...>]

* Supplement

I run open-jTalk on macOS Mojave (ver 10.14.5).
I build all software from sources.



2019年5月28日(火) 7:16 <ope...@li...>:

> I tried it, but that command does not produce speech.  It only makes a loud
> noise.
>
> Also, even though I downloaded open_jtalk-1.11.tar.gz from sourceforge,
> when I
> run open_jtalk with no options, it says:
>
> The Japanese TTS System "Open JTalk"
> Version 1.10 (http://open-jtalk.sourceforge.net/)
>
> Maybe the 1.11 file is really version 1.10 by mistake?
>
>
> On Mon, May 27, 2019 at 03:59:25AM +0900,
> ope...@li... wrote:
> > Hello,
> >
> > sorry,  "-b 0,40" is just only an example.
> > I tried to adjusted sound by options.
> >
> > * options
> > "-b 0.55 -jm 20.0 -r 1.0 -u 0.1 -a 0.55 -jf 50.0 -fm 6.5 -g 14.0"
> >
> > * voice file
> > https://www.dropbox.com/s/66hei9gffndn6sw/output.wav?dl=0
> >
> > * My perfect checking command(please change paths)
> > /bin/echo "その夜までのやる気がでないんだよ" |
> > ./open_jtalk/open_jtalk-1.11/bin/open_jtalk -b 0.55 -jm 20.0 -r 1.0 -u
> > 0.1 -a 0.55 -jf 50.0 -fm 6.5 -z 0 -g 14.0 -x
> > ./open_jtalk/open_jtalk_dic_utf_8-1.11 -m
> > ./open_jtalk/MMDAgent_Example-1.8/Voice/mei/mei_normal.htsvoice -ow
> > output.wav
> >
> > Regards
> >
> >
> >
> >
> >
> > 2019年5月27日(月) 0:53 <ope...@li...>:
> >
> > > Thank you for your response.
> > >
> > > Unfortunately, it still does not sound the same.
> > >
> > > I tried:
> > >
> > >   open_jtalk -m /usr/share/open-jtalk/voices/mei_bashful.htsvoice \
> > >     -x /usr/share/open-jtalk/dic -ow test.wav -g -20 -b 0.40 test.txt
> > >
> > > And got this result:
> > >
> > >   http://xezlec.com/test-b0.40.wav
> > >
> > > Compare to the result from the demo page:
> > >
> > >   http://xezlec.com/20190511210807_10698.wav
> > >
> > >
> > > On Sat, May 25, 2019 at 11:14:35PM +0900,
> > > ope...@li... wrote:
> > > > Hello,
> > > >
> > > > Maybe,  you must add the option of '-b' for quality of sound.
> > > >
> > > > ex : open_jtalk -m /usr/share/open-jtalk/voices/mei_bashful.htsvoice
> -x
> > > > /usr/share/open-jtalk/dic -ow test.wav -g -20 -b 0.40 test.txt
> > > >
> > > > Please try it.
> > > >
> > > > Regards
> > >
> > >
> > > > _______________________________________________
> > > > open-jtalk-users mailing list
> > > > ope...@li...
> > > > https://lists.sourceforge.net/lists/listinfo/open-jtalk-users
> > >
> > >
> > >
> > > _______________________________________________
> > > open-jtalk-users mailing list
> > > ope...@li...
> > > https://lists.sourceforge.net/lists/listinfo/open-jtalk-users
> > >
>
>
> > _______________________________________________
> > open-jtalk-users mailing list
> > ope...@li...
> > https://lists.sourceforge.net/lists/listinfo/open-jtalk-users
>
>
>
> _______________________________________________
> open-jtalk-users mailing list
> ope...@li...
> https://lists.sourceforge.net/lists/listinfo/open-jtalk-users
>

Re: [open-jtalk-users] Does Not Sound Like Demo Page

[<ope...@li...>]

Hello.

> Maybe the 1.11 file is really version 1.10 by mistake?
It is not mistake. I use open_jtalk-1.11 (25 December, 2018)

Recently, I build all software are updated to newest.
My software versions are like this.
---
MMDAgent_Example-1.8
hts_engine_API-1.10
open_jtalk-1.11
open_jtalk_dic_utf_8-1.11
MMDAgent_Example-1.8
hts_engine_API-1.10
---

So, If you use some old version software.
Please try to rebuild.

Regards

==============
My English is not good.I write together Japanese comment.

> Maybe the 1.11 file is really version 1.10 by mistake?
open-jtalk は 1.11 を使用しており、1.10 ではありません。

最近、私は open-jtalk を最新版で構築しました。
私のソフトのバージョンはこのようになります。
---
MMDAgent_Example-1.8
hts_engine_API-1.10
open_jtalk-1.11
open_jtalk_dic_utf_8-1.11
MMDAgent_Example-1.8
hts_engine_API-1.10
---

もし、古いソフトを使用している場合。
再構築を試してください。

以上。


2019年5月28日(火) 7:16 <ope...@li...>:

> I tried it, but that command does not produce speech.  It only makes a loud
> noise.
>
> Also, even though I downloaded open_jtalk-1.11.tar.gz from sourceforge,
> when I
> run open_jtalk with no options, it says:
>
> The Japanese TTS System "Open JTalk"
> Version 1.10 (http://open-jtalk.sourceforge.net/)
>
> Maybe the 1.11 file is really version 1.10 by mistake?
>
>
> On Mon, May 27, 2019 at 03:59:25AM +0900,
> ope...@li... wrote:
> > Hello,
> >
> > sorry,  "-b 0,40" is just only an example.
> > I tried to adjusted sound by options.
> >
> > * options
> > "-b 0.55 -jm 20.0 -r 1.0 -u 0.1 -a 0.55 -jf 50.0 -fm 6.5 -g 14.0"
> >
> > * voice file
> > https://www.dropbox.com/s/66hei9gffndn6sw/output.wav?dl=0
> >
> > * My perfect checking command(please change paths)
> > /bin/echo "その夜までのやる気がでないんだよ" |
> > ./open_jtalk/open_jtalk-1.11/bin/open_jtalk -b 0.55 -jm 20.0 -r 1.0 -u
> > 0.1 -a 0.55 -jf 50.0 -fm 6.5 -z 0 -g 14.0 -x
> > ./open_jtalk/open_jtalk_dic_utf_8-1.11 -m
> > ./open_jtalk/MMDAgent_Example-1.8/Voice/mei/mei_normal.htsvoice -ow
> > output.wav
> >
> > Regards
> >
> >
> >
> >
> >
> > 2019年5月27日(月) 0:53 <ope...@li...>:
> >
> > > Thank you for your response.
> > >
> > > Unfortunately, it still does not sound the same.
> > >
> > > I tried:
> > >
> > >   open_jtalk -m /usr/share/open-jtalk/voices/mei_bashful.htsvoice \
> > >     -x /usr/share/open-jtalk/dic -ow test.wav -g -20 -b 0.40 test.txt
> > >
> > > And got this result:
> > >
> > >   http://xezlec.com/test-b0.40.wav
> > >
> > > Compare to the result from the demo page:
> > >
> > >   http://xezlec.com/20190511210807_10698.wav
> > >
> > >
> > > On Sat, May 25, 2019 at 11:14:35PM +0900,
> > > ope...@li... wrote:
> > > > Hello,
> > > >
> > > > Maybe,  you must add the option of '-b' for quality of sound.
> > > >
> > > > ex : open_jtalk -m /usr/share/open-jtalk/voices/mei_bashful.htsvoice
> -x
> > > > /usr/share/open-jtalk/dic -ow test.wav -g -20 -b 0.40 test.txt
> > > >
> > > > Please try it.
> > > >
> > > > Regards
> > >
> > >
> > > > _______________________________________________
> > > > open-jtalk-users mailing list
> > > > ope...@li...
> > > > https://lists.sourceforge.net/lists/listinfo/open-jtalk-users
> > >
> > >
> > >
> > > _______________________________________________
> > > open-jtalk-users mailing list
> > > ope...@li...
> > > https://lists.sourceforge.net/lists/listinfo/open-jtalk-users
> > >
>
>
> > _______________________________________________
> > open-jtalk-users mailing list
> > ope...@li...
> > https://lists.sourceforge.net/lists/listinfo/open-jtalk-users
>
>
>
> _______________________________________________
> open-jtalk-users mailing list
> ope...@li...
> https://lists.sourceforge.net/lists/listinfo/open-jtalk-users
>

Re: [open-jtalk-users] Does Not Sound Like Demo Page

[<ope...@li...>]

I tried it, but that command does not produce speech.  It only makes a loud
noise.

Also, even though I downloaded open_jtalk-1.11.tar.gz from sourceforge, when I
run open_jtalk with no options, it says:

The Japanese TTS System "Open JTalk"
Version 1.10 (http://open-jtalk.sourceforge.net/)

Maybe the 1.11 file is really version 1.10 by mistake?


