New release of QTAU(piano roll editor) and espeak-sg (singing generator)


Tobias Platen
 

Today I just made a new release of my piano roll editor for singing
synthesis called QTAU, including its dependencies Sekai and Sinsy.
QTAU can make espeak-ng sing, using a patch I had written some time
ago, future versions will also support Festival and MBROLA voices.
The singing generator source code is at
https://notabug.org/isengaara/sekai/src/master/bin/espeak-sg.cpp and
currently does not work with klatt voices yet. In the next few month I
plan to write a patch to support singing with klatt voices too.

https://notabug.org/isengaara/qtau
https://notabug.org/isengaara/sinsy
https://notabug.org/isengaara/sekai


Russell James
 

Hi Tobias

This sounds awesome!

I always wondered what it would take to make he speak sing instead of talk :-)

Can you share an example song?

I imagine when people hear this it will go viral! :-)

Thank you for sharing

Russ


On Sat, May 1, 2021, 7:11 AM Tobias Platen <tplaten@...> wrote:
Today I just made a new release of my piano roll editor for singing
synthesis called QTAU, including its dependencies Sekai and Sinsy.
QTAU can make espeak-ng sing, using a patch I had written some time
ago, future versions will also support Festival and MBROLA voices.
The singing generator source code is at
https://notabug.org/isengaara/sekai/src/master/bin/espeak-sg.cpp and
currently does not work with klatt voices yet. In the next few month I
plan to write a patch to support singing with klatt voices too.

https://notabug.org/isengaara/qtau
https://notabug.org/isengaara/sinsy
https://notabug.org/isengaara/sekai







Tobias Platen
 

On Sat, 2021-05-01 at 07:32 -0400, Russell James wrote:
Hi Tobias
This sounds awesome!
I always wondered what it would take to make he speak sing instead of
talk :-)
Can you share an example song?
I imagine when people hear this it will go viral! :-)
Thank you for sharing
Russ

On Sat, May 1, 2021, 7:11 AM Tobias Platen <tplaten@posteo.de> wrote:
Today I just made a new release of my piano roll editor for singing
synthesis called QTAU, including its dependencies Sekai and Sinsy.
QTAU can make espeak-ng sing, using a patch I had written some time
ago, future versions will also support Festival and MBROLA voices.
The singing generator source code is at
https://notabug.org/isengaara/sekai/src/master/bin/espeak-sg.cpp
and
currently does not work with klatt voices yet. In the next few
month I
plan to write a patch to support singing with klatt voices too.

https://notabug.org/isengaara/qtau
https://notabug.org/isengaara/sinsy
https://notabug.org/isengaara/sekai
The first one is using the MBROLA algorithm on an UTAU voice,
the second one is espeak without MBROLA.
https://archive.org/details/kasane-teto-vo-samp-sakura-no-ame
https://archive.org/details/ten_suchdolsky_rybnik_espeak_vosamp





Russell James
 

Hi Tobias

This is very cool!

What is it like to create these with your editor?

Russ

The first one is using the MBROLA algorithm on an UTAU voice,
the second one is espeak without MBROLA.
https://archive.org/details/kasane-teto-vo-samp-sakura-no-ame
https://archive.org/details/ten_suchdolsky_rybnik_espeak_vosamp


On Sun, May 2, 2021 at 4:14 AM Tobias Platen <tplaten@...> wrote:
On Sat, 2021-05-01 at 07:32 -0400, Russell James wrote:
> Hi Tobias
> This sounds awesome!
> I always wondered what it would take to make he speak sing instead of
> talk :-)
> Can you share an example song?
> I imagine when people hear this it will go viral! :-)
> Thank you for sharing
> Russ
>
> On Sat, May 1, 2021, 7:11 AM Tobias Platen <tplaten@...> wrote:
> > Today I just made a new release of my piano roll editor for singing
> > synthesis called QTAU, including its dependencies Sekai and Sinsy.
> > QTAU can make espeak-ng sing, using a patch I had written some time
> > ago, future versions will also support Festival and MBROLA voices.
> > The singing generator source code is at
> > https://notabug.org/isengaara/sekai/src/master/bin/espeak-sg.cpp
> > and
> > currently does not work with klatt voices yet. In the next few
> > month I
> > plan to write a patch to support singing with klatt voices too.
> >
> > https://notabug.org/isengaara/qtau
> > https://notabug.org/isengaara/sinsy
> > https://notabug.org/isengaara/sekai
> >
The first one is using the MBROLA algorithm on an UTAU voice,
the second one is espeak without MBROLA.
https://archive.org/details/kasane-teto-vo-samp-sakura-no-ame
https://archive.org/details/ten_suchdolsky_rybnik_espeak_vosamp
> >
> >
> >
> >
> >
>
>







Tobias Platen
 

On Sun, 2021-05-02 at 19:18 -0400, Russell James wrote:
Hi Tobias

This is very cool!

What is it like to create these with your editor?
You can enter notes in the editor. I also plan something to replace
festival with espeak+mbrola+sinsy as a replacement for the
https://freebsoft.org/singing-computer
I plan to do a workshop soon:
https://www.qtau.de/index.php/sinsycon-2021

Russ

The first one is using the MBROLA algorithm on an UTAU voice,
the second one is espeak without MBROLA.
https://archive.org/details/kasane-teto-vo-samp-sakura-no-ame
https://archive.org/details/ten_suchdolsky_rybnik_espeak_vosamp


On Sun, May 2, 2021 at 4:14 AM Tobias Platen <tplaten@posteo.de>
wrote:
On Sat, 2021-05-01 at 07:32 -0400, Russell James wrote:
Hi Tobias
This sounds awesome!
I always wondered what it would take to make he speak sing
instead of
talk :-)
Can you share an example song?
I imagine when people hear this it will go viral! :-)
Thank you for sharing
Russ

On Sat, May 1, 2021, 7:11 AM Tobias Platen <tplaten@posteo.de>
wrote:
Today I just made a new release of my piano roll editor for
singing
synthesis called QTAU, including its dependencies Sekai and
Sinsy.
QTAU can make espeak-ng sing, using a patch I had written some
time
ago, future versions will also support Festival and MBROLA
voices.
The singing generator source code is at
https://notabug.org/isengaara/sekai/src/master/bin/espeak-sg.cpp
and
currently does not work with klatt voices yet. In the next few
month I
plan to write a patch to support singing with klatt voices too.

https://notabug.org/isengaara/qtau
https://notabug.org/isengaara/sinsy
https://notabug.org/isengaara/sekai
The first one is using the MBROLA algorithm on an UTAU voice,
the second one is espeak without MBROLA.
https://archive.org/details/kasane-teto-vo-samp-sakura-no-ame
https://archive.org/details/ten_suchdolsky_rybnik_espeak_vosamp