[mary-users] Fwd: HMMVoiceCreation without GUI

[Vojtěch Hudeček]

Hello,

I have successfully created new HMM Voice for marytts-5.2
I used the tutorial https://github.com/marytts/marytts/wiki/HMMVoiceCreation
and cmu-us-slt-arctic data.

However, I need to automate this process, so I have created Makefile which
does it for me.
I am able to run the Makefile and all the steps complete without errors,
the new voice is compiled and looks ok.
However, when I synthesize some utterance using this voice, the result is
very bad,
the speech is very incomprehensible and it sounds really weird.

Is there any chance, that I somehow omitted some important step that is
normally done by the Voice Import tools?
Can you please suggest some way to track the problems?

Thank you.

Regards,
Vojtech Hudecek
-------------- next part --------------
An HTML attachment was scrubbed...
URL: http://www.dfki.de/pipermail/mary-users/attachments/20161007/b331cdad/attachment.htm