On Mon, May 27, 2019 at 03:59:25AM +0900, ope...@li... wrote:
> Hello,
> 
> sorry,  "-b 0,40" is just only an example.
> I tried to adjusted sound by options.
> 
> * options
> "-b 0.55 -jm 20.0 -r 1.0 -u 0.1 -a 0.55 -jf 50.0 -fm 6.5 -g 14.0"
> 
> * voice file
> https://www.dropbox.com/s/66hei9gffndn6sw/output.wav?dl=0
> 
> * My perfect checking command(please change paths)
> /bin/echo "その夜までのやる気がでないんだよ" |
> ./open_jtalk/open_jtalk-1.11/bin/open_jtalk -b 0.55 -jm 20.0 -r 1.0 -u
> 0.1 -a 0.55 -jf 50.0 -fm 6.5 -z 0 -g 14.0 -x
> ./open_jtalk/open_jtalk_dic_utf_8-1.11 -m
> ./open_jtalk/MMDAgent_Example-1.8/Voice/mei/mei_normal.htsvoice -ow
> output.wav
> 
> Regards
> 
> 
> 
> 
> 
> 2019年5月27日(月) 0:53 <ope...@li...>:
> 
> > Thank you for your response.
> >
> > Unfortunately, it still does not sound the same.
> >
> > I tried:
> >
> >   open_jtalk -m /usr/share/open-jtalk/voices/mei_bashful.htsvoice \
> >     -x /usr/share/open-jtalk/dic -ow test.wav -g -20 -b 0.40 test.txt
> >
> > And got this result:
> >
> >   http://xezlec.com/test-b0.40.wav
> >
> > Compare to the result from the demo page:
> >
> >   http://xezlec.com/20190511210807_10698.wav
> >
> >
> > On Sat, May 25, 2019 at 11:14:35PM +0900,
> > ope...@li... wrote:
> > > Hello,
> > >
> > > Maybe,  you must add the option of '-b' for quality of sound.
> > >
> > > ex : open_jtalk -m /usr/share/open-jtalk/voices/mei_bashful.htsvoice -x
> > > /usr/share/open-jtalk/dic -ow test.wav -g -20 -b 0.40 test.txt
> > >
> > > Please try it.
> > >
> > > Regards
> >
> >
> > > _______________________________________________
> > > open-jtalk-users mailing list
> > > ope...@li...
> > > https://lists.sourceforge.net/lists/listinfo/open-jtalk-users
> >
> >
> >
> > _______________________________________________
> > open-jtalk-users mailing list
> > ope...@li...
> > https://lists.sourceforge.net/lists/listinfo/open-jtalk-users
> >


> _______________________________________________
> open-jtalk-users mailing list
> ope...@li...
> https://lists.sourceforge.net/lists/listinfo/open-jtalk-users

Re: [open-jtalk-users] Does Not Sound Like Demo Page

[<ope...@li...>]

Hello,

sorry,  "-b 0,40" is just only an example.
I tried to adjusted sound by options.

* options
"-b 0.55 -jm 20.0 -r 1.0 -u 0.1 -a 0.55 -jf 50.0 -fm 6.5 -g 14.0"

* voice file
https://www.dropbox.com/s/66hei9gffndn6sw/output.wav?dl=0

* My perfect checking command(please change paths)
/bin/echo "その夜までのやる気がでないんだよ" |
./open_jtalk/open_jtalk-1.11/bin/open_jtalk -b 0.55 -jm 20.0 -r 1.0 -u
0.1 -a 0.55 -jf 50.0 -fm 6.5 -z 0 -g 14.0 -x
./open_jtalk/open_jtalk_dic_utf_8-1.11 -m
./open_jtalk/MMDAgent_Example-1.8/Voice/mei/mei_normal.htsvoice -ow
output.wav

Regards





2019年5月27日(月) 0:53 <ope...@li...>:

> Thank you for your response.
>
> Unfortunately, it still does not sound the same.
>
> I tried:
>
>   open_jtalk -m /usr/share/open-jtalk/voices/mei_bashful.htsvoice \
>     -x /usr/share/open-jtalk/dic -ow test.wav -g -20 -b 0.40 test.txt
>
> And got this result:
>
>   http://xezlec.com/test-b0.40.wav
>
> Compare to the result from the demo page:
>
>   http://xezlec.com/20190511210807_10698.wav
>
>
> On Sat, May 25, 2019 at 11:14:35PM +0900,
> ope...@li... wrote:
> > Hello,
> >
> > Maybe,  you must add the option of '-b' for quality of sound.
> >
> > ex : open_jtalk -m /usr/share/open-jtalk/voices/mei_bashful.htsvoice -x
> > /usr/share/open-jtalk/dic -ow test.wav -g -20 -b 0.40 test.txt
> >
> > Please try it.
> >
> > Regards
>
>
> > _______________________________________________
> > open-jtalk-users mailing list
> > ope...@li...
> > https://lists.sourceforge.net/lists/listinfo/open-jtalk-users
>
>
>
> _______________________________________________
> open-jtalk-users mailing list
> ope...@li...
> https://lists.sourceforge.net/lists/listinfo/open-jtalk-users
>

Re: [open-jtalk-users] Does Not Sound Like Demo Page

[<ope...@li...>]

Thank you for your response.

Unfortunately, it still does not sound the same.

I tried:

  open_jtalk -m /usr/share/open-jtalk/voices/mei_bashful.htsvoice \
    -x /usr/share/open-jtalk/dic -ow test.wav -g -20 -b 0.40 test.txt

And got this result:

  http://xezlec.com/test-b0.40.wav

Compare to the result from the demo page:

  http://xezlec.com/20190511210807_10698.wav


On Sat, May 25, 2019 at 11:14:35PM +0900, ope...@li... wrote:
> Hello,
> 
> Maybe,  you must add the option of '-b' for quality of sound.
> 
> ex : open_jtalk -m /usr/share/open-jtalk/voices/mei_bashful.htsvoice -x
> /usr/share/open-jtalk/dic -ow test.wav -g -20 -b 0.40 test.txt
> 
> Please try it.
> 
> Regards


> _______________________________________________
> open-jtalk-users mailing list
> ope...@li...
> https://lists.sourceforge.net/lists/listinfo/open-jtalk-users

Re: [open-jtalk-users] Does Not Sound Like Demo Page

[<ope...@li...>]

Hello,

Maybe,  you must add the option of '-b' for quality of sound.

ex : open_jtalk -m /usr/share/open-jtalk/voices/mei_bashful.htsvoice -x
/usr/share/open-jtalk/dic -ow test.wav -g -20 -b 0.40 test.txt

Please try it.

Regards

[open-jtalk-users] Does Not Sound Like Demo Page

[<ope...@li...>]

Hello,

I listened to the open-jtalk demo page and it sounded very good!  But when I
installed the program, it sounded very bad, very primitive.  Why don't they
sound the same?  Does the web page use a different program?

Here is the sound I got from the web page:
http://xezlec.com/20190511210807_10698.wav

And here is the sound I got from the open_jtalk program:
http://xezlec.com/test.wav

I am using open-jtalk version 1.11 and hts-engine version 1.10.
This is the command I used:
open_jtalk -m /usr/share/open-jtalk/voices/mei_bashful.htsvoice \
  -x /usr/share/open-jtalk/dic -ow test.wav -g -20 test.txt

What have I done wrong?

Re: [open-jtalk-users] Open JTalk-1.11

[<ope...@li...>]

On Wed, Dec 26, 2018 at 11:22 AM <ope...@li...>
wrote:

> Hi,
>
> There are three major updates in version 1.11.
>
>  1. add UniDic.
> New dictionary includes unidic-csj-2.2.0 in https://unidic.ninjal.ac.jp/
>
>  2. remove long vowels of naturalized words in dictionary.
> (example) "HEY" = /h/ /e/ /e/ -> "HEY" = /h/ /e/ /i/
>
>  3. add digit rules for digits including comma.
> (example) "1,000円" = /i/ /ch/ /i/ /pau/ /z/ /e/ /r/ /o/ /e/ /N/ ->
> "1,000円" = /s/ /e/ /N/ /e/ /N/
>
> Regards,
> Keiichiro Oura
>
>
Hi,
thanks for the details.
Regards,
Takeshi

Re: [open-jtalk-users] Open JTalk-1.11

[<ope...@li...>]

Hi,

There are three major updates in version 1.11.

 1. add UniDic.
New dictionary includes unidic-csj-2.2.0 in https://unidic.ninjal.ac.jp/

 2. remove long vowels of naturalized words in dictionary.
(example) "HEY" = /h/ /e/ /e/ -> "HEY" = /h/ /e/ /i/

 3. add digit rules for digits including comma.
(example) "1,000円" = /i/ /ch/ /i/ /pau/ /z/ /e/ /r/ /o/ /e/ /N/ ->
"1,000円" = /s/ /e/ /N/ /e/ /N/

Regards,
Keiichiro Oura

2018年12月26日(水) 9:26 <ope...@li...>:
>
>
> On Tue, Dec 25, 2018 at 7:16 PM <ope...@li...> wrote:
>>
>> We are please to inform you the release of
>>
>>                      The Open JTalk version 1.11
>>                       release December 25, 2018
>>
>> for Japanese text-to-speech synthesis.  We apologize if you receive
>> multiple copies.  Please check
>>
>>  http://open-jtalk.sourceforge.net/
>>
>> We would appreciate it if you would distribute this email to anyone
>> who would be interested in this software.
>>
>> Best regards,
>> Keiichi Tokuda
>> in behalf of HTS Working Group
>> to...@ni...
>> http://www.sp.nitech.ac.jp/~tokuda/
>
>
> Hi,
> thanks for this release.
> But I don't see details about what changes were done.
> It would be nice get a Release Notes every time a new version goes out.
>
> Takeshi
> _______________________________________________
> open-jtalk-users mailing list
> ope...@li...
> https://lists.sourceforge.net/lists/listinfo/open-jtalk-users

Re: [open-jtalk-users] Open JTalk-1.11

[<ope...@li...>]

On Tue, Dec 25, 2018 at 7:16 PM <ope...@li...>
wrote:

> We are please to inform you the release of
>
>                      The Open JTalk version 1.11
>                       release December 25, 2018
>
> for Japanese text-to-speech synthesis.  We apologize if you receive
> multiple copies.  Please check
>
>  http://open-jtalk.sourceforge.net/
>
> We would appreciate it if you would distribute this email to anyone
> who would be interested in this software.
>
> Best regards,
> Keiichi Tokuda
> in behalf of HTS Working Group
> to...@ni...
> http://www.sp.nitech.ac.jp/~tokuda/


Hi,
thanks for this release.
But I don't see details about what changes were done.
It would be nice get a Release Notes every time a new version goes out.

