Hi Nikolay, Thanks for the clarification. I have re-read the documentation and the following is my understanding:
1. In case a word is found in the lexicon, use the phonetic transcription of the same and synthesize the voice from the lpc coefficients and residuals. 2. In case the word is not found, use the lts_rules to predict its pronunciation and then synthesize as above. 3. The lpc coefficients are stored in the group files in each voice directory. There are ofcourse, a number of intermediate steps such as f0 contour or prosody etc., but I have omitted that. Please correct me if my understanding is incorrect. Thanks, Sriram On Thu, Dec 1, 2011 at 2:47 AM, Nickolay V. Shmyrev <nshmy...@yandex.ru>wrote: > Hello > > You can download kal_diphone voice database source here > > http://festvox.org/dbs/dbs_kal.html > > Also you might want to understand how diphone voice is built. Your > "out-of-vocab" doesn't have much sense for diphone voices. > > On Wed, Nov 30, 2011 at 2:55 PM, Sriram Shankar <sri...@voxomos.com> > wrote: > > Hi, > > > > I am trying to add some wav files (in my voice) to try and synthesize > some > > words that are currently out-of-vocab for voice_kal_diphone. I am not > able > > to figure out the directory where these wav files may be stored. I have > > executed system-wide searches and still not found anything. Any help in > > figuring out where the wav files for voice_kal_diphone are stored would > be > > most welcome. >
_______________________________________________ Festlang-talk mailing list Festlang-talk@lists.berlios.de https://lists.berlios.de/mailman/listinfo/festlang-talk
