HMM-based Speech Synthesis: Fundamentals and Its Recent Advances