Hi I tried the usage of the "Speaker adaptation/adaptive training demo" for applying emotions (sad/happy .... ) on normal speech. Using normal speech for training and emotion speech for adaptation, all speech from the same speaker. And Using SAT method. Is this the right / best method? The synthesized emotion speech quality is somehow lower than expected (lower than the normal speech). Sometimes I can sense the emotions, however sometimes the speech itself is not clear. Sometimes the emotion is vague. What do you recommend? Regards |