Dear All, i read that we can build a language db using the HTS in labels we can start without phoneme boundaries (Flat Starting-Symbolic Label Files) i read that the HCompV can compute the global means and variance and set this to all phonemes model i'am wondering 1. i can't imagine how the calculation of global means and variance can be done.more explanation plz 2. how the HERest can determine the phoneme boundaries (this automatic segmentation ) which algorithm can do this and how it do? With Regards, Marc Sobhy Connect to the next generation of MSN Messenger Get it now! |