Hello everyone.
Currently I am working on HTS on an attempt to tweak it and make it run in streaming. What i have achieved
until now is to make it run, phoneme by phoneme, or n-phonemes by n-phonemes, with n-shift of phonemes
each time.
The problem is, that each time i try to build my computed speech frame, I get some zeros at the beginning of
the computed frame, so i have many discontinuities in my results. I have been debugging quiet a lot in order to
find where these zeros come from, but unfortunately without any luck.
Could you suggest to how to fix that, look for that bug or anything???
Any suggestion would be great!!! Thank you!!!