From ftm
- CataRT ... data-driven concatenative sound synthesis in real time based on unit selection from large databases for Max/MSP/FTM
- Gesture Follower ... gesture recognition and online-alignement (following) using HMMs
For simpler examples using FTM see here.