Takeshi

Re: [open-jtalk-users] open-jtalk-users Digest, Vol 12, Issue 3

[<ope...@li...>]

ありがとう。 メリークリスマス

On Tue, Dec 25, 2018, 7:07 AM <
ope...@li... wrote:

> Send open-jtalk-users mailing list submissions to
>         ope...@li...
>
> To subscribe or unsubscribe via the World Wide Web, visit
>         https://lists.sourceforge.net/lists/listinfo/open-jtalk-users
> or, via email, send a message with subject or body 'help' to
>         ope...@li...
>
> You can reach the person managing the list at
>         ope...@li...
>
> When replying, please edit your Subject line so it is more specific
> than "Re: Contents of open-jtalk-users digest..."
>
>
> Today's Topics:
>
>    1. Open JTalk-1.11 (ope...@li...)
>
>
> ----------------------------------------------------------------------
>
> Message: 1
> Date: Tue, 25 Dec 2018 19:15:44 +0900
> From: ope...@li...
> To: open-jtalk-users <ope...@li...>
> Cc: Keiichi Tokuda <to...@ni...>, Keiichiro Oura
>         <ur...@ni...>
> Subject: [open-jtalk-users] Open JTalk-1.11
> Message-ID:
>         <
> mai...@li...>
> Content-Type: text/plain; charset="UTF-8"
>
> We are please to inform you the release of
>
>                      The Open JTalk version 1.11
>                       release December 25, 2018
>
> for Japanese text-to-speech synthesis.  We apologize if you receive
> multiple copies.  Please check
>
>  http://open-jtalk.sourceforge.net/
>
> We would appreciate it if you would distribute this email to anyone
> who would be interested in this software.
>
> Best regards,
> Keiichi Tokuda
> in behalf of HTS Working Group
> to...@ni...
> http://www.sp.nitech.ac.jp/~tokuda/
>
>
>
> ------------------------------
>
>
>
> ------------------------------
>
> Subject: Digest Footer
>
> _______________________________________________
> open-jtalk-users mailing list
> ope...@li...
> https://lists.sourceforge.net/lists/listinfo/open-jtalk-users
>
>
> ------------------------------
>
> End of open-jtalk-users Digest, Vol 12, Issue 3
> ***********************************************
>

[open-jtalk-users] Open JTalk-1.11

[<ope...@li...>]

We are please to inform you the release of

                     The Open JTalk version 1.11
                      release December 25, 2018

for Japanese text-to-speech synthesis.  We apologize if you receive
multiple copies.  Please check

 http://open-jtalk.sourceforge.net/

We would appreciate it if you would distribute this email to anyone
who would be interested in this software.

Best regards,
Keiichi Tokuda
in behalf of HTS Working Group
to...@ni...
http://www.sp.nitech.ac.jp/~tokuda/

[open-jtalk-users] Open JTalk

[<ope...@li...>]

We plan to release the next version of Open JTalk after spring 2018.

Best regards,
Keiichi Tokuda
in behalf of HTS Working Group
to...@ni...
http://www.sp.nitech.ac.jp/~tokuda/

[open-jtalk-users] Fwd: hts_engine API

[<ope...@li...>]

We plan to release the next version of hts_engine API after spring 2018.

Best regards,
Keiichi Tokuda
in behalf of HTS Working Group
to...@ni...
http://www.sp.nitech.ac.jp/~tokuda/

[open-jtalk-users] Fwd: HTS-2.3.2

[<ope...@li...>]

We are please to inform you the release of

          The HMM/DNN-based Speech Synthesis System (HTS)
             version 2.3.2  release December 25, 2017

for HMM-based speech synthesis.  We apologize if you receive multiple
copies.  Please check

 http://hts.sp.nitech.ac.jp/

Trajectory training considering global variance based on DNN and
speaker adaptive training for DNN were added.

We would appreciate it if you would distribute this email to anyone
who would be interested in this software toolkit.

Best regards,
Keiichi Tokuda
in behalf of HTS Working Group
to...@ni...
http://www.sp.nitech.ac.jp/~tokuda/

[open-jtalk-users] Fwd: Open JTalk-1.10

[<ope...@li...>]

We are please to inform you the release of

                     The Open JTalk version 1.10
                      release December 25, 2016

for Japanese text-to-speech synthesis.  We apologize if you receive
multiple copies.  Please check

 http://open-jtalk.sourceforge.net/

We would appreciate it if you would distribute this email to anyone
who would be interested in this software.

Best regards,
Keiichi Tokuda
in behalf of HTS Working Group
to...@ni...
http://www.sp.nitech.ac.jp/~tokuda/

[open-jtalk-users] Fwd: Flite+hts_engine-1.07

[<ope...@li...>]

We are please to inform you the release of

                    Flite+hts_engine version 1.07
                      release December 25, 2016

for HMM-based speech synthesis.  We apologize if you receive multiple
copies.  Please check

 http://hts-engine.sourceforge.net/

We would appreciate it if you would distribute this email to anyone
who would be interested in this software.

Best regards,
Keiichi Tokuda
in behalf of HTS Working Group
to...@ni...
http://www.sp.nitech.ac.jp/~tokuda/

[open-jtalk-users] Fwd: HTS-2.3.1

[<ope...@li...>]

We are please to inform you the release of

            The HMM-based Speech Synthesis System (HTS)
             version 2.3.1  release December 25, 2016

for HMM-based speech synthesis.  We apologize if you receive multiple
copies.  Please check

 http://hts.sp.nitech.ac.jp/

One of new features is a DNN-based demo script.

We would appreciate it if you would distribute this email to anyone
who would be interested in this software toolkit.

Best regards,
Keiichi Tokuda
in behalf of HTS Working Group
to...@ni...
http://www.sp.nitech.ac.jp/~tokuda/

Re: [open-jtalk-users] How to make wav from HTS demo?

[<ope...@li...>]

Hi,

If demo scripts run correctly, gen/qst001/ver1/hts_engine/*.wav are
generated.
It seems that demo scripts in your environments don't run correctly.
Please check log/warning messages of demo scripts.

hts-users ML and its archives can help you.
http://hts.sp.nitech.ac.jp/?Mailing%20List

Regards,
Keiichiro Oura





2016-04-05 17:47 GMT+09:00 <ope...@li...>:

