[mary-users] Can I use triphone DB and triphone modules for making new Korean TTS?

[황원준]

If I have audio files made for triphone unit selection, can I only focus on Synthesis components in maryTTS without concerning NLP components?


Thank you for your help.


-----Original Message-----
From: "Ingmar Steiner"<ingmar.steiner at dfki.de> 
To: "황원준"<hwj1593 at naver.com>; 
Cc: <mary-users at dfki.de>; 
Sent: 2017-08-09 (수) 17:24:49
Subject: Re: [mary-users] Can I use triphone DB and triphone modules for making new Korean TTS?
 
Dear 황원준,


the unit selection algorithm in MaryTTS will select larger chunks of 
halfphones if suitable candidates are found in the unit database. While 
there is no explicit triphone unit class in the code, this should not 
make a huge difference in practice. And if required, it would be 
possible to add features and tune weights to bias the selection toward 
triphones.


Having said that, I'm not aware of anyone having built Korean NLP 
components for MaryTTS, but you're welcome to try. Please have a look at 
the wiki and issue tracker on our main GitHub repository and ask 
questions there as appropriate.


Best wishes,


-Ingmar


On 08.08.17 04:32, 황원준 wrote:
> Hello! I am looking for triphone concatenation TTS systems for Korean.
> 
> I was wondering that MaryTTS not only supports conversion Text to 
> Allophone but also supports Text to speech using triphone DB in unit 
> selection.
> 
> 
> 
> _______________________________________________
> Mary-users mailing list
> Mary-users at dfki.de
> http://www.dfki.de/mailman/cgi-bin/listinfo/mary-users
> 

-------------- next part --------------
An HTML attachment was scrubbed...
URL: http://www.dfki.de/pipermail/mary-users/attachments/20170810/8bbc459f/attachment.htm