Currently we can use only TTS voice per callflow, but we would ideally, like to use more than one voice per call flow. This will require more advanced SSML (Speech Synthesis Markup Language) capabilities.
Actually, we are currently limited to one voice per Halef instance as it is configured on the speech server. A changed requires a Halef restart.
The way to implement dynamic voices is to enable SSML support. This should not be too difficult as MARY (and I believe Festival, too) support SSML. It is therefore only a matter of changing JVXML and Cairo to propagate SSML code found in the VXML instructions to the synthesizer. If, and to which extent, OpenVXML will have to be changed, too, will have to be seen later.
Last edit: David Suendermann-Oeft 2014-12-04
If you would like to refer to this comment somewhere else in this project, copy and paste the following link:
I am sorry for the delayed reply, I am very busy at the moment....
Most of my exams will be over end of February and during the semester
break there should be time if not already earlier! Please keep me
updated if something will be done in the meanwhile!
Best regards
Martin
Am 04.12.2014 um 23:42 schrieb David Suendermann-Oeft:
assigned_to: Vikram Ramanarayanan --> Martin Mory - Comment:
Martin,
Will you have cycles to help us with this? Yours,
DSO
** [tickets:#17] Allow more than one TTS voice to be used during
one particular callflow (requires SSML support)**
Status: open Milestone: 1.0 Created: Thu Dec 04, 2014
09:24 PM UTC by Vikram Ramanarayanan Last Updated: Thu Dec 04,
2014 10:38 PM UTC Owner: Martin Mory
Currently we can use only TTS voice per callflow, but we would
ideally, like to use more than one voice per call flow. This will
require more advanced SSML (Speech Synthesis Markup Language)
capabilities.
Actually, we are currently limited to one voice per Halef instance as it is configured on the speech server. A changed requires a Halef restart.
The way to implement dynamic voices is to enable SSML support. This should not be too difficult as MARY (and I believe Festival, too) support SSML. It is therefore only a matter of changing JVXML and Cairo to propagate SSML code found in the VXML instructions to the synthesizer. If, and to which extent, OpenVXML will have to be changed, too, will have to be seen later.
Last edit: David Suendermann-Oeft 2014-12-04
-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1
Hi David,
I am sorry for the delayed reply, I am very busy at the moment....
Most of my exams will be over end of February and during the semester
break there should be time if not already earlier! Please keep me
updated if something will be done in the meanwhile!
Best regards
Martin
Am 04.12.2014 um 23:42 schrieb David Suendermann-Oeft:
GIT/M/O d s++: a-- c++ UL++++ P L+++ E W++ N o? K- w++ O? M? V? PS+ PE
Y? PGP++ t? 5? X? R !tv b+ DI+ D-- G e>+++ h-- r>+++ y?
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v2
iQEcBAEBAgAGBQJUhiOqAAoJEFdK58cB1JR+GL4H/2U3jG9Wqf6lOURp1ZBSPn5Z
vlmCAmu03zgYa8e30mMD53MDx//ixS8G/Aq/VGfK8Wm64vFHl/RzoxGDrKyklJpy
IYGpX3YeUkxMJfb2UrlXXTPrq22A8wt5Wz8HfowvNUzq9q5A+cgrQfgU57da8+A+
qDjowUmvy1kaQ03XzUItM9OZ/Hzhf7AmdQ+mLbjo/n0YrjRBf1Z+2S1OtyfdKQ1B
xfErp1eVdxtv+y53lpV2lNjOAJnXIJyKbCFfl6uawpslnLfM3Qf7AMC1/MdVupE0
bupe6BUpcr5+2Wz2FcezMctGAIZ9mYV7DWnA7y7cXHcXbEpiSLwe+AZV0riQPkI=
=XT+C
-----END PGP SIGNATURE-----
Related
Tickets: #17
Martin,
Will you have cycles to help us with this?
Yours,
DSO