> Thanks.
>
> There are no waveforms instead of
>
> .raw(/data/raw/nitech_jp_atr503_m001_*.raw)
>
> and
>
> .lab(/data/labels/mono/nitech_jp_atr503_m001_*.lab)
>
> In execution, only raw file used.
>
> Where are wav files in HTS-demo?
>
> Regards.
>
>
>
> 2016-04-05 16:36 GMT+09:00 <ope...@li...>:
>
>> Hi,
>>
>> The version seems ok.
>> Let me know configure options&logs of demo scripts.
>> Do you use little endian machine?
>> Were waveforms (gen/qst001/ver1/hts_engine/*.wav) in demo scripts
>> synthesized correctly?
>>
>> Regards,
>> Keiichiro Oura
>>
>>
>>
>> 2016-04-05 15:11 GMT+09:00 <ope...@li...>:
>>
>>> Thanks.
>>>
>>> These are the version
>>>
>>> HTK-3.4.1
>>> HDecode-3.4.1
>>> HTS-2.3_for_HTK-3.4.1.patch
>>> SPTK-3.9
>>> hts_engine_API-1.10
>>>
>>> and I couldn't find out how to know the version of demo scripts,
>>> but I downloaded the latest version. Here
>>>
>>> http://hts.sp.nitech.ac.jp/?Download
>>> Speaker dependent training demo
>>>
>>>    - Japanese [image: file]tar.bz2
>>>    <http://hts.sp.nitech.ac.jp/archives/2.3/HTS-demo_NIT-ATR503-M001.tar.bz2> (126
>>>    MB)
>>>
>>>
>>> Regards.
>>>
>>>
>>> 2016-04-05 12:57 GMT+09:00 <ope...@li...>:
>>>
>>>> Hi,
>>>>
>>>> Let me know the version of HTS, SPTK, hts_engine API, and demo
>>>> scripts, respectively.
>>>>
>>>> Regards,
>>>> Keiichiro Oura
>>>>
>>>>
>>>> 2016-04-05 12:51 GMT+09:00  <ope...@li...>:
>>>> > Thanks.
>>>> >
>>>> > I could generate wav file by open_jtalk with this
>>>> > hts_voice_nitech_jp_atr503_m001.
>>>> > But I want to use HTS voice generated by
>>>> > HTS_demo-NIT-ATR503-M001(http://hts.sp.nitech.ac.jp/?Download)
>>>> >
>>>> > I could generate nitech_jp_atr503_m001.htsvoice from
>>>> > HTS_demo-NIT-ATR503-M001, but this error occured when open_jtalk
>>>> worked
>>>> >
>>>> > Error: HTS_Model_load_pdf: # of pdfs at 2-th state should be positive.
>>>> >
>>>> > Regards.
>>>> >
>>>> >
>>>> > 2016-04-05 12:12 GMT+09:00 <ope...@li...>:
>>>> >>
>>>> >> Hi,
>>>> >>
>>>> >> The link of HTS voice is shown in http://open-jtalk.sourceforge.net/
>>>> >>
>>>> >>
>>>> http://downloads.sourceforge.net/open-jtalk/hts_voice_nitech_jp_atr503_m001-1.05.tar.gz
>>>> >>
>>>> >> Regards,
>>>> >> Keiichiro Oura
>>>> >>
>>>> >>
>>>> >> 2016-04-05 11:45 GMT+09:00  <ope...@li...
>>>> >:
>>>> >> > Thanks.
>>>> >> >
>>>> >> > I installed
>>>> >> >
>>>> >> > hts_engine_API-1.10
>>>> >> > open_jtalk-1.09
>>>> >> >
>>>> >> > but I don't know which HTS voice I should install.
>>>> >> > I couldn't download HTS voice version 1.05 from
>>>> >> > hts_engine.sourceforge.net.
>>>> >> >
>>>> >> > Regards.
>>>> >> >
>>>> >> >
>>>> >> > 2016-04-04 10:37 GMT+09:00 <ope...@li...
>>>> >:
>>>> >> >>
>>>> >> >> Hi,
>>>> >> >>
>>>> >> >> It seems that incorrect HTS voice was created.
>>>> >> >> Could you try to use latest Open JTalk, hts_engine API, and HTS
>>>> voice
>>>> >> >> released in sourceforge?
>>>> >> >>
>>>> >> >> Regards,
>>>> >> >> Keiichiro Oura
>>>> >> >>
>>>> >> >>
>>>> >> >>
>>>> >> >> 2016-04-03 16:52 GMT+09:00  <
>>>> ope...@li...>:
>>>> >> >> > I made voice datas  by HTS-demo_NIT-ATR503-M001, then I could
>>>> >> >> > generate
>>>> >> >> > these
>>>> >> >> > files.
>>>> >> >> >
>>>> >> >> >
>>>> >> >> > $ ls
>>>> >> >> >
>>>> >> >> > dur.pdf     lf0.win1  lpf.win1  mgc.win3
>>>> >> >> > tree-gv-mgc.inf
>>>> >> >> >
>>>> >> >> > gv-lf0.pdf  lf0.win2  mgc.pdf   nitech_jp_atr503_m001.htsvoice
>>>> >> >> > tree-lf0.inf
>>>> >> >> >
>>>> >> >> > gv-mgc.pdf  lf0.win3  mgc.win1  tree-dur.inf
>>>> >> >> > tree-lpf.inf
>>>> >> >> >
>>>> >> >> > lf0.pdf     lpf.pdf   mgc.win2  tree-gv-lf0.inf
>>>> >> >> > tree-mgc.inf
>>>> >> >> >
>>>> >> >> >
>>>> >> >> > First, I tried to generate a wav file with this script.
>>>> >> >> >
>>>> >> >> >
>>>> >> >> > VOICE=/home/HTS/open_jtalk/bin
>>>> >> >> >
>>>> >> >> >
>>>> >> >> > echo $1 |nkf -e| /home/HTS/open_jtalk/bin/open_jtalk \
>>>> >> >> >
>>>> >> >> > -td $VOICE/tree-dur.inf \
>>>> >> >> >
>>>> >> >> > -tf $VOICE/tree-lf0.inf \
>>>> >> >> >
>>>> >> >> > -tm $VOICE/tree-mgc.inf \
>>>> >> >> >
>>>> >> >> > -tl $VOICE/tree-lpf.inf \
>>>> >> >> >
>>>> >> >> > -md $VOICE/dur.pdf \
>>>> >> >> >
>>>> >> >> > -mf $VOICE/lf0.pdf \
>>>> >> >> >
>>>> >> >> > -mm $VOICE/mgc.pdf \
>>>> >> >> >
>>>> >> >> > -ml $VOICE/lpf.pdf \
>>>> >> >> >
>>>> >> >> > -df $VOICE/lf0.win1 \
>>>> >> >> >
>>>> >> >> > -df $VOICE/lf0.win2 \
>>>> >> >> >
>>>> >> >> > -df $VOICE/lf0.win3 \
>>>> >> >> >
>>>> >> >> > -dm $VOICE/mgc.win1 \
>>>> >> >> >
>>>> >> >> > -dm $VOICE/mgc.win2 \
>>>> >> >> >
>>>> >> >> > -dm $VOICE/mgc.win3 \
>>>> >> >> >
>>>> >> >> > -dl $VOICE/lpf.win1 \
>>>> >> >> >
>>>> >> >> > -ef $VOICE/tree-gv-lf0.inf \
>>>> >> >> >
>>>> >> >> > -em $VOICE/tree-gv-mgc.inf \
>>>> >> >> >
>>>> >> >> > -cf $VOICE/gv-lf0.pdf \
>>>> >> >> >
>>>> >> >> > -cm $VOICE/gv-mgc.pdf \
>>>> >> >> >
>>>> >> >> > -k  $VOICE/gv-switch.inf \
>>>> >> >> >
>>>> >> >> > -x /home/HTS/open_jtalk/dic \
>>>> >> >> >
>>>> >> >> > -ow /home/HTS/public_html/sound/complex16_$2.wav \
>>>> >> >> >
>>>> >> >> > -ot /home/HTS/open_jtalk/out.log \
>>>> >> >> >
>>>> >> >> > -s 16000 -p 80 -a 0.0 \
>>>> >> >> >
>>>> >> >> > -g 1 -jm 0.55 -l
>>>> >> >> >
>>>> >> >> >
>>>> >> >> > but I couldn’t generate WAV, because this error occurred
>>>> >> >> >
>>>> >> >> >
>>>> >> >> > $ sh talk.sh
>>>> >> >> >
>>>> >> >> > Error: Dictionary or HTS voice cannot be loaded.
>>>> >> >> >
>>>> >> >> >
>>>> >> >> > so I found second way to generate wav with using htsvoice. This
>>>> >> >> > script
>>>> >> >> > is
>>>> >> >> > here
>>>> >> >> >
>>>> >> >> >
>>>> >> >> > VOICE=/home/HTS/open_jtalk/16kmodels/complex_16k
>>>> >> >> >
>>>> >> >> >
>>>> >> >> > echo $1 |nkf -e| /home/HTS/open_jtalk/bin/open_jtalk \
>>>> >> >> >
>>>> >> >> > -m nitech_jp_atr503_m001.htsvoice \
>>>> >> >> >
>>>> >> >> > -x /home/HTS/open_jtalk/dic \
>>>> >> >> >
>>>> >> >> > -ow /home/HTS/public_html/sound/complex16_$2.wav \
>>>> >> >> >
>>>> >> >> > -ot /home/HTS/open_jtalk/out.log \
>>>> >> >> >
>>>> >> >> > -s 16000 -p 80 -a 0.0 \
>>>> >> >> >
>>>> >> >> > -g 1 -jm 0.55 -l
>>>> >> >> >
>>>> >> >> >
>>>> >> >> > but I couldn’t generate it. The error is here.
>>>> >> >> >
>>>> >> >> >
>>>> >> >> > Error: HTS_Model_load_pdf: # of pdfs at 2-th state should be
>>>> >> >> > positive.
>>>> >> >> >
>>>> >> >> >
>>>> >> >> > How can I generate a WAV file?
>>>> >> >> >
>>>> >> >> >
>>>> >> >> >
>>>> >> >> >
>>>> >> >> >
>>>> ------------------------------------------------------------------------------
>>>> >> >> > Transform Data into Opportunity.
>>>> >> >> > Accelerate data analysis in your applications with
>>>> >> >> > Intel Data Analytics Acceleration Library.
>>>> >> >> > Click to learn more.
>>>> >> >> >
>>>> http://pubads.g.doubleclick.net/gampad/clk?id=278785471&iu=/4140
>>>> >> >> > _______________________________________________
>>>> >> >> > open-jtalk-users mailing list
>>>> >> >> > ope...@li...
>>>> >> >> > https://lists.sourceforge.net/lists/listinfo/open-jtalk-users
>>>> >> >> >
>>>> >> >>
>>>> >> >>
>>>> >> >>
>>>> >> >>
>>>> ------------------------------------------------------------------------------
>>>> >> >> _______________________________________________
>>>> >> >> open-jtalk-users mailing list
>>>> >> >> ope...@li...
>>>> >> >> https://lists.sourceforge.net/lists/listinfo/open-jtalk-users
>>>> >> >
>>>> >> >
>>>> >> >
>>>> >> >
>>>> >> >
>>>> ------------------------------------------------------------------------------
>>>> >> >
>>>> >> > _______________________________________________
>>>> >> > open-jtalk-users mailing list
>>>> >> > ope...@li...
>>>> >> > https://lists.sourceforge.net/lists/listinfo/open-jtalk-users
>>>> >> >
>>>> >>
>>>> >>
>>>> >>
>>>> ------------------------------------------------------------------------------
>>>> >> _______________________________________________
>>>> >> open-jtalk-users mailing list
>>>> >> ope...@li...
>>>> >> https://lists.sourceforge.net/lists/listinfo/open-jtalk-users
>>>> >
>>>> >
>>>> >
>>>> >
>>>> ------------------------------------------------------------------------------
>>>> >
>>>> > _______________________________________________
>>>> > open-jtalk-users mailing list
>>>> > ope...@li...
>>>> > https://lists.sourceforge.net/lists/listinfo/open-jtalk-users
>>>> >
>>>>
>>>>
>>>> ------------------------------------------------------------------------------
>>>> _______________________________________________
>>>> open-jtalk-users mailing list
>>>> ope...@li...
>>>> https://lists.sourceforge.net/lists/listinfo/open-jtalk-users
>>>>
>>>
>>>
>>>
>>> ------------------------------------------------------------------------------
>>>
>>> _______________________________________________
>>> open-jtalk-users mailing list
>>> ope...@li...
>>> https://lists.sourceforge.net/lists/listinfo/open-jtalk-users
>>>
>>>
>>
>>
>> ------------------------------------------------------------------------------
>>
>> _______________________________________________
>> open-jtalk-users mailing list
>> ope...@li...
>> https://lists.sourceforge.net/lists/listinfo/open-jtalk-users
>>
>>
>
>
> ------------------------------------------------------------------------------
>
> _______________________________________________
> open-jtalk-users mailing list
> ope...@li...
> https://lists.sourceforge.net/lists/listinfo/open-jtalk-users
>
>

Re: [open-jtalk-users] How to make wav from HTS demo?

[<ope...@li...>]

Thanks.

There are no waveforms instead of

.raw(/data/raw/nitech_jp_atr503_m001_*.raw)

and

.lab(/data/labels/mono/nitech_jp_atr503_m001_*.lab)

In execution, only raw file used.

Where are wav files in HTS-demo?

Regards.



