Thanks a lot for your reading my letter. I meet a problem during running the HTS Speaker dependent training demo.
After correctly build and running the Speaker dependent training demo with raw voice data and utterance (from http://hts.sp.nitech.ac.jp/archives/2.1/HTS-demo_CMU-ARCTIC-SLT.tar.bz2) I could get the correct synthesis voice.
Now I am trying to use my own voice database. According to this document(http://festvox.org/bsv/x3082.html), I have already built my own raw voice and utterance. So I build my own database correctly. But I found nothing helpful to generate/edit the input synthesis source label files (under /Demo_Root_Folder/data/labels/gen).
I am very appreciation for anyone who have had worked on this topic could give me some tips.
Thanks a lot