2016-04-05 16:36 GMT+09:00 <ope...@li...>:

> Hi,
>
> The version seems ok.
> Let me know configure options&logs of demo scripts.
> Do you use little endian machine?
> Were waveforms (gen/qst001/ver1/hts_engine/*.wav) in demo scripts
> synthesized correctly?
>
> Regards,
> Keiichiro Oura
>
>
>
> 2016-04-05 15:11 GMT+09:00 <ope...@li...>:
>
>> Thanks.
>>
>> These are the version
>>
>> HTK-3.4.1
>> HDecode-3.4.1
>> HTS-2.3_for_HTK-3.4.1.patch
>> SPTK-3.9
>> hts_engine_API-1.10
>>
>> and I couldn't find out how to know the version of demo scripts,
>> but I downloaded the latest version. Here
>>
>> http://hts.sp.nitech.ac.jp/?Download
>> Speaker dependent training demo
>>
>>    - Japanese [image: file]tar.bz2
>>    <http://hts.sp.nitech.ac.jp/archives/2.3/HTS-demo_NIT-ATR503-M001.tar.bz2> (126
>>    MB)
>>
>>
>> Regards.
>>
>>
>> 2016-04-05 12:57 GMT+09:00 <ope...@li...>:
>>
>>> Hi,
>>>
>>> Let me know the version of HTS, SPTK, hts_engine API, and demo
>>> scripts, respectively.
>>>
>>> Regards,
>>> Keiichiro Oura
>>>
>>>
>>> 2016-04-05 12:51 GMT+09:00  <ope...@li...>:
>>> > Thanks.
>>> >
>>> > I could generate wav file by open_jtalk with this
>>> > hts_voice_nitech_jp_atr503_m001.
>>> > But I want to use HTS voice generated by
>>> > HTS_demo-NIT-ATR503-M001(http://hts.sp.nitech.ac.jp/?Download)
>>> >
>>> > I could generate nitech_jp_atr503_m001.htsvoice from
>>> > HTS_demo-NIT-ATR503-M001, but this error occured when open_jtalk worked
>>> >
>>> > Error: HTS_Model_load_pdf: # of pdfs at 2-th state should be positive.
>>> >
>>> > Regards.
>>> >
>>> >
>>> > 2016-04-05 12:12 GMT+09:00 <ope...@li...>:
>>> >>
>>> >> Hi,
>>> >>
>>> >> The link of HTS voice is shown in http://open-jtalk.sourceforge.net/
>>> >>
>>> >>
>>> http://downloads.sourceforge.net/open-jtalk/hts_voice_nitech_jp_atr503_m001-1.05.tar.gz
>>> >>
>>> >> Regards,
>>> >> Keiichiro Oura
>>> >>
>>> >>
>>> >> 2016-04-05 11:45 GMT+09:00  <ope...@li...>:
>>> >> > Thanks.
>>> >> >
>>> >> > I installed
>>> >> >
>>> >> > hts_engine_API-1.10
>>> >> > open_jtalk-1.09
>>> >> >
>>> >> > but I don't know which HTS voice I should install.
>>> >> > I couldn't download HTS voice version 1.05 from
>>> >> > hts_engine.sourceforge.net.
>>> >> >
>>> >> > Regards.
>>> >> >
>>> >> >
>>> >> > 2016-04-04 10:37 GMT+09:00 <ope...@li...
>>> >:
>>> >> >>
>>> >> >> Hi,
>>> >> >>
>>> >> >> It seems that incorrect HTS voice was created.
>>> >> >> Could you try to use latest Open JTalk, hts_engine API, and HTS
>>> voice
>>> >> >> released in sourceforge?
>>> >> >>
>>> >> >> Regards,
>>> >> >> Keiichiro Oura
>>> >> >>
>>> >> >>
>>> >> >>
>>> >> >> 2016-04-03 16:52 GMT+09:00  <
>>> ope...@li...>:
>>> >> >> > I made voice datas  by HTS-demo_NIT-ATR503-M001, then I could
>>> >> >> > generate
>>> >> >> > these
>>> >> >> > files.
>>> >> >> >
>>> >> >> >
>>> >> >> > $ ls
>>> >> >> >
>>> >> >> > dur.pdf     lf0.win1  lpf.win1  mgc.win3
>>> >> >> > tree-gv-mgc.inf
>>> >> >> >
>>> >> >> > gv-lf0.pdf  lf0.win2  mgc.pdf   nitech_jp_atr503_m001.htsvoice
>>> >> >> > tree-lf0.inf
>>> >> >> >
>>> >> >> > gv-mgc.pdf  lf0.win3  mgc.win1  tree-dur.inf
>>> >> >> > tree-lpf.inf
>>> >> >> >
>>> >> >> > lf0.pdf     lpf.pdf   mgc.win2  tree-gv-lf0.inf
>>> >> >> > tree-mgc.inf
>>> >> >> >
>>> >> >> >
>>> >> >> > First, I tried to generate a wav file with this script.
>>> >> >> >
>>> >> >> >
>>> >> >> > VOICE=/home/HTS/open_jtalk/bin
>>> >> >> >
>>> >> >> >
>>> >> >> > echo $1 |nkf -e| /home/HTS/open_jtalk/bin/open_jtalk \
>>> >> >> >
>>> >> >> > -td $VOICE/tree-dur.inf \
>>> >> >> >
>>> >> >> > -tf $VOICE/tree-lf0.inf \
>>> >> >> >
>>> >> >> > -tm $VOICE/tree-mgc.inf \
>>> >> >> >
>>> >> >> > -tl $VOICE/tree-lpf.inf \
>>> >> >> >
>>> >> >> > -md $VOICE/dur.pdf \
>>> >> >> >
>>> >> >> > -mf $VOICE/lf0.pdf \
>>> >> >> >
>>> >> >> > -mm $VOICE/mgc.pdf \
>>> >> >> >
>>> >> >> > -ml $VOICE/lpf.pdf \
>>> >> >> >
>>> >> >> > -df $VOICE/lf0.win1 \
>>> >> >> >
>>> >> >> > -df $VOICE/lf0.win2 \
>>> >> >> >
>>> >> >> > -df $VOICE/lf0.win3 \
>>> >> >> >
>>> >> >> > -dm $VOICE/mgc.win1 \
>>> >> >> >
>>> >> >> > -dm $VOICE/mgc.win2 \
>>> >> >> >
>>> >> >> > -dm $VOICE/mgc.win3 \
>>> >> >> >
>>> >> >> > -dl $VOICE/lpf.win1 \
>>> >> >> >
>>> >> >> > -ef $VOICE/tree-gv-lf0.inf \
>>> >> >> >
>>> >> >> > -em $VOICE/tree-gv-mgc.inf \
>>> >> >> >
>>> >> >> > -cf $VOICE/gv-lf0.pdf \
>>> >> >> >
>>> >> >> > -cm $VOICE/gv-mgc.pdf \
>>> >> >> >
>>> >> >> > -k  $VOICE/gv-switch.inf \
>>> >> >> >
>>> >> >> > -x /home/HTS/open_jtalk/dic \
>>> >> >> >
>>> >> >> > -ow /home/HTS/public_html/sound/complex16_$2.wav \
>>> >> >> >
>>> >> >> > -ot /home/HTS/open_jtalk/out.log \
>>> >> >> >
>>> >> >> > -s 16000 -p 80 -a 0.0 \
>>> >> >> >
>>> >> >> > -g 1 -jm 0.55 -l
>>> >> >> >
>>> >> >> >
>>> >> >> > but I couldn’t generate WAV, because this error occurred
>>> >> >> >
>>> >> >> >
>>> >> >> > $ sh talk.sh
>>> >> >> >
>>> >> >> > Error: Dictionary or HTS voice cannot be loaded.
>>> >> >> >
>>> >> >> >
>>> >> >> > so I found second way to generate wav with using htsvoice. This
>>> >> >> > script
>>> >> >> > is
>>> >> >> > here
>>> >> >> >
>>> >> >> >
>>> >> >> > VOICE=/home/HTS/open_jtalk/16kmodels/complex_16k
>>> >> >> >
>>> >> >> >
>>> >> >> > echo $1 |nkf -e| /home/HTS/open_jtalk/bin/open_jtalk \
>>> >> >> >
>>> >> >> > -m nitech_jp_atr503_m001.htsvoice \
>>> >> >> >
>>> >> >> > -x /home/HTS/open_jtalk/dic \
>>> >> >> >
>>> >> >> > -ow /home/HTS/public_html/sound/complex16_$2.wav \
>>> >> >> >
>>> >> >> > -ot /home/HTS/open_jtalk/out.log \
>>> >> >> >
>>> >> >> > -s 16000 -p 80 -a 0.0 \
>>> >> >> >
>>> >> >> > -g 1 -jm 0.55 -l
>>> >> >> >
>>> >> >> >
>>> >> >> > but I couldn’t generate it. The error is here.
>>> >> >> >
>>> >> >> >
>>> >> >> > Error: HTS_Model_load_pdf: # of pdfs at 2-th state should be
>>> >> >> > positive.
>>> >> >> >
>>> >> >> >
>>> >> >> > How can I generate a WAV file?
>>> >> >> >
>>> >> >> >
>>> >> >> >
>>> >> >> >
>>> >> >> >
>>> ------------------------------------------------------------------------------
>>> >> >> > Transform Data into Opportunity.
>>> >> >> > Accelerate data analysis in your applications with
>>> >> >> > Intel Data Analytics Acceleration Library.
>>> >> >> > Click to learn more.
>>> >> >> > http://pubads.g.doubleclick.net/gampad/clk?id=278785471&iu=/4140
>>> >> >> > _______________________________________________
>>> >> >> > open-jtalk-users mailing list
>>> >> >> > ope...@li...
>>> >> >> > https://lists.sourceforge.net/lists/listinfo/open-jtalk-users
>>> >> >> >
>>> >> >>
>>> >> >>
>>> >> >>
>>> >> >>
>>> ------------------------------------------------------------------------------
>>> >> >> _______________________________________________
>>> >> >> open-jtalk-users mailing list
>>> >> >> ope...@li...
>>> >> >> https://lists.sourceforge.net/lists/listinfo/open-jtalk-users
>>> >> >
>>> >> >
>>> >> >
>>> >> >
>>> >> >
>>> ------------------------------------------------------------------------------
>>> >> >
>>> >> > _______________________________________________
>>> >> > open-jtalk-users mailing list
>>> >> > ope...@li...
>>> >> > https://lists.sourceforge.net/lists/listinfo/open-jtalk-users
>>> >> >
>>> >>
>>> >>
>>> >>
>>> ------------------------------------------------------------------------------
>>> >> _______________________________________________
>>> >> open-jtalk-users mailing list
>>> >> ope...@li...
>>> >> https://lists.sourceforge.net/lists/listinfo/open-jtalk-users
>>> >
>>> >
>>> >
>>> >
>>> ------------------------------------------------------------------------------
>>> >
>>> > _______________________________________________
>>> > open-jtalk-users mailing list
>>> > ope...@li...
>>> > https://lists.sourceforge.net/lists/listinfo/open-jtalk-users
>>> >
>>>
>>>
>>> ------------------------------------------------------------------------------
>>> _______________________________________________
>>> open-jtalk-users mailing list
>>> ope...@li...
>>> https://lists.sourceforge.net/lists/listinfo/open-jtalk-users
>>>
>>
>>
>>
>> ------------------------------------------------------------------------------
>>
>> _______________________________________________
>> open-jtalk-users mailing list
>> ope...@li...
>> https://lists.sourceforge.net/lists/listinfo/open-jtalk-users
>>
>>
>
>
> ------------------------------------------------------------------------------
>
> _______________________________________________
> open-jtalk-users mailing list
> ope...@li...
> https://lists.sourceforge.net/lists/listinfo/open-jtalk-users
>
>

Re: [open-jtalk-users] How to make wav from HTS demo?

[<ope...@li...>]

Hi,

The version seems ok.
Let me know configure options&logs of demo scripts.
Do you use little endian machine?
Were waveforms (gen/qst001/ver1/hts_engine/*.wav) in demo scripts
synthesized correctly?

Regards,
Keiichiro Oura



2016-04-05 15:11 GMT+09:00 <ope...@li...>:

> Thanks.
>
> These are the version
>
> HTK-3.4.1
> HDecode-3.4.1
> HTS-2.3_for_HTK-3.4.1.patch
> SPTK-3.9
> hts_engine_API-1.10
>
> and I couldn't find out how to know the version of demo scripts,
> but I downloaded the latest version. Here
>
> http://hts.sp.nitech.ac.jp/?Download
> Speaker dependent training demo
>
>    - Japanese [image: file]tar.bz2
>    <http://hts.sp.nitech.ac.jp/archives/2.3/HTS-demo_NIT-ATR503-M001.tar.bz2> (126
>    MB)
>
>
> Regards.
>
>
> 2016-04-05 12:57 GMT+09:00 <ope...@li...>:
>
>> Hi,
>>
>> Let me know the version of HTS, SPTK, hts_engine API, and demo
>> scripts, respectively.
>>
>> Regards,
>> Keiichiro Oura
>>
>>
>> 2016-04-05 12:51 GMT+09:00  <ope...@li...>:
>> > Thanks.
>> >
>> > I could generate wav file by open_jtalk with this
>> > hts_voice_nitech_jp_atr503_m001.
>> > But I want to use HTS voice generated by
>> > HTS_demo-NIT-ATR503-M001(http://hts.sp.nitech.ac.jp/?Download)
>> >
>> > I could generate nitech_jp_atr503_m001.htsvoice from
>> > HTS_demo-NIT-ATR503-M001, but this error occured when open_jtalk worked
>> >
>> > Error: HTS_Model_load_pdf: # of pdfs at 2-th state should be positive.
>> >
>> > Regards.
>> >
>> >
>> > 2016-04-05 12:12 GMT+09:00 <ope...@li...>:
>> >>
>> >> Hi,
>> >>
>> >> The link of HTS voice is shown in http://open-jtalk.sourceforge.net/
>> >>
>> >>
>> http://downloads.sourceforge.net/open-jtalk/hts_voice_nitech_jp_atr503_m001-1.05.tar.gz
>> >>
>> >> Regards,
>> >> Keiichiro Oura
>> >>
>> >>
>> >> 2016-04-05 11:45 GMT+09:00  <ope...@li...>:
>> >> > Thanks.
>> >> >
>> >> > I installed
>> >> >
>> >> > hts_engine_API-1.10
>> >> > open_jtalk-1.09
>> >> >
>> >> > but I don't know which HTS voice I should install.
>> >> > I couldn't download HTS voice version 1.05 from
>> >> > hts_engine.sourceforge.net.
>> >> >
>> >> > Regards.
>> >> >
>> >> >
>> >> > 2016-04-04 10:37 GMT+09:00 <ope...@li...>:
>> >> >>
>> >> >> Hi,
>> >> >>
>> >> >> It seems that incorrect HTS voice was created.
>> >> >> Could you try to use latest Open JTalk, hts_engine API, and HTS
>> voice
>> >> >> released in sourceforge?
>> >> >>
>> >> >> Regards,
>> >> >> Keiichiro Oura
>> >> >>
>> >> >>
>> >> >>
>> >> >> 2016-04-03 16:52 GMT+09:00  <ope...@li...
>> >:
>> >> >> > I made voice datas  by HTS-demo_NIT-ATR503-M001, then I could
>> >> >> > generate
>> >> >> > these
>> >> >> > files.
>> >> >> >
>> >> >> >
>> >> >> > $ ls
>> >> >> >
>> >> >> > dur.pdf     lf0.win1  lpf.win1  mgc.win3
>> >> >> > tree-gv-mgc.inf
>> >> >> >
>> >> >> > gv-lf0.pdf  lf0.win2  mgc.pdf   nitech_jp_atr503_m001.htsvoice
>> >> >> > tree-lf0.inf
>> >> >> >
>> >> >> > gv-mgc.pdf  lf0.win3  mgc.win1  tree-dur.inf
>> >> >> > tree-lpf.inf
>> >> >> >
>> >> >> > lf0.pdf     lpf.pdf   mgc.win2  tree-gv-lf0.inf
>> >> >> > tree-mgc.inf
>> >> >> >
>> >> >> >
>> >> >> > First, I tried to generate a wav file with this script.
>> >> >> >
>> >> >> >
>> >> >> > VOICE=/home/HTS/open_jtalk/bin
>> >> >> >
>> >> >> >
>> >> >> > echo $1 |nkf -e| /home/HTS/open_jtalk/bin/open_jtalk \
>> >> >> >
>> >> >> > -td $VOICE/tree-dur.inf \
>> >> >> >
>> >> >> > -tf $VOICE/tree-lf0.inf \
>> >> >> >
>> >> >> > -tm $VOICE/tree-mgc.inf \
>> >> >> >
>> >> >> > -tl $VOICE/tree-lpf.inf \
>> >> >> >
>> >> >> > -md $VOICE/dur.pdf \
>> >> >> >
>> >> >> > -mf $VOICE/lf0.pdf \
>> >> >> >
>> >> >> > -mm $VOICE/mgc.pdf \
>> >> >> >
>> >> >> > -ml $VOICE/lpf.pdf \
>> >> >> >
>> >> >> > -df $VOICE/lf0.win1 \
>> >> >> >
>> >> >> > -df $VOICE/lf0.win2 \
>> >> >> >
>> >> >> > -df $VOICE/lf0.win3 \
>> >> >> >
>> >> >> > -dm $VOICE/mgc.win1 \
>> >> >> >
>> >> >> > -dm $VOICE/mgc.win2 \
>> >> >> >
>> >> >> > -dm $VOICE/mgc.win3 \
>> >> >> >
>> >> >> > -dl $VOICE/lpf.win1 \
>> >> >> >
>> >> >> > -ef $VOICE/tree-gv-lf0.inf \
>> >> >> >
>> >> >> > -em $VOICE/tree-gv-mgc.inf \
>> >> >> >
>> >> >> > -cf $VOICE/gv-lf0.pdf \
>> >> >> >
>> >> >> > -cm $VOICE/gv-mgc.pdf \
>> >> >> >
>> >> >> > -k  $VOICE/gv-switch.inf \
>> >> >> >
>> >> >> > -x /home/HTS/open_jtalk/dic \
>> >> >> >
>> >> >> > -ow /home/HTS/public_html/sound/complex16_$2.wav \
>> >> >> >
>> >> >> > -ot /home/HTS/open_jtalk/out.log \
>> >> >> >
>> >> >> > -s 16000 -p 80 -a 0.0 \
>> >> >> >
>> >> >> > -g 1 -jm 0.55 -l
>> >> >> >
>> >> >> >
>> >> >> > but I couldn’t generate WAV, because this error occurred
>> >> >> >
>> >> >> >
>> >> >> > $ sh talk.sh
>> >> >> >
>> >> >> > Error: Dictionary or HTS voice cannot be loaded.
>> >> >> >
>> >> >> >
>> >> >> > so I found second way to generate wav with using htsvoice. This
>> >> >> > script
>> >> >> > is
>> >> >> > here
>> >> >> >
>> >> >> >
>> >> >> > VOICE=/home/HTS/open_jtalk/16kmodels/complex_16k
>> >> >> >
>> >> >> >
>> >> >> > echo $1 |nkf -e| /home/HTS/open_jtalk/bin/open_jtalk \
>> >> >> >
>> >> >> > -m nitech_jp_atr503_m001.htsvoice \
>> >> >> >
>> >> >> > -x /home/HTS/open_jtalk/dic \
>> >> >> >
>> >> >> > -ow /home/HTS/public_html/sound/complex16_$2.wav \
>> >> >> >
>> >> >> > -ot /home/HTS/open_jtalk/out.log \
>> >> >> >
>> >> >> > -s 16000 -p 80 -a 0.0 \
>> >> >> >
>> >> >> > -g 1 -jm 0.55 -l
>> >> >> >
>> >> >> >
>> >> >> > but I couldn’t generate it. The error is here.
>> >> >> >
>> >> >> >
>> >> >> > Error: HTS_Model_load_pdf: # of pdfs at 2-th state should be
>> >> >> > positive.
>> >> >> >
>> >> >> >
>> >> >> > How can I generate a WAV file?
>> >> >> >
>> >> >> >
>> >> >> >
>> >> >> >
>> >> >> >
>> ------------------------------------------------------------------------------
>> >> >> > Transform Data into Opportunity.
>> >> >> > Accelerate data analysis in your applications with
>> >> >> > Intel Data Analytics Acceleration Library.
>> >> >> > Click to learn more.
>> >> >> > http://pubads.g.doubleclick.net/gampad/clk?id=278785471&iu=/4140
>> >> >> > _______________________________________________
>> >> >> > open-jtalk-users mailing list
>> >> >> > ope...@li...
>> >> >> > https://lists.sourceforge.net/lists/listinfo/open-jtalk-users
>> >> >> >
>> >> >>
>> >> >>
>> >> >>
>> >> >>
>> ------------------------------------------------------------------------------
>> >> >> _______________________________________________
>> >> >> open-jtalk-users mailing list
>> >> >> ope...@li...
>> >> >> https://lists.sourceforge.net/lists/listinfo/open-jtalk-users
>> >> >
>> >> >
>> >> >
>> >> >
>> >> >
>> ------------------------------------------------------------------------------
>> >> >
>> >> > _______________________________________________
>> >> > open-jtalk-users mailing list
>> >> > ope...@li...
>> >> > https://lists.sourceforge.net/lists/listinfo/open-jtalk-users
>> >> >
>> >>
>> >>
>> >>
>> ------------------------------------------------------------------------------
>> >> _______________________________________________
>> >> open-jtalk-users mailing list
>> >> ope...@li...
>> >> https://lists.sourceforge.net/lists/listinfo/open-jtalk-users
>> >
>> >
>> >
>> >
>> ------------------------------------------------------------------------------
>> >
>> > _______________________________________________
>> > open-jtalk-users mailing list
>> > ope...@li...
>> > https://lists.sourceforge.net/lists/listinfo/open-jtalk-users
>> >
>>
>>
>> ------------------------------------------------------------------------------
>> _______________________________________________
>> open-jtalk-users mailing list
>> ope...@li...
>> https://lists.sourceforge.net/lists/listinfo/open-jtalk-users
>>
>
>
>
> ------------------------------------------------------------------------------
>
> _______________________________________________
> open-jtalk-users mailing list
> ope...@li...
> https://lists.sourceforge.net/lists/listinfo/open-jtalk-users
>
>

Re: [open-jtalk-users] How to make wav from HTS demo?

[<ope...@li...>]

Thanks.

These are the version

HTK-3.4.1
HDecode-3.4.1
HTS-2.3_for_HTK-3.4.1.patch
SPTK-3.9
hts_engine_API-1.10

and I couldn't find out how to know the version of demo scripts,
but I downloaded the latest version. Here

http://hts.sp.nitech.ac.jp/?Download
Speaker dependent training demo

   - Japanese [image: file]tar.bz2
   <http://hts.sp.nitech.ac.jp/archives/2.3/HTS-demo_NIT-ATR503-M001.tar.bz2>
(126
   MB)


Regards.


2016-04-05 12:57 GMT+09:00 <ope...@li...>:

> Hi,
>
> Let me know the version of HTS, SPTK, hts_engine API, and demo
> scripts, respectively.
>
> Regards,
> Keiichiro Oura
>
>
> 2016-04-05 12:51 GMT+09:00  <ope...@li...>:
> > Thanks.
> >
> > I could generate wav file by open_jtalk with this
> > hts_voice_nitech_jp_atr503_m001.
> > But I want to use HTS voice generated by
> > HTS_demo-NIT-ATR503-M001(http://hts.sp.nitech.ac.jp/?Download)
> >
> > I could generate nitech_jp_atr503_m001.htsvoice from
> > HTS_demo-NIT-ATR503-M001, but this error occured when open_jtalk worked
> >
> > Error: HTS_Model_load_pdf: # of pdfs at 2-th state should be positive.
> >
> > Regards.
> >
> >
> > 2016-04-05 12:12 GMT+09:00 <ope...@li...>:
> >>
> >> Hi,
> >>
> >> The link of HTS voice is shown in http://open-jtalk.sourceforge.net/
> >>
> >>
> http://downloads.sourceforge.net/open-jtalk/hts_voice_nitech_jp_atr503_m001-1.05.tar.gz
> >>
> >> Regards,
> >> Keiichiro Oura
> >>
> >>
> >> 2016-04-05 11:45 GMT+09:00  <ope...@li...>:
> >> > Thanks.
> >> >
> >> > I installed
> >> >
> >> > hts_engine_API-1.10
> >> > open_jtalk-1.09
> >> >
> >> > but I don't know which HTS voice I should install.
> >> > I couldn't download HTS voice version 1.05 from
> >> > hts_engine.sourceforge.net.
> >> >
> >> > Regards.
> >> >
> >> >
> >> > 2016-04-04 10:37 GMT+09:00 <ope...@li...>:
> >> >>
> >> >> Hi,
> >> >>
> >> >> It seems that incorrect HTS voice was created.
> >> >> Could you try to use latest Open JTalk, hts_engine API, and HTS voice
> >> >> released in sourceforge?
> >> >>
> >> >> Regards,
> >> >> Keiichiro Oura
> >> >>
> >> >>
> >> >>
> >> >> 2016-04-03 16:52 GMT+09:00  <ope...@li...
> >:
> >> >> > I made voice datas  by HTS-demo_NIT-ATR503-M001, then I could
> >> >> > generate
> >> >> > these
> >> >> > files.
> >> >> >
> >> >> >
> >> >> > $ ls
> >> >> >
> >> >> > dur.pdf     lf0.win1  lpf.win1  mgc.win3
> >> >> > tree-gv-mgc.inf
> >> >> >
> >> >> > gv-lf0.pdf  lf0.win2  mgc.pdf   nitech_jp_atr503_m001.htsvoice
> >> >> > tree-lf0.inf
> >> >> >
> >> >> > gv-mgc.pdf  lf0.win3  mgc.win1  tree-dur.inf
> >> >> > tree-lpf.inf
> >> >> >
> >> >> > lf0.pdf     lpf.pdf   mgc.win2  tree-gv-lf0.inf
> >> >> > tree-mgc.inf
> >> >> >
> >> >> >
> >> >> > First, I tried to generate a wav file with this script.
> >> >> >
> >> >> >
> >> >> > VOICE=/home/HTS/open_jtalk/bin
> >> >> >
> >> >> >
> >> >> > echo $1 |nkf -e| /home/HTS/open_jtalk/bin/open_jtalk \
> >> >> >
> >> >> > -td $VOICE/tree-dur.inf \
> >> >> >
> >> >> > -tf $VOICE/tree-lf0.inf \
> >> >> >
> >> >> > -tm $VOICE/tree-mgc.inf \
> >> >> >
> >> >> > -tl $VOICE/tree-lpf.inf \
> >> >> >
> >> >> > -md $VOICE/dur.pdf \
> >> >> >
> >> >> > -mf $VOICE/lf0.pdf \
> >> >> >
> >> >> > -mm $VOICE/mgc.pdf \
> >> >> >
> >> >> > -ml $VOICE/lpf.pdf \
> >> >> >
> >> >> > -df $VOICE/lf0.win1 \
> >> >> >
> >> >> > -df $VOICE/lf0.win2 \
> >> >> >
> >> >> > -df $VOICE/lf0.win3 \
> >> >> >
> >> >> > -dm $VOICE/mgc.win1 \
> >> >> >
> >> >> > -dm $VOICE/mgc.win2 \
> >> >> >
> >> >> > -dm $VOICE/mgc.win3 \
> >> >> >
> >> >> > -dl $VOICE/lpf.win1 \
> >> >> >
> >> >> > -ef $VOICE/tree-gv-lf0.inf \
> >> >> >
> >> >> > -em $VOICE/tree-gv-mgc.inf \
> >> >> >
> >> >> > -cf $VOICE/gv-lf0.pdf \
> >> >> >
> >> >> > -cm $VOICE/gv-mgc.pdf \
> >> >> >
> >> >> > -k  $VOICE/gv-switch.inf \
> >> >> >
> >> >> > -x /home/HTS/open_jtalk/dic \
> >> >> >
> >> >> > -ow /home/HTS/public_html/sound/complex16_$2.wav \
> >> >> >
> >> >> > -ot /home/HTS/open_jtalk/out.log \
> >> >> >
> >> >> > -s 16000 -p 80 -a 0.0 \
> >> >> >
> >> >> > -g 1 -jm 0.55 -l
> >> >> >
> >> >> >
> >> >> > but I couldn’t generate WAV, because this error occurred
> >> >> >
> >> >> >
> >> >> > $ sh talk.sh
> >> >> >
> >> >> > Error: Dictionary or HTS voice cannot be loaded.
> >> >> >
> >> >> >
> >> >> > so I found second way to generate wav with using htsvoice. This
> >> >> > script
> >> >> > is
> >> >> > here
> >> >> >
> >> >> >
> >> >> > VOICE=/home/HTS/open_jtalk/16kmodels/complex_16k
> >> >> >
> >> >> >
> >> >> > echo $1 |nkf -e| /home/HTS/open_jtalk/bin/open_jtalk \
> >> >> >
> >> >> > -m nitech_jp_atr503_m001.htsvoice \
> >> >> >
> >> >> > -x /home/HTS/open_jtalk/dic \
> >> >> >
> >> >> > -ow /home/HTS/public_html/sound/complex16_$2.wav \
> >> >> >
> >> >> > -ot /home/HTS/open_jtalk/out.log \
> >> >> >
> >> >> > -s 16000 -p 80 -a 0.0 \
> >> >> >
> >> >> > -g 1 -jm 0.55 -l
> >> >> >
> >> >> >
> >> >> > but I couldn’t generate it. The error is here.
> >> >> >
> >> >> >
> >> >> > Error: HTS_Model_load_pdf: # of pdfs at 2-th state should be
> >> >> > positive.
> >> >> >
> >> >> >
> >> >> > How can I generate a WAV file?
> >> >> >
> >> >> >
> >> >> >
> >> >> >
> >> >> >
> ------------------------------------------------------------------------------
> >> >> > Transform Data into Opportunity.
> >> >> > Accelerate data analysis in your applications with
> >> >> > Intel Data Analytics Acceleration Library.
> >> >> > Click to learn more.
> >> >> > http://pubads.g.doubleclick.net/gampad/clk?id=278785471&iu=/4140
> >> >> > _______________________________________________
> >> >> > open-jtalk-users mailing list
> >> >> > ope...@li...
> >> >> > https://lists.sourceforge.net/lists/listinfo/open-jtalk-users
> >> >> >
> >> >>
> >> >>
> >> >>
> >> >>
> ------------------------------------------------------------------------------
> >> >> _______________________________________________
> >> >> open-jtalk-users mailing list
> >> >> ope...@li...
> >> >> https://lists.sourceforge.net/lists/listinfo/open-jtalk-users
> >> >
> >> >
> >> >
> >> >
> >> >
> ------------------------------------------------------------------------------
> >> >
> >> > _______________________________________________
> >> > open-jtalk-users mailing list
> >> > ope...@li...
> >> > https://lists.sourceforge.net/lists/listinfo/open-jtalk-users
> >> >
> >>
> >>
> >>
> ------------------------------------------------------------------------------
> >> _______________________________________________
> >> open-jtalk-users mailing list
> >> ope...@li...
> >> https://lists.sourceforge.net/lists/listinfo/open-jtalk-users
> >
> >
> >
> >
> ------------------------------------------------------------------------------
> >
> > _______________________________________________
> > open-jtalk-users mailing list
> > ope...@li...
> > https://lists.sourceforge.net/lists/listinfo/open-jtalk-users
> >
>
>
> ------------------------------------------------------------------------------
> _______________________________________________
> open-jtalk-users mailing list
> ope...@li...
> https://lists.sourceforge.net/lists/listinfo/open-jtalk-users
>

Re: [open-jtalk-users] How to make wav from HTS demo?

[<ope...@li...>]

Hi,

Let me know the version of HTS, SPTK, hts_engine API, and demo
scripts, respectively.

Regards,
Keiichiro Oura


2016-04-05 12:51 GMT+09:00  <ope...@li...>:
> Thanks.
>
> I could generate wav file by open_jtalk with this
> hts_voice_nitech_jp_atr503_m001.
> But I want to use HTS voice generated by
> HTS_demo-NIT-ATR503-M001(http://hts.sp.nitech.ac.jp/?Download)
>
> I could generate nitech_jp_atr503_m001.htsvoice from
> HTS_demo-NIT-ATR503-M001, but this error occured when open_jtalk worked
>
> Error: HTS_Model_load_pdf: # of pdfs at 2-th state should be positive.
>
> Regards.
>
>
> 2016-04-05 12:12 GMT+09:00 <ope...@li...>:
>>
>> Hi,
>>
>> The link of HTS voice is shown in http://open-jtalk.sourceforge.net/
>>
>> http://downloads.sourceforge.net/open-jtalk/hts_voice_nitech_jp_atr503_m001-1.05.tar.gz
>>
>> Regards,
>> Keiichiro Oura
>>
>>
>> 2016-04-05 11:45 GMT+09:00  <ope...@li...>:
>> > Thanks.
>> >
>> > I installed
>> >
>> > hts_engine_API-1.10
>> > open_jtalk-1.09
>> >
>> > but I don't know which HTS voice I should install.
>> > I couldn't download HTS voice version 1.05 from
>> > hts_engine.sourceforge.net.
>> >
>> > Regards.
>> >
>> >
>> > 2016-04-04 10:37 GMT+09:00 <ope...@li...>:
>> >>
>> >> Hi,
>> >>
>> >> It seems that incorrect HTS voice was created.
>> >> Could you try to use latest Open JTalk, hts_engine API, and HTS voice
>> >> released in sourceforge?
>> >>
>> >> Regards,
>> >> Keiichiro Oura
>> >>
>> >>
>> >>
>> >> 2016-04-03 16:52 GMT+09:00  <ope...@li...>:
>> >> > I made voice datas  by HTS-demo_NIT-ATR503-M001, then I could
>> >> > generate
>> >> > these
>> >> > files.
>> >> >
>> >> >
>> >> > $ ls
>> >> >
>> >> > dur.pdf     lf0.win1  lpf.win1  mgc.win3
>> >> > tree-gv-mgc.inf
>> >> >
>> >> > gv-lf0.pdf  lf0.win2  mgc.pdf   nitech_jp_atr503_m001.htsvoice
>> >> > tree-lf0.inf
>> >> >
>> >> > gv-mgc.pdf  lf0.win3  mgc.win1  tree-dur.inf
>> >> > tree-lpf.inf
>> >> >
>> >> > lf0.pdf     lpf.pdf   mgc.win2  tree-gv-lf0.inf
>> >> > tree-mgc.inf
>> >> >
>> >> >
>> >> > First, I tried to generate a wav file with this script.
>> >> >
>> >> >
>> >> > VOICE=/home/HTS/open_jtalk/bin
>> >> >
>> >> >
>> >> > echo $1 |nkf -e| /home/HTS/open_jtalk/bin/open_jtalk \
>> >> >
>> >> > -td $VOICE/tree-dur.inf \
>> >> >
>> >> > -tf $VOICE/tree-lf0.inf \
>> >> >
>> >> > -tm $VOICE/tree-mgc.inf \
>> >> >
>> >> > -tl $VOICE/tree-lpf.inf \
>> >> >
>> >> > -md $VOICE/dur.pdf \
>> >> >
>> >> > -mf $VOICE/lf0.pdf \
>> >> >
>> >> > -mm $VOICE/mgc.pdf \
>> >> >
>> >> > -ml $VOICE/lpf.pdf \
>> >> >
>> >> > -df $VOICE/lf0.win1 \
>> >> >
>> >> > -df $VOICE/lf0.win2 \
>> >> >
>> >> > -df $VOICE/lf0.win3 \
>> >> >
>> >> > -dm $VOICE/mgc.win1 \
>> >> >
>> >> > -dm $VOICE/mgc.win2 \
>> >> >
>> >> > -dm $VOICE/mgc.win3 \
>> >> >
>> >> > -dl $VOICE/lpf.win1 \
>> >> >
>> >> > -ef $VOICE/tree-gv-lf0.inf \
>> >> >
>> >> > -em $VOICE/tree-gv-mgc.inf \
>> >> >
>> >> > -cf $VOICE/gv-lf0.pdf \
>> >> >
>> >> > -cm $VOICE/gv-mgc.pdf \
>> >> >
>> >> > -k  $VOICE/gv-switch.inf \
>> >> >
>> >> > -x /home/HTS/open_jtalk/dic \
>> >> >
>> >> > -ow /home/HTS/public_html/sound/complex16_$2.wav \
>> >> >
>> >> > -ot /home/HTS/open_jtalk/out.log \
>> >> >
>> >> > -s 16000 -p 80 -a 0.0 \
>> >> >
>> >> > -g 1 -jm 0.55 -l
>> >> >
>> >> >
>> >> > but I couldn’t generate WAV, because this error occurred
>> >> >
>> >> >
>> >> > $ sh talk.sh
>> >> >
>> >> > Error: Dictionary or HTS voice cannot be loaded.
>> >> >
>> >> >
>> >> > so I found second way to generate wav with using htsvoice. This
>> >> > script
>> >> > is
>> >> > here
>> >> >
>> >> >
>> >> > VOICE=/home/HTS/open_jtalk/16kmodels/complex_16k
>> >> >
>> >> >
>> >> > echo $1 |nkf -e| /home/HTS/open_jtalk/bin/open_jtalk \
>> >> >
>> >> > -m nitech_jp_atr503_m001.htsvoice \
>> >> >
>> >> > -x /home/HTS/open_jtalk/dic \
>> >> >
>> >> > -ow /home/HTS/public_html/sound/complex16_$2.wav \
>> >> >
>> >> > -ot /home/HTS/open_jtalk/out.log \
>> >> >
>> >> > -s 16000 -p 80 -a 0.0 \
>> >> >
>> >> > -g 1 -jm 0.55 -l
>> >> >
>> >> >
>> >> > but I couldn’t generate it. The error is here.
>> >> >
>> >> >
>> >> > Error: HTS_Model_load_pdf: # of pdfs at 2-th state should be
>> >> > positive.
>> >> >
>> >> >
>> >> > How can I generate a WAV file?
>> >> >
>> >> >
>> >> >
>> >> >
>> >> > ------------------------------------------------------------------------------
>> >> > Transform Data into Opportunity.
>> >> > Accelerate data analysis in your applications with
>> >> > Intel Data Analytics Acceleration Library.
>> >> > Click to learn more.
>> >> > http://pubads.g.doubleclick.net/gampad/clk?id=278785471&iu=/4140
>> >> > _______________________________________________
>> >> > open-jtalk-users mailing list
>> >> > ope...@li...
>> >> > https://lists.sourceforge.net/lists/listinfo/open-jtalk-users
>> >> >
>> >>
>> >>
>> >>
>> >> ------------------------------------------------------------------------------
>> >> _______________________________________________
>> >> open-jtalk-users mailing list
>> >> ope...@li...
>> >> https://lists.sourceforge.net/lists/listinfo/open-jtalk-users
>> >
>> >
>> >
>> >
>> > ------------------------------------------------------------------------------
>> >
>> > _______________________________________________
>> > open-jtalk-users mailing list
>> > ope...@li...
>> > https://lists.sourceforge.net/lists/listinfo/open-jtalk-users
>> >
>>
>>
>> ------------------------------------------------------------------------------
>> _______________________________________________
>> open-jtalk-users mailing list
>> ope...@li...
>> https://lists.sourceforge.net/lists/listinfo/open-jtalk-users
>
>
>
> ------------------------------------------------------------------------------
>
> _______________________________________________
> open-jtalk-users mailing list
> ope...@li...
> https://lists.sourceforge.net/lists/listinfo/open-jtalk